wehub/stefan-jansen--machine-learning-for-trading
1
0
Fork 0
Code Issues Pull Requests Actions 3 Deployments Models Agent Notes Packages Projects Releases Wiki Activity
Files
main
stefan-jansen--machine-lear…/05_synthetic_data/03_sigcwgan_signatures.ipynb
T
wehub-resource-sync 0249a267e1 chore: import upstream snapshot with attribution
2026-07-13 13:26:28 +08:00

9283 lines
656 KiB
Plaintext
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
{
"cells": [
{
"cell_type": "markdown",
"id": "deacd2c7",
"metadata": {
"papermill": {
"duration": 0.006591,
"end_time": "2026-06-14T16:58:43.604375+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:43.597784+00:00",
"status": "completed"
}
},
"source": [
"# Chapter 5: Sig-WGAN - Signature-Based Wasserstein GANs\n",
"\n",
"**Chapter 5: Synthetic Data Generation**\n",
"**Section Reference**: Section 5.4 (GANs for Time Series)\n",
"\n",
"**Docker image**: `ml4t-gpu`\n",
"\n",
"> **Docker required**: This notebook uses `signatory` and `esig`, which are x86-only\n",
"> packages not included in the default environment. Run with:\n",
"> ```bash\n",
"> docker compose --profile py312 run --rm py312 python 05_synthetic_data/03_sigcwgan_signatures.py\n",
"> ```\n",
"\n",
"## Purpose\n",
"\n",
"This notebook implements the **unconditional Sig-Wasserstein GAN** — the core of\n",
"Sig-CWGAN (Ni et al., Mathematical Finance 2024) — which uses **path signatures**\n",
"as a principled distance metric for comparing time-series distributions. We build\n",
"and train the unconditional variant; the conditional extension P(future | past) is\n",
"explained conceptually but not implemented here.\n",
"\n",
"## Learning Objectives\n",
"\n",
"By completing this notebook, you will:\n",
"- Transform financial time series into path signatures using the `signatory` library\n",
"- Implement an (unconditional) Sig-Wasserstein GAN that matches expected signatures\n",
"- Use the expected signature (Sig-W1) as the training objective instead of a\n",
" trained discriminator\n",
"- **Understand how the conditional Sig-CWGAN extends this** with a Monte Carlo\n",
" expected-signature objective (the key paper insight; the conditional path is\n",
" described, not built, in this notebook)\n",
"- Evaluate generative fidelity using signature-based metrics and TSTR\n",
"\n",
"## Cross-References\n",
"\n",
"- **Upstream**: ETF Universe loader (`data`)\n",
"- **Downstream**: Conditional generation for scenario analysis\n",
"- **Book**: Section 5.2 discusses signature-based methods\n",
"\n",
"---\n",
"\n",
"## The Evolution from TimeGAN\n",
"\n",
"TimeGAN (2019) introduced embedding spaces and stepwise supervision. But its\n",
"discriminator still uses a binary classifier to distinguish real/fake - which\n",
"doesn't provide a smooth gradient landscape for generator training.\n",
"\n",
"Sig-CWGAN (2020→2024) introduces two innovations:\n",
"\n",
"1. **Signature Distance**: Instead of a trained discriminator, use the\n",
" **expected signature** to characterize time series distributions. The\n",
" signature is a universal feature set that uniquely identifies probability\n",
" measures on path space.\n",
"\n",
"2. **Conditional W1 Metric** (the paper's full method): learn the conditional\n",
" distribution P(future | past) rather than the unconditional law — more useful\n",
" for finance (given what happened yesterday, what might happen tomorrow?). This\n",
" notebook implements the unconditional Sig-W1 core; the conditional metric is\n",
" described here but left as an extension.\n",
"\n",
"## What Are Path Signatures?\n",
"\n",
"The signature of a path is an infinite series of iterated integrals that\n",
"completely characterizes the path up to reparameterization:\n",
"\n",
"```\n",
"Sig(X)_t = (1, ∫dX_s, ∫∫dX_s⊗dX_u, ∫∫∫dX_s⊗dX_u⊗dX_v, ...)\n",
"```\n",
"\n",
"**Key Properties**:\n",
"- **Uniqueness**: The expected signature E[Sig(X)] uniquely determines\n",
" the law of the stochastic process X (under mild conditions)\n",
"- **Universality**: Any continuous function of paths can be approximated\n",
" by a linear function of the signature\n",
"- **Computability**: Truncated signatures are tractable to compute\n",
"\n",
"## Architecture Overview\n",
"\n",
"Sig-CWGAN consists of three main components:\n",
"1. **Path Augmentation**: Transform raw returns into signature-ready paths\n",
"2. **Signature Computation**: Compute truncated path signatures\n",
"3. **AR-FNN Generator**: Autoregressive feedforward network conditioned on past\n",
"\n",
"See `figures/figure_5_10_sigcwgan_architecture.png` for the full architecture diagram.\n",
"\n",
"**Key Insight**: The loss is on **expected signatures**, not single-sample signatures.\n",
"For each past, we generate m futures with different noise, compute their signatures,\n",
"and average before comparing to the target. This is the core Sig-CWGAN algorithm.\n",
"\n",
"## References\n",
"\n",
"- **Paper**: Ni, H., Szpruch, L., Wiese, M., Liao, S., & Xiao, B. (2024).\n",
" \"Sig-Wasserstein GANs for Conditional Time Series Generation.\"\n",
" Mathematical Finance, 34(4), 1179-1227.\n",
" https://onlinelibrary.wiley.com/doi/full/10.1111/mafi.12423\n",
"- **ArXiv (2020)**: https://arxiv.org/abs/2006.05421\n",
"- **Code**: https://github.com/SigCGANs/Conditional-Sig-Wasserstein-GANs\n",
"\n",
"---\n",
"\n",
"## Paper-Exact Data Source\n",
"\n",
"**This notebook now uses the EXACT same data as the paper**:\n",
"- **Source**: S&P 500 index daily close prices\n",
"- **Returns**: Log returns: `log(close_t) - log(close_{t-1})`\n",
"- **Period**: 2005-01-01 to 2020-06-01 (paper period)\n",
"- **Single asset**: Paper uses `.SPX` only (not multiple assets)\n",
"\n",
"The original paper used Oxford MAN Realized Library (now defunct), which\n",
"provided the same S&P 500 index prices, loaded via `load_sp500_index()`.\n",
"\n",
"## WARNING: Signature Dimension Scaling\n",
"\n",
"Signature dimension scales as O(d^depth) where d = feature count after augmentations.\n",
"\n",
"| Assets | After Augmentations | Depth 4 Sig Dim | Feasibility |\n",
"|--------|---------------------|-----------------|-------------|\n",
"| 1 | 5 features | 780 dims | [OK] Fast |\n",
"| 2 | 7 features | 2,800 dims | [OK] Feasible |\n",
"| 5 | 13 features | 30,940 dims | WARNING: Slow |\n",
"| 10 | 23 features | 303,600 dims | [FAIL] OOM |\n",
"\n",
"**This notebook uses 1 asset (S&P 500) to match the paper exactly.**"
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "3e8ad34d",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:43.616446Z",
"iopub.status.busy": "2026-06-14T16:58:43.616084Z",
"iopub.status.idle": "2026-06-14T16:58:45.431081Z",
"shell.execute_reply": "2026-06-14T16:58:45.430520Z"
},
"papermill": {
"duration": 1.822067,
"end_time": "2026-06-14T16:58:45.431955+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:43.609888+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"\"\"\"Sig-CWGAN — Signature-based conditional Wasserstein GAN for time series generation.\"\"\"\n",
"\n",
"import json\n",
"import math\n",
"from datetime import UTC, datetime\n",
"from pathlib import Path\n",
"\n",
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"import plotly.graph_objects as go\n",
"import polars as pl\n",
"import torch\n",
"import torch.nn as nn\n",
"import torch.optim as optim"
]
},
{
"cell_type": "markdown",
"id": "1c054612",
"metadata": {
"lines_to_next_cell": 0,
"papermill": {
"duration": 0.005245,
"end_time": "2026-06-14T16:58:45.443310+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:45.438065+00:00",
"status": "completed"
}
},
"source": [
"## Sig-CWGAN Architecture\n",
"\n",
"The architecture uses path signatures as an analytic discriminator, replacing\n",
"trained neural networks with a mathematically principled distance metric."
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "5060e708",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:45.454730Z",
"iopub.status.busy": "2026-06-14T16:58:45.454532Z",
"iopub.status.idle": "2026-06-14T16:58:49.126527Z",
"shell.execute_reply": "2026-06-14T16:58:49.125941Z"
},
"papermill": {
"duration": 3.678595,
"end_time": "2026-06-14T16:58:49.127195+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:45.448600+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"from IPython.display import Image, display\n",
"from plotly.subplots import make_subplots\n",
"from scipy import stats\n",
"from tqdm import tqdm\n",
"\n",
"from utils.paths import get_chapter_dir, get_output_dir\n",
"from utils.reproducibility import set_global_seeds\n",
"from utils.style import COLORS, plot_fidelity_comparison\n",
"\n",
"ASSETS_DIR = get_chapter_dir(5) / \"assets\"\n",
"if (ASSETS_DIR / \"sigcwgan_architecture.jpeg\").exists():\n",
" display(Image(ASSETS_DIR / \"sigcwgan_architecture.jpeg\", width=800))\n",
"\n",
"# Checkpoint path for model persistence\n",
"CHECKPOINT_PATH = get_output_dir(5, \"sigcwgan\") / \"checkpoints\" / \"sigcwgan_model.pt\""
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "4e3929ea",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.184048Z",
"iopub.status.busy": "2026-06-14T16:58:49.183884Z",
"iopub.status.idle": "2026-06-14T16:58:49.186363Z",
"shell.execute_reply": "2026-06-14T16:58:49.185931Z"
},
"papermill": {
"duration": 0.009171,
"end_time": "2026-06-14T16:58:49.186821+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.177650+00:00",
"status": "completed"
},
"tags": [
"parameters"
]
},
"outputs": [],
"source": [
"N_LAGS = 16 # Path length (paper: 16)\n",
"SIG_DEPTH = 4 # Signature truncation depth (paper: 4)\n",
"LSTM_HIDDEN_DIM = 50 # LSTM hidden dimension\n",
"LSTM_N_LAYERS = 2 # LSTM layers\n",
"TOTAL_STEPS = 2500 # Gradient steps (paper: 2500)\n",
"BATCH_SIZE = 2000 # Batch size (paper: 2000)\n",
"MC_SAMPLES_TRAIN = 16 # Monte Carlo samples during training\n",
"MC_SAMPLES_EVAL = 500 # Monte Carlo samples during evaluation\n",
"RETRAIN = False # Set True to force retraining even if checkpoint exists\n",
"SEED = 42"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "a985c8b8",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.199025Z",
"iopub.status.busy": "2026-06-14T16:58:49.198864Z",
"iopub.status.idle": "2026-06-14T16:58:49.418812Z",
"shell.execute_reply": "2026-06-14T16:58:49.418224Z"
},
"papermill": {
"duration": 0.227063,
"end_time": "2026-06-14T16:58:49.419271+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.192208+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"set_global_seeds(SEED)"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "013ed315",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.431208Z",
"iopub.status.busy": "2026-06-14T16:58:49.431076Z",
"iopub.status.idle": "2026-06-14T16:58:49.438280Z",
"shell.execute_reply": "2026-06-14T16:58:49.437809Z"
},
"papermill": {
"duration": 0.013874,
"end_time": "2026-06-14T16:58:49.438748+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.424874+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"# =============================================================================\n",
"# SIGNATURE LIBRARY SETUP\n",
"# =============================================================================\n",
"# SigCWGAN training requires differentiable path signatures (autograd support).\n",
"# signatory provides GPU-accelerated differentiable signatures.\n",
"# Install: pip install --no-build-isolation signatory (requires torch first).\n",
"# =============================================================================\n",
"\n",
"import signatory"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "bf1b145d",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.450326Z",
"iopub.status.busy": "2026-06-14T16:58:49.450200Z",
"iopub.status.idle": "2026-06-14T16:58:49.452408Z",
"shell.execute_reply": "2026-06-14T16:58:49.451885Z"
},
"papermill": {
"duration": 0.008675,
"end_time": "2026-06-14T16:58:49.452942+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.444267+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def compute_signature_gpu(paths: torch.Tensor, depth: int) -> torch.Tensor:\n",
" \"\"\"Compute signatures on GPU (differentiable).\"\"\"\n",
" return signatory.signature(paths, depth)"
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "4b86a50f",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.471057Z",
"iopub.status.busy": "2026-06-14T16:58:49.470824Z",
"iopub.status.idle": "2026-06-14T16:58:49.473222Z",
"shell.execute_reply": "2026-06-14T16:58:49.472679Z"
},
"papermill": {
"duration": 0.010657,
"end_time": "2026-06-14T16:58:49.473728+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.463071+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def compute_signature_np(path: np.ndarray, depth: int) -> np.ndarray:\n",
" \"\"\"Compute signature of a single path (numpy).\"\"\"\n",
" path_tensor = torch.tensor(path, dtype=torch.float32).unsqueeze(0)\n",
" sig = signatory.signature(path_tensor, depth)\n",
" return sig.squeeze(0).numpy()"
]
},
{
"cell_type": "markdown",
"id": "63099046",
"metadata": {
"papermill": {
"duration": 0.005412,
"end_time": "2026-06-14T16:58:49.487730+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.482318+00:00",
"status": "completed"
}
},
"source": [
"### Configuration (Ni et al., Mathematical Finance 2024)\n",
"\n",
"Following the original paper exactly:\n",
"- `configs/STOCKS/SigWGAN.json` specifies the exact settings\n",
"- LSTMGenerator with Brownian motion noise input\n",
"- Augmentations: Scale(2, dim=0) → AddTime → LeadLag → VisiTrans(\"I\")\n",
"- Factorial normalization on signature levels\n",
"- batch_size=2000, depth=4, lr=1e-3, n_gradient_steps=2500\n",
"\n",
"**Data source**: S&P 500 index daily close prices (matching Oxford MAN source).\n",
"Paper uses log returns of close prices for a single asset (.SPX),\n",
"period 2005-01-01 to 2020-06-01.\n",
"\n",
"Signature dimension scales as $O(d^{\\text{depth}})$ where $d$ is the feature count.\n",
"With depth=4 and proper augmentations, we match the paper's methodology."
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "7d8e0475",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.499882Z",
"iopub.status.busy": "2026-06-14T16:58:49.499643Z",
"iopub.status.idle": "2026-06-14T16:58:49.502861Z",
"shell.execute_reply": "2026-06-14T16:58:49.502366Z"
},
"papermill": {
"duration": 0.01011,
"end_time": "2026-06-14T16:58:49.503356+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.493246+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"CONFIG = {\n",
" # Data - PAPER EXACT: Single asset (S&P 500), log returns\n",
" \"data_source\": \"sp500\", # Use bundled S&P 500 index data\n",
" \"start_date\": \"2005-01-01\", # Paper start date\n",
" \"end_date\": \"2020-06-01\", # Paper end date (for training)\n",
" \"holdout_start\": \"2018-01-01\", # Our holdout for TSTR evaluation\n",
" # Paper uses n_lags=16 for path length\n",
" \"n_lags\": N_LAGS,\n",
" # Signature - PAPER USES DEPTH 4\n",
" \"sig_depth\": SIG_DEPTH,\n",
" # Architecture - PAPER USES LSTM GENERATOR\n",
" \"lstm_hidden_dim\": LSTM_HIDDEN_DIM,\n",
" \"lstm_n_layers\": LSTM_N_LAYERS,\n",
" \"noise_dim\": 1, # Single asset (paper convention: noise_dim = input_dim)\n",
" # Training - MATCH PAPER EXACTLY\n",
" \"total_steps\": TOTAL_STEPS,\n",
" \"batch_size\": BATCH_SIZE,\n",
" \"learning_rate\": 1e-3, # Paper: 1e-3 (NOT 1e-2)\n",
" \"lr_scheduler_step\": 128, # Paper: StepLR(128, 0.95)\n",
" \"lr_scheduler_gamma\": 0.95,\n",
" # Augmentations (from configs/STOCKS/SigWGAN.json)\n",
" \"scale_factor\": 2.0,\n",
" \"scale_dim\": 0,\n",
" # MC samples for expected signature (conditional version)\n",
" \"mc_samples_train\": MC_SAMPLES_TRAIN,\n",
" \"mc_samples_eval\": MC_SAMPLES_EVAL,\n",
" # Factorial normalization (CRITICAL - paper uses this)\n",
" \"normalise_sig\": True,\n",
"}"
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "c027dfe0",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.516704Z",
"iopub.status.busy": "2026-06-14T16:58:49.516507Z",
"iopub.status.idle": "2026-06-14T16:58:49.519285Z",
"shell.execute_reply": "2026-06-14T16:58:49.518766Z"
},
"papermill": {
"duration": 0.009465,
"end_time": "2026-06-14T16:58:49.519735+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.510270+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Using device: cuda\n"
]
}
],
"source": [
"device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
"print(f\"Using device: {device}\")"
]
},
{
"cell_type": "markdown",
"id": "c97c13f0",
"metadata": {
"papermill": {
"duration": 0.005511,
"end_time": "2026-06-14T16:58:49.534045+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.528534+00:00",
"status": "completed"
}
},
"source": [
"## Data Integrity and TSTR Methodology\n",
"\n",
"**Critical**: This notebook implements a proper train/holdout split to enable\n",
"unbiased Train-Synthetic-Test-Real (TSTR) evaluation.\n",
"\n",
"### Normalization: Preserving Sign (CRITICAL)\n",
"\n",
"**IMPORTANT**: We use z-score normalization (mean 0, std 1), NOT MinMax to [0,1].\n",
"\n",
"Why? The path signature uses cumulative sums. If returns are mapped to [0,1],\n",
"every coordinate becomes monotonically increasing - the path can only \"go up\".\n",
"This destroys the sign structure of returns, which is essential for meaningful\n",
"financial paths and signatures.\n",
"\n",
"```\n",
"WRONG: returns ∈ [0,1] → cumsum always increases → no \"down\" moves\n",
"RIGHT: returns ∈ ℝ (centered) → cumsum goes up and down → realistic paths\n",
"```"
]
},
{
"cell_type": "markdown",
"id": "3bdb9718",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005406,
"end_time": "2026-06-14T16:58:49.544900+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.539494+00:00",
"status": "completed"
}
},
"source": [
"## 1. Load Financial Data\n",
"\n",
"**PAPER-EXACT DATA SOURCE**: S&P 500 index daily close prices\n",
"\n",
"The paper uses the Oxford MAN Realized Library which provided S&P 500 index\n",
"prices (now defunct). We use bundled S&P 500 data which is equivalent.\n",
"\n",
"**Key difference from previous implementation**:\n",
"- **Before**: ETF percent returns (SPY, QQQ)\n",
"- **Now**: S&P 500 index LOG returns (single asset, paper-exact)"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "a7d00f28",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.556853Z",
"iopub.status.busy": "2026-06-14T16:58:49.556632Z",
"iopub.status.idle": "2026-06-14T16:58:49.581280Z",
"shell.execute_reply": "2026-06-14T16:58:49.580826Z"
},
"papermill": {
"duration": 0.031604,
"end_time": "2026-06-14T16:58:49.581833+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.550229+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Loaded 3871 days of S&P 500 LOG returns\n",
"Date range: 2005-01-04 to 2020-06-01\n",
"Return type: LOG returns (paper-exact)\n",
"Data shape: (3871, 1) (days, assets)\n",
"Number of assets: 1 (paper uses 1 = S&P 500 only)\n",
"\n",
"TSTR Data Split (temporal at holdout_start):\n",
" Training: 3,264 days (2005-01-04 to 2017-12-29)\n",
" Holdout: 607 days (2018-01-02 to 2020-06-01)\n",
"\n",
"Paper reference period: 2005-01-01 to 2020-06-01 (~3,800 days)\n",
"\n",
"Using RAW log returns (paper-exact, no normalization):\n",
" Mean: 0.000245\n",
" Std: 0.011910\n",
" Range: [-0.094695, 0.109572]\n",
" NOTE: Paper uses raw log returns to keep signature magnitudes small\n"
]
}
],
"source": [
"def load_sp500_log_returns(\n",
" start_date: str, end_date: str | None = None\n",
") -> tuple[np.ndarray, np.ndarray]:\n",
" \"\"\"\n",
" Load S&P 500 index LOG returns (PAPER-EXACT).\n",
"\n",
" The paper uses: log(close_t) - log(close_{t-1})\n",
" NOT percent returns: (close_t - close_{t-1}) / close_{t-1}\n",
"\n",
" For small returns these are nearly identical, but log returns have\n",
" nicer properties (additive over time, symmetric).\n",
"\n",
" Args:\n",
" start_date: Start date (YYYY-MM-DD)\n",
" end_date: End date (YYYY-MM-DD), or None for all data\n",
"\n",
" Returns:\n",
" Tuple of (log_returns, timestamps) where:\n",
" - log_returns: np.ndarray of shape (n_days, 1)\n",
" - timestamps: np.ndarray of datetime64 timestamps\n",
" \"\"\"\n",
" from data import load_sp500_index\n",
"\n",
" df = load_sp500_index()\n",
"\n",
" start_dt = pl.lit(start_date).str.to_date()\n",
" df = df.filter(pl.col(\"timestamp\") >= start_dt)\n",
"\n",
" if end_date is not None:\n",
" end_dt = pl.lit(end_date).str.to_date()\n",
" df = df.filter(pl.col(\"timestamp\") <= end_dt)\n",
"\n",
" df = df.sort(\"timestamp\")\n",
"\n",
" # Compute LOG returns (PAPER-EXACT)\n",
" # log(close_t) - log(close_{t-1}) = log(close_t / close_{t-1})\n",
" df = df.with_columns(pl.col(\"close\").log().diff().alias(\"log_return\")).drop_nulls()\n",
"\n",
" timestamps = df.select(\"timestamp\").to_numpy().flatten()\n",
" log_returns = df.select(\"log_return\").to_numpy().astype(np.float32)\n",
"\n",
" print(f\"Loaded {len(log_returns)} days of S&P 500 LOG returns\")\n",
" print(f\"Date range: {timestamps[0]} to {timestamps[-1]}\")\n",
" print(\"Return type: LOG returns (paper-exact)\")\n",
"\n",
" return log_returns, timestamps\n",
"\n",
"\n",
"all_returns, all_timestamps = load_sp500_log_returns(CONFIG[\"start_date\"], CONFIG.get(\"end_date\"))\n",
"n_total_days, n_assets = all_returns.shape\n",
"print(f\"Data shape: {all_returns.shape} (days, assets)\")\n",
"print(f\"Number of assets: {n_assets} (paper uses 1 = S&P 500 only)\")\n",
"\n",
"# CRITICAL: Split at holdout_start date\n",
"holdout_dt = np.datetime64(CONFIG[\"holdout_start\"])\n",
"train_mask = all_timestamps < holdout_dt\n",
"\n",
"returns_train_raw = all_returns[train_mask]\n",
"holdout_returns_raw = all_returns[~train_mask]\n",
"train_timestamps = all_timestamps[train_mask]\n",
"holdout_timestamps = all_timestamps[~train_mask]\n",
"n_days = len(returns_train_raw)\n",
"\n",
"print(\"\\nTSTR Data Split (temporal at holdout_start):\")\n",
"print(\n",
" f\" Training: {len(returns_train_raw):,} days ({train_timestamps[0]} to {train_timestamps[-1]})\"\n",
")\n",
"print(\n",
" f\" Holdout: {len(holdout_returns_raw):,} days ({holdout_timestamps[0]} to {holdout_timestamps[-1]})\"\n",
")\n",
"print(\"\\nPaper reference period: 2005-01-01 to 2020-06-01 (~3,800 days)\")\n",
"\n",
"# =============================================================================\n",
"# PAPER-EXACT: NO NORMALIZATION\n",
"# =============================================================================\n",
"# The paper uses RAW log returns without any normalization.\n",
"# Daily log returns are tiny (~0.001), which keeps signature values small.\n",
"# This is critical for matching the paper's Sig-W1 values (~2.76).\n",
"# =============================================================================\n",
"\n",
"# Use raw log returns (NO z-score normalization - paper-exact)\n",
"returns = returns_train_raw\n",
"holdout_returns = holdout_returns_raw\n",
"\n",
"# Store statistics for reference (but don't apply normalization)\n",
"train_mean = returns_train_raw.mean(axis=0)\n",
"train_std = returns_train_raw.std(axis=0)\n",
"\n",
"print(\"\\nUsing RAW log returns (paper-exact, no normalization):\")\n",
"print(f\" Mean: {train_mean[0]:.6f}\")\n",
"print(f\" Std: {train_std[0]:.6f}\")\n",
"print(f\" Range: [{returns.min():.6f}, {returns.max():.6f}]\")\n",
"print(\" NOTE: Paper uses raw log returns to keep signature magnitudes small\")\n",
"\n",
"# Keep original scale for output\n",
"returns_original = returns_train_raw\n",
"holdout_returns_original = holdout_returns_raw"
]
},
{
"cell_type": "markdown",
"id": "f89232cf",
"metadata": {
"papermill": {
"duration": 0.005341,
"end_time": "2026-06-14T16:58:49.594425+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.589084+00:00",
"status": "completed"
}
},
"source": [
"## 2. Path Augmentation and Signature Computation\n",
"\n",
"Raw return paths need **augmentation** before signature computation.\n",
"\n",
"**PAPER PIPELINE** (configs/STOCKS/SigWGAN.json):\n",
"1. **Scale(2, dim=0)**: Scale first dimension by 2\n",
"2. **AddTime**: Prepend time coordinate [0,1]\n",
"3. **LeadLag**: Lead-lag transform for quadratic variation\n",
"4. **VisiTrans(\"I\")**: I-visibility transform (adds 2 rows + 1 column)\n",
"\n",
"**CRITICAL**: Our previous implementation used cumsum_concat + lag_added,\n",
"which is NOT what the paper uses. The paper's augmentations are specifically\n",
"designed for their signature-based loss.\n",
"\n",
"The signature depth of 4 with these augmentations matches the paper methodology."
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "c4d36455",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.605816Z",
"iopub.status.busy": "2026-06-14T16:58:49.605691Z",
"iopub.status.idle": "2026-06-14T16:58:49.608259Z",
"shell.execute_reply": "2026-06-14T16:58:49.607764Z"
},
"papermill": {
"duration": 0.008812,
"end_time": "2026-06-14T16:58:49.608508+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.599696+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"# =============================================================================\n",
"# PAPER-EXACT PATH TRANSFORMS (from lib/augmentations.py)\n",
"# =============================================================================\n",
"# These transforms MUST match the paper exactly for comparable results.\n",
"# Reference: configs/STOCKS/SigWGAN.json\n",
"# =============================================================================\n",
"\n",
"\n",
"def scale_transform(path: np.ndarray, scale: float = 2.0, dim: int = 0) -> np.ndarray:\n",
" \"\"\"\n",
" Scale a specific dimension of the path.\n",
"\n",
" Paper config: {\"name\": \"Scale\", \"scale\": 2, \"dim\": 0}\n",
"\n",
" Args:\n",
" path: Shape (seq_len, d) or (batch, seq_len, d)\n",
" scale: Scaling factor\n",
" dim: Dimension to scale (0 = first asset)\n",
"\n",
" Returns:\n",
" Scaled path with same shape\n",
" \"\"\"\n",
" result = path.copy()\n",
" result[..., dim] = scale * result[..., dim]\n",
" return result"
]
},
{
"cell_type": "code",
"execution_count": 12,
"id": "0df6bed2",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.619994Z",
"iopub.status.busy": "2026-06-14T16:58:49.619850Z",
"iopub.status.idle": "2026-06-14T16:58:49.622066Z",
"shell.execute_reply": "2026-06-14T16:58:49.621676Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.008586,
"end_time": "2026-06-14T16:58:49.622452+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.613866+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def add_time_transform(path: np.ndarray) -> np.ndarray:\n",
" \"\"\"\n",
" Prepend time coordinate [0,1] to the path.\n",
"\n",
" Paper config: {\"name\": \"AddTime\"}\n",
"\n",
" This makes time explicit in the signature, allowing it to capture\n",
" the rate of change, not just the path shape.\n",
"\n",
" Args:\n",
" path: Shape (seq_len, d)\n",
"\n",
" Returns:\n",
" Shape (seq_len, d+1) with time as first coordinate\n",
" \"\"\"\n",
" seq_len = path.shape[0]\n",
" time = np.linspace(0, 1, seq_len).reshape(-1, 1).astype(path.dtype)\n",
" return np.concatenate([time, path], axis=-1)"
]
},
{
"cell_type": "markdown",
"id": "dd3146dd",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005298,
"end_time": "2026-06-14T16:58:49.633167+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.627869+00:00",
"status": "completed"
}
},
"source": [
"### Lead-Lag Transform\n",
"\n",
"The lead-lag transform recovers quadratic variation from discrete observations\n",
"by doubling the path with offset copies. This is critical because discrete\n",
"sampling loses continuous-path information that signatures need."
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "34260a79",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.644830Z",
"iopub.status.busy": "2026-06-14T16:58:49.644704Z",
"iopub.status.idle": "2026-06-14T16:58:49.646968Z",
"shell.execute_reply": "2026-06-14T16:58:49.646615Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.008573,
"end_time": "2026-06-14T16:58:49.647249+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.638676+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def lead_lag_transform(path: np.ndarray) -> np.ndarray:\n",
" \"\"\"\n",
" Apply lead-lag transformation (PAPER VERSION).\n",
"\n",
" Paper config: {\"name\": \"LeadLag\"}\n",
"\n",
" This is crucial for signatures because it recovers quadratic variation\n",
" information that would otherwise be lost in discrete approximations.\n",
"\n",
" Implementation from lib/augmentations.py:\n",
" x_rep = repeat_interleave(x, repeats=2, dim=1)\n",
" x_ll = cat([x_rep[:, :-1], x_rep[:, 1:]], dim=2)\n",
"\n",
" Args:\n",
" path: Shape (seq_len, d)\n",
"\n",
" Returns:\n",
" Shape (2 * seq_len - 1, 2 * d)\n",
" \"\"\"\n",
" # Paper implementation: repeat each point twice, then concatenate\n",
" # shifted versions to create lead-lag pairs\n",
" seq_len, d = path.shape\n",
"\n",
" # Repeat each point twice along sequence dimension\n",
" x_rep = np.repeat(path, 2, axis=0) # (2*seq_len, d)\n",
"\n",
" # Create lead-lag: [x_rep[:-1], x_rep[1:]]\n",
" x_ll = np.concatenate([x_rep[:-1], x_rep[1:]], axis=1) # (2*seq_len-1, 2*d)\n",
"\n",
" return x_ll"
]
},
{
"cell_type": "markdown",
"id": "ce2a8240",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005313,
"end_time": "2026-06-14T16:58:49.657957+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.652644+00:00",
"status": "completed"
}
},
"source": [
"### Visibility Transform\n",
"\n",
"The I-visibility transform adds boundary rows and a visibility marker column,\n",
"ensuring signature terms handle path initialization correctly."
]
},
{
"cell_type": "code",
"execution_count": 14,
"id": "8c45d55f",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.669501Z",
"iopub.status.busy": "2026-06-14T16:58:49.669366Z",
"iopub.status.idle": "2026-06-14T16:58:49.671869Z",
"shell.execute_reply": "2026-06-14T16:58:49.671528Z"
},
"papermill": {
"duration": 0.008813,
"end_time": "2026-06-14T16:58:49.672130+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.663317+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def visibility_transform_I(path: np.ndarray) -> np.ndarray:\n",
" \"\"\"\n",
" I-visibility transform (PAPER VERSION).\n",
"\n",
" Paper config: {\"name\": \"VisiTrans\", \"type\": \"I\"}\n",
"\n",
" Implementation from lib/augmentations.py:\n",
" - Adds 2 rows at start (zeros, copy of first row)\n",
" - Adds 1 column (visibility marker: 0,0,1,1,1,...,1)\n",
"\n",
" Args:\n",
" path: Shape (seq_len, d)\n",
"\n",
" Returns:\n",
" Shape (seq_len + 2, d + 1)\n",
" \"\"\"\n",
" seq_len, d = path.shape\n",
"\n",
" # Row 0: zeros\n",
" # Row 1: copy of original row 0\n",
" # Rows 2+: original path\n",
" init_zeros = np.zeros((1, d), dtype=path.dtype)\n",
" init_copy = path[:1, :] # First row of original path\n",
" main_path = np.concatenate([init_zeros, init_copy, path], axis=0) # (seq_len+2, d)\n",
"\n",
" # New column: visibility marker [0, 0, 1, 1, 1, ..., 1]\n",
" new_col = np.concatenate([np.zeros(2), np.ones(seq_len)]).reshape(-1, 1).astype(path.dtype)\n",
"\n",
" return np.concatenate([main_path, new_col], axis=1) # (seq_len+2, d+1)"
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "bdbc07d7",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.683806Z",
"iopub.status.busy": "2026-06-14T16:58:49.683685Z",
"iopub.status.idle": "2026-06-14T16:58:49.685703Z",
"shell.execute_reply": "2026-06-14T16:58:49.685287Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.008739,
"end_time": "2026-06-14T16:58:49.686360+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.677621+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def cumulative_sum_transform(path: np.ndarray) -> np.ndarray:\n",
" \"\"\"\n",
" Convert returns to cumulative sum (log-price path).\n",
"\n",
" NOTE: This is NOT used in the paper's STOCKS config.\n",
" Kept for backwards compatibility but not part of the paper pipeline.\n",
" \"\"\"\n",
" return np.cumsum(path, axis=0)"
]
},
{
"cell_type": "markdown",
"id": "62bf9467",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005451,
"end_time": "2026-06-14T16:58:49.698180+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.692729+00:00",
"status": "completed"
}
},
"source": [
"### Paper Augmentation Pipeline\n",
"\n",
"The full pipeline from `configs/STOCKS/SigWGAN.json`:\n",
"Scale(2, dim=0) → AddTime → LeadLag → VisiTrans(\"I\").\n",
"This specific ordering is critical for reproducing paper results."
]
},
{
"cell_type": "code",
"execution_count": 16,
"id": "96c28f34",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.709864Z",
"iopub.status.busy": "2026-06-14T16:58:49.709728Z",
"iopub.status.idle": "2026-06-14T16:58:49.712091Z",
"shell.execute_reply": "2026-06-14T16:58:49.711665Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.008852,
"end_time": "2026-06-14T16:58:49.712390+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.703538+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def augment_path_paper(path: np.ndarray, config: dict) -> np.ndarray:\n",
" \"\"\"\n",
" Apply PAPER-EXACT augmentation pipeline.\n",
"\n",
" Pipeline from configs/STOCKS/SigWGAN.json:\n",
" 1. Scale(2, dim=0)\n",
" 2. AddTime\n",
" 3. LeadLag\n",
" 4. VisiTrans(\"I\")\n",
"\n",
" Args:\n",
" path: Shape (seq_len, d) - raw returns or prices\n",
" config: Configuration dict with scale_factor, scale_dim\n",
"\n",
" Returns:\n",
" Augmented path ready for signature computation\n",
" \"\"\"\n",
" result = path.copy()\n",
"\n",
" # Step 1: Scale(2, dim=0) - scale first dimension by 2\n",
" result = scale_transform(result, scale=config[\"scale_factor\"], dim=config[\"scale_dim\"])\n",
"\n",
" # Step 2: AddTime - prepend time coordinate\n",
" result = add_time_transform(result)\n",
"\n",
" # Step 3: LeadLag - lead-lag transform\n",
" result = lead_lag_transform(result)\n",
"\n",
" # Step 4: VisiTrans(\"I\") - I-visibility transform\n",
" result = visibility_transform_I(result)\n",
"\n",
" return result"
]
},
{
"cell_type": "markdown",
"id": "f8d08304",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005271,
"end_time": "2026-06-14T16:58:49.723068+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.717797+00:00",
"status": "completed"
}
},
"source": [
"### Signature Dimension Helpers\n",
"\n",
"Path and signature dimensions scale exponentially with augmentation count\n",
"and signature depth. These helpers compute exact sizes for verification."
]
},
{
"cell_type": "code",
"execution_count": 17,
"id": "9df4b864",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.735203Z",
"iopub.status.busy": "2026-06-14T16:58:49.735066Z",
"iopub.status.idle": "2026-06-14T16:58:49.737131Z",
"shell.execute_reply": "2026-06-14T16:58:49.736701Z"
},
"papermill": {
"duration": 0.008857,
"end_time": "2026-06-14T16:58:49.737393+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.728536+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def compute_path_dim_paper(input_dim: int) -> int:\n",
" \"\"\"\n",
" Compute path dimension after paper augmentations.\n",
"\n",
" Scale(2, dim=0): d → d (no change)\n",
" AddTime: d → d+1\n",
" LeadLag: d+1 → 2*(d+1)\n",
" VisiTrans(\"I\"): 2*(d+1) → 2*(d+1)+1\n",
"\n",
" Total: d → 2*(d+1)+1 = 2d+3\n",
" \"\"\"\n",
" d = input_dim\n",
" d = d + 1 # AddTime\n",
" d = d * 2 # LeadLag\n",
" d = d + 1 # VisiTrans adds 1 column\n",
" return d"
]
},
{
"cell_type": "code",
"execution_count": 18,
"id": "99bd4e27",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.750050Z",
"iopub.status.busy": "2026-06-14T16:58:49.749911Z",
"iopub.status.idle": "2026-06-14T16:58:49.752176Z",
"shell.execute_reply": "2026-06-14T16:58:49.751634Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.009417,
"end_time": "2026-06-14T16:58:49.752630+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.743213+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def compute_sig_dim_paper(input_dim: int, depth: int) -> int:\n",
" \"\"\"\n",
" Compute signature dimension after paper augmentations.\n",
"\n",
" Signature dimension = sum_{k=1}^{depth} path_dim^k\n",
" \"\"\"\n",
" path_dim = compute_path_dim_paper(input_dim)\n",
" return sum(path_dim**k for k in range(1, depth + 1))"
]
},
{
"cell_type": "markdown",
"id": "6e9da390",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005659,
"end_time": "2026-06-14T16:58:49.764042+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.758383+00:00",
"status": "completed"
}
},
"source": [
"### Factorial-Normalized Signatures\n",
"\n",
"Without normalization, signature level $k$ has magnitude $O(1/k!)$, making\n",
"higher levels invisible to the loss. The paper multiplies each level\n",
"by $k!$ to bring all levels to comparable scale."
]
},
{
"cell_type": "code",
"execution_count": 19,
"id": "8d162c15",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.775942Z",
"iopub.status.busy": "2026-06-14T16:58:49.775782Z",
"iopub.status.idle": "2026-06-14T16:58:49.778950Z",
"shell.execute_reply": "2026-06-14T16:58:49.778270Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.009847,
"end_time": "2026-06-14T16:58:49.779353+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.769506+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def compute_signature_with_factorial_norm(\n",
" path: np.ndarray, depth: int, normalise: bool = True\n",
") -> np.ndarray:\n",
" \"\"\"\n",
" Compute signature WITH FACTORIAL NORMALIZATION (paper method).\n",
"\n",
" From lib/trainers/sig_wgan.py:\n",
" for i in range(depth):\n",
" expected_signature[count:count + dim**(i+1)] *= math.factorial(i+1)\n",
"\n",
" This rescales signature levels to comparable magnitudes:\n",
" - Level 1 terms × 1!\n",
" - Level 2 terms × 2!\n",
" - Level 3 terms × 3!\n",
" - Level 4 terms × 4!\n",
"\n",
" Without this, higher-level terms have much smaller magnitudes due to\n",
" the iterated integral structure, causing training instability.\n",
"\n",
" Args:\n",
" path: Augmented path, shape (seq_len, d)\n",
" depth: Signature truncation depth\n",
" normalise: Whether to apply factorial normalization (paper: True)\n",
"\n",
" Returns:\n",
" Signature vector (possibly normalized)\n",
" \"\"\"\n",
" sig = compute_signature_np(path, depth)\n",
"\n",
" if normalise:\n",
" dim = path.shape[-1] # Path dimension after augmentation\n",
" count = 0\n",
" for level in range(depth):\n",
" level_size = dim ** (level + 1)\n",
" # Multiply level (i+1) terms by (i+1)!\n",
" sig[count : count + level_size] *= math.factorial(level + 1)\n",
" count += level_size\n",
"\n",
" return sig"
]
},
{
"cell_type": "markdown",
"id": "2680b85f",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005444,
"end_time": "2026-06-14T16:58:49.790570+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.785126+00:00",
"status": "completed"
}
},
"source": [
"### Augment and Compute Signature\n",
"\n",
"Convenience function combining the full augmentation pipeline with\n",
"factorial-normalized signature computation."
]
},
{
"cell_type": "code",
"execution_count": 20,
"id": "a79e7918",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.802468Z",
"iopub.status.busy": "2026-06-14T16:58:49.802249Z",
"iopub.status.idle": "2026-06-14T16:58:49.805627Z",
"shell.execute_reply": "2026-06-14T16:58:49.804754Z"
},
"papermill": {
"duration": 0.010008,
"end_time": "2026-06-14T16:58:49.806029+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.796021+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def augment_and_signature_paper(\n",
" path: np.ndarray, depth: int, config: dict, normalise: bool = True\n",
") -> np.ndarray:\n",
" \"\"\"\n",
" Full paper pipeline: augment + compute signature with factorial normalization.\n",
"\n",
" Args:\n",
" path: Raw path, shape (seq_len, d)\n",
" depth: Signature truncation depth\n",
" config: Configuration dict\n",
" normalise: Whether to apply factorial normalization\n",
"\n",
" Returns:\n",
" Signature vector\n",
" \"\"\"\n",
" augmented = augment_path_paper(path, config)\n",
" return compute_signature_with_factorial_norm(augmented, depth, normalise=normalise)"
]
},
{
"cell_type": "markdown",
"id": "91345ae0",
"metadata": {
"papermill": {
"duration": 0.005383,
"end_time": "2026-06-14T16:58:49.817042+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.811659+00:00",
"status": "completed"
}
},
"source": [
"### Verify Augmentation Pipeline\n",
"\n",
"Confirm dimension calculations match actual signature output."
]
},
{
"cell_type": "code",
"execution_count": 21,
"id": "d7a85b81",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.829069Z",
"iopub.status.busy": "2026-06-14T16:58:49.828864Z",
"iopub.status.idle": "2026-06-14T16:58:49.835416Z",
"shell.execute_reply": "2026-06-14T16:58:49.834763Z"
},
"papermill": {
"duration": 0.013564,
"end_time": "2026-06-14T16:58:49.835954+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.822390+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"=== Paper-Exact Signature Specification ===\n",
"Pipeline: Scale(2, dim=0) → AddTime → LeadLag → VisiTrans('I')\n",
"Input dimension: 1\n",
"Path dimension after augmentations: 5\n",
"Signature dimension (depth=4): 780\n",
"Factorial normalization: True\n",
"\n",
"Test path shape: (16, 1)\n",
"Augmented path shape: (33, 5)\n",
"Signature dimension: 780\n",
"Expected signature dimension: 780\n",
"[OK] Signature dimensions verified\n"
]
}
],
"source": [
"# =============================================================================\n",
"# VERIFY AUGMENTATION PIPELINE\n",
"# =============================================================================\n",
"\n",
"print(\"\\n=== Paper-Exact Signature Specification ===\")\n",
"print(\"Pipeline: Scale(2, dim=0) → AddTime → LeadLag → VisiTrans('I')\")\n",
"print(f\"Input dimension: {n_assets}\")\n",
"path_dim = compute_path_dim_paper(n_assets)\n",
"sig_dim = compute_sig_dim_paper(n_assets, CONFIG[\"sig_depth\"])\n",
"print(f\"Path dimension after augmentations: {path_dim}\")\n",
"print(f\"Signature dimension (depth={CONFIG['sig_depth']}): {sig_dim}\")\n",
"print(f\"Factorial normalization: {CONFIG['normalise_sig']}\")\n",
"\n",
"# Test signature computation\n",
"test_path = returns[: CONFIG[\"n_lags\"]]\n",
"test_augmented = augment_path_paper(test_path, CONFIG)\n",
"test_sig = compute_signature_with_factorial_norm(\n",
" test_augmented, CONFIG[\"sig_depth\"], normalise=CONFIG[\"normalise_sig\"]\n",
")\n",
"\n",
"print(f\"\\nTest path shape: {test_path.shape}\")\n",
"print(f\"Augmented path shape: {test_augmented.shape}\")\n",
"print(f\"Signature dimension: {len(test_sig)}\")\n",
"print(f\"Expected signature dimension: {sig_dim}\")\n",
"\n",
"# Verify dimensions match\n",
"assert len(test_sig) == sig_dim, f\"Signature dim {len(test_sig)} != expected {sig_dim}\"\n",
"print(\"[OK] Signature dimensions verified\")"
]
},
{
"cell_type": "markdown",
"id": "39748a6c",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.007029,
"end_time": "2026-06-14T16:58:49.853484+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.846455+00:00",
"status": "completed"
}
},
"source": [
"## 3. Create Training Dataset\n",
"\n",
"**Paper Approach**: The paper trains on UNCONDITIONAL generation - matching\n",
"the distribution of full paths, not conditional on past.\n",
"\n",
"We create rolling windows of length `n_lags` for signature computation."
]
},
{
"cell_type": "code",
"execution_count": 22,
"id": "b3c8672f",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.865501Z",
"iopub.status.busy": "2026-06-14T16:58:49.865362Z",
"iopub.status.idle": "2026-06-14T16:58:49.869842Z",
"shell.execute_reply": "2026-06-14T16:58:49.869263Z"
},
"papermill": {
"duration": 0.011198,
"end_time": "2026-06-14T16:58:49.870174+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.858976+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Created 3249 training windows of length 16\n",
"Training windows shape: (3249, 16, 1)\n",
"Created 592 holdout windows for evaluation\n",
"Holdout windows shape: (592, 16, 1)\n"
]
}
],
"source": [
"def create_rolling_windows(data: np.ndarray, n_lags: int) -> np.ndarray:\n",
" \"\"\"\n",
" Create overlapping windows of fixed length.\n",
"\n",
" Paper uses rolling windows for training data - each window is a\n",
" complete path for signature computation.\n",
"\n",
" Args:\n",
" data: Shape (n_days, n_assets)\n",
" n_lags: Window length (paper: 16)\n",
"\n",
" Returns:\n",
" Shape (n_windows, n_lags, n_assets)\n",
" \"\"\"\n",
" n_windows = len(data) - n_lags + 1\n",
" windows = np.zeros((n_windows, n_lags, data.shape[1]), dtype=data.dtype)\n",
"\n",
" for i in range(n_windows):\n",
" windows[i] = data[i : i + n_lags]\n",
"\n",
" return windows\n",
"\n",
"\n",
"# Create training windows\n",
"n_lags = CONFIG[\"n_lags\"]\n",
"train_windows = create_rolling_windows(returns, n_lags)\n",
"holdout_windows = create_rolling_windows(holdout_returns, n_lags)\n",
"\n",
"print(f\"Created {len(train_windows)} training windows of length {n_lags}\")\n",
"print(f\"Training windows shape: {train_windows.shape}\")\n",
"print(f\"Created {len(holdout_windows)} holdout windows for evaluation\")\n",
"print(f\"Holdout windows shape: {holdout_windows.shape}\")"
]
},
{
"cell_type": "markdown",
"id": "4e025f82",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.005535,
"end_time": "2026-06-14T16:58:49.881376+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.875841+00:00",
"status": "completed"
}
},
"source": [
"## 4. Compute Expected Signature of Real Data\n",
"\n",
"**Paper Approach (UNCONDITIONAL)**: The Sig-W1 metric compares expected signatures:\n",
"\n",
"```\n",
"Sig-W1(μ, ν) = ||E[Sig(X)]_μ - E[Sig(X)]_ν||_2\n",
"```\n",
"\n",
"We precompute the expected signature of the REAL training data. During training,\n",
"we minimize the distance between generated and real expected signatures.\n",
"\n",
"**FACTORIAL NORMALIZATION**: The paper multiplies level k signature terms by k!\n",
"to rescale them to comparable magnitudes (higher levels have smaller raw values)."
]
},
{
"cell_type": "code",
"execution_count": 23,
"id": "aee9a2ac",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:49.893577Z",
"iopub.status.busy": "2026-06-14T16:58:49.893357Z",
"iopub.status.idle": "2026-06-14T16:58:50.243543Z",
"shell.execute_reply": "2026-06-14T16:58:50.243000Z"
},
"papermill": {
"duration": 0.357016,
"end_time": "2026-06-14T16:58:50.243874+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:49.886858+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Computing signatures of real training windows...\n",
"Using factorial normalization: True\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Real signatures: 0%| | 0/3249 [00:00<?, ?it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Real signatures: 31%|███ | 1000/3249 [00:00<00:00, 9993.81it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Real signatures: 62%|██████▏ | 2000/3249 [00:00<00:00, 9937.59it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Real signatures: 92%|█████████▏| 2994/3249 [00:00<00:00, 9791.54it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Real signatures: 100%|██████████| 3249/3249 [00:00<00:00, 9753.03it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"Real signatures shape: (3249, 780)\n",
"Expected signature shape: (780,)\n",
"Expected signature norm: 20.0355\n",
"\n",
"=== Signature Dimension Check ===\n",
" Input dimension (n_assets): 1\n",
" Path dim after paper augmentations: 5\n",
" Expected signature dim: 780\n",
" Actual signature dim: 780\n",
" [OK] Dimensions match as expected\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"def compute_signatures_batch_paper(\n",
" paths: np.ndarray, depth: int, config: dict, normalise: bool = True, desc: str = \"Computing\"\n",
") -> np.ndarray:\n",
" \"\"\"\n",
" Compute signatures for a batch of paths using paper's pipeline.\n",
"\n",
" Args:\n",
" paths: Shape (n_paths, seq_len, n_assets)\n",
" depth: Signature truncation depth\n",
" config: Configuration dict with augmentation params\n",
" normalise: Whether to apply factorial normalization\n",
" desc: Progress bar description\n",
"\n",
" Returns:\n",
" Shape (n_paths, sig_dim)\n",
" \"\"\"\n",
" signatures = []\n",
" for path in tqdm(paths, desc=desc):\n",
" sig = augment_and_signature_paper(path, depth, config, normalise=normalise)\n",
" signatures.append(sig)\n",
" return np.array(signatures)\n",
"\n",
"\n",
"# Compute expected signature of real training data\n",
"print(\"Computing signatures of real training windows...\")\n",
"print(f\"Using factorial normalization: {CONFIG['normalise_sig']}\")\n",
"\n",
"real_signatures = compute_signatures_batch_paper(\n",
" train_windows,\n",
" CONFIG[\"sig_depth\"],\n",
" CONFIG,\n",
" normalise=CONFIG[\"normalise_sig\"],\n",
" desc=\"Real signatures\",\n",
")\n",
"\n",
"# Expected signature = mean over all training windows\n",
"expected_sig_real = real_signatures.mean(axis=0)\n",
"\n",
"print(f\"\\nReal signatures shape: {real_signatures.shape}\")\n",
"print(f\"Expected signature shape: {expected_sig_real.shape}\")\n",
"print(f\"Expected signature norm: {np.linalg.norm(expected_sig_real):.4f}\")\n",
"\n",
"# =============================================================================\n",
"# SIGNATURE DIMENSION VERIFICATION\n",
"# =============================================================================\n",
"expected_sig_dim = compute_sig_dim_paper(n_assets, CONFIG[\"sig_depth\"])\n",
"actual_sig_dim = real_signatures.shape[1]\n",
"\n",
"print(\"\\n=== Signature Dimension Check ===\")\n",
"print(f\" Input dimension (n_assets): {n_assets}\")\n",
"print(f\" Path dim after paper augmentations: {compute_path_dim_paper(n_assets)}\")\n",
"print(f\" Expected signature dim: {expected_sig_dim}\")\n",
"print(f\" Actual signature dim: {actual_sig_dim}\")\n",
"\n",
"if actual_sig_dim != expected_sig_dim:\n",
" print(\" WARNING: Dimension mismatch! Check augmentations.\")\n",
"else:\n",
" print(\" [OK] Dimensions match as expected\")"
]
},
{
"cell_type": "markdown",
"id": "bdcc8202",
"metadata": {
"papermill": {
"duration": 0.006166,
"end_time": "2026-06-14T16:58:50.256882+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.250716+00:00",
"status": "completed"
}
},
"source": [
"## 5. Generator Architecture (PAPER: LSTMGenerator)\n",
"\n",
"**CRITICAL FIX**: The paper uses an **LSTMGenerator** with **Brownian motion noise**,\n",
"NOT an AR-FNN with independent Gaussian noise.\n",
"\n",
"From `lib/networks/generators.py`:\n",
"```python\n",
"z = (0.1 * torch.randn(batch_size, n_lags, input_dim)).cumsum(1) # Brownian!\n",
"h1, _ = self.rnn(z, (h0, c0)) # LSTM processes Brownian path\n",
"x = self.linear(h1) # Linear projection to output\n",
"```\n",
"\n",
"The Brownian motion noise is KEY - it provides smooth, continuous input paths\n",
"to the LSTM, which then transforms them into realistic output paths."
]
},
{
"cell_type": "markdown",
"id": "449c9c4d",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.00614,
"end_time": "2026-06-14T16:58:50.269007+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.262867+00:00",
"status": "completed"
}
},
"source": [
"### Residual Feedforward Network\n",
"\n",
"Helper network for initializing LSTM hidden states from noise input."
]
},
{
"cell_type": "code",
"execution_count": 24,
"id": "6dd7b512",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.281860Z",
"iopub.status.busy": "2026-06-14T16:58:50.281725Z",
"iopub.status.idle": "2026-06-14T16:58:50.284925Z",
"shell.execute_reply": "2026-06-14T16:58:50.284384Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.010416,
"end_time": "2026-06-14T16:58:50.285227+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.274811+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"class ResFNN(nn.Module):\n",
" \"\"\"\n",
" Feedforward network with residual connections and Tanh activation.\n",
"\n",
" Used for initializing LSTM hidden states (paper implementation).\n",
" \"\"\"\n",
"\n",
" def __init__(self, input_dim: int, output_dim: int, hidden_dims: list[int]):\n",
" super().__init__()\n",
"\n",
" layers = []\n",
" dims = [input_dim] + hidden_dims\n",
"\n",
" for i in range(len(dims) - 1):\n",
" layers.append(nn.Linear(dims[i], dims[i + 1]))\n",
" layers.append(nn.ReLU())\n",
"\n",
" layers.append(nn.Linear(dims[-1], output_dim))\n",
" self.net = nn.Sequential(*layers)\n",
"\n",
" # Residual projection if dimensions don't match\n",
" if input_dim != output_dim:\n",
" self.residual = nn.Linear(input_dim, output_dim)\n",
" else:\n",
" self.residual = nn.Identity()\n",
"\n",
" def forward(self, x: torch.Tensor) -> torch.Tensor:\n",
" return self.net(x) + self.residual(x)"
]
},
{
"cell_type": "markdown",
"id": "ba08469a",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.006571,
"end_time": "2026-06-14T16:58:50.297690+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.291119+00:00",
"status": "completed"
}
},
"source": [
"### LSTM Generator\n",
"\n",
"The paper's generator processes Brownian motion noise (cumulative sum of\n",
"Gaussian) through an LSTM, producing smooth output paths. This is different\n",
"from AR-FNN approaches that use independent noise at each step."
]
},
{
"cell_type": "code",
"execution_count": 25,
"id": "304f4ccf",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.310919Z",
"iopub.status.busy": "2026-06-14T16:58:50.310765Z",
"iopub.status.idle": "2026-06-14T16:58:50.315503Z",
"shell.execute_reply": "2026-06-14T16:58:50.314831Z"
},
"papermill": {
"duration": 0.012353,
"end_time": "2026-06-14T16:58:50.315976+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.303623+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"class LSTMGenerator(nn.Module):\n",
" \"\"\"\n",
" LSTM Generator from the paper (lib/networks/generators.py).\n",
"\n",
" Key differences from AR-FNN:\n",
" 1. Uses BROWNIAN MOTION as input noise (cumsum of Gaussian)\n",
" 2. LSTM processes the entire noise sequence at once\n",
" 3. Linear projection to output dimension\n",
"\n",
" This is the EXACT architecture from the paper.\n",
" \"\"\"\n",
"\n",
" def __init__(\n",
" self,\n",
" input_dim: int,\n",
" output_dim: int,\n",
" hidden_dim: int = 50,\n",
" n_layers: int = 2,\n",
" init_fixed: bool = True,\n",
" ):\n",
" super().__init__()\n",
" self.input_dim = input_dim\n",
" self.output_dim = output_dim\n",
" self.hidden_dim = hidden_dim\n",
" self.n_layers = n_layers\n",
" self.init_fixed = init_fixed\n",
"\n",
" # LSTM processes Brownian motion noise\n",
" self.rnn = nn.LSTM(\n",
" input_size=input_dim, hidden_size=hidden_dim, num_layers=n_layers, batch_first=True\n",
" )\n",
"\n",
" # Linear projection to output dimension (no bias, per paper)\n",
" self.linear = nn.Linear(hidden_dim, output_dim, bias=False)\n",
"\n",
" # Optional: NN to initialize h0 from noise (if init_fixed=False)\n",
" if not init_fixed:\n",
" self.initial_nn = nn.Sequential(\n",
" ResFNN(input_dim, hidden_dim * n_layers, [hidden_dim, hidden_dim]), nn.Tanh()\n",
" )\n",
"\n",
" # Initialize weights\n",
" self._init_weights()\n",
"\n",
" def _init_weights(self):\n",
" \"\"\"Initialize weights following paper conventions.\"\"\"\n",
" for name, param in self.named_parameters():\n",
" if \"weight\" in name:\n",
" nn.init.xavier_uniform_(param)\n",
" elif \"bias\" in name:\n",
" nn.init.zeros_(param)\n",
"\n",
" def forward(self, batch_size: int, n_lags: int, device: torch.device) -> torch.Tensor:\n",
" \"\"\"\n",
" Generate paths using LSTM on Brownian motion noise.\n",
"\n",
" Args:\n",
" batch_size: Number of paths to generate\n",
" n_lags: Length of each path\n",
" device: Torch device\n",
"\n",
" Returns:\n",
" Generated paths, shape (batch_size, n_lags, output_dim)\n",
" \"\"\"\n",
" # =================================================================\n",
" # CRITICAL: BROWNIAN MOTION NOISE (paper implementation)\n",
" # =================================================================\n",
" # z = (0.1 * randn(...)).cumsum(1)\n",
" # First point is fixed at 0\n",
" # =================================================================\n",
" z = 0.1 * torch.randn(batch_size, n_lags, self.input_dim, device=device)\n",
" z[:, 0, :] = 0 # First point fixed at origin\n",
" z = z.cumsum(dim=1) # Cumulative sum → Brownian path\n",
"\n",
" # Initialize hidden states\n",
" if self.init_fixed:\n",
" h0 = torch.zeros(self.n_layers, batch_size, self.hidden_dim, device=device)\n",
" else:\n",
" z0 = torch.randn(batch_size, self.input_dim, device=device)\n",
" h0 = self.initial_nn(z0)\n",
" h0 = h0.view(batch_size, self.n_layers, self.hidden_dim).permute(1, 0, 2).contiguous()\n",
"\n",
" c0 = torch.zeros_like(h0)\n",
"\n",
" # Process Brownian path through LSTM\n",
" h1, _ = self.rnn(z, (h0, c0))\n",
"\n",
" # Project to output dimension\n",
" x = self.linear(h1)\n",
"\n",
" assert x.shape[1] == n_lags\n",
" return x"
]
},
{
"cell_type": "markdown",
"id": "7f07bad4",
"metadata": {
"papermill": {
"duration": 0.006398,
"end_time": "2026-06-14T16:58:50.328390+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.321992+00:00",
"status": "completed"
}
},
"source": [
"### Initialize Generator\n",
"\n",
"Instantiate the LSTM generator with paper hyperparameters."
]
},
{
"cell_type": "code",
"execution_count": 26,
"id": "348b504a",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.341141Z",
"iopub.status.busy": "2026-06-14T16:58:50.340893Z",
"iopub.status.idle": "2026-06-14T16:58:50.514760Z",
"shell.execute_reply": "2026-06-14T16:58:50.514126Z"
},
"papermill": {
"duration": 0.181535,
"end_time": "2026-06-14T16:58:50.515694+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.334159+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"=== LSTM Generator (Paper Architecture) ===\n",
"Input dim (noise): 1\n",
"Output dim (assets): 1\n",
"Hidden dim: 50\n",
"LSTM layers: 2\n",
"Total parameters: 31,050\n"
]
}
],
"source": [
"# Initialize generator (PAPER ARCHITECTURE)\n",
"generator = LSTMGenerator(\n",
" input_dim=CONFIG[\"noise_dim\"],\n",
" output_dim=n_assets,\n",
" hidden_dim=CONFIG[\"lstm_hidden_dim\"],\n",
" n_layers=CONFIG[\"lstm_n_layers\"],\n",
" init_fixed=True, # Paper default\n",
").to(device)\n",
"\n",
"print(\"\\n=== LSTM Generator (Paper Architecture) ===\")\n",
"print(f\"Input dim (noise): {CONFIG['noise_dim']}\")\n",
"print(f\"Output dim (assets): {n_assets}\")\n",
"print(f\"Hidden dim: {CONFIG['lstm_hidden_dim']}\")\n",
"print(f\"LSTM layers: {CONFIG['lstm_n_layers']}\")\n",
"print(f\"Total parameters: {sum(p.numel() for p in generator.parameters()):,}\")"
]
},
{
"cell_type": "markdown",
"id": "5266d74d",
"metadata": {
"papermill": {
"duration": 0.006872,
"end_time": "2026-06-14T16:58:50.528918+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.522046+00:00",
"status": "completed"
}
},
"source": [
"## 6. Sig-W1 Loss (PAPER IMPLEMENTATION)\n",
"\n",
"**Paper Loss**: The Sig-W1 metric is the L2 norm (NOT squared MSE) of the\n",
"difference between expected signatures:\n",
"\n",
"```\n",
"Sig-W1(μ, ν) = ||E[Sig(X)]_μ - E[Sig(X)]_ν||_2\n",
"```\n",
"\n",
"For training:\n",
"- Generate a batch of fake paths\n",
"- Compute their expected signature (mean over batch)\n",
"- Compare to precomputed expected signature of real data\n",
"- Loss = L2 norm of difference (RMSE style)\n",
"\n",
"**FACTORIAL NORMALIZATION**: Applied to signatures during training too!"
]
},
{
"cell_type": "code",
"execution_count": 27,
"id": "deab370a",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.542794Z",
"iopub.status.busy": "2026-06-14T16:58:50.542623Z",
"iopub.status.idle": "2026-06-14T16:58:50.545622Z",
"shell.execute_reply": "2026-06-14T16:58:50.544979Z"
},
"papermill": {
"duration": 0.010455,
"end_time": "2026-06-14T16:58:50.545993+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.535538+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"# =============================================================================\n",
"# GPU PATH TRANSFORMS (PAPER-EXACT, Differentiable)\n",
"# =============================================================================\n",
"# These MUST match the numpy transforms exactly:\n",
"# Scale(2, dim=0) → AddTime → LeadLag → VisiTrans(\"I\")\n",
"# =============================================================================\n",
"\n",
"\n",
"def scale_transform_gpu(paths: torch.Tensor, scale: float = 2.0, dim: int = 0) -> torch.Tensor:\n",
" \"\"\"\n",
" Scale a specific dimension (GPU, differentiable).\n",
"\n",
" Paper config: {\"name\": \"Scale\", \"scale\": 2, \"dim\": 0}\n",
" \"\"\"\n",
" result = paths.clone()\n",
" result[..., dim] = scale * result[..., dim]\n",
" return result"
]
},
{
"cell_type": "code",
"execution_count": 28,
"id": "19bb884d",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.560599Z",
"iopub.status.busy": "2026-06-14T16:58:50.560384Z",
"iopub.status.idle": "2026-06-14T16:58:50.563595Z",
"shell.execute_reply": "2026-06-14T16:58:50.562962Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.011131,
"end_time": "2026-06-14T16:58:50.563973+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.552842+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def add_time_transform_gpu(paths: torch.Tensor) -> torch.Tensor:\n",
" \"\"\"\n",
" Prepend time coordinate [0,1] (GPU, differentiable).\n",
"\n",
" Paper config: {\"name\": \"AddTime\"}\n",
" \"\"\"\n",
" batch_size, seq_len, dim = paths.shape\n",
" time = torch.linspace(0, 1, seq_len, device=paths.device, dtype=paths.dtype)\n",
" time = time.view(1, -1, 1).expand(batch_size, -1, 1)\n",
" return torch.cat([time, paths], dim=-1)"
]
},
{
"cell_type": "markdown",
"id": "7c099aa4",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.006041,
"end_time": "2026-06-14T16:58:50.576625+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.570584+00:00",
"status": "completed"
}
},
"source": [
"### GPU Lead-Lag and Visibility Transforms\n",
"\n",
"Differentiable versions of the structural transforms, needed for\n",
"backpropagation through the signature computation during training."
]
},
{
"cell_type": "code",
"execution_count": 29,
"id": "c7fdf917",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.589946Z",
"iopub.status.busy": "2026-06-14T16:58:50.589752Z",
"iopub.status.idle": "2026-06-14T16:58:50.592932Z",
"shell.execute_reply": "2026-06-14T16:58:50.592272Z"
},
"papermill": {
"duration": 0.010722,
"end_time": "2026-06-14T16:58:50.593373+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.582651+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def lead_lag_transform_gpu(paths: torch.Tensor) -> torch.Tensor:\n",
" \"\"\"\n",
" Lead-lag transformation (GPU, differentiable).\n",
"\n",
" Paper config: {\"name\": \"LeadLag\"}\n",
"\n",
" Implementation from lib/augmentations.py:\n",
" x_rep = repeat_interleave(x, repeats=2, dim=1)\n",
" x_ll = cat([x_rep[:, :-1], x_rep[:, 1:]], dim=2)\n",
" \"\"\"\n",
" # Repeat each point twice along sequence dimension\n",
" x_rep = paths.repeat_interleave(2, dim=1) # (batch, 2*seq_len, dim)\n",
"\n",
" # Create lead-lag: [x_rep[:-1], x_rep[1:]]\n",
" x_ll = torch.cat([x_rep[:, :-1], x_rep[:, 1:]], dim=2) # (batch, 2*seq_len-1, 2*dim)\n",
"\n",
" return x_ll"
]
},
{
"cell_type": "code",
"execution_count": 30,
"id": "e17b6817",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.606679Z",
"iopub.status.busy": "2026-06-14T16:58:50.606446Z",
"iopub.status.idle": "2026-06-14T16:58:50.610503Z",
"shell.execute_reply": "2026-06-14T16:58:50.609828Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.011285,
"end_time": "2026-06-14T16:58:50.610950+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.599665+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def visibility_transform_I_gpu(paths: torch.Tensor) -> torch.Tensor:\n",
" \"\"\"\n",
" I-visibility transform (GPU, differentiable).\n",
"\n",
" Paper config: {\"name\": \"VisiTrans\", \"type\": \"I\"}\n",
"\n",
" Implementation from lib/augmentations.py:\n",
" - Adds 2 rows at start (zeros, copy of first row)\n",
" - Adds 1 column (visibility marker: 0,0,1,1,1,...,1)\n",
" \"\"\"\n",
" batch_size, seq_len, dim = paths.shape\n",
" device = paths.device\n",
" dtype = paths.dtype\n",
"\n",
" # Row 0: zeros, Row 1: copy of original row 0, Rows 2+: original path\n",
" init_zeros = torch.zeros(batch_size, 1, dim, device=device, dtype=dtype)\n",
" init_copy = paths[:, :1, :] # First row of original path\n",
" main_path = torch.cat([init_zeros, init_copy, paths], dim=1) # (batch, seq_len+2, dim)\n",
"\n",
" # New column: visibility marker [0, 0, 1, 1, 1, ..., 1]\n",
" vis_marker = torch.cat(\n",
" [\n",
" torch.zeros(2, device=device, dtype=dtype),\n",
" torch.ones(seq_len, device=device, dtype=dtype),\n",
" ]\n",
" )\n",
" vis_marker = vis_marker.view(1, -1, 1).expand(batch_size, -1, 1)\n",
"\n",
" return torch.cat([main_path, vis_marker], dim=2) # (batch, seq_len+2, dim+1)"
]
},
{
"cell_type": "markdown",
"id": "e97cf4e8",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.00574,
"end_time": "2026-06-14T16:58:50.623199+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.617459+00:00",
"status": "completed"
}
},
"source": [
"### GPU Augmentation Pipeline and Expected Signature\n",
"\n",
"The full differentiable pipeline: augment → compute signature → average →\n",
"normalize. This is used during training to compute the Sig-W1 loss."
]
},
{
"cell_type": "code",
"execution_count": 31,
"id": "6f092782",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.636581Z",
"iopub.status.busy": "2026-06-14T16:58:50.636332Z",
"iopub.status.idle": "2026-06-14T16:58:50.639801Z",
"shell.execute_reply": "2026-06-14T16:58:50.639155Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.01137,
"end_time": "2026-06-14T16:58:50.640303+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.628933+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def augment_paths_gpu_paper(paths: torch.Tensor, config: dict) -> torch.Tensor:\n",
" \"\"\"\n",
" Apply PAPER-EXACT augmentation pipeline on GPU (differentiable).\n",
"\n",
" Pipeline from configs/STOCKS/SigWGAN.json:\n",
" 1. Scale(2, dim=0)\n",
" 2. AddTime\n",
" 3. LeadLag\n",
" 4. VisiTrans(\"I\")\n",
" \"\"\"\n",
" result = paths\n",
"\n",
" # Step 1: Scale(2, dim=0)\n",
" result = scale_transform_gpu(result, scale=config[\"scale_factor\"], dim=config[\"scale_dim\"])\n",
"\n",
" # Step 2: AddTime\n",
" result = add_time_transform_gpu(result)\n",
"\n",
" # Step 3: LeadLag\n",
" result = lead_lag_transform_gpu(result)\n",
"\n",
" # Step 4: VisiTrans(\"I\")\n",
" result = visibility_transform_I_gpu(result)\n",
"\n",
" return result"
]
},
{
"cell_type": "markdown",
"id": "349b9503",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.006123,
"end_time": "2026-06-14T16:58:50.652670+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.646547+00:00",
"status": "completed"
}
},
"source": [
"### Expected Signature with Factorial Normalization (GPU)\n",
"\n",
"Compute the mean signature over a batch of augmented paths, applying\n",
"factorial rescaling per level to equalize magnitude across depths."
]
},
{
"cell_type": "code",
"execution_count": 32,
"id": "f3a6dda1",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.666298Z",
"iopub.status.busy": "2026-06-14T16:58:50.666059Z",
"iopub.status.idle": "2026-06-14T16:58:50.669886Z",
"shell.execute_reply": "2026-06-14T16:58:50.669253Z"
},
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.011481,
"end_time": "2026-06-14T16:58:50.670334+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.658853+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def compute_expected_signature_gpu(\n",
" paths: torch.Tensor, depth: int, config: dict, normalise: bool = True\n",
") -> torch.Tensor:\n",
" \"\"\"\n",
" Compute expected signature WITH FACTORIAL NORMALIZATION (GPU, differentiable).\n",
"\n",
" Paper implementation from lib/trainers/sig_wgan.py:\n",
" expected_signature = signatory.signature(x_path_augmented, depth).mean(0)\n",
" if normalise:\n",
" for i in range(depth):\n",
" expected_signature[count:count + dim**(i+1)] *= math.factorial(i+1)\n",
"\n",
" Args:\n",
" paths: Shape (batch, seq_len, dim) - raw generated paths\n",
" depth: Signature truncation depth\n",
" config: Configuration dict\n",
" normalise: Whether to apply factorial normalization\n",
"\n",
" Returns:\n",
" Expected signature vector, shape (sig_dim,)\n",
" \"\"\"\n",
" # Augment paths\n",
" augmented = augment_paths_gpu_paper(paths, config)\n",
"\n",
" # Compute signatures\n",
" sigs = compute_signature_gpu(augmented, depth) # (batch, sig_dim)\n",
"\n",
" # Expected signature = mean over batch\n",
" expected_sig = sigs.mean(dim=0)\n",
"\n",
" # Factorial normalization\n",
" if normalise:\n",
" dim = augmented.shape[-1] # Path dimension after augmentation\n",
" count = 0\n",
" for level in range(depth):\n",
" level_size = dim ** (level + 1)\n",
" expected_sig[count : count + level_size] = expected_sig[\n",
" count : count + level_size\n",
" ] * math.factorial(level + 1)\n",
" count += level_size\n",
"\n",
" return expected_sig"
]
},
{
"cell_type": "markdown",
"id": "4d6e3e08",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.00584,
"end_time": "2026-06-14T16:58:50.682174+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.676334+00:00",
"status": "completed"
}
},
"source": [
"## 7. Training Loop (PAPER: Unconditional Sig-W1)\n",
"\n",
"**Paper Algorithm** (lib/trainers/sig_wgan.py):\n",
"\n",
"```python\n",
"for step in range(n_gradient_steps):\n",
" x_fake = G(batch_size, n_lags, device) # Generate batch\n",
" loss = sig_w1_metric(x_fake) # Sig-W1 loss\n",
" loss.backward()\n",
" optimizer.step()\n",
" scheduler.step()\n",
"```\n",
"\n",
"**Sig-W1 Loss**:\n",
"```\n",
"loss = ||E[Sig(fake)] - E[Sig(real)]||_2 (RMSE, not MSE!)\n",
"```\n",
"\n",
"The generator produces paths unconditionally - no conditioning on past."
]
},
{
"cell_type": "code",
"execution_count": 33,
"id": "5e98cb99",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.694874Z",
"iopub.status.busy": "2026-06-14T16:58:50.694727Z",
"iopub.status.idle": "2026-06-14T16:58:50.699383Z",
"shell.execute_reply": "2026-06-14T16:58:50.698933Z"
},
"papermill": {
"duration": 0.012168,
"end_time": "2026-06-14T16:58:50.700217+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.688049+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def train_sigwgan_paper(\n",
" generator: nn.Module,\n",
" expected_sig_real: np.ndarray,\n",
" config: dict,\n",
" device: torch.device,\n",
") -> list[float]:\n",
" \"\"\"\n",
" Train Sig-WGAN generator using PAPER-EXACT Sig-W1 objective.\n",
"\n",
" This is UNCONDITIONAL generation - the generator learns to produce\n",
" paths whose expected signature matches the real data's expected signature.\n",
"\n",
" Paper reference: lib/trainers/sig_wgan.py, SigWGANTrainer.fit()\n",
" \"\"\"\n",
" # Adam optimizer (paper default)\n",
" optimizer = optim.Adam(generator.parameters(), lr=config[\"learning_rate\"])\n",
"\n",
" # StepLR scheduler: gamma=0.95, step_size=128 (from paper)\n",
" scheduler = optim.lr_scheduler.StepLR(\n",
" optimizer,\n",
" step_size=config[\"lr_scheduler_step\"],\n",
" gamma=config[\"lr_scheduler_gamma\"],\n",
" )\n",
"\n",
" # Convert expected sig to tensor (target)\n",
" target_sig = torch.tensor(expected_sig_real, dtype=torch.float32, device=device)\n",
"\n",
" # Training params\n",
" total_steps = config[\"total_steps\"]\n",
" batch_size = config[\"batch_size\"]\n",
" n_lags = config[\"n_lags\"]\n",
" sig_depth = config[\"sig_depth\"]\n",
" normalise = config[\"normalise_sig\"]\n",
"\n",
" losses = []\n",
" best_loss = None\n",
" best_state = None\n",
"\n",
" pbar = tqdm(range(total_steps), desc=f\"Training ({total_steps} steps)\")\n",
"\n",
" for step in pbar:\n",
" optimizer.zero_grad()\n",
"\n",
" # =================================================================\n",
" # PAPER ALGORITHM: Generate batch, compute Sig-W1 loss\n",
" # =================================================================\n",
" # x_fake = G(batch_size, n_lags, device)\n",
" # expected_sig_fake = compute_expected_signature(x_fake, ...)\n",
" # loss = rmse(expected_sig_real, expected_sig_fake)\n",
" # =================================================================\n",
"\n",
" # Generate batch of fake paths (unconditional!)\n",
" x_fake = generator(batch_size, n_lags, device)\n",
"\n",
" # Compute expected signature of fake paths (with factorial normalization)\n",
" expected_sig_fake = compute_expected_signature_gpu(\n",
" x_fake, sig_depth, config, normalise=normalise\n",
" )\n",
"\n",
" # Sig-W1 loss: RMSE (L2 norm), NOT squared MSE\n",
" loss = torch.sqrt(torch.sum((expected_sig_fake - target_sig) ** 2))\n",
"\n",
" loss.backward()\n",
" optimizer.step()\n",
" scheduler.step()\n",
"\n",
" loss_val = loss.item()\n",
" losses.append(loss_val)\n",
"\n",
" # Track best model\n",
" if best_loss is None or loss_val < best_loss:\n",
" best_loss = loss_val\n",
" best_state = {k: v.cpu().clone() for k, v in generator.state_dict().items()}\n",
"\n",
" # Logging (every 100 steps, matching paper)\n",
" if (step + 1) % 100 == 0:\n",
" current_lr = scheduler.get_last_lr()[0]\n",
" pbar.set_postfix({\"Sig-W1\": f\"{loss_val:.2e}\", \"LR\": f\"{current_lr:.2e}\"})\n",
" print(f\" Step {step + 1}: sig-w1 loss: {loss_val:.2e}\", flush=True)\n",
"\n",
" # Load best model (paper does this)\n",
" if best_state is not None:\n",
" generator.load_state_dict(best_state)\n",
" print(f\"\\nLoaded best model (loss: {best_loss:.2e})\")\n",
"\n",
" return losses"
]
},
{
"cell_type": "markdown",
"id": "a9c37e84",
"metadata": {
"papermill": {
"duration": 0.008856,
"end_time": "2026-06-14T16:58:50.720334+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.711478+00:00",
"status": "completed"
}
},
"source": [
"### Run Training\n",
"\n",
"Execute the Sig-W1 optimization loop with paper-exact settings."
]
},
{
"cell_type": "code",
"execution_count": 34,
"id": "6ef559c7",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.733265Z",
"iopub.status.busy": "2026-06-14T16:58:50.733121Z",
"iopub.status.idle": "2026-06-14T16:58:50.742274Z",
"shell.execute_reply": "2026-06-14T16:58:50.741779Z"
},
"papermill": {
"duration": 0.016419,
"end_time": "2026-06-14T16:58:50.742801+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.726382+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"======================================================================\n",
"LOADING CHECKPOINT (set RETRAIN=True to force retraining)\n",
"======================================================================\n",
"\n",
"Checkpoint path: 05_synthetic_data/output/sigcwgan/checkpoints/sigcwgan_model.pt\n",
"Loaded generator weights from checkpoint\n",
"Training history: 2500 steps\n",
"Final loss: 0.0673\n",
"Best loss: 0.0540\n"
]
}
],
"source": [
"# Check if checkpoint exists and RETRAIN is False\n",
"SKIP_TRAINING = CHECKPOINT_PATH.exists() and not RETRAIN\n",
"\n",
"if SKIP_TRAINING:\n",
" print(\"\\n\" + \"=\" * 70)\n",
" print(\"LOADING CHECKPOINT (set RETRAIN=True to force retraining)\")\n",
" print(\"=\" * 70)\n",
" print(f\"\\nCheckpoint path: {CHECKPOINT_PATH}\")\n",
"\n",
" # Load checkpoint\n",
" checkpoint = torch.load(CHECKPOINT_PATH, map_location=device, weights_only=False)\n",
" generator.load_state_dict(checkpoint[\"generator_state_dict\"])\n",
" expected_sig_real = checkpoint[\"expected_sig_real\"]\n",
" training_losses = checkpoint.get(\"training_history\", [])\n",
"\n",
" print(\"Loaded generator weights from checkpoint\")\n",
" print(f\"Training history: {len(training_losses)} steps\")\n",
" if training_losses:\n",
" print(f\"Final loss: {training_losses[-1]:.4f}\")\n",
" print(f\"Best loss: {min(training_losses):.4f}\")\n",
"else:\n",
" # Train the generator\n",
" print(\"\\n\" + \"=\" * 70)\n",
" print(\"TRAINING SIG-WGAN (PAPER-EXACT IMPLEMENTATION)\")\n",
" print(\"=\" * 70)\n",
" print(\"\\nSettings from configs/STOCKS/SigWGAN.json:\")\n",
" print(f\" Total gradient steps: {CONFIG['total_steps']} (paper: 2500)\")\n",
" print(f\" Batch size: {CONFIG['batch_size']} (paper: 2000)\")\n",
" print(f\" Learning rate: {CONFIG['learning_rate']} (paper: 1e-3)\")\n",
" print(\n",
" f\" LR scheduler: StepLR(step={CONFIG['lr_scheduler_step']}, gamma={CONFIG['lr_scheduler_gamma']})\"\n",
" )\n",
" print(f\" Signature depth: {CONFIG['sig_depth']} (paper: 4)\")\n",
" print(f\" Factorial normalization: {CONFIG['normalise_sig']}\")\n",
" print(f\" Path length (n_lags): {CONFIG['n_lags']} (paper: 16)\")\n",
" print(\"\\nAugmentation pipeline:\")\n",
" print(\" Scale(2, dim=0) → AddTime → LeadLag → VisiTrans('I')\")\n",
" print(\"\\nGenerator: LSTMGenerator with Brownian motion noise\")\n",
" print(f\" Device: {device}\")\n",
"\n",
" training_losses = train_sigwgan_paper(generator, expected_sig_real, CONFIG, device)"
]
},
{
"cell_type": "code",
"execution_count": 35,
"id": "f025aaa3",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.755636Z",
"iopub.status.busy": "2026-06-14T16:58:50.755427Z",
"iopub.status.idle": "2026-06-14T16:58:50.757731Z",
"shell.execute_reply": "2026-06-14T16:58:50.757417Z"
},
"papermill": {
"duration": 0.009391,
"end_time": "2026-06-14T16:58:50.758129+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.748738+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"# Save checkpoint after training (skip if loaded from checkpoint)\n",
"if not SKIP_TRAINING:\n",
" print(\"\\nSaving checkpoint...\")\n",
" CHECKPOINT_PATH.parent.mkdir(parents=True, exist_ok=True)\n",
" checkpoint = {\n",
" \"generator_state_dict\": generator.state_dict(),\n",
" \"expected_sig_real\": expected_sig_real,\n",
" \"training_history\": training_losses,\n",
" \"config\": CONFIG,\n",
" \"train_mean\": train_mean.tolist(),\n",
" \"train_std\": train_std.tolist(),\n",
" }\n",
" torch.save(checkpoint, CHECKPOINT_PATH)\n",
" print(f\"Checkpoint saved to: {CHECKPOINT_PATH}\")"
]
},
{
"cell_type": "markdown",
"id": "9d347344",
"metadata": {
"papermill": {
"duration": 0.005693,
"end_time": "2026-06-14T16:58:50.769552+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.763859+00:00",
"status": "completed"
}
},
"source": [
"## 8. Training Progress Visualization"
]
},
{
"cell_type": "code",
"execution_count": 36,
"id": "006624d9",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:50.781598Z",
"iopub.status.busy": "2026-06-14T16:58:50.781467Z",
"iopub.status.idle": "2026-06-14T16:58:51.721196Z",
"shell.execute_reply": "2026-06-14T16:58:51.720781Z"
},
"papermill": {
"duration": 0.947334,
"end_time": "2026-06-14T16:58:51.722517+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:50.775183+00:00",
"status": "completed"
}
},
"outputs": [
{
"data": {
"application/vnd.plotly.v1+json": {
"config": {
"plotlyServerURL": "https://plot.ly"
},
"data": [
{
"line": {
"color": "#0a1628"
},
"mode": "lines",
"name": "Sig-W1 Loss",
"type": "scatter",
"y": [
0.437832236289978,
0.8962365984916687,
0.4821290671825409,
0.5870509147644043,
0.6769847869873047,
0.5935327410697937,
0.4681537449359894,
0.44864657521247864,
0.5219295620918274,
0.5511184334754944,
0.5211339592933655,
0.4653891921043396,
0.4369567334651947,
0.455264687538147,
0.48342204093933105,
0.4916815757751465,
0.47478678822517395,
0.4481937289237976,
0.43678921461105347,
0.4452924430370331,
0.45964035391807556,
0.46458008885383606,
0.4566849172115326,
0.4444758892059326,
0.4365823268890381,
0.43925732374191284,
0.4468919038772583,
0.45319393277168274,
0.4468694031238556,
0.44069185853004456,
0.4361659288406372,
0.4367850422859192,
0.43956252932548523,
0.44174715876579285,
0.44118624925613403,
0.43879926204681396,
0.43442627787590027,
0.43441152572631836,
0.4355638325214386,
0.4356424808502197,
0.43558862805366516,
0.43373945355415344,
0.43021097779273987,
0.4300106465816498,
0.43146786093711853,
0.4283180832862854,
0.4266139566898346,
0.42423924803733826,
0.4254913926124573,
0.4236954152584076,
0.4214121997356415,
0.42501121759414673,
0.42063239216804504,
0.43368101119995117,
0.4184776246547699,
0.4233109951019287,
0.4205649197101593,
0.4265788197517395,
0.4140113890171051,
0.421701043844223,
0.41951024532318115,
0.4109867811203003,
0.40996983647346497,
0.40914154052734375,
0.4103967845439911,
0.40674111247062683,
0.4147895574569702,
0.4094524383544922,
0.40186581015586853,
0.40174761414527893,
0.39847761392593384,
0.3944268822669983,
0.3958699107170105,
0.393818199634552,
0.3869667053222656,
0.3972095251083374,
0.3763138949871063,
0.3762418031692505,
0.3664822280406952,
0.35766950249671936,
0.3505052328109741,
0.3434787690639496,
0.3378547728061676,
0.3267984986305237,
0.32184115052223206,
0.30745136737823486,
0.30109652876853943,
0.30219435691833496,
0.29069650173187256,
0.27920445799827576,
0.2616289556026459,
0.25915729999542236,
0.25352388620376587,
0.23779228329658508,
0.21898670494556427,
0.2239319533109665,
0.2015332579612732,
0.18858680129051208,
0.21512769162654877,
0.1565280556678772,
0.1554562747478485,
0.14323461055755615,
0.13804440200328827,
0.15673640370368958,
0.11765323579311371,
0.14065806567668915,
0.097946397960186,
0.10510856658220291,
0.1281883716583252,
0.09194289147853851,
0.11078361421823502,
0.09598270058631897,
0.10239788889884949,
0.1468258649110794,
0.1037612035870552,
0.08028257638216019,
0.08120130002498627,
0.11657073348760605,
0.08581690490245819,
0.1012314036488533,
0.11014903336763382,
0.10355764627456665,
0.08819052577018738,
0.10414668172597885,
0.062253691256046295,
0.11016334593296051,
0.07610439509153366,
0.11080986261367798,
0.09714754670858383,
0.09341773390769958,
0.08219220489263535,
0.09481977671384811,
0.0898096114397049,
0.07840403914451599,
0.08277668803930283,
0.08299069106578827,
0.07421600073575974,
0.10265131294727325,
0.08828774094581604,
0.0865086019039154,
0.09570814669132233,
0.06227052956819534,
0.09174319356679916,
0.07717062532901764,
0.07788776606321335,
0.07200467586517334,
0.08008322864770889,
0.06232469528913498,
0.07837748527526855,
0.06621423363685608,
0.07051864266395569,
0.05919961631298065,
0.1025613322854042,
0.07366140186786652,
0.06451558321714401,
0.06842593103647232,
0.0850491151213646,
0.07690058648586273,
0.06928244978189468,
0.06242000684142113,
0.0645078793168068,
0.06089935451745987,
0.0786331370472908,
0.058234766125679016,
0.06942635029554367,
0.06938242167234421,
0.060976285487413406,
0.07492289692163467,
0.06647646427154541,
0.06762121617794037,
0.06395582854747772,
0.06551235169172287,
0.05684678629040718,
0.061220504343509674,
0.06424698233604431,
0.0635797530412674,
0.06177527457475662,
0.061556316912174225,
0.05983782187104225,
0.08400193601846695,
0.0695645809173584,
0.05884791538119316,
0.06271377950906754,
0.07869835197925568,
0.057804401963949203,
0.07144881039857864,
0.06794853508472443,
0.06921331584453583,
0.08370111882686615,
0.0696977898478508,
0.05908718705177307,
0.06679128110408783,
0.08296730369329453,
0.05947145074605942,
0.07773926109075546,
0.06736842542886734,
0.07182032614946365,
0.0617709644138813,
0.06690981239080429,
0.060209810733795166,
0.07683904469013214,
0.06476429849863052,
0.0669490247964859,
0.06450147181749344,
0.08008066564798355,
0.06973280012607574,
0.07822203636169434,
0.07550856471061707,
0.06529189646244049,
0.08870775252580643,
0.07384025305509567,
0.0746915265917778,
0.07794007658958435,
0.06037232652306557,
0.06991751492023468,
0.06150133162736893,
0.06361604481935501,
0.05896862596273422,
0.07117467373609543,
0.07563824206590652,
0.06157906353473663,
0.0661199688911438,
0.07435552775859833,
0.06627994030714035,
0.07438719272613525,
0.06805682182312012,
0.0617876797914505,
0.05830467864871025,
0.06350552290678024,
0.07404232025146484,
0.06953263282775879,
0.07087798416614532,
0.08478901535272598,
0.07991771399974823,
0.06336842477321625,
0.061176031827926636,
0.058059919625520706,
0.06903108954429626,
0.05986545607447624,
0.05713120847940445,
0.06724543869495392,
0.05919589102268219,
0.05849592015147209,
0.05794982612133026,
0.07592393457889557,
0.06047304719686508,
0.07485372573137283,
0.0609610415995121,
0.06048998981714249,
0.07618183642625809,
0.07076998800039291,
0.0598403662443161,
0.06206756830215454,
0.05978131666779518,
0.06355717033147812,
0.05739562585949898,
0.06655241549015045,
0.07390301674604416,
0.062414079904556274,
0.05859082192182541,
0.08403397351503372,
0.08069455623626709,
0.08656850457191467,
0.07586228847503662,
0.06343124061822891,
0.08161251991987228,
0.07006805390119553,
0.0789855346083641,
0.057542167603969574,
0.06627796590328217,
0.07944993674755096,
0.06732218712568283,
0.07947457581758499,
0.06171618402004242,
0.07030973583459854,
0.06720597296953201,
0.07972268015146255,
0.058880459517240524,
0.06237548217177391,
0.07041935622692108,
0.07547149807214737,
0.05627689138054848,
0.08086735755205154,
0.05819658190011978,
0.06606005132198334,
0.06062295287847519,
0.06802528351545334,
0.062450896948575974,
0.08032263070344925,
0.08275106549263,
0.055713336914777756,
0.06405936181545258,
0.06545405834913254,
0.05954550951719284,
0.07075654715299606,
0.05555669590830803,
0.059323906898498535,
0.07794544845819473,
0.06404975056648254,
0.07492072135210037,
0.056851379573345184,
0.057658854871988297,
0.07524659484624863,
0.062381256371736526,
0.06049196794629097,
0.07029504328966141,
0.06941389292478561,
0.05897950008511543,
0.061814919114112854,
0.08139289170503616,
0.08209892362356186,
0.06220695748925209,
0.05888303369283676,
0.06872887909412384,
0.060584984719753265,
0.05817389488220215,
0.05732070654630661,
0.06345631182193756,
0.061594463884830475,
0.06522916257381439,
0.06521397083997726,
0.05883292853832245,
0.06145549938082695,
0.06936143338680267,
0.0595025010406971,
0.05924932286143303,
0.07535907626152039,
0.06460767984390259,
0.05845043435692787,
0.06268292665481567,
0.0633445531129837,
0.056879423558712006,
0.06564629822969437,
0.06133414804935455,
0.058736298233270645,
0.06774023175239563,
0.06074047461152077,
0.06054971367120743,
0.07107219845056534,
0.07021693140268326,
0.06589995324611664,
0.05795795097947121,
0.061633989214897156,
0.11395511031150818,
0.06965883821249008,
0.0635368674993515,
0.06980235129594803,
0.09997952729463577,
0.08566690236330032,
0.062393996864557266,
0.07236282527446747,
0.07021228224039078,
0.08790194988250732,
0.05923822522163391,
0.056467942893505096,
0.05769193544983864,
0.0563354566693306,
0.06506264209747314,
0.05614475533366203,
0.059152111411094666,
0.05773762986063957,
0.07754523307085037,
0.06104866787791252,
0.06290482729673386,
0.06587358564138412,
0.07077935338020325,
0.08674142509698868,
0.06637464463710785,
0.06953844428062439,
0.0657811313867569,
0.06729835271835327,
0.05816062167286873,
0.0637681856751442,
0.07516482472419739,
0.0766066461801529,
0.09246960282325745,
0.08660726249217987,
0.06732770055532455,
0.06353782117366791,
0.07424642145633698,
0.07099068909883499,
0.07966919243335724,
0.07555398344993591,
0.06152556464076042,
0.06192314624786377,
0.06967893242835999,
0.0629807859659195,
0.08388500660657883,
0.062246084213256836,
0.0581635981798172,
0.07460502535104752,
0.06392531096935272,
0.060261070728302,
0.059902530163526535,
0.05942017585039139,
0.06797686219215393,
0.06037429720163345,
0.05835525318980217,
0.060071323066949844,
0.0600581131875515,
0.05767878517508507,
0.06246443837881088,
0.05585138499736786,
0.06274287402629852,
0.058936890214681625,
0.05695313215255737,
0.056509215384721756,
0.06554771214723587,
0.0571114756166935,
0.05398167669773102,
0.06060997396707535,
0.07086219638586044,
0.07230839878320694,
0.05901104584336281,
0.060609519481658936,
0.0663817971944809,
0.07241939008235931,
0.0550057627260685,
0.05849001184105873,
0.07011236995458603,
0.06451631337404251,
0.05578864365816116,
0.06287357211112976,
0.06631006300449371,
0.05980433151125908,
0.05986763909459114,
0.060433849692344666,
0.05609939992427826,
0.06608432531356812,
0.06432729214429855,
0.07644631713628769,
0.06803137809038162,
0.06733982264995575,
0.0606103390455246,
0.06824561953544617,
0.06106226518750191,
0.0628097876906395,
0.06552748382091522,
0.06295516341924667,
0.07253248989582062,
0.060151685029268265,
0.0688948929309845,
0.07367383688688278,
0.06937448680400848,
0.0606064572930336,
0.06052098423242569,
0.06833566725254059,
0.07229563593864441,
0.06551839411258698,
0.0708916112780571,
0.06172587350010872,
0.07720992714166641,
0.06909119337797165,
0.0628032386302948,
0.06073031574487686,
0.05959305167198181,
0.06764543801546097,
0.062446229159832,
0.058878540992736816,
0.06015873700380325,
0.06415887922048569,
0.06354953348636627,
0.06012840196490288,
0.06968380510807037,
0.06609894335269928,
0.06092086806893349,
0.06428232789039612,
0.07764935493469238,
0.05824434384703636,
0.060407303273677826,
0.06322674453258514,
0.06864453107118607,
0.05956840515136719,
0.07216387242078781,
0.05784676969051361,
0.07216362655162811,
0.057821124792099,
0.06990185379981995,
0.058840278536081314,
0.057678885757923126,
0.06668280065059662,
0.07027623802423477,
0.08424356579780579,
0.07778072357177734,
0.09057828783988953,
0.07024085521697998,
0.06251756846904755,
0.055912334471940994,
0.060846660286188126,
0.06184736266732216,
0.06363782286643982,
0.0660509541630745,
0.05929873511195183,
0.06126939505338669,
0.09175089746713638,
0.07840190082788467,
0.08481904119253159,
0.08127772063016891,
0.06992270797491074,
0.06802546977996826,
0.07432565838098526,
0.05619509890675545,
0.0660247877240181,
0.08449965715408325,
0.0676693543791771,
0.07567520439624786,
0.08484958857297897,
0.08802550286054611,
0.06946028023958206,
0.08019877970218658,
0.06041237339377403,
0.06336231529712677,
0.06867711246013641,
0.08096809685230255,
0.07070107012987137,
0.06886991113424301,
0.06455029547214508,
0.07049310207366943,
0.06290407478809357,
0.06802473217248917,
0.06628604978322983,
0.05850031599402428,
0.058276478201150894,
0.057813894003629684,
0.0698937475681305,
0.06964606046676636,
0.06334629654884338,
0.06629534810781479,
0.06171725317835808,
0.07470228523015976,
0.08139543980360031,
0.06481476128101349,
0.05804816633462906,
0.058621156960725784,
0.06611902266740799,
0.05919837951660156,
0.05628272145986557,
0.07182192802429199,
0.06100941449403763,
0.062434058636426926,
0.06455160677433014,
0.0959658995270729,
0.0575711615383625,
0.07602665573358536,
0.08419487625360489,
0.07657523453235626,
0.058528508991003036,
0.08227460831403732,
0.07220108807086945,
0.07370401918888092,
0.06463856250047684,
0.05814911425113678,
0.05614234879612923,
0.08414217829704285,
0.07902425527572632,
0.06850364804267883,
0.06227171793580055,
0.05822494626045227,
0.05787653848528862,
0.06360818445682526,
0.10342076420783997,
0.07576505839824677,
0.0785997062921524,
0.08033862709999084,
0.0639340877532959,
0.06860605627298355,
0.09128958731889725,
0.06895945221185684,
0.07953537255525589,
0.06427919119596481,
0.05927848443388939,
0.06246394291520119,
0.060235898941755295,
0.07005278021097183,
0.07165227085351944,
0.06293335556983948,
0.059123728424310684,
0.06277191638946533,
0.06262434273958206,
0.06653480231761932,
0.085601307451725,
0.05868114158511162,
0.08210989832878113,
0.07328189164400101,
0.07499444484710693,
0.06098286807537079,
0.07441937178373337,
0.09712281823158264,
0.06956121325492859,
0.05939982831478119,
0.07345566898584366,
0.06204703822731972,
0.07234145700931549,
0.0751131996512413,
0.05552535504102707,
0.06063816696405411,
0.055633656680583954,
0.06171436980366707,
0.05772789567708969,
0.07144394516944885,
0.08385664969682693,
0.07363162189722061,
0.06039750203490257,
0.08098942041397095,
0.0619812048971653,
0.07590896636247635,
0.06814315170049667,
0.06012757122516632,
0.05908045172691345,
0.06204090267419815,
0.07650851458311081,
0.062314216047525406,
0.06040022894740105,
0.06027308851480484,
0.06393547356128693,
0.07393420487642288,
0.07581938058137894,
0.07067691534757614,
0.0592774972319603,
0.06654839217662811,
0.059278108179569244,
0.06610003858804703,
0.07163883000612259,
0.06023164838552475,
0.05907067656517029,
0.05898650363087654,
0.0646602213382721,
0.06567580252885818,
0.07484801858663559,
0.07742781937122345,
0.06557991355657578,
0.06220630928874016,
0.08430440723896027,
0.06338748335838318,
0.07250598818063736,
0.06264365464448929,
0.06887228041887283,
0.062148597091436386,
0.06039414182305336,
0.07578092068433762,
0.061190005391836166,
0.08011554181575775,
0.059144724160432816,
0.05937863886356354,
0.06935480237007141,
0.060260336846113205,
0.05669182166457176,
0.06297176331281662,
0.06489711999893188,
0.07092999666929245,
0.061066195368766785,
0.07167430967092514,
0.06012486293911934,
0.0826382115483284,
0.0598379522562027,
0.0637255609035492,
0.07193543761968613,
0.06756044924259186,
0.06348297744989395,
0.08842987567186356,
0.06487695127725601,
0.05839531496167183,
0.06512816250324249,
0.07374314218759537,
0.0804043859243393,
0.0599287785589695,
0.07000208646059036,
0.06809181720018387,
0.07635471224784851,
0.07317611575126648,
0.07944032549858093,
0.07776639610528946,
0.068963423371315,
0.08690851926803589,
0.06551789492368698,
0.06104905158281326,
0.06082293391227722,
0.06984708458185196,
0.05799757316708565,
0.06787490844726562,
0.0707324966788292,
0.0663188248872757,
0.06322538107633591,
0.06676321476697922,
0.06272528320550919,
0.05602695792913437,
0.05785926431417465,
0.059633031487464905,
0.06693246960639954,
0.0600951611995697,
0.09319966286420822,
0.06580182909965515,
0.061208486557006836,
0.05775294452905655,
0.06890569627285004,
0.06221655383706093,
0.06772686541080475,
0.08874434232711792,
0.061352457851171494,
0.05732254683971405,
0.061071280390024185,
0.06542027741670609,
0.06202518939971924,
0.059372276067733765,
0.0569482296705246,
0.07681290060281754,
0.05928587168455124,
0.06063772365450859,
0.08300842344760895,
0.07467453181743622,
0.07968322932720184,
0.0661931037902832,
0.06848163902759552,
0.05960891395807266,
0.06623706966638565,
0.06762393563985825,
0.06777694076299667,
0.05534960329532623,
0.057035814970731735,
0.07575216889381409,
0.057562973350286484,
0.05716409906744957,
0.06038404628634453,
0.060216739773750305,
0.06179792061448097,
0.07668726146221161,
0.056275974959135056,
0.0690552145242691,
0.05875371769070625,
0.057152390480041504,
0.07826562970876694,
0.08345485478639603,
0.05871588736772537,
0.07123308628797531,
0.05851433798670769,
0.07481137663125992,
0.05766959860920906,
0.061494022607803345,
0.06203139200806618,
0.06915109604597092,
0.08493730425834656,
0.0581301711499691,
0.05986456200480461,
0.061301253736019135,
0.06195393577218056,
0.062009070068597794,
0.058398108929395676,
0.05716949701309204,
0.07339459657669067,
0.05702946335077286,
0.0594412162899971,
0.059190068393945694,
0.07713770866394043,
0.05876510217785835,
0.05667910352349281,
0.0792238637804985,
0.06329076737165451,
0.059095606207847595,
0.06161458417773247,
0.06637627631425858,
0.06391884386539459,
0.07141894847154617,
0.06387600302696228,
0.06623221933841705,
0.07708217948675156,
0.08204412460327148,
0.0723966434597969,
0.0737038105726242,
0.0598476342856884,
0.058690302073955536,
0.06428730487823486,
0.06280449777841568,
0.06354393810033798,
0.055703625082969666,
0.061799097806215286,
0.05955120921134949,
0.058719635009765625,
0.06947487592697144,
0.06328768283128738,
0.06636861711740494,
0.06100135296583176,
0.06512518227100372,
0.08096789568662643,
0.059846218675374985,
0.06828982383012772,
0.07949142158031464,
0.07455021888017654,
0.05842357501387596,
0.058922164142131805,
0.05699598044157028,
0.05683247372508049,
0.059093039482831955,
0.0633532702922821,
0.05810466781258583,
0.0598379522562027,
0.057799749076366425,
0.06096349284052849,
0.06030259281396866,
0.060571178793907166,
0.06751982122659683,
0.05945105105638504,
0.09046332538127899,
0.06864739209413528,
0.05708928778767586,
0.059357814490795135,
0.06284589320421219,
0.06022144481539726,
0.06081819161772728,
0.05844118073582649,
0.05670944228768349,
0.08203212171792984,
0.06341119110584259,
0.07000736147165298,
0.06572305411100388,
0.06021227315068245,
0.056216005235910416,
0.06788460165262222,
0.07426583021879196,
0.05933179333806038,
0.06320211291313171,
0.058769259601831436,
0.07395214587450027,
0.07192866504192352,
0.0652981549501419,
0.06393091380596161,
0.073581263422966,
0.06472543627023697,
0.0844288170337677,
0.06170310080051422,
0.05839213356375694,
0.06976484507322311,
0.06546185910701752,
0.089297816157341,
0.05781356990337372,
0.0916052907705307,
0.0664740726351738,
0.07167693972587585,
0.058932892978191376,
0.07308465987443924,
0.05875576287508011,
0.08940223604440689,
0.08661504834890366,
0.06410243362188339,
0.10288198292255402,
0.06414221972227097,
0.07796776294708252,
0.06151888519525528,
0.06200270727276802,
0.05706306919455528,
0.07246015220880508,
0.06344262510538101,
0.06232461333274841,
0.06498843431472778,
0.07072184234857559,
0.05821896716952324,
0.06902004778385162,
0.07375402003526688,
0.07243921607732773,
0.059112608432769775,
0.057406697422266006,
0.06156216189265251,
0.06448404490947723,
0.07561000436544418,
0.05918160080909729,
0.061171818524599075,
0.059880711138248444,
0.07513880729675293,
0.07329072803258896,
0.0662933737039566,
0.09262578934431076,
0.05887468904256821,
0.06275869160890579,
0.06806938350200653,
0.05998268350958824,
0.06293079257011414,
0.06832589209079742,
0.06168711185455322,
0.07754795253276825,
0.06802890449762344,
0.05991340056061745,
0.06161804497241974,
0.07337391376495361,
0.06278050690889359,
0.07964591681957245,
0.05990666523575783,
0.07520708441734314,
0.0583016574382782,
0.07012289762496948,
0.05685046687722206,
0.07691221684217453,
0.05858481302857399,
0.08140182495117188,
0.0821366235613823,
0.06876866519451141,
0.062417272478342056,
0.0680936723947525,
0.05769994482398033,
0.07330328226089478,
0.05432485044002533,
0.06368598341941833,
0.05837409198284149,
0.06067393720149994,
0.06811361759901047,
0.07231462746858597,
0.06657006591558456,
0.059851232916116714,
0.061626676470041275,
0.06841515749692917,
0.09088427573442459,
0.07363425195217133,
0.07111933082342148,
0.06112469732761383,
0.07147762924432755,
0.06896978616714478,
0.07624974846839905,
0.062483906745910645,
0.08142697066068649,
0.061630476266145706,
0.07759924232959747,
0.06549807637929916,
0.08243977278470993,
0.07082604616880417,
0.0700119212269783,
0.07532168924808502,
0.06411338597536087,
0.06335662305355072,
0.07207271456718445,
0.08719411492347717,
0.07745660096406937,
0.06893163919448853,
0.05839356407523155,
0.06785739213228226,
0.061380039900541306,
0.06651460379362106,
0.06197100877761841,
0.06591523438692093,
0.059371329843997955,
0.058772701770067215,
0.0737895667552948,
0.05561641976237297,
0.06596948206424713,
0.06570024788379669,
0.05512300133705139,
0.057111240923404694,
0.057342298328876495,
0.055860526859760284,
0.07079173624515533,
0.07213633507490158,
0.05645810812711716,
0.05809115245938301,
0.05672023817896843,
0.06690440326929092,
0.05921070650219917,
0.05775982886552811,
0.061317987740039825,
0.05770082771778107,
0.05951511859893799,
0.0628383606672287,
0.05715440958738327,
0.06385277211666107,
0.06182459369301796,
0.058164745569229126,
0.05782179906964302,
0.10729732364416122,
0.06843145936727524,
0.06233211234211922,
0.08769728988409042,
0.06129436194896698,
0.08315455913543701,
0.05932497978210449,
0.05970948189496994,
0.061967670917510986,
0.06726232916116714,
0.06048867106437683,
0.07432378083467484,
0.06124400720000267,
0.05749020725488663,
0.07448694854974747,
0.06370271742343903,
0.06272789090871811,
0.05897647887468338,
0.06203538551926613,
0.059710294008255005,
0.058871518820524216,
0.055867891758680344,
0.07078123837709427,
0.07754696905612946,
0.06652575731277466,
0.06630755215883255,
0.08630232512950897,
0.06213290989398956,
0.06679277122020721,
0.07375526428222656,
0.06638406962156296,
0.06404333561658859,
0.05723235011100769,
0.07129605859518051,
0.062131043523550034,
0.059488695114851,
0.0598134808242321,
0.0629759132862091,
0.05641498416662216,
0.06667695194482803,
0.08120688050985336,
0.06824934482574463,
0.07483504712581635,
0.05751161277294159,
0.07850846648216248,
0.061849236488342285,
0.05911802127957344,
0.05797918140888214,
0.06218530237674713,
0.05906757712364197,
0.0628775805234909,
0.058759141713380814,
0.061808325350284576,
0.063055120408535,
0.05889759585261345,
0.06350741535425186,
0.06162913516163826,
0.05714832618832588,
0.07032599300146103,
0.058073535561561584,
0.061070095747709274,
0.0564495287835598,
0.06335549801588058,
0.05789443850517273,
0.08033383637666702,
0.06534034013748169,
0.06279150396585464,
0.05815257504582405,
0.060852762311697006,
0.0674903392791748,
0.06841767579317093,
0.07328929007053375,
0.06213989108800888,
0.06178078427910805,
0.05763719230890274,
0.058201778680086136,
0.056405775249004364,
0.056089770048856735,
0.05660499259829521,
0.05616096407175064,
0.05781461298465729,
0.05740443617105484,
0.059251587837934494,
0.07829823344945908,
0.059116460382938385,
0.05622539669275284,
0.06109243631362915,
0.06411817669868469,
0.06880834698677063,
0.07424599677324295,
0.06357090175151825,
0.11953315883874893,
0.059298835694789886,
0.06226976588368416,
0.06389924883842468,
0.059519048780202866,
0.06559472531080246,
0.06195956841111183,
0.061612699180841446,
0.05666176974773407,
0.05951736867427826,
0.08120126277208328,
0.06072587892413139,
0.06264124065637589,
0.06820331513881683,
0.05656053125858307,
0.05844602361321449,
0.06385188549757004,
0.06337455660104752,
0.05910288915038109,
0.061254117637872696,
0.055579423904418945,
0.06260528415441513,
0.08026356995105743,
0.08460286259651184,
0.059490371495485306,
0.05662756413221359,
0.06053265184164047,
0.05744277313351631,
0.06454108655452728,
0.061128243803977966,
0.07377778738737106,
0.06039074435830116,
0.06704956293106079,
0.06405342370271683,
0.07138118147850037,
0.059892937541007996,
0.06908324360847473,
0.05804835259914398,
0.06332975625991821,
0.06045527383685112,
0.06145423650741577,
0.06916333734989166,
0.06068763509392738,
0.07243027538061142,
0.060972489416599274,
0.05861353501677513,
0.06363414973020554,
0.059973426163196564,
0.06787928193807602,
0.05967271700501442,
0.05817728117108345,
0.06946825236082077,
0.06551321595907211,
0.0596977174282074,
0.06499185413122177,
0.06473325937986374,
0.07150053977966309,
0.059993427246809006,
0.059577032923698425,
0.0633438229560852,
0.058757368475198746,
0.059207186102867126,
0.06342311948537827,
0.08378607779741287,
0.0607624277472496,
0.0665815994143486,
0.06648310273885727,
0.07121109962463379,
0.06936542689800262,
0.06068916991353035,
0.06026707962155342,
0.05741250142455101,
0.055831603705883026,
0.07864847034215927,
0.06053852289915085,
0.057751335203647614,
0.05753566324710846,
0.056782640516757965,
0.06454668194055557,
0.06372480094432831,
0.06994949281215668,
0.06739617884159088,
0.06346980482339859,
0.06350810825824738,
0.06387195736169815,
0.05701638013124466,
0.05664225295186043,
0.06285717338323593,
0.06139212101697922,
0.05600503459572792,
0.06083469092845917,
0.0669136717915535,
0.06336604058742523,
0.07416634261608124,
0.05848151445388794,
0.05695964768528938,
0.06766722351312637,
0.06222335249185562,
0.06030607968568802,
0.0582917183637619,
0.06232573464512825,
0.057729125022888184,
0.056405890733003616,
0.06353073567152023,
0.07654081284999847,
0.06360853463411331,
0.05862804129719734,
0.06096538528800011,
0.05726458877325058,
0.059343934059143066,
0.0663246363401413,
0.05776461958885193,
0.06381092220544815,
0.05807046964764595,
0.06164790317416191,
0.057197704911231995,
0.06957287341356277,
0.06777171790599823,
0.06453131139278412,
0.06353569030761719,
0.08867461234331131,
0.058561038225889206,
0.06422510743141174,
0.07099202275276184,
0.057550933212041855,
0.07184688746929169,
0.06010831892490387,
0.05793619900941849,
0.05964534729719162,
0.07141504436731339,
0.061235520988702774,
0.05502285808324814,
0.05755080282688141,
0.05844425782561302,
0.06426119804382324,
0.06367553025484085,
0.055734939873218536,
0.06956849247217178,
0.07119426876306534,
0.06029297411441803,
0.06642650812864304,
0.06080073490738869,
0.05968576669692993,
0.06399738043546677,
0.07649010419845581,
0.06444092839956284,
0.06036689877510071,
0.07358455657958984,
0.0587586835026741,
0.0657922551035881,
0.07592718303203583,
0.05889536812901497,
0.06347689777612686,
0.055905673652887344,
0.08014988899230957,
0.08030232042074203,
0.057251814752817154,
0.06064005196094513,
0.05965602025389671,
0.060980819165706635,
0.05907832458615303,
0.05624833330512047,
0.07010401040315628,
0.058218490332365036,
0.05567796528339386,
0.060678113251924515,
0.058563098311424255,
0.06594698876142502,
0.06138532981276512,
0.10026749968528748,
0.06324837356805801,
0.060616713017225266,
0.0689089447259903,
0.0590314120054245,
0.08774620294570923,
0.0606573112308979,
0.0603896789252758,
0.07183530926704407,
0.06335204839706421,
0.060091160237789154,
0.05691837891936302,
0.07423362135887146,
0.07174155116081238,
0.0571679063141346,
0.05689491704106331,
0.05948253720998764,
0.05948112532496452,
0.0642954632639885,
0.06365139037370682,
0.06026681512594223,
0.057800792157649994,
0.062145985662937164,
0.05656718090176582,
0.05829911306500435,
0.06190628185868263,
0.06863883882761002,
0.05729065462946892,
0.05820280686020851,
0.07709836214780807,
0.05770768225193024,
0.05876152589917183,
0.058212630450725555,
0.05733190104365349,
0.07540430873632431,
0.06665773689746857,
0.06945761293172836,
0.05537303909659386,
0.07670070230960846,
0.06985140591859818,
0.0585063099861145,
0.08357882499694824,
0.06105436757206917,
0.07447946071624756,
0.06329622864723206,
0.09060648828744888,
0.06985368579626083,
0.062241267412900925,
0.06970168650150299,
0.0692460760474205,
0.08877084404230118,
0.06332218647003174,
0.0688018947839737,
0.05924199894070625,
0.077580526471138,
0.07443837821483612,
0.061378113925457,
0.05870237201452255,
0.05980649217963219,
0.06789368391036987,
0.06772404164075851,
0.05900375917553902,
0.057068370282649994,
0.060027025640010834,
0.06159627437591553,
0.06232204660773277,
0.06732948124408722,
0.0729713961482048,
0.06794315576553345,
0.06981537491083145,
0.05793016403913498,
0.07031544297933578,
0.06551490724086761,
0.0766129344701767,
0.06557658314704895,
0.08888103812932968,
0.06489712744951248,
0.07241253554821014,
0.061966001987457275,
0.05964796990156174,
0.06062939763069153,
0.07305625826120377,
0.06490505486726761,
0.06577173620462418,
0.061815597116947174,
0.05880681052803993,
0.08796054124832153,
0.06195143610239029,
0.05862331762909889,
0.06261483579874039,
0.06810659915208817,
0.0649336650967598,
0.05932345613837242,
0.06762661784887314,
0.07585637271404266,
0.05916460230946541,
0.06477835029363632,
0.06858276575803757,
0.056867897510528564,
0.0603468157351017,
0.08221316337585449,
0.0601692795753479,
0.060433026403188705,
0.07078666239976883,
0.058480650186538696,
0.06206759065389633,
0.05956202372908592,
0.06659301370382309,
0.060410015285015106,
0.06538624316453934,
0.06734685599803925,
0.06869835406541824,
0.058533769100904465,
0.0641859620809555,
0.06316844373941422,
0.05601366609334946,
0.0698704868555069,
0.05870702117681503,
0.06464505940675735,
0.06427322328090668,
0.06481506675481796,
0.06110440194606781,
0.060084495693445206,
0.05835452675819397,
0.07592835277318954,
0.06864926964044571,
0.056893490254879,
0.06606585532426834,
0.06876634061336517,
0.06027957424521446,
0.06552382558584213,
0.06422533094882965,
0.05470859259366989,
0.07330994307994843,
0.055685654282569885,
0.058811549097299576,
0.057100821286439896,
0.061222661286592484,
0.05765119194984436,
0.06897583603858948,
0.05716584622859955,
0.06015670672059059,
0.06812373548746109,
0.05649334192276001,
0.06921061128377914,
0.06572431325912476,
0.05653199180960655,
0.05911150947213173,
0.07662328332662582,
0.055046629160642624,
0.05887928977608681,
0.061128776520490646,
0.06795001775026321,
0.08186126500368118,
0.06880809366703033,
0.061966776847839355,
0.06760966032743454,
0.06627614796161652,
0.06268381327390671,
0.06574629247188568,
0.06620992720127106,
0.07928936928510666,
0.06065228581428528,
0.05790950357913971,
0.06478565186262131,
0.0607563816010952,
0.0844525694847107,
0.0587107390165329,
0.08608021587133408,
0.0618438795208931,
0.06395396590232849,
0.07978548109531403,
0.05833549425005913,
0.0882612094283104,
0.06675764173269272,
0.06777303665876389,
0.08315547555685043,
0.0602383017539978,
0.06603708118200302,
0.07139597833156586,
0.062188927084207535,
0.07763306051492691,
0.06585214287042618,
0.07000406086444855,
0.05849943682551384,
0.06229104474186897,
0.059811633080244064,
0.05614129826426506,
0.06600773334503174,
0.05779251083731651,
0.0597110316157341,
0.061622679233551025,
0.05603473633527756,
0.056652314960956573,
0.0589173287153244,
0.05575447902083397,
0.06274167448282242,
0.05998487398028374,
0.06641416996717453,
0.056985124945640564,
0.06047734245657921,
0.056761711835861206,
0.06252418458461761,
0.05611371621489525,
0.06367643922567368,
0.06408941000699997,
0.060909856110811234,
0.05724536255002022,
0.05792533606290817,
0.07148800045251846,
0.056464701890945435,
0.08174218237400055,
0.05920235812664032,
0.060274817049503326,
0.057280588895082474,
0.058020565658807755,
0.057536691427230835,
0.05974484980106354,
0.05579910799860954,
0.08243515342473984,
0.06392909586429596,
0.057439401745796204,
0.0718333050608635,
0.056106049567461014,
0.060391899198293686,
0.06373430788516998,
0.06564545631408691,
0.06882417947053909,
0.06485944986343384,
0.0637911930680275,
0.058602944016456604,
0.06357134878635406,
0.06265124678611755,
0.060742758214473724,
0.06278187781572342,
0.062366120517253876,
0.06388279050588608,
0.06757285445928574,
0.07270358502864838,
0.07284651696681976,
0.0739532932639122,
0.0830734595656395,
0.06905874609947205,
0.07426995784044266,
0.07057613134384155,
0.05771636962890625,
0.059503763914108276,
0.056669458746910095,
0.08025676757097244,
0.056762516498565674,
0.06462860107421875,
0.07960512489080429,
0.07176311314105988,
0.05757337063550949,
0.06980125606060028,
0.06405583024024963,
0.05607187747955322,
0.06020018085837364,
0.07520496845245361,
0.060699719935655594,
0.06528694927692413,
0.057686932384967804,
0.08126845210790634,
0.07960201054811478,
0.07816626876592636,
0.05979442223906517,
0.06482920050621033,
0.06878339499235153,
0.07053259015083313,
0.05760347843170166,
0.05972076579928398,
0.05953334644436836,
0.05663483589887619,
0.05724947899580002,
0.0588894747197628,
0.05937005952000618,
0.05744237080216408,
0.05630819872021675,
0.058435723185539246,
0.06547816097736359,
0.05934123694896698,
0.05807151272892952,
0.07008587568998337,
0.07124530524015427,
0.06131167337298393,
0.05730034410953522,
0.061392899602651596,
0.06412868946790695,
0.05700787156820297,
0.06115531548857689,
0.05683152377605438,
0.059104058891534805,
0.058743882924318314,
0.07022470235824585,
0.06087040528655052,
0.06104867905378342,
0.06902024149894714,
0.05752410739660263,
0.0588136687874794,
0.07826944440603256,
0.07226814329624176,
0.06107697635889053,
0.0732237696647644,
0.06092633306980133,
0.06391336023807526,
0.058359161019325256,
0.07739417999982834,
0.06414002180099487,
0.0645621120929718,
0.07205046713352203,
0.07520148903131485,
0.057242006063461304,
0.06774181872606277,
0.06324174255132675,
0.08326205611228943,
0.08479594439268112,
0.06151650846004486,
0.07276402413845062,
0.05757783353328705,
0.07536569982767105,
0.0610942579805851,
0.06405643373727798,
0.07457911968231201,
0.06560676544904709,
0.07220511883497238,
0.07919272780418396,
0.07142286747694016,
0.06829412281513214,
0.0680619329214096,
0.06399902701377869,
0.06299365311861038,
0.05844544619321823,
0.07456111162900925,
0.06372451037168503,
0.05875256657600403,
0.06295489519834518,
0.06580614298582077,
0.06669024378061295,
0.07059706002473831,
0.0658557340502739,
0.05762182176113129,
0.06448626518249512,
0.07420538365840912,
0.06158186495304108,
0.06861633062362671,
0.06655073165893555,
0.05801079794764519,
0.06736969947814941,
0.06128492206335068,
0.057887639850378036,
0.06274155527353287,
0.06300321221351624,
0.06866748631000519,
0.0677344799041748,
0.05734378844499588,
0.05859893187880516,
0.08711656183004379,
0.06247682124376297,
0.058366455137729645,
0.06291761249303818,
0.06459589302539825,
0.06345174461603165,
0.07403500378131866,
0.06464932858943939,
0.06546933203935623,
0.06064514443278313,
0.0706913098692894,
0.07192075252532959,
0.0701521709561348,
0.058430880308151245,
0.06715824455022812,
0.0658407062292099,
0.0611584447324276,
0.06283833086490631,
0.05852697789669037,
0.05679520219564438,
0.05879354476928711,
0.05796557664871216,
0.059244558215141296,
0.06350667029619217,
0.06142726168036461,
0.05641688033938408,
0.08925821632146835,
0.07442236691713333,
0.06337419152259827,
0.05803447216749191,
0.06097307428717613,
0.07359246909618378,
0.05769148841500282,
0.061914555728435516,
0.05697203800082207,
0.06026427820324898,
0.05897535756230354,
0.07421012222766876,
0.0633913055062294,
0.059556521475315094,
0.06523447483778,
0.05863088369369507,
0.06307121366262436,
0.06381551921367645,
0.06144099310040474,
0.06506194919347763,
0.06059112772345543,
0.08323647081851959,
0.06370292603969574,
0.06100071594119072,
0.05968464910984039,
0.07775619626045227,
0.06932523846626282,
0.0669391006231308,
0.0656697154045105,
0.06463515013456345,
0.057570166885852814,
0.06555288285017014,
0.0812022015452385,
0.059440210461616516,
0.05592459812760353,
0.06269743293523788,
0.05863290652632713,
0.061153195798397064,
0.060181159526109695,
0.05942210555076599,
0.07090552896261215,
0.06005191057920456,
0.06353950500488281,
0.06497838348150253,
0.05722671374678612,
0.06663119047880173,
0.05887838080525398,
0.057001400738954544,
0.05931796506047249,
0.061275944113731384,
0.06463660299777985,
0.06786423921585083,
0.06457483768463135,
0.06033027544617653,
0.060411639511585236,
0.059403665363788605,
0.07528996467590332,
0.056980885565280914,
0.057326678186655045,
0.06076650694012642,
0.07622896134853363,
0.05846816301345825,
0.05718320608139038,
0.06596291810274124,
0.06184028089046478,
0.059085626155138016,
0.06514144688844681,
0.06373683363199234,
0.06212160736322403,
0.06550564616918564,
0.06317748129367828,
0.06077578291296959,
0.06315594911575317,
0.05807927995920181,
0.05770833417773247,
0.06546661257743835,
0.06061359867453575,
0.05839979648590088,
0.05900469794869423,
0.06923314183950424,
0.06105522811412811,
0.06996679306030273,
0.0568247064948082,
0.077634796500206,
0.06048303097486496,
0.05987635254859924,
0.06764432787895203,
0.06296714395284653,
0.059783052653074265,
0.072914257645607,
0.0771830752491951,
0.058235086500644684,
0.06384214758872986,
0.06619685143232346,
0.06398963183164597,
0.060037486255168915,
0.07352937757968903,
0.058386657387018204,
0.05909562110900879,
0.08135686069726944,
0.06569398194551468,
0.06799900531768799,
0.0672113448381424,
0.05722682923078537,
0.06847714632749557,
0.06281189620494843,
0.06371433287858963,
0.06324677169322968,
0.0685548409819603,
0.07750118523836136,
0.06954994797706604,
0.06037984415888786,
0.058238644152879715,
0.06388363242149353,
0.06135205924510956,
0.06183440983295441,
0.05769994482398033,
0.05704418569803238,
0.06679882109165192,
0.0611739382147789,
0.06028001010417938,
0.06357614696025848,
0.06379695236682892,
0.05525320768356323,
0.07706809043884277,
0.07159658521413803,
0.06502825766801834,
0.07519721239805222,
0.060577500611543655,
0.06419560313224792,
0.06001439318060875,
0.07832616567611694,
0.0655483528971672,
0.05797438696026802,
0.059232745319604874,
0.05667591467499733,
0.05748980492353439,
0.07500548660755157,
0.08619647473096848,
0.06534840911626816,
0.06411920487880707,
0.06254556030035019,
0.05969015136361122,
0.07928501069545746,
0.06205378845334053,
0.05650786682963371,
0.0584208182990551,
0.07304830849170685,
0.06072394177317619,
0.06331820785999298,
0.07694445550441742,
0.058460693806409836,
0.05728711560368538,
0.058767396956682205,
0.05841322988271713,
0.056521378457546234,
0.059216391295194626,
0.08115322142839432,
0.061842724680900574,
0.059207454323768616,
0.06598517298698425,
0.07598137855529785,
0.06197427585721016,
0.057866860181093216,
0.061670754104852676,
0.06214696913957596,
0.06701798737049103,
0.05515339598059654,
0.0637967586517334,
0.07236655056476593,
0.07145416736602783,
0.0688251480460167,
0.06098748743534088,
0.060776788741350174,
0.057823676615953445,
0.0641704797744751,
0.05958939343690872,
0.06784689426422119,
0.06655621528625488,
0.07478360086679459,
0.0605073906481266,
0.0609530471265316,
0.0575064942240715,
0.08619136363267899,
0.0638718381524086,
0.07109677791595459,
0.06975160539150238,
0.0863998532295227,
0.06637400388717651,
0.07977721840143204,
0.061677880585193634,
0.06300540268421173,
0.060859885066747665,
0.060833659023046494,
0.06374207884073257,
0.060327447950839996,
0.05849696323275566,
0.07085788249969482,
0.0608210489153862,
0.07007230818271637,
0.06744550168514252,
0.060884155333042145,
0.09401986002922058,
0.07668048143386841,
0.060785628855228424,
0.060786329209804535,
0.06988002359867096,
0.06895063817501068,
0.057313885539770126,
0.06298820674419403,
0.06945887953042984,
0.06390134990215302,
0.0642218366265297,
0.06703659892082214,
0.07980572432279587,
0.05893761292099953,
0.06274846941232681,
0.06283698976039886,
0.06499093025922775,
0.06581245362758636,
0.06674586981534958,
0.05944749340415001,
0.06234254688024521,
0.0620596706867218,
0.05781518667936325,
0.06762173026800156,
0.06730428338050842,
0.06217896565794945,
0.07224985957145691,
0.060493096709251404,
0.06715478003025055,
0.0647277757525444,
0.07314172387123108,
0.06718700379133224,
0.06546218693256378,
0.06777281314134598,
0.057991258800029755,
0.06710764020681381,
0.06464485824108124,
0.05599307641386986,
0.06709365546703339,
0.07027929276227951,
0.056765731424093246,
0.05591728165745735,
0.0688624307513237,
0.06240088865160942,
0.05916460230946541,
0.0652926042675972,
0.08996525406837463,
0.07221236824989319,
0.055483873933553696,
0.06508228182792664,
0.0658484399318695,
0.06712695211172104,
0.06163083389401436,
0.06196114793419838,
0.060588911175727844,
0.06071050465106964,
0.056305114179849625,
0.05679723992943764,
0.05751112848520279,
0.07563158869743347,
0.05917641520500183,
0.07367242872714996,
0.09127327799797058,
0.06114397197961807,
0.06674090772867203,
0.06118585541844368,
0.05619725584983826,
0.05920475721359253,
0.058415766805410385,
0.05845160782337189,
0.06549958139657974,
0.07227012515068054,
0.0684426873922348,
0.056484464555978775,
0.06178153678774834,
0.0587494894862175,
0.06309499591588974,
0.06184149160981178,
0.05575172230601311,
0.056808896362781525,
0.05637131258845329,
0.05744941905140877,
0.06443772464990616,
0.05615289881825447,
0.05993719398975372,
0.06339256465435028,
0.05800918489694595,
0.0598299503326416,
0.06618223339319229,
0.0581543929874897,
0.06716618686914444,
0.06063525378704071,
0.07807965576648712,
0.05684481933712959,
0.06160999834537506,
0.0619305856525898,
0.059752997010946274,
0.05806198716163635,
0.06469094008207321,
0.06856141239404678,
0.0601821094751358,
0.06290949136018753,
0.05959838628768921,
0.058364998549222946,
0.06631797552108765,
0.05517958849668503,
0.055494774132966995,
0.060192499309778214,
0.0683564767241478,
0.06361056864261627,
0.0588672049343586,
0.05552633851766586,
0.0585198849439621,
0.056374162435531616,
0.06690778583288193,
0.06723847985267639,
0.0659223422408104,
0.06926744431257248,
0.06049048900604248,
0.05898473784327507,
0.07719410210847855,
0.05676357075572014,
0.05573897063732147,
0.0725347027182579,
0.05618390813469887,
0.05556076765060425,
0.06599680334329605,
0.06191345676779747,
0.0731152594089508,
0.057211168110370636,
0.059494227170944214,
0.058435965329408646,
0.06070593744516373,
0.06259196996688843,
0.0620281919836998,
0.06165178492665291,
0.056521039456129074,
0.05843997746706009,
0.062316980212926865,
0.05852144956588745,
0.06469656527042389,
0.060396190732717514,
0.06303072720766068,
0.06103653460741043,
0.05437937006354332,
0.06082785874605179,
0.07060053199529648,
0.08732984215021133,
0.05793095752596855,
0.06329011917114258,
0.06425786018371582,
0.05994216725230217,
0.057082463055849075,
0.08927694708108902,
0.07159019261598587,
0.057864051312208176,
0.06142333149909973,
0.09145152568817139,
0.06183703616261482,
0.08075301349163055,
0.06698117405176163,
0.06344950199127197,
0.06257250905036926,
0.0652865543961525,
0.06362559646368027,
0.06133025884628296,
0.059930555522441864,
0.06144530326128006,
0.05858614295721054,
0.06263396888971329,
0.06646948307752609,
0.05713609978556633,
0.0648808628320694,
0.060937874019145966,
0.06146419048309326,
0.0696353167295456,
0.05859651044011116,
0.06060481071472168,
0.07965920865535736,
0.05978751927614212,
0.06527625024318695,
0.059783175587654114,
0.05856098234653473,
0.06387346982955933,
0.06653684377670288,
0.059379518032073975,
0.0623338557779789,
0.0594169981777668,
0.057141922414302826,
0.07941555976867676,
0.05974632501602173,
0.0692623034119606,
0.059486545622348785,
0.07226838171482086,
0.06476675719022751,
0.060590796172618866,
0.06493193656206131,
0.0585501603782177,
0.06544004380702972,
0.06612936407327652,
0.057241156697273254,
0.05757598578929901,
0.05723418667912483,
0.06566401571035385,
0.061844706535339355,
0.061528727412223816,
0.06876502931118011,
0.060769133269786835,
0.06344827264547348,
0.058576442301273346,
0.055767133831977844,
0.0597340352833271,
0.06138647720217705,
0.06375885754823685,
0.059285566210746765,
0.06738823652267456,
0.05839797109365463,
0.06848839670419693,
0.059567417949438095,
0.05666476488113403,
0.060953203588724136,
0.06527874618768692,
0.0615830235183239,
0.08249107748270035,
0.06550393253564835,
0.07578440010547638,
0.06229466199874878,
0.06289573013782501,
0.06046444550156593,
0.0625322088599205,
0.060501400381326675,
0.06904244422912598,
0.06468359380960464,
0.06896170228719711,
0.05621562525629997,
0.06509561091661453,
0.05497635528445244,
0.05964590236544609,
0.059737034142017365,
0.05769484117627144,
0.06124766916036606,
0.06167520582675934,
0.05788852274417877,
0.06639967858791351,
0.05750693008303642,
0.06076805666089058,
0.07027177512645721,
0.07538502663373947,
0.067345030605793,
0.06363442540168762,
0.06224413961172104,
0.061682749539613724,
0.0630112811923027,
0.07231791317462921,
0.0679212212562561,
0.08762545138597488,
0.05715407803654671,
0.05450328439474106,
0.0675831288099289,
0.0780695229768753,
0.06339375674724579,
0.05835174769163132,
0.06990254670381546,
0.0746440589427948,
0.06305292993783951,
0.070757657289505,
0.05786628648638725,
0.05857512354850769,
0.057580895721912384,
0.06417793780565262,
0.06274006515741348,
0.061599165201187134,
0.05849495902657509,
0.05966687947511673,
0.05892038717865944,
0.0577663779258728,
0.05631422996520996,
0.058101147413253784,
0.05953484773635864,
0.0657445564866066,
0.06545165181159973,
0.0711073949933052,
0.06431517004966736,
0.059811897575855255,
0.06674955040216446,
0.0702826976776123,
0.05952949449419975,
0.06747495383024216,
0.06040548160672188,
0.07097367197275162,
0.06488184630870819,
0.07157248258590698,
0.06000060588121414,
0.06697439402341843,
0.0661972239613533,
0.06028654798865318,
0.06091951206326485,
0.057872235774993896,
0.060999754816293716,
0.059064485132694244,
0.05855536088347435,
0.10209396481513977,
0.07057790458202362,
0.06606462597846985,
0.05800531432032585,
0.05893383547663689,
0.06410185992717743,
0.060987960547208786,
0.06032022461295128,
0.08307808637619019,
0.0580887608230114,
0.056602843105793,
0.07508042454719543,
0.07418139278888702,
0.05749616026878357,
0.06257693469524384,
0.06657706946134567,
0.054361581802368164,
0.06530396640300751,
0.057197678834199905,
0.06894759088754654,
0.067872054874897,
0.05640261247754097,
0.06168398633599281,
0.05962805077433586,
0.05957464501261711,
0.07855463773012161,
0.06327655166387558,
0.056512702256441116,
0.05633966252207756,
0.06777863204479218,
0.0687953531742096,
0.06000074744224548,
0.05809706822037697,
0.05973527580499649,
0.057100631296634674,
0.06484733521938324,
0.05817033722996712,
0.06036219000816345,
0.058866117149591446,
0.06806633621454239,
0.07236925512552261,
0.06123414635658264,
0.058161526918411255,
0.05535965412855148,
0.056216392666101456,
0.07669295370578766,
0.059429481625556946,
0.0602201446890831,
0.07349781692028046,
0.07168897986412048,
0.06787940859794617,
0.06426358222961426,
0.08316757529973984,
0.059087567031383514,
0.05689283832907677,
0.07041408866643906,
0.060691557824611664,
0.07351211458444595,
0.07028193026781082,
0.07075964659452438,
0.05655835196375847,
0.05831659585237503,
0.05937304347753525,
0.058270394802093506,
0.061264000833034515,
0.07857178896665573,
0.055860478430986404,
0.06347303837537766,
0.06643235683441162,
0.06735134869813919,
0.06307021528482437,
0.05626387149095535,
0.058376092463731766,
0.05671636760234833,
0.05581330135464668,
0.059582315385341644,
0.06188497319817543,
0.05987522751092911,
0.06602036207914352,
0.05845363065600395,
0.062234483659267426,
0.05488927662372589,
0.06339552998542786,
0.07095084339380264,
0.065340556204319,
0.057432010769844055,
0.05724683776497841,
0.06310717761516571,
0.08557689934968948,
0.06813833117485046,
0.056219588965177536,
0.06509975343942642,
0.0628313422203064,
0.060528069734573364,
0.06065072864294052,
0.06939130276441574,
0.06178755685687065,
0.05852657929062843,
0.05600744113326073,
0.06482324749231339,
0.059018783271312714,
0.06470882892608643,
0.06188017502427101,
0.06560879200696945,
0.07895632088184357,
0.05500146001577377,
0.061781082302331924,
0.0620616115629673,
0.05980274826288223,
0.05595971643924713,
0.06467469781637192,
0.07354192435741425,
0.05750612914562225,
0.059660155326128006,
0.05543701723217964,
0.05756091699004173,
0.06513208150863647,
0.06350742280483246,
0.06388760358095169,
0.059433549642562866,
0.05747736990451813,
0.07480011880397797,
0.0591496117413044,
0.055449120700359344,
0.05943894758820534,
0.05838528648018837,
0.061615847051143646,
0.06575734168291092,
0.06779500097036362,
0.06112801656126976,
0.07080327719449997,
0.06778804212808609,
0.06561806797981262,
0.054735369980335236,
0.06518440693616867,
0.0565037801861763,
0.05982808396220207,
0.06283903867006302,
0.05880964919924736,
0.06397459656000137,
0.060774266719818115,
0.055492497980594635,
0.05943884328007698,
0.061088401824235916,
0.05707283318042755,
0.07537294179201126,
0.0656702071428299,
0.06086309999227524,
0.06819276511669159,
0.0650283694267273,
0.059436433017253876,
0.06291937083005905,
0.05822581425309181,
0.06889462471008301,
0.06148342415690422,
0.06302402913570404,
0.06352151185274124,
0.06046031787991524,
0.06696823239326477,
0.091937355697155,
0.05909105762839317,
0.057868216186761856,
0.061452098190784454,
0.055977750569581985,
0.059626661241054535,
0.05750139430165291,
0.07347916066646576,
0.057693179696798325,
0.05843140184879303,
0.07046180963516235,
0.06564074009656906,
0.06353752315044403,
0.06358343362808228,
0.0677260011434555,
0.08281043916940689,
0.07932344079017639,
0.07341055572032928,
0.0625871792435646,
0.06573368608951569,
0.057756174355745316,
0.06658010929822922,
0.07175806909799576,
0.062230758368968964,
0.07447569817304611,
0.09438151121139526,
0.07051154226064682,
0.07028365135192871,
0.06057070195674896,
0.061266396194696426,
0.05707686021924019,
0.07036862522363663,
0.07546283304691315,
0.08433912694454193,
0.07859573513269424,
0.05518055334687233,
0.06531360000371933,
0.0669928640127182,
0.057121388614177704,
0.06337909400463104,
0.05661369860172272,
0.08051921427249908,
0.05653433874249458,
0.0676204115152359,
0.0645650327205658,
0.06251278519630432,
0.060977306216955185,
0.05958693474531174,
0.05601762980222702,
0.058682881295681,
0.05743210390210152,
0.060221925377845764,
0.07870494574308395,
0.06247570365667343,
0.06338773667812347,
0.05577077344059944,
0.05811229348182678,
0.055448200553655624,
0.05700879544019699,
0.06005067378282547,
0.06317397952079773,
0.06534821540117264,
0.0658082664012909,
0.061410706490278244,
0.0558442547917366,
0.06304047256708145,
0.06754761934280396,
0.05847933143377304,
0.054382532835006714,
0.05453287810087204,
0.06999831646680832,
0.059032976627349854,
0.0748315081000328,
0.06193497031927109,
0.06070122495293617,
0.06828451156616211,
0.0574902780354023,
0.08301455527544022,
0.056130580604076385,
0.060004815459251404,
0.08078175038099289,
0.05862082540988922,
0.05606425553560257,
0.0620858296751976,
0.05594566464424133,
0.06614255160093307,
0.054970916360616684,
0.06426341831684113,
0.05816094949841499,
0.06725841015577316,
0.07465846836566925,
0.06339402496814728,
0.061592523008584976,
0.05973048135638237,
0.05980953201651573,
0.0638720691204071,
0.05790233984589577,
0.10134735703468323,
0.07176757603883743,
0.06119561567902565,
0.06526646763086319,
0.0695897713303566,
0.06799367815256119,
0.06368294358253479,
0.0903024673461914,
0.059214524924755096,
0.061195939779281616,
0.06306684017181396,
0.06941746920347214,
0.057874757796525955,
0.06708740442991257,
0.058318864554166794,
0.05978729948401451,
0.0701908990740776,
0.06567864120006561,
0.06043601408600807,
0.05822688341140747,
0.05815248191356659,
0.056445639580488205,
0.07598666846752167,
0.06540040671825409,
0.0664394199848175,
0.06156004220247269,
0.06643612682819366,
0.05483365058898926,
0.06197347864508629,
0.0614471472799778,
0.06079690158367157,
0.0775088518857956,
0.06565781682729721,
0.07386654615402222,
0.057120297104120255,
0.056882597506046295,
0.07825477421283722,
0.06965917348861694,
0.07548344880342484,
0.06150782108306885,
0.06852643191814423,
0.06014005467295647,
0.06528343260288239,
0.061814308166503906,
0.0586201436817646,
0.06420645862817764,
0.067235067486763,
0.06323446333408356,
0.0682930275797844,
0.07318655401468277,
0.06005240976810455,
0.060839854180812836,
0.058449454605579376,
0.0731891319155693,
0.06178953871130943,
0.06741740554571152,
0.06442710757255554,
0.06650936603546143,
0.057006195187568665,
0.06715325266122818,
0.05591083690524101,
0.08250389993190765,
0.06420943886041641,
0.06820151954889297,
0.07037217915058136,
0.059491898864507675,
0.05859022215008736,
0.059891391545534134,
0.060971278697252274,
0.05875962972640991,
0.058329544961452484,
0.063262939453125,
0.07286055386066437,
0.061000384390354156,
0.06593567878007889,
0.06197730451822281,
0.05965700373053551,
0.06629227846860886,
0.05938404053449631,
0.056490447372198105,
0.06203078851103783,
0.07788753509521484,
0.06300835311412811,
0.0596415139734745,
0.059186846017837524,
0.09686442464590073,
0.07047535479068756,
0.06682425737380981,
0.06150004640221596,
0.05580094829201698,
0.05925916135311127,
0.07485292851924896,
0.09848517924547195,
0.06432590633630753,
0.06496814638376236,
0.05670277774333954,
0.05842829495668411,
0.06579073518514633,
0.057248979806900024,
0.05435609444975853,
0.06361985951662064,
0.06707675009965897,
0.05861910432577133,
0.10109905898571014,
0.05654747411608696,
0.05609507858753204,
0.0555749237537384,
0.06111624464392662,
0.05559597536921501,
0.05707908421754837,
0.060660701245069504,
0.05716215819120407,
0.05951845273375511,
0.064541757106781,
0.0698176920413971,
0.05716194212436676,
0.06380718946456909,
0.06184648722410202,
0.059877313673496246,
0.06400177627801895,
0.06191052123904228,
0.05565802752971649,
0.06185721978545189,
0.07522857189178467,
0.0624237097799778,
0.06685299426317215,
0.07108581066131592,
0.07991273701190948,
0.05669805407524109,
0.05897118151187897,
0.06358261406421661,
0.06911467015743256,
0.06322716176509857,
0.05924277380108833,
0.0792098417878151,
0.06731507182121277
]
}
],
"layout": {
"template": {
"layout": {
"colorway": [
"#0a1628",
"#D4A84B",
"#1a2d4a",
"#C87533",
"#10b981",
"#ef4444"
],
"font": {
"color": "#334155",
"family": "DM Sans, DejaVu Sans, sans-serif",
"size": 11
},
"hoverlabel": {
"bgcolor": "white",
"font": {
"family": "DM Sans, DejaVu Sans, sans-serif",
"size": 11
}
},
"legend": {
"bgcolor": "rgba(255,255,255,0.8)",
"bordercolor": "#e8e8e6",
"borderwidth": 1,
"font": {
"size": 10
}
},
"paper_bgcolor": "#FAFAF9",
"plot_bgcolor": "white",
"title": {
"font": {
"color": "#0a1628",
"size": 14
},
"x": 0.5,
"xanchor": "center"
},
"xaxis": {
"gridcolor": "#e8e8e6",
"gridwidth": 0.5,
"linecolor": "#e8e8e6",
"showgrid": true,
"tickfont": {
"size": 10
},
"title": {
"font": {
"size": 11
}
}
},
"yaxis": {
"gridcolor": "#e8e8e6",
"gridwidth": 0.5,
"linecolor": "#e8e8e6",
"showgrid": true,
"tickfont": {
"size": 10
},
"title": {
"font": {
"size": 11
}
}
}
}
},
"title": {
"text": "Sig-WGAN Training Progress (2500 Steps, Paper Implementation)"
},
"xaxis": {
"title": {
"text": "Generator Update Step"
}
},
"yaxis": {
"title": {
"text": "Sig-W1 Distance (RMSE)"
}
}
}
},
"image/png": "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"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"Final Sig-W1 loss: 0.0673\n",
"Best Sig-W1 loss: 0.0540\n",
"Paper reference: ~2.76 (S&P 500 log returns, 2005-2020)\n",
"\n",
"Using paper-exact data source: S&P 500 index log returns\n"
]
}
],
"source": [
"fig = go.Figure()\n",
"fig.add_trace(\n",
" go.Scatter(y=training_losses, mode=\"lines\", name=\"Sig-W1 Loss\", line=dict(color=COLORS[\"blue\"]))\n",
")\n",
"fig.update_layout(\n",
" title=f\"Sig-WGAN Training Progress ({CONFIG['total_steps']} Steps, Paper Implementation)\",\n",
" xaxis_title=\"Generator Update Step\",\n",
" yaxis_title=\"Sig-W1 Distance (RMSE)\",\n",
" template=\"ml4t\",\n",
")\n",
"fig.show()\n",
"\n",
"print(f\"\\nFinal Sig-W1 loss: {training_losses[-1]:.4f}\")\n",
"print(f\"Best Sig-W1 loss: {min(training_losses):.4f}\")\n",
"print(\"Paper reference: ~2.76 (S&P 500 log returns, 2005-2020)\")\n",
"print(\"\\nUsing paper-exact data source: S&P 500 index log returns\")"
]
},
{
"cell_type": "markdown",
"id": "13946a29",
"metadata": {
"papermill": {
"duration": 0.008262,
"end_time": "2026-06-14T16:58:51.741270+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:51.733008+00:00",
"status": "completed"
}
},
"source": [
"**Interpretation**: The Sig-W1 loss measures the L2 distance between expected\n",
"signatures of real and generated paths. A decreasing loss curve confirms the\n",
"generator is learning to produce paths whose signature statistics match the\n",
"training data. Compare the final value to the paper's reference of ~2.76 on\n",
"S&P 500 log returns."
]
},
{
"cell_type": "markdown",
"id": "951d7659",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.008223,
"end_time": "2026-06-14T16:58:51.757615+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:51.749392+00:00",
"status": "completed"
}
},
"source": [
"## 9. Generate Synthetic Paths\n",
"\n",
"The LSTM generator produces paths UNCONDITIONALLY - no conditioning on past.\n",
"Each call generates a fresh batch of paths from Brownian motion noise."
]
},
{
"cell_type": "code",
"execution_count": 37,
"id": "345c4c62",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:51.774739Z",
"iopub.status.busy": "2026-06-14T16:58:51.774601Z",
"iopub.status.idle": "2026-06-14T16:58:51.850875Z",
"shell.execute_reply": "2026-06-14T16:58:51.850323Z"
},
"papermill": {
"duration": 0.085525,
"end_time": "2026-06-14T16:58:51.851260+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:51.765735+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Generated 3249 synthetic paths\n",
"Synthetic paths shape: (3249, 16, 1)\n",
"Generated 592 synthetic paths for holdout comparison\n"
]
}
],
"source": [
"def generate_paths(\n",
" generator: nn.Module,\n",
" n_samples: int,\n",
" n_lags: int,\n",
" device: torch.device,\n",
") -> np.ndarray:\n",
" \"\"\"\n",
" Generate synthetic paths (UNCONDITIONAL).\n",
"\n",
" The generator takes batch_size, n_lags, and device - no conditioning input.\n",
"\n",
" Args:\n",
" generator: Trained LSTMGenerator\n",
" n_samples: Number of paths to generate\n",
" n_lags: Length of each path\n",
" device: Torch device\n",
"\n",
" Returns:\n",
" Shape (n_samples, n_lags, n_assets)\n",
" \"\"\"\n",
" generator.eval()\n",
"\n",
" with torch.no_grad():\n",
" paths = generator(n_samples, n_lags, device)\n",
"\n",
" return paths.cpu().numpy()\n",
"\n",
"\n",
"# Generate synthetic paths\n",
"n_synthetic = len(train_windows) # Match training set size\n",
"synthetic_paths = generate_paths(generator, n_synthetic, CONFIG[\"n_lags\"], device)\n",
"print(f\"Generated {n_synthetic} synthetic paths\")\n",
"print(f\"Synthetic paths shape: {synthetic_paths.shape}\")\n",
"\n",
"# Also generate for holdout comparison\n",
"n_holdout_gen = len(holdout_windows)\n",
"synthetic_holdout = generate_paths(generator, n_holdout_gen, CONFIG[\"n_lags\"], device)\n",
"print(f\"Generated {n_holdout_gen} synthetic paths for holdout comparison\")"
]
},
{
"cell_type": "markdown",
"id": "19bc0d28",
"metadata": {
"papermill": {
"duration": 0.008487,
"end_time": "2026-06-14T16:58:51.868624+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:51.860137+00:00",
"status": "completed"
}
},
"source": [
"## 10. Evaluation\n",
"\n",
"### 10.1 Fidelity: Visual Comparison with PCA and t-SNE\n",
"\n",
"We project both real and synthetic paths into 2D to assess whether the\n",
"generator covers the same regions of the data manifold."
]
},
{
"cell_type": "code",
"execution_count": 38,
"id": "24de22d4",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:51.886246Z",
"iopub.status.busy": "2026-06-14T16:58:51.886099Z",
"iopub.status.idle": "2026-06-14T16:58:54.254596Z",
"shell.execute_reply": "2026-06-14T16:58:54.254098Z"
},
"papermill": {
"duration": 2.378023,
"end_time": "2026-06-14T16:58:54.255034+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:51.877011+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"utils/style.py:764: UserWarning: The figure layout has changed to tight\n",
" plt.tight_layout()\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAABJ8AAAH6CAYAAACpjO0/AAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjksIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvJkbTWQAAAAlwSFlzAAAPYQAAD2EBqD+naQABAABJREFUeJzs3Xd4VGX2wPHvlEwy6b2SQkkg9Cq9SRNQEBWx69pX17723XV3dV397Vp2XXvvBRsqKFU60kILJCS0EEJ6m0wymcmU3x/DDJlkEtImCXA+z+Nj5p079763ZLj35LznVZhMRhtCCCGEEEIIIYQQQniAsqs7IIQQQgghhBBCCCHOXRJ8EkIIIYQQQgghhBAeI8EnIYQQQgghhBBCCOExEnwSQgghhBBCCCGEEB6j7uoOCCGE6BhLV6znmRfecr7esvyTLuyNEK7e+fgb3v3kOwCio8L57qOXu7ZD54n8gmIuu/EB5+tX/+8Jhg/p34U9Ojuuhe74fZq25wB3P/Ks8/W3H75ETHQEAE//+02WrdwAwLDB/XjtX3/qkj7WdzacZyGEEJ1Hgk9CCNFNrVy7haUr1pN1OAddVTU+3hoCA/yJjgojuVcCY0cNZczIwV3dTQBsNhtbtu9h5dot7M88TGl5BUZjHYEBfvRK6sHo4YOYPX0C4WEhLF6yghdf+8j52aVfvkpocJDz9fZd6dz72HPO17+75lJuv/EKl+1de8djHDl2AoBpk0bzzJP3uLxvtVrZ8Fsaa9ZvIyPrCGXlldQaTWh9vImKDCO5VyKjhg1g8viR+Plq3e7Tmg3bePKZ/7q0PXjXDSycP9Pt8vUf/gBSU3rx3it/d1mm4b7Xf3hsjwU33E9BYUmjdm9vDWEhQfTv25tL51zIiKFdG3Q4G6VnHOKr75ezLyObsrJKlEolgQF+hIUG0btnAv379mLB3Gld0jcJLDWtYfAIQK1WofXxJigwgPi4aEYM6c/cmZMIDgrwaF/q/37ect0Cbr3+co9urzN0l/MshBDi7CHBJyGE6Ib+9n9v8MvqjS5t1TUGqmsM5BcWs2tvJgWFpS7Bp9SUXvzh1qs7u6sUFpXy1HOvsmd/VqP3yit07Nx9gJ27D3As9yR//uMdDB3Y12WZ3fsOcuHEC5yv96QfdHm/4etKXRVHc/Kcr4cO6ufy/tGcPJ567lWyjxxv1B99dQ36ozUcPprLL6s3UlJWzg2L5rndr6Ur1jVqW7ZyQ5PBp4Yyso6wdtN2powf1aLlPcFoNHGyoJiTBcWsWvcbj953M5fOubDL+nO2+eHnX3nuP+9hs9lc2muNRopKysjIOsrajdu7LPh0trpg+CC0Pj4A+Pv5dtp2zWYLVfoaqvQ1nDhZyJbte3j742+4745rG53Drvo+bU5cTJRLnwID/LqwN2fWVedZCCFE9yTBJyGE6Ga2bN/jEnjql9yT0SMGodX6UFGp4+ChY6RnHGr0uV5JPeiV1KMzu0pZeSV3PfwMJwuKnW2x0RFMGDOc0JAgqvTV7M845BKY6t0zngB/X6r0NQDs3pfpEnzavc812LT/4GHq6sx4edn/ydqdftAlGFA/mHXs+El+/8enqdTpXfozdtRQIsJDMJnqyM3LZ3f6QYpLypvcr9KyCrbu2NeoPTP7KIeP5dI7Kf6Mxwbg7Q+/YdLYESiVnVdiMS4mkgVzp1FnNnP4aC6r1291Hq833v+KeRdN6dT+nK0qdXpefP1j57GLDA9l6sQLCAkOpKbGwJGcE42uVdEygwekMHhASqduc8HcacTGRKCrquZA5mHS9mZgs9kwGk3833/fp0pf7RKI7orv06YYamvx1miIigzj2oVzu7o7LdYV51kIIUT3JcEnIYToZralnQ569IiN4p3//A2VyjVYUF1dw6FTw84czlSj5NDRXN54/yt27csAYHD/FH5/8yLWb97R5uETL7/xsUvg6bKLp/HAXTegVqlclsvNK+DAwcMAKJVKBg/oy6atuwB7MMnBbDazP9O+XHhoMCVlFRiNJjKyjjgfYnbvy3QuH+DvR++epwNBz770tkvg6aar53Pr9Zc3On42m41d+zIxmy1u9+vn1RuxWK0A+Gp98PXVUlJqD1YtXbGBe2+/piWHhyM5J/hl9SbmzJjYouXHzrrO+XNbh+dERoQ2ekBdte43wB5QKa/QERYa7PJ+WXklX373C1u27yEvvwiz2UJEeAijRw7m+isvJjoy3GX5rMM5LFm2hszsYxSVlKGr0oMNQkMCGdCvDwvnz2RIgwy31mo4/PKbD18kNjrS+dpqtTL/2nspKasA4I4bF3LTNfMB++/C0pXrOXLsBPpqA75ab0KCA0nulciwwf24/JIZZ9z+3v0HMRpNztdvvPgXYqJcj4PZYmHn7gPO1zt3H+APj56uyfPlu/8ioUeMS58vueYeysorAbjr5kVcv+iSRr+76358n8++XsayVRsoKCohOCiA6ZPHcudNC9FovAD3Qy3r1wNqru7P2k3b+XTxUrKPHEfjpWbksIHcd/u1REWGNVo2+3AOX36/nN37MikuLUelVNIjLpoLJ17AogWznJktDesRARQUlrhc03966Hbmzpx0xiFbhtpalixby7rNOzhy7AQ1BgOBAf4k9Ihh2qTRXDHvzOevoemTR7sMSdyzP4tH//qi8/vizQ8WM2H0cGfAqbnv04rKKj768kd+276H/MISzBYzgQH+REWE0b9vby6aNp6BqX0aDcUFePeT75z7Xn+9DYfmjRo2kHc+/paMrCNU1xhY8c2bZB/OabLmU0NlFZW88f5XbNq6G72+hqTEWK5beDEzpox1Wa65IYHuzlNHnmddVTWLl6xg429p5OYVYDSZCA4KZEC/3lw2dxoXjBjksnxbfk+EEEJ0HxJ8EkKIbsZisTp/1lfXkF9YTI/YKJdl/Px8GdKKvyhnZB3hD488S42h1tn22469pO3NaDQMrqVKSstZtW6r83Vy70QeuvtGt1k18XHRxMdFO18PHXQ6+HT46HH01TX4+/mSmX2MWqMRgKsvn8Mrb38G2IfenQ4+nQ5WDR6QgkKhAGB/5iH2Hch2vjfugqHccdNCt31XKBQMH5za5L4tW3H6gdGexRXIF9/+AsDyNZu465ZFjQJs9fl4e6NUKaipqeXdT75l5tSxqNVd809uRFiI82elUkGAv+tQnX0Hsnnkry9SUVnl0p6XX8S3P65ixZrN/OtvD7oMb9yTfpBvf1rdaFsFRaUUFJXa62U9eBtzZ05qc79HDh1AdFS488F45drfuPGq05kpaXsznIEnpVLBnBkTANcHZgfHUKvjJwrYtS+zRcGn+r+HAIeO5DQKPqlVKkbXe0AeMbQ/vZJ6OOuR/fDLWpdhUrv2ZjoDTyqlktnTJ7jd9r2P/dMlW7C4pJzPv1lGeYWOpx6584x9b85HX/7I1p2nA9xGo4lfN2zj0JHjfPzGs3hrNM73vv1xFS++/jEWi2uQNvtwDtmHc1ixZjOvPP94o2Bme+TlF/HAk/9Hbl6BS3tZeSVl5ZVUV9e0KfjU0JABKTxy783Oum5Wq42vliznsftuafZzRpOJOx78O8dP5LvtX0bWEXy13gxM7dPmvm3dmc4Hny1xBsBbq6Kiilvv/Sv5haf/MJB1KIe//PNVSkrLufryOW3uW0c5djyP+x5/nqKSMpf2ktJy1m3awbpNO7jy0lk88Pvrm1yHJ39PhBBCdDwJPgkhRDfTt0+S8+eKyioW3fJHknslkprSi37JPRk+JNUlkHMmNpuNf7z4tkvgacaUscTFRLJ6/Va2paW3qZ8792S4DH+bM31ii4dzDasXyLBabexJP8j40cNcsppmXTiOr39YSX5hMbvTM7l+0SVU1xjIPpzjXGbooNOBs+279rtsY97sKa3dJQD2Zx7m6PHTNaWmTxlDaHCQM/hUVl7Jlu17mDhmeJPr8Pb2YuH8mbzz8becLCjm+2W/dsgDc2uYzWYOnRp25zBx7AiXjIDq6hoe/dtLzsBTdFQ40yeNwdvbi183bOdIzgn01TU88fR/+er9fzvrtmi8vBiY2ofkXokEBfqj1XqjrzawY9d+MrKOYLPZ+O9bnzFt8hh8vDW0hUKhYM70ibz3qT2QtOLXzS7BpxW/bnb+PGrYQCIj7Fk79YNio4YNZPjgVAy1tRSVlLEnPQuj6XQ2U3OSeyeiUCic1/gjf32JuJhIBvTrQ98+SQwZ2Jf+fXs5g58OV8ybwf/9930Afl61kTtvWugMPK7ZcPpcjB45mPB6gcH69uzPYvL4kfRMiGP5ms3OIMKKX+2Bz4iwEG66aj75hcV8+MUPzs8tmDuNuBh7dlhUROMsJoCtO/eRmtKL0SMHkbYng72nHt5z8wpYv3mnMzNm7/4sXnjtQ6xW+/4PTO3DmJGDqa6p5edVG6iorOLo8Tz+/q83+M8/H3PWI9qWts/5nRLg7+dyzlJTejV7zC0WK4/+7SWXwFNqSi9GDhuA1Wplf+ZhqmsMza6jNaaMH0mAvx9V+moA0vZknPEzaXsynIEnjcaLS2ZNISI8hLLyCk6cLGTX3tPfYdMnj6FXYg8+/OIH5zYuGD6QC4YPcrtuh/SMbHy8vbl42jgiwkLJOnysVUNljx7Pw9/Pl6suuwgFCn5asc45zPm1975kwpjhrfo3pL6OOM9mi4VH//ayM/CkUiq5aNp4IiJCWb95pzN4+9X3y+nbJ6nJzNGW/J4IIYToPiT4JIQQ3cysaeP55seVZGQdBezBmYOHjnHw0DHnMkMGpPDQ3TeS3DvxjOvbn3mYw0dzna+vv/Ji7rrlKsCeXXTFTQ86H4xao7jU9S/WifExTSzZWN/knmh9vDHU2rOcdu87FXw6NQSvR2wUYaHBDBvcj/yVxezdn43VamXv/iyXbIChA08HsYpLXWs4JfZw7c+t9z3lHNJXX8PhictWrnf+HODvx5gRg/HyUhMXE0lefpF9mRUbmg0+AVx12Wy+/mElFZVVfPD591zcjiyg1ti1N9NlCIzDmJGDefz+W13alq7cQHmFDrDv6wf/e4agQH8Arl04l8tveJDySh3llTqWrdzAlZfOAmD+nKnMnzOVQ0eOc/hYLpU6PSqlikljR5CRdQQAXZWezKwjjQrCt8bcmRN5/7PvsdlsHDl2wllvq67OzNqNO5zLXTxrsvNnk6nO+fNTj9zZKCvHcQ7PJC4mkisvncWX3/3i8tm8/CJn4Cs2OoK7b73apWbZRdPG8/p7X1Klr6GsvJKNv+1iyoRRWCxW1m463edL6vW5oUULLuL+O+3n8MKJF3DDXU8C9u+CzOyjRISFMH/OVPILXINPDYeWudO/b2/efPHPqNVqzGYz866913kNZBw84gw+ffbNMmfgafjgVF55/nFnAGT65NHccu9TAGxLS+fQkeP06ZXAtQvnYqitdQYl/Py0rapRtHn7bpfvqvlzpvLovTe7BPhaev5aQqlUEh8X7RwS3FwdOIf6wcthg/rxxz/c6PK+yVRHhc4ezB07aghjRw3h6x9XOr9jB/VPPuMxUSmVvPHCn+ib3LNV+1PfC0//0ZktOmXCKO540D7zptlsYdnKDU1mhZ6Jo+5Ue87zpq27XDLHHrzrBi67ZDoAN141n6tve8SZ8fjFtz83GXxqye+JEEKI7kOCT0II0c2oVSpeef4JPvriB35asd45TKe+PfuzuO/x5/n0recICQ5sdn2OYIDD7Omnb+QDA/yYNHY4SxvUJQF75kP9YWwOg/ont7uIrFqlYmBqMtt32R9eHEXEHVkYjoDF0EH9WLZyA/pq+wx19TOjfLy96Zec1MxWFM28557JVMfKtb85X0+ZMNJZ6Hz65DHOB/1N23ZRqasiKLDpKdr9fLXcsOgS/vvWZ5SWVfLVkuXO+jhNaRgI6yg9YqO47YbLnYElh731hqxU6au5aGHTQ1X2Hch2Bp8OZh/l7/96kyM5J5pcHmg0pKa1YqMjGT44lZ177HWVVv66hd6/i+e3HXucD/OBAf5MGjvC+ZkhA/uyedtuAK694zEG9OtNj9hoeiX2aHXW4H13XEvPhDi++n652309WVDMn/7xCq889zgjhtqDPlofHy6eOZnPv/0ZgCU//8qUCaPYve/0kLvgoAAmjBnW5HYvP/UgDpDQIKhbVdX6QHF98y6a4szEUqvVxEZHOINPunpB6PrXRtreDMbPvqHJde49kE2fXgnt6hfA3nTXGTNvv/GKRplljsyujtJwJsMz6Z/SC42XF6a6Orbu3Mc1tz1Kn17xxMfFkNInkZFDBxAZHtquPo0ZNaRdgae4mEiX7+jBA1KIjY5w1ufLzD7arv61V/oB1wkzZs84PfzUx1vDtEmj+XTxUsBeq7C21oiPj3ej9Xjy90QIIUTHk+luhBCiG/Lz1fL7mxfx0+f/49M3n+OJB29jzoyJ+PqeDl6UV+pcZsVrir66xuV1aEiQ6+sm6rVsS9vH/975vNF/joLoEWGuD1g5ufnuVtOk+kPmMrOPcODgYWdAwVGHqn49qt3pB9lVL/g0ILW3Sx2lhn/lbliTZeH8Wfzh1qsZNrjpTJz1m3e6ZIFNn3y6OO+Mqad/rqszs3zNZs7kskumOx9EP/lqaaNz4QlxMZH84darufLSmfj5agE4cbKQPzzyLEdz8lyW1bXiAa280h6gqDWa+ONfXjhj4Ansx6m9Lp51OmNs5dotAKz4dYuzbebUsS5DCR+553fOejuVOj2bt+3hq++X89x/3uXKm//In/7xCtYW1tJRKBTMnzOVT996jiWf/IdnnryHRQsuIrpe7SebzcYX3/3s8rkr5s9AqbQHTbal7aOwqNRl+OOsC8c3WwOsfm0pjZdr4eTWBkuaWzeAV7311193a66NilPXRnvpqk5PFuDj7U1ocFAzS7ef1WrlxMlC5+uI8DNnykRGhPGnP95OcJA98Hz0eB4r1/7Ge59+x2N/e5l519zjvE7bqjVZpO64+4NE/e/9pr6HGl5apg74/XWn/nn21fo0CsrXP+82m42qJvrryd8TIYQQHU8yn4QQohtTKBTOKb8vmTWZW6+7jCt+96BzOExuXuEZ1oCzTo9DeYXOJQOm7FTR5tYaMSTVpSbOz6s3cOWlM1tcm6T+cCyz2cInXy11vnbMlBYfF014WAglpeVs3bnP5S/2DQuljxo2gDc/WOx8vXTleiaNO50RM+vCcQCUlle61GWpb2m9IXcA9z3+nNvlHMs6MoGa4q3RcPO1C3juP+9Spa/mmx9XNrt8R6g/292Fk8Zw1x+fxmq1Yag18sJrH/K/559wLhsYcLr4eHhoMFddNrvJ9TpqCO3el+ks9A32oZs3LLqE4KAAamuNTJ3ffMHm1poy4QL+/b8Pqa4xcLKgmG1p6Wz8bZfz/YbDGaMiw3j75b+emmHxCCdOFnD4aC7rt6RhsVhYvX4rY0YOdhmq1xKREWFMiwhj2qTR3H3LVdxw1xMcO34SaPx7GBsdybgLhrLxt11YrTaW/Pwrazdtb7LPDdUPTDXM/Gkvtdq1UH5Tqw8M8HNmRA0ZkMLEetllDQ3qn9whfQsMOP29VGs0UlZR6dEA1LpNO1yCzSPOMGTRYcaUsUydMIr9B49w+GguJ/IK2Ln3AFmHcqgx1PLsi+8wfvQwfLXNZzo2Resmy6c1HOetvvoZtPX/TVDWuwAa1kM70aDoe0epf55rDLUYamtdAlBlFaf7qlAoCGjwb5iDJ39PhBBCdDwJPgkhRDezdOV6TKY6Zk4Zi1+Dm24fH28UCiVgn33K39/9TXl9DYu/rly7hdtusE+nrauqZv2WNLefu/X6y12m3W4oPCyEaZNGs2qdfZha1qEcXnr9Y+6/83pUKtcAlD0QcJhZF453tg3o19s5fAVg3eYdzvXWn91v2KC+rFz7G5u37Xb5a3bDWkID+vVhYGof0jPsQzrWb97Jx1/+yHVXXtyiB5Pi0nK2t6L4etahHGetm+bMnTWJT79eSm5eAaVljYdQ1le/VlPDac/bYsiAFC6aNsE53fvO3QdI25vhnOlvUP9kZ0ZOeWUVo0cMarQ/NpuN7bv20yPWPtzJMTW9w6wLxzmzQOpn93QUH28N06eMYcmyXwF4/j/vOmdETO6V0Gh4UvbhHHr3jG80w+IjT73Iht/s1/rBQ8fOGHzKzD7Kuk07uHTOhURFuhbvVqmULtlWAW5+DxfOm+kMkn369VJnLap+yT07ZIgagKpBIKnW2LJi6i0xqH8y6zfvBOwB20vnTG30fVRrNLFm/VaXIV5q1elbS2Nt6/ozeGAKnI4f885H3/LwPTe5/P7mF5Y0yt5qi/SMQ/zfK+87XyuVijMGk8F+/dcYaomJCmfIgBTnrKO6qmpmXXEHYA+cHT+RT79T12b9mTE78hw1JS+/iL37s5znZe/+LOeQO8DZL3D9N+TAwcPYbDYUCgWHjuaycevpIG9D7TnPDYOVP6/c6Kz5VGs0uXyP9OmV4HbInRBCiLOPBJ+EEKKbyS8o5t1PvuPlNz5hyIC+JPdOIDDAn0qdnl83bnOZ9nzMyMFnXN/A1D707hnvLOT7/mffc7KgmOjIMFav39qmYuMO991xLemZh5zFYb/+YSVbtu9hwpjhhIYEoquqZn/mYfakH2T29AkuwSdvjYbUlJ7OqbIdgaWGGU1DB6Wycu1vLoEntVrFwH6NpzJ/4oHbuOPBv7nM7PTL6k2MHjGI4OAAe7Bt845GnwP4edUGl2LmE8YMw8fb9aHHarWyZsM25+ufVqx3Frxtilql4rYbLucv/3y12eU85YZFl/DL6o3ObLkPP1/iDD7NnTmJDz5fQkVlFRaLhTse/DtTJ15Aj9go6urMHD+RT9reDMrKK3n1/54gNjqy0ZCgvz3/OtMmjyG/sJhfVm/yyD5cPHOyM/hU/yF6rpsMoj89+z+qa2oYPrg/EeEhBAb4k5dfyObte5zLtCRoW1Nj4IPPl/DhFz/QLzmJ/n37EB4WjMlUx/Zd6WQdOj3rorvfw1HDB5IYH0tO7kmXIuju+txWIUGBqNUqzGb7d8IbHywm+8hx1CoVw4eknnHWseZcc/kcNmxJw2azceJkIdfe8TiTx48kNCSI6uoaDh/LZdfeTAy1RpeC0PWHrpVX6njm32+SlBiHAgWXz5vR7OyH40YNdfmu+m7parIOH2PEkP7YgKxDxyir0PHRa/9o9f6sWreVA1lHqKqqZv/Bw6Q1mK3zrpuvomdi3BnXk5tXwG33/5XUlF4k90ogPCwElUrJbzv2uixXP7soIjzEObxv2YoNeGs0+Gp9iIuNZMr4Ua3el5Z46M//5uJZk5yz3TmoVCqX85Wa0st5Le/am8mt9/2V8LBgduza3+yw2Xad59FDSegR4xwa/eJrH3Eg6wgR4SGs37zT+e8JwFULLmr9zgshhOiWJPgkhBDdlOMh11GUu6H5s6c6gwhn8uSDt/GHR56lxlCLzWZz1orSeHkxYmh/du62F3RWtnLoQnhYCK/960/85Z+vkp5hL06el1/kMkNYc4YO6ucMPjkMaRB8GuZmtrR+yT3d/jW8Z2Icr/7fk/z5n6+Sk2sfEnUk50ST9YnqD/9YtvJ0/az4uGj+9beH3H7m9w897ZyVb8Wvm/nDbVe7ZDa4M33yGD7+8keyjxxvdjlPSIyPZfL4Ufx6Kmi2LS2d/ZmHGdCvN/5+vjz/1AM8+reXqKisosZQy9IV65tdX7/knowZOdj5sH30eB7vfPwNAHNmTHRmWXWkgal9SEqIdQ5zA/DyUnPRtPFuly8tq2yy7k5ggD/zLpra4m3bbDYyso46Z59sqG+fJLfDFRUKBVfMm8ELr37obNN4eTFz6rgWb/tMvLzUjB89jHWnZtHLPpxD9mF7IOEPt17druDTkIF9eejuG3jp9U+wWCwUFpfy1ffLz/i5MSMH4+Pt7cxOqz+ZwdyZk5oNSqhUSp5/6gHuf+J5Z7Bmf+Zhl1kqk9uYNfbd0tVu2328vXng99cxb3bLrwmwT+TQcDIHhynjR7lkb04ZP8o51Le8Usd7n34HwLgLhnok+BQfF42h1sgX3zb+Hr7zpoUuGYGLLr2In1dudGagOmb+8/bWMHxwKml7M9xuoz3nWa1S8fxT93Pf489TVFKGxWp1+71z5aUzm5zpTgghxNlHgk9CCNHNLFpwEb2T4tmx5wCZWUcpK6+gvLIKq9VKcFAA/fr0ZM6MiUyZ0PKHltSUXrz50lO8/t6Xzhnj+vftzZ2/u5Kflq9zBp9akhHSUExUOG+99Bc2bd3NqnVb2J95mNKyCkx1ZgID/OjTM4HJ40cwtd509A5DB/VzmSoeGmc+9UyMIyQo0FnwGhoHqOpL7p3IJ288y+r121i3aTsZ2Ucpr9BhsVjw89USGx1Jv+SeXDBiIONGDQXsQ3AcwSpoPjtl7sxJzuBTeYWOzVt3u9SWckehUHDHTQv5419eaHY5T7nxqnnO4BPAB59/7wyuDR6QwqdvPcfXS+xZa7l5BdTWGvH19SEuJoqBqX2YOHaEyzDHf/75Pt74YDGr1/1GZZWeqIgwLp41mesWXuyR4BPYj/ur73zhfD1h9DC3sw3+/uYr2bYznYysIxSXllGp06NWqYmKDGPE0P5cd8XcFg3bGtQ/hVeef5yduw+wd38WRSVllFfoqDWa8PfT0jOxB1PGj2TB3GkuQ/DqmzNjIm+8/xXVNQYAJo0b4VJnqyM8fv8t+Plq+W3HXioqdc4Mt45w+SUzGDqoH4uXrGDX3gyKisupM5sJCvQnMT6GoQP7cWGD3+uw0GD+9fcHefujb8g+nIOh1tiqbcbFRPLR6//g+2W/sm7jdo7k5FFjqCXA35eEHjFMnzymzfujUqnQ+ngTHBRAfFw0o4YNYM6Mic3OWtlQQo8Y7r39Gvbuz+ZITi5l5ToMhlp8fbX0TIxj2qTRXHbxNJfPXH7JDKr01fy8aiMFRaUu2aueEB4WzF8fvYs3P1jMlm17qK4xkJgQy7VXzHXWvnPomRjHf597jDfe/4qM7KOoVSqGDe7HHTddydqN25oMPrX3PCclxPHR68/y9Q8r2LAljdy8AoymOkKCAhiQ2ocFc6cxesSgNh8DIYQQ3Y/CZDLKdBBCCHGOq6szo1IpGxUDrzHUcu0djzmHOcyfPZXH7u/YgtFCnM+uuvURZ2DzpX880qKhskIIIYQQ5xrJfBJCiPPA0eN5PPLUi8y8cBw9E+II8Pcjv7CY75eucQaelEoFl58q+iqEaLuswzlUVOjYtG23M/CU0CNaMjmEEEIIcd6S4JMQQpwnCotL+fjLH92+5+Wl5uF7fkdy78RO7pUQ556X3/jYWeMH7MMu7739WpkOXgghhBDnLQk+CSHEeSAqIpSrLruItL0ZFBaVoq824K3xIiY6guGDU7ns4ukkJcR2dTeFOKf4eHvTMzGWm66+lPGjh3V1d4QQQgghuozUfBJCCCGEEEIIIYQQHqM88yJCCCGEEEIIIYQQQrSNBJ+EEEIIIYQQQgghhMdI8EkIIYQQQgghhBBCeIwEn4QQQgghhBBCCCGEx0jwSQghhBBCCCGEEEJ4jASfhBBCCCGEEEIIIYTHSPBJCCGEEEIIIYQQQniMBJ+EEEIIIYQQQgghhMdI8EkIIYQQQgghhBBCeIwEn4QQQgghhBBCCCGEx0jwSQghhBBCCCGEEEJ4jASfhBBCCCGEEEIIIYTHSPBJCCGEEEIIIYQQQniMBJ+EEEIIIYQQQgghhMdI8EkIIYQQQgghhBBCeIwEn4QQQgghhBBCCCGEx0jwSQghhBBCCCGEEEJ4jASfhBBCCCGEEEIIIYTHSPBJCCGEEEIIIYQQQniMBJ+EEEIIIYQQQgghhMdI8EkIIYQQQgghhBBCeIwEn4QQQgghhBBCCCGEx0jwSQghhBBCCCGEEEJ4jASfhBBCCCGEEEIIIYTHSPBJCCGEEEIIIYQQQniMBJ+EEEIIIYQQQgghhMdI8EkIIYQQQgghhBBCeIwEn4QQQgghhBBCCCGEx0jwSQghhBBCCCGEEEJ4jASfhBBCCCGEEEIIIYTHSPBJCCGEEEIIIYQQQniMBJ+EEEIIIYQQQgghhMdI8EkIIYQQQgghhBBCeIwEn4QQQgghhBBCCCGEx0jwSQghhBBCCCGEEEJ4jASfhBBCCCGEEEIIIYTHSPDpPPDOx99ww++f8Ph2xs66jnWbd3h8O0IIIYQQXWnBDffzxbe/eHQbS1esZ8Zlt3t0G0IIIURnUZhMRltXd6K7ePrfb7Js5QYA1GoVURFhzJ4+kRuvnodapQLAZrOx5Odf+fGXdRzNOYFKpaJHbBSzLhzPpXOm4uPj7VxfUXEpl9/0IAlxMXz61nMdsv22qDHUUldXR1BgQJvXUd87H3/D+s07+ej1Z13aS8sqCPD3Q6Px6pDtCCGEEKL7uevhZ0julcgDv7/+jMsuWfYrX/+wkrz8QlQqFTHREUybNJobr5oH2O8p3v3kOy6dcyGP3nez83NZh3O48a4n+fbDl4iJjiC/oJjLbnzA7TbefvmvDEzt4/a9sbOuc/7s56ulV1IPbr/xCkYOHdCaXW6kvEKH1sfb5b6vPRbccD+LLr2Iqy67yNlWazRRYzAQGhzUIdsQQgghupK6qzvQ3YwZOZg/PXQ7pjozW7bt5t+vfoharXLeJP3t/15n7aYd3HT1fB66+wZCggLJPnKcL7/7hZjocCaPG+lc19KVG5g2aTS79x1kf+YhBvRzf2PUmu3XV1dnxsvrzKfQV+sDWp9WHIW2CQsN9vg2hBBCCHF2+HH5Ol5+4xMeuOt6hg3qR12dmUNHj3Pk2AmX5TQaL35cvo5rrphDfFx0s+v873OP0Suxh0tbUKB/s5/500O3M2bkYCp0Vbz5/mL++JcX+PTN54iLiWy0rNlsRq0+871VSHDgGZdpLx9vDT7eGo9vRwghhOgMEnxqQOPl5QyiXHbJdNZt3sHG39K48ap5rFr3G8vXbOb5px5g0rgRzs/EREcwcexwqmsMzjabzcbSFev54x9uIjI8lB9/Wdei4FNz23/632+i19eQmtKLb35ciZeXF99+9BKHjuby8usfsy8jGx9vb6ZOGMW9d1xrDzrhPlPph59/5bNvfia/oJjoqHCuvHQml18yw/l+UXEp/3vnc37bsY+6OjNJCbH88Q83cuz4Sd795Dvg9F8T//TQ7cydOYmxs67juafudwbgztQvx/4MHpDC598uo67OwvQpY3jgzutadOMnhBBCiM719L/fZNfeTHbtzeSr75cDOLOTGtq4JY1pky5g3kVTnG29kno0Wi6xRwwhwYG88f5X/ONP9za7/aDAgFb/scvf35ew0GDCQoN5+N7fMe+ae9iWto8Fc6cxdtZ1PPyHm9iyYw87dh3g2oVzuPX6y/n2x1V89s0yCotLiY2O4KarL2X29AnOdTbMVKrSV/PK25+xYUsapro6UpN7cd8d15LcO9H5mQ2/pfH+p99x+OgJtFpvhgzsy/NPPcBdDz9DQWEJ/3nzE/7z5icAbFn+CUtXrOflNz5h5bdvOddxpn6NnXUdj99/C5u27Wbrzn1EhIVw7+3XMHHs6ftWIYQQoivIE/4ZeGs0VOr0AKxYs5mEHjEugScHhUKBv5+v8/XOPQeoNRoZNWwgEeEh3P7A37jvzmvR+rQuA6n+9gF27N6Pr6+W//zzMQAMtbU88MTzDExN5r1X/k55hY5/vvQOL7z6IX/+4x1u17l8zSbe/ugbHrr7RlL6JJJ1KId/vvwuPj7ezJ0xiRpDLXc9/A8iwkL4198eJDQkiIOHjmG12pg2eQyHj51g6469/Pc5ex/86u23Q0v7tXPPAcJCg/nf809y4mQhf372f6T0SmT+nKmtOk5CCCGE8LwHfn89x08U0DupB7fdcDkAwUHus4BCQ4LYtS+T/MISYqLCm13vXTcv4uZ7/0JG1hFSU3p1eL8dvDX2TCKz2eJse+eTb7nr5kXcf+f1qFRK1m7azktvfMz9d17HqGED2bR1F/944S0iw0MZMbS/2/U++cwreHt78eIzD+Pv58t3S9dwz2P/5Mt3/01QoD+btu7i8b+9zI1Xz+cvD99JXZ2Zzdv3APDPP9/PDb9/gvlzpjJ/dtP3Py3t17uffMfdt17FH269mq9/WMFfn3+dbz96+YwZYkIIIYQnSfCpCTabje279rN15z6umG/PCMo9WUBij5gWff7HX9YxffJYVColvZPiiYuOZM36bcydOanN2wfw8fHmiQdudQ63W7LsV4ymOv7yyB3OwNZDd9/Iw0+9wN23XEVoSOM6AW9/9A333H4NUyaMAiA2OpKjx/P4fumvzJ0xiRW/bqa8Use7//2780alfhq8r9YblUrZ7F8eV6zZ0qJ+BQT48dDdN6JSKUlKiGXcBUPYsXu/BJ+EEEKIbsjfzxcvLxXe3pozZiDdct0CHn/6P1x2w/0k9IhmYGoyY0cN4cKJF6BUus550ze5J9MmjebVd7/gf883PUnK7Q/8DaVC4dK2Zsm7Lep7ba2RNz9cjEqpZNigfs72mVPHcfGsyc7Xf/nnq8ydMcmZEZ7QI4b0jEN89s1St8GnPekHOXDwMMu+fM1Z9/Le269h/eYd/LpxG5fOuZAPPv+B6VPGOAN2gDMrKijQH6VKia9W2+wx/ezrZS3q15yZE5k5dRwAd/7uSr76fgUHDh5m7KghLTpOQgghhCdI8KmBTVt3ceH8WzBbLFitNmZOHcut118GgK2Fpdmr9NWs3bSdN1/8i7Nt1oXj+fGXtWcMPjW3fYDeSfEudZ6O5eaR3CvBJaNq8IAUrFYbOSfyGwWfDLW15OUX8exL7/Dcy6dv1iwWK35+WgCyD+eQ0jupXX8ha2m/eiXGoVKdvgENDwvm8NETjdYnhBBCiO7rmtsepaCoBIAhA/vy0j8eITwshLdf/iuHj+Wye18m+w5k8/S/3+THX9by0j8eaRSAuv3GhVx92yNs3bmvyZpKTz/xB5LiY1vVt7/881VUSiVGk4ngoEAef/BW+vRKcL6fmtLTZfljx082+iPY4AEpzmGGDWUfOY6htpaLFt7p0m40mcg7WXRqmRzmz57Sqn431NJ+9el5et+0Pj74+Wopr9C1a9tCCCFEe0nwqYHhQ/rz8D034eWlJjwsxGWWuYS4aHJOnDzjOlb8uhmTqY5b733K2WbDhtVq4/iJfBKayZ5qbvsA2nbOqmIwGAF4/P5b6N+3t8t7jiCQIyW9M6hVDS9BBVabtdO2L4QQQoj2e+GZhzGbzQB4NyiS3Tspnt5J8Vx+yQwWzD3InQ89za69mY2yiHrERjFv9lRee+8LnnjgNrfbiYoIO2NR8obuu+M6Rg0bgL+fr9ugVntnrDMYagkLDebV/3uy0XsB/n5AJ99bqV3vHRUKBbaW/gVVCCGE8BDlmRc5v2h9vImPiyY6MrxR4Gfm1HEcP1HA+s07G33OZrOhr64B7EPurr58Dh++/g/nfx+9/ixDB/blp+Xr2rx9d5Li45x/cXPYuz8LpVLhdohgaEgQ4WEh5OUXER8X7fJfbLR91pfeveLJPpzjUmuqPrVajcXafICotf0SQgghxNnBS63G2uA+ICYq3Hk/ERke2uRnkxLiADDUGt2+f8u1C8g9UcCqtb91WH/DQoOIj4tu8Qx1SQmx7Nuf7dK2d3+Ws+8N9e2TRFlZJWqVqtG9VXBQAAB9esazY/f+Jrfp7pi2t19CCCFEdyLBp1aYNnk00yeP4S/PvcoHny8hI+sI+YUlbPxtF/c89k927jlA1uEcDh46xryLpjj/0uf4b8aUsSxbtQGzxXLmjbXQrAvH4a3x4ul/vcnhY7ns3H2AF1/7iIumTXBb7wngtusv46Mvf+Sr75dz/EQ+h47m8tPydXz+zTIAZk4ZR1hoEI/97SX27M8iL7+IXzdsY98B+w1PTFQE+QXFZB3OoaKyCpOprkP6JYQQQojuLyYqgv2Zh8kvKKaisqrJoMn//fd93vv0O/bszyK/sIT0jEP8/V9vEBIUyKD+7mcADg0J4qrLZvPVEvdD3Cp1VZSWVbj8ZzSZOmzfAK5dOJelK9fz7Y+ryM0r4PNvlrFu0w6uvWKO2+VHDR/IwNQ+PPq3l9i6cx/5BcXs3Z/FG+9/RUbWEcBe/2rl2i28/dE3HDuex6GjuXz85Y/OdcREhbN7XyZFJWVUVFZ1SL+EEEKI7kSG3bWCQqHgb4/dxffLfuWn5ev48PMfUKmUxMdFM3v6BEaPGMyr73xOz4Q4khIa1yOYPH4kL7z2IVu27e6wKW99fLx56dlHefn1j7n5nr/g4+3N1AmjuPeOa5v8zLzZU/H29ubTr5fyv3c+x8fbm94941m0YBYAXl5qXn72MV5561Me+tO/sFis9EyM5aG7bwJg6oRRrNu0nT888g+q9DX86aHbG9Wyaku/hBBCCNH9XXPFHJ7+95tcffujGI0mvv3wJWKiIxotN2rYAH5asY7vflpNZZWeoMAABqX24b/PP05QYECT67/2ijl8t3S12z9u3fvYc43a/v743cyYMrZ9O1XP5HEjeeDO6/nsm2W89MbHxEZH8ORDtzN8iPuZ7hQKBS888zBvfrCYZ154i4pKHWEhwQwd1JfQYPsf3IYP6c8/nryX9z77no+/+hE/Xy1DB54uen7bDVfw/H/eY+FND2Gqq2PL8k/a3S8hhBCiO1GYTEYZBH6Oe+29L9mTftClALoQQgghhGibi6++m9tvuIJ5s2V2XiGEEKIlJPPpHGaz2cjLL2LHrv2k9Ens6u4IIYQQQpzVamuN7N2fRVm5jp6JPbq6O0IIIcRZQ2o+ncP01TVcc/ujeHmpufGq+V3dHSGEEEKIs9r3y37lz/98lUULZjGof3JXd0cIIYQ4a8iwOyGEEEIIIYQQQgjhMZL5JIQQQgghhBBCCCE8RoJPQgghhBBCCCGEEMJjJPgkhBBCCCGEEEIIITxGgk/YZ4Uzm83YbFL+SgghhBDCHblfEkIIIURbSfAJsFgsTJx7ExaLpau7IoQQQgjRLcn9khBCCCHaSoJPQgghhBBCCCGEEMJjJPgkhBBCCCGEEEIIITxGgk9CCCGEEEIIIYQQwmMk+CSEEEIIIYQQQgghPEbd1R2oLzevgL//600qdVX4+2n500N30Cuph8syO3bv57V3v8RQW4sCBeNGD+WumxehVNrjaBt/28Urb3+G1Wqld894/vzQ7fj5+XbF7gghhBBCCCGEEEKc97pV5tPz/3mPS+dM5av3/s11V17CMy+82WiZAH8/nn7iD3z+9v/x/qtPs+9ANj+v2ghAjaGWZ196m+efeoDF779AeGgI7332fSfvhRBCCCGEEEIIIYRw6DbBp7KKSjKyjzBr2ngApk4YRWFxGbl5BS7L9e2TRFxMJADeGg3JvRLJLywGYMv2PaT0TiQpIRaAyy+Zzsq1WzpxL4QQQgghhBBCCCFEfd1m2F1RcRnhocGoVSoAFAoFURFhFBaXEh8X7fYzpWUV/LpxG//++0MAFBaVEh0V7nw/Jiqc0rIKzBaLc70OX/+wkm9+XAmAzWbzxC4JIYQQQgghhBBCnPe6TfCptaqra3j4qRe4buFcUlN6tfrzV8ybwRXzZgBgNpuZOPemDu6hEEIIIYQQQgghhOg2wafIiFBK6mUp2Ww2CotLiYoIa7RsdY2B+5/8FxPHjuDqy+c426Miw9iWts/5Or+whLB62VRCCCGE6BgLbrgfjZcab40Go8nE3JmTuGHRvDavb+ys61jxzZsE+Pt1YC+FEEIIIbqO3C+d1m2CT6HBQfTtk8Ty1ZuYO3MSv27cTmR4aKMhdzWGWh548v8YM3Iwv7vmUpf3xowczAv/+5Bjx0+SlBDLNz+uYvrksZ24F0IIIUT3UVGhY+fufdQYaumX0pvk3kkduv6nn7iHlN6JFJWUcc1tjzJiyAAG9OvdodsQQgghhPCk0rJydu5Kx1RXx4B+yfRMiu/Q9cv9kl23CT4BPHrvzTzzwlt8+MUP+PlqefKh2wF49qW3mThmOBPHjuCr75Zz4OARamuNrNu0HYALJ47mpmvm4+er5fEHbuHRv72ExWKhd1I8f374jq7cJSGEEKJLpB/I4oNPv8ZsNgOweu0mRgwbxHWLLkWhUHTotiLDQ0mMj6WgqIToyDBefO0j8gtLMJpMTBo7gjtuWgjAf9/6jN37MjCbLfj5anns/ltIjI/t0L4IIYQQQrTUzl37+PSrJVitVgBW/bqRCWNHccWlszt8W+f7/VK3Cj4lxsfy9st/bdT+xAO3OX++6Zr53HTN/CbXMXHsCCaOHeGJ7gkhxDlPX11DTY2BiPDQDg9QiM5jNpv54psfnYEnh5279jF0cH8G9e/bods7dvwklboqhg9O5W//9zo3XDWP4YNTMVss/PHP/2b1+q1MmzSa66+8mHtvvwaAlWu38NLrH/Pys492aF+EEEI0VltrRFelJyQ4EC8vr67ujhDdQm2tka++W+YMPDls3LKdoYNT6dMrqUO3d77fL3Wr4JMQQoiuUWMw8NU3S9m7PxOr1UpoSDDzL57BkIGpXd010QbHjp9Ar692+96+/Qc7LPj052dfQaFQcvxEPvfdcS0+Php27NpPWXmlc5kaQy3HT+QDsC0tna9/WEFNTS1WmxVdlfs+CiGE6BhWq5UlS1ey+bed1JnN+PpqmT5lPBdOHtfVXROiy2UfOYbRaHT73r79Bzss+CT3S3YSfBJCCMFnX/1A+oGDztdl5RV8+Ok3PHjPrfSIjW7mk6I7Uqub/ue9IyfhcNQw2JaWzsNPvcCIoQMAePs/f8Vbo3FZtqCohBde/ZD3Xvk7PWKjOHTkOL//4zMd1hchhBCN/bJqPes2bnW+rqkx8MOyVQQG+DNy+OAu7JkQXa+5e6Lm7qVaS+6X7JRd3QEhhBBdq7y8kv0ZWY3arVYrm3/b2QU9Eu2VGB9HeHio2/dGDhvU4du7YPhALrt4Gm99sJjhQ/rz8Zc/Ot8rLi2nqLgUfbUBtVpFeGgwNpuNr39Y2eH9EEIIcVpJaRmffPEdR47lUlZegc1mc763ccuOLuyZEN1DSp+eBAYENGpXKBSMGDqww7d3vt8vSfBJCCHOczq93uWG1OW9Kn0n90Z0BIVCwU3XXkFQ4OkbKqVSydxZF9KrZ4JHtvm7axawZ38Wd9x0BSdOFnLt7Y9x7R2P8fjfX6ZSp6dPz3hmTBnDNbc/xs33/IWoyDCP9EMIIQTs3Z/JP/71KoeO5FBQWExm1hEys444/72Xf9/PzGw2Y7FYurobwoNUKhW/u+4K/P39nG1qtZoFl8wiNibKI9s8n++XFCaT0f0Tx3nEbDYzce5NbFj6QYem1wkhxNnAZKrjqWdfxmAwNHpv3pzpUhfiLGaxWMg4eBiDoZaU5J4uwSghWkvul4To/vILivjplzV88uX3YIPqGgNK5ekJRJJ7JxERHsqIYYO4/qoFXdfRbqy0rJxvf1jOgcxslEolQwamsmDeLALqBSjEuaWuro6MrMOYjCb6pvSWc+0hcucghBDnOY3Gizkzp/DNkp9d2qMiIxh7wfAu6pXoCCqVioH9U7q6G0IIITpBWXkF/33jQwqLijEYagEwm+uoM5vx8/UFoLyiksSEOGZNm9SVXW01Q0kGmsB4VBp/Z5vFpMeky0Ub3nGTo9TV1fHqWx9TVl5h34bFQtqedAqLS/jjvbfJTMDnKC8vLwYP6NfV3TjnSfBJCCEEE8eNIjIijC1b06iqrialT08mjhuFVuvT1V0TQgghRAts2LQdg8GASnW6sopGo0GlVhMeFoLJVMeA1GT+eO9thIYEd11HW8lQkoHu+FpU3oGEJM9HpfHHYtJTnr0Ei1EH0GEBqN17M5yBp/ryThaQdegofZN7dch2hDgfSfBJCCEEAH2Te8lNlRBCCHGWyi8sAsDP1xc/Xy3VNfbh9CqlkrjYaPx8tdx587VnVeAJsGc8eQdiMeooz15CYOI0dDmrsRh1qLwD0QTGd9i2SsrKmn6vtJy+yR22KSHOO1JwXAghhBBCCCHOclGR4c6fU/r0ROvjDYBSqcDXV8u8OdPp3Suxq7rXZiqNvz3jyRGAyvrOGXhyZEJ1lB5xMU2+F9+j6feEEGcmmU9CCHEOOZh9hOWr1pOXX0hEeCjTJo9j2JABXd0tIYQQQnjYhLEj+W37boxGI1qtD0MH90dXpWfooP7ccuMi/P18PbZtT9dkUmn8CUycRnnWd862wMRpHRp4AhjQL5mE+DiO5+a5tA/s35eEHrEdui0hzjcSfBJCiHNE1qGjvPneZ1itVgBO5OXz4WffYLFYGDl8cBf3TgghhBCeFBEext23X8+y5b9y6EgO/n6+XHzRNKZPHe/RQtmdUZPJYtKjy1nt0qbLWd3hmU9KpZK7br2OVWs3sWdfBiqVihFDBzJ10pgO24YQ5ysJPgkhxDli1dpNzsBTfctXb5DgkxBCCHEeSOgRy523XNup2/R0Tab6gSyVd6DL+suzl3R4AMrHx5uLL7qQiy+6sMPWKYSQmk9CCHHOKCgodtteXFLqNiglzm2GkgwsJr1Lm8Wkx1CS0WHbWLtxOzfd/Sdu+P0TLLrlYf7wyLNtvtaq9NV89OUPLm13PfwM6zbvaPW6lq5Yz7HjJ52vN2zZyctvfNKmfgkhhGiep2symXS5LuvT+Ee7bM+ky+2gPRHnI7lf6rz7Jcl8EkKIc0RUVDi6qqpG7RHhYSiV8reG80lnDIEoKS3nuf+8y/v/e4aYKHuR24PZR9s8tEOvr+HjL3/ihkXz2tUvgKUr1+Pv70tSgr0+x8SxI5g4dkS71yuEEMI9T9Zkcvx7Vb+mlCPg1VE1pcT5Se6XOvd+SYJPQghxjpg+ZTyHDh/DZrO5tM+4cEIX9Uh0lc6YlrqsQodSqSQwwM/Z1je5J2s2bGPJsjX855+PAWCxWLnipgd48ZlHKK+o5IVXP2LooL7s3Z+NxWLhzw/fQWpKL57/73vUGAzc8PsnUKlUvP+/pwHYk57F518vo6SsglHDBvLofTcDUF1j4L9vfkr2keOY6uoY2K8PD919Iz+v2kBm1lH+88anvPvxt9z5uyspr9CxfvNOnv/rAwD8tHwdX32/HJsN1GoVz/7pXmKiI9p9TIQQ4nzl6ZpM7gIAKo2/BJ5Eu8j9UufeL8mfwoUQ4hzRN7kXt910FUkJPfBSq4mNieL6qxZwwYghXd010ck6Y1rqPj3jGTKgLwuuv59H//YSnyz+iaKSMiaPG8nxvAJycu1p3Bt+20lcbBQ9E+MAyMk9yZwZE/n4jWe5Yv5M3vxgMQCP3nszvlotH73+rPNGCiAvv5D//etJPnnzObbu3Me+A9kAvPLWZwwZ2Jf3Xvk7H7/+LFablS+//4V5s6fSL6Un9915LR+9/izjLhjq0u+0PQd479PvefGZh/n4jWd5/d9/IiQ4sN3HQwghzlcNazKFpCxweaBvOKRJiO5C7pc6935JMp+EEOIc0r9fMv37JXd1N0Q34OlpqZVKJf/8y30cO36SXfsy+G37Xj78/Afee+XvXH7xdL75cRUP3nUD3/ywiivmzXB+Li42igH9+gAwKLUPn329tNntTJ88BrVKhVqlIrl3Ann5hQzqn8z6zTvZl5HNF9/+DIDRZGrR8NJN23Zz0bTxhIeFAPbCskIIIdquYU0mxwO9IyAlQ+NEdyb3S+554n5Jgk9CCCHEOaizpqVOSoglKSGWBXOncf8Tz7PxtzTmzZ7KNbc/wuzpEziRX8jEMafrB3hrvJw/K1VKLGcouKmpt7xKqcRisS9vw8Y//3wfCT1iOmxfhBBCtJ7UZOp6FYd/xjdqOBr/KGebSV9ITWEawb1nd2HPuj+5X+o8MuxOCCGEOMd0xhCIopIy9uzPcr7WVVWTX1hMXEwUgQF+TBwzgsf+9jKXzrkQlerMtxu+vlqMJhN1deYWbX/S2BF8/NVPmC0W5/Zz8woA8PPVoq+ucfu5CWOGs3zNJkpKywGorTVSW2ts0TaFEEK4pw1PbfSgLjWZOkfF4Z+pOLyMgh0vY9IXAvbAU8GOl6k4vIyKwz93cQ+7L7lf6tz7Jcl8EkIIIc4xnTEEwmKx8v6n33GyoBgfbw0Wi5XZ0ycyaZz9r3bz50xl2coNzJ89tUXrCwr0Z/a0CVx35+P4an1c6hi4c9+d1/Hau19y4++fRKFUoFIpufuWq4mPi2b+7At55e1P+fLbX7jzd1e6fG7YoH7cfN0C7n/y/1CgwMtLxT/+dB8xMvxOCCHEWcg3aji63PVYTXoKdrxMWP9rKD3wGVaTHqXGH9+o4V3dxW5L7pc6935JYTIZbWde7NxmNpuZOPcmNiz9ALVa4nFCCCHOfoaSDJchEGD/C19nDYH4dPFSjuWe5MkHb/P4tkTnkPslIYTonhyZTtZ6mTpKjT/RI+93GYonGpP7pc4jdw5CCCHEOagrp6W+5rZHUSgUvPSPhz2+LSHEuc1isaBSqbq6G0J0axr/KML6X0Px7recbWH9r5HAUwvI/VLnkeCTEEIIITrUZ28/39VdEEKc5bIOHeWnX9ZwPDcPf38/JowdycwLJ7ZoliYhzjcmfSGlBz5zaSs98BlevpL51J2db/dL8u0thBBCCCGE6DZOnCzgrfc/53huHgB6fTW/rFzHT7+s6eKeCdH91B9yp9T4EzH0dpQaf2cNKEcRciG6mgSfhBBCCCGEEN3Guo1bMZsbz+S0ccsOjEZTF/RIiO6rpjDNGXiKHnk/fpGDiB55vzMAVVOY1tVdFAKQYXdCCCGEEEKIbqSkpMxtu8lkQldVRYR3WCf3SIjuK7j3bMA+651jiJ3GP4rokfdTU5jmfF+IriaZT0KccvRYLmm70ykpdX/DI4QQQgghPC8uNtptu6+vlpDgoE7uTWOGkgws9WYVA/vsWIaSjC7q0dnHZju/Jlz39DUT3Ht2o9pOGv8oCTyJbkUyn8R5T1el550Pv3TWFVAoFIwZNYwrL5uLQqHo4t4JIYQQQpxfpkwczc7d6RgMBpf2mRdORK3u2scXQ0kGuuNrUXkHEpI8H5XGH4tJT3n2EixGHeB+9ixhd/zESX78eTWHDh9Dq/Vh3OgRXDR9UpefV0+Sa0YIu3P3t1yIFvr6+2XOwBPY/xKzZVsa8T1iGTd6eBf2TAghhGi7F1/7iA2/pVFQWMKHr/2DlN6JACy44X40Xmq8NRoAblg0j+lTxnRlV4VwER4Wyn2/v4kVqzdw5NhxggIDmDhuFCOHD+7qrqEJjEflHYjFqKM8ewmBidPQ5azGYtSh8g5EExjf1V3stsrKK3j1rY8xGo0A1NQYWPXrRqr01Vx9xSVd3DvPkWtGCDsJPonzWm2tkfQDWW7f25G2V4JPQgghzlpTJ17AdQvncsdDTzd67+kn7nEGo4TojqKjIrjhmsu6uhuNqDT+hCTPd2atlGd9Z2+vl9Ui3LMXjDc2at++cw9zZ00lMODcPHZyzQhhJzWfxHnNYrVgtVrdvlfnZpaV+ioqdazbtJW1G36jtKzcE90TQggh2mzYoH5ERkhhZiE6mkrjT2DiNJe2wMRpEkQ4g6bqqlqtVkpLz+17ablmhJDMJ3Ge8/P1JSmhB8eOn2j03oDU5CY/t33nXr745kcsFgsAS5au5NKLZzJ5wmiP9VUIIYToKH//1xvYbDb69+3NXTcvIiQ4sKu7JMRZw2LSo8tZ7dKmy1ktWSxnEBsTxd70zEbtarWayMjWBcoNJRlYzDVoQ/s6j7nFpKe2NAull7bb1VCSa0YIyXwSgisWzMHXV+vSlhAfx5QJ7utfVOmr+fLbn5yBJ7DXifr+pxUUl5R6tK9CCCFEe73+7z/xyRv/5MNXnyE4MICn//1mk8t+/cNKrr7tEa6+7RGuu/PxTuylEJ7V1tnWLCY9RXvexVxTah82lbIAlXcg5ppSiva822hGM3HauNEjCPBvHGiZOHYUfr6+LV6PoSSDsqzvKN3/GaWZi7GY9FhMekozF1Ny4FPKsr7vVjMP1i8ubrWYCOw5y6UGlElf2K36K4SnSOaTOO/1iI3myYfvZkfaPsorKkmMj2PwwH6oVCq3yx/IzMbsZkiezWZjb3om06aM93SXhRBCiDaLjgwH7NkGixbMYtEtDze57BXzZnDFvBkAmM1mJs69qTO6KITHbNi8nV/Xb6GsvILYmCgumjGZwQP6tfjzVbkbMJYfQqnxJzR1ERr/KAITp1Ow42WsJj1VuRu6dHp7q9XKpt92snvfARQKBUMH9Wfc6OEolV2fcxAY4M99d93EL6vWk5V9FD8/LWMvGM7EcaNatR5NYDxe/jGYDaUYitMpsZgAqC2z13H18o/uVkW8TbpcZ+BJoVBSfXILgYnT0eWswlxTSmnGFyhV9gkgulvGlhAdSYJPQmAfftcRQ+YUCkUH9EYIIYTwDENtLWazhQB/PwBWrt0ihcfFeWP9pm18+8Mvztcn8wt5/+PF3HnLtfRN7tWidQTET8RQloVCoUSXs8o5c5nGPxabzUpA/ERPdb9F3v/ka/btPz207dDhY2QfPsrvrlvYhb06LTwslOsWXdqudag0/oT1W0gpYChOx1CcDoBCpUEbMZCwfgu71VA2R0BJ5ROKLmcVFqMOXc4q/GPHUH7oJ5Qqjcx6J84L3Sr4lJtXwN//9SaVuir8/bT86aE76JXUw2WZ/IJinn7hTbIO5RAbHcFHrz/rfC9tzwEe+NO/SOwR42x76+W/4uOt6bR9EOe+gf1T8FKrGxUkVygUDBkkf60QQgjRPTz3n3fZvG03ZWWV3P/E8/hqffjPPx/j8af/g9VqxWazERcdyV8evrOruypEp1izbnOjNpvNxup1m1scfFJp/IkcckujmcvU2tAur99z9FguaXvSOZFXQGlZOUqFkrCwECxWK1MnjSUpoceZV3KWUGn8Ce41G7OhDFPlMQC8AuII7jW7WwWeHBwBqPqz3lUeXeEMPHX1tSNEZ+hWwafn//Mel86ZytyZk1izYRvPvPAm773iOj2wr6+WO25ciL66hjc/WNxoHYk9YlwCUkJ0ND9fX665cj6ffrXEOfxOqVRy+fzZhIWGdHHvhBBCCLvH7rvFbftHr/2jk3siRNerq6ujolLn9r2SEvezsDXFMXOZI/AE3WPmssPHjrMvPZPikjJMdXUoFErKKyvRV1dzLOdEu4NPNpuNnbv2sXufvT7RsMH9GT50YJdk/ltMeiqO/Eyd/qSzra4qj4ojP3e7zKf6uuu1I0Rn6DbBp7KKSjKyj/DyPx8FYOqEUbzw6ofk5hUQHxftXC4o0J8hA/uStudAV3VVCIYNGUCf3kns238Qq9XKwP4pBAfJTEFCCCGEEN2Rl5cXEeFhbieHiY2JcnltKMlAExjvEhCwmPSYdLlow1O77cxlJ08WcuJkARaL9VSLhbq6OiwWK9XVhnav/5MvvycvewvlBh+MFjXpBw5y8NBRFs2/0HlsOoOjuLhjuJ02YiBgr/lkKE6nFLptAKq7XjtCdIaurzx3SlFxGeGhwahPFXlWKBRERYRRWNy62cPy8ou48e4nufmeP/PNjys90VUhAAjw92Pc6OFMGDtSAk9CCCGEEN3cRdMnNWpTq9VMn3p6shhDSQa642vtQ6NOzVznmK1Md3wt1flpzmFT9We7c8xc1pWz3Z04WYC7Sfxqa2vRaNqXc5CTm0de9hb6hpczNKYYb5U9+3/Prh3kpH2K7vjaTpuxzaTLpU6fD9gDT+EDriV8wLXOIFSdvgCTLrdT+tIa9We9627XjhCdodtkPnWEvn2SWPLpf/H386WouJQH//xvggIDmD55TKNlv/5hpTM41dapVoUQQgghhBBnhxHDBqHx1vDrui2UlJYTFxvFrGmTXIajaQLjXQICjoLijoCBTWE7HTxIno9V4U2VzyiUNevxMeqcQY/mMqc8xWQyERIchK6qiro6e3DIy8uLwAB/lE3M4txSh4/kUG7wodasxkdtZmhMMRnFoaRGlFFZpiIwsU+HFMy22WxnHManDU8lFLCYa9CG9nUe57B+C6mNGILSS9thx9lms3Eg8xD7M7Lw8lIzfOhAEuPj2rQux6x39Ws81a8B1ZnZY0J0hW4TfIqMCKWkrAKzxYJapcJms1FYXEpURFiL1+Hn51tvfWHMmDKWPekH3QafZOpgIYQQQgghzi+D+vdlUP++Tb7fMCDgqM1TP2CgUvuiCYwnbd8RvvtpOTU1BrxVZgb2DmZWWDW20u0uy9fPeAE8FmAYMiiVNeu3oFarsFrtQ++USiUB/n4MbeekOAH+/hgtanbnRzA0phgftZlhMUUAqDugYPaJkwX8sGwV2YeOotX6MHrkMObOmoJa7f5x1d0xVGn88YsZ3uY+uPPpV0vYkbbX+Xrdxq0suGRWm2bJdvS5fmDScb1J4EmcD7rNsLvQ4CD69kli+epNAPy6cTuR4aEu9Z7OpKS03PlFW11jYNPWXaT0SfJEd4UQQgghhBDnIEdR6PrqF4XWhqeSV6Tjs8VLqKmx11IyWtTszNLz9c9pLplTJn2By1CrjsgOasqEsSMZNrg/4WEhqNUq1Go1kRFhLLxsTrsnxRk8sB/+/n4YLWoyikOd7V5eXvQecUW7Ak8VFTr+9+ZHZGUfwWazUVNj4Nf1m/nim5/a1ef2yjp0lA2bt5NzPI9DR3IoLinDZrPx48+rqdJXt2md2vDURsdKpfGXwJM4L3SbzCeAR++9mWdeeIsPv/gBP18tTz50OwDPvvQ2E8cMZ+LYEdTWGrnylj9SV2dGX13DvGvv4aJpE7jr5kX8unE73/20GpVKhcVi4cJJF3DxzMZju4UQQgghhBCd60yFvLuLlhSF3rp9t9vSHVlHCzDPuxF1+cYmM6c8JSI8jHvvvInvl64k98RJvNRqRg4fzKUXz2z3ur29Ndxx8zUs/vprYpT2iZ98fbUk906iJm893r5t37dNW3dSW1vbqD1tdzoXz7qQ4OCuqa36y6p17N57wHmei4pLKSwuoX/fZLIPHWX40IFd0i8hzlbdKviUGB/L2y//tVH7Ew/c5vzZx8ebHz59xe3nF86fycL57f9yFUIIIYQQQnQcRyHvrhiO1hoNi0LXr/lUnr3E2ffq6pom12EwKYlLnOYMPIFr5pQn9e6VyEP33EqNwYCXWo2Xl1eHrTs2IoAbpoeir+yLwsufmNSL3R6b1iopKXPbbrVaKSkr65Lgk9VqZUfa3kYBRp1OT3FJKT4+3p3eJyHOdt1m2J0QQgghhBDi3NSwkHdnDkdrjYZFoTX+0fagyqm+OwqKpyT3cvt5X18tMRH+bjOnOnM2M1+ttkMDT3D62PgHRdBjyFVNHpvWio2JctuuVquJioxoT5fb7MTJAny1WrfFz2uNJvo2cf6FEE2T4JMQQgghhBDCoxyFlZ0BqKzvGs381R1ow1MJTJji0idH3wMTpjizs0YNH0TPRNeAmUKhYP5F49EfW3Z631IWuATdOjMA1dFaemxaa8wFwwgKDGjUPn7MCAL8/drV57bS+njj4+NNcu8k1OrTMwX6eGuYOmkMqnbOHng+0lfXYDabu7obogt1q2F3QgghhBBCiHOTo5B3VwxHa42mZlKr3+7l5cVdt13H9rR9HMw+jK9Wy5gLhhGp1aE7vsclqFZ/9rzuVt+qtVpybForwN+P++76Hb+sWs/BrMNotVrGjh7GpHEXtKer7RIRHuYMLoaGBKGr0qNUKAgI8GfOzCld1q+zjaEkg8P5Rn5csYXComI0Gg3jRqYyfUxv/KOkZtb5RoJPQgghhBBCnGUMJRlY6wz4hKU4gzcWkx5D2UFUat9uGeBoSSHvs4mXlxfjRg9n3Ojh9VrjAFwKqzsCUGd74MmTQkOCGXfBcBLj4wgPCyGlT0+3Q9460w1XX8a7H3/Fibx8goMC8VKrmXvRhfTpldSl/TpbGEoyyMtYRvqeQ1SURgBqFJYaDMd/Ya8lmGGTbpDfh/OMBJ+EEEIIIYQ4ixhKMijL+h6zoQRtxEDC+i0EoDRzMYbidNTaMELpHgW8HVpayLs1PD173vETJ9Hp9CQmxDU5/KtKX83uvQcwmUz075dMTHRkm7ODzpbZADtaXV0d7328mIyDh5xt8T1iuePma/D38+2yfoWEBPHHe28jJzeP6uoakhJ74KvVdll/2qMrri1NYDwnCnR4q8wMjSkmoziU1IgyfNRmcvMrGazpmnpeoutI8EkIIYQQQoiziCYwHi//aMyGEgzF6ZRYTADUlmUB4OUf020KeDs0LOTd3uFonpw9r1JXxbsffcXx3DzAXvh62pRxzJ4xxWW59ANZfPjp19SdqmPz48+rmTppHPPnTm/1Ns+W2QA9YfW6zS6BJ4DcEyf5/qcVXLfo0q7pVD2J8XFd3YU2cQScTLpc57UVmDgdS20ZmsB4j19bKo0/GeU9CDYX4aM2MyymCIBas5rd+eFMNSo4+/IdRXtIwXEhhBBCCCHOIiqNP2H9FqKNsNdMMRSnYyhOB3BmQnW3YWwdXazak7Pnfb74B2fgCcBsNrN81Xr27s90ttXV1fHZ4iXOwJPDr+s3c+To8VZv82yZDdATdu5Kd9u+e+8BbDZbJ/fm7GYoybAPvz0VzCzPXgIqb6wWE+aaUgp2vEz5kZ8p2vNup1xbsXE9ySgOdWnLKA7FSxtEaEiQx7YruicJPgkhhBBCCHGWUWn8Ce41G6+A01kZXv6xBPea3e0CTw7a8NRGfWtrsWpPzZ5XqaviYPYRt+9t27HH+XP24WPU1BjcLrcnPaPV2z1bZgP0hKYCTFar9bwIPjkCRvU5AkitXY8j4KTyCUXlHYjZUEbxnrexmWuprTiC1aSnrioPBYpOubYmjh7AkLgql7bUiDIumjoKtVoGYZ1vJPgkhBBCCCHEWcZi0lNx5Gfqqk5n6NTpT1Jx5OdGD7LnKsfsefW1d/Y8k8nUZMDDaDI5f26uGHZbC2V7Yn/OBoMG9HPbPrB/Ckql5x9XOyr409ZtOwJGjj44hlvqjq9tVR/qZ8/pclbhFzsWk/6kPeBUU4QmIAaFSoN3YCIKlZfHry2LSY+1cA2jhvQmPCqePHMqfoHhDB2QSL+Q3PPme0qcJsEnIYQQQgghziIWk95ZXBzsQ+3qD8ErzVx8XjzYNTV7Xnv2PTwslIjwMLfv9e/bx/lzcu8k/JsoQj50cP82bdsT+3M2mHnhBOJ7xLq0hYWGcOnFMz2+7Y4M/rRFRw63bJg9pzu6HLU2HIVKg1dAHHXVRXj5x6JQeQGev7Ycdd58A8IZd/GD/OGeB5m28EkiouKddd7E+UWCT0IIIYQQQpxFTLpc6vQFgD3wFD7gWsIHXOsMQNXp8zv9wa6zs0cazp4XkrLA5SG+rQ/VCoWChQvm4NVgSFBSQg/GjR7hfK1Wq7nh6svw9vZ2+eycmVNJSujRbfbnbKDV+vDA3Tdz8w1XMmv6JK5bdCmPPXgnoSHBHt92V9faqh8wKivOY8+vr3Hi+GEsSm2bhsTVz56zWeqo059E7RdFXVUeNosJs6GEwJ6zOuXa6ug6b+LspzCZjOf+QNozMJvNTJx7ExuWfiBjT4UQQggh3JD7pe7FUJKBtc6AT1iK88HOYtJjKDuISu3bqQ92Z5qpzRMPmp7eZll5BVt37Ean09MrKYFhQ/q7ve5ra43s23+QWqOR/v36EBYa0i33RzSt4ayCQKfW2rJarXz5+Qf46LY42/aX9uCaa39Hr54JrVqXY1/MNaUYdTnYLCZQqND4x2I2lqPxj0WtDSUwcTq6nFVybYlOJcEn5GZKCCGEEOJM5H5JNKVh1k5g4jT7kB4PF8x2TCVff90Wkx6TLvesfJjurP05147bmdhstjPW4TLpCyjP+s75OiRlARr/6FZtp63Hdfu2rRza+gE+6tMzJ9aa1eSa+vLwg/e1uIZY/d9Dq8WEzWbBXFOMxi8GGzaC+1xC9cktzoCTJjD+nD3nonuSOwchhBBCCCFEmzmG0jgefB0P8Z7OHnH30NzW2fPOF2fKsAL3x/VsU1trZPWyT9iyN58aIwxITWH+3OmEBGgaBVyaqrXVmmu3rcfVYtJTefgHfNRmas1qMopDSY0ow0dtJp6D5OcdI7ZHzxb1wVFjSeUdSHjyfEy6XFQ+oc4MJyxGQk61O/pyLpxrOP8CqmcrCT4JIYQQQggh2sVRa6Z+9sj5MFNbRynP20350TX4BoQTkXqZx4JCjhpH5tpKyrK+JyhpukuWmqdrHHWWbz5/Dd/aAwwIU7M7P4J9+zMpOJnDLbOjUFiqAfvxbC5rrzx7SZMBqIbBDk1gPAq1D8bKHMqzlzTK/mvquJp0uWiUJmrN9n4aLfb/D40ptmdC1RYCLQs+Oa4PR78cr8/VgJPD+RJQPRdI8EkIIYQQQgjRLh2RPXI+slqtfPfjcnbu3M6giHz8vLOIys1j2JTfeSQoVFltZulOG966fWi9LIQe2EtSYjxa/9AWn6vunmVy/MRJ9maXMzRGjY/azNCYYjKKQ+nll09hfi1x8b2cx7N+tpBj/+tn8bnbJ3fBDgezoRSgxdl/2vBUQnpdyNI9WzBa7I/mjgBU3wQ/YvuMadW+n4/ZgBZzDQq1jzNgGJg4jYojP1Onz8c7KPGcCaieC2S2OyGEEEIIIUSbnc8ztbXX8tXr2bB5OzVG2J0fQbVRwcncI2RuervDa2bV1dXx6lsfs2v/EQ4UhWK12igpLSfjYDYBCRe2OPCkO77W5bw6zr/u+FqPzW7YGsXFpc4ATq3ZHoAaFlOEj9qMvhaX49mWGdmamiHPZq7FJzQFldfp49iS7L9Bo2YzZux4l9pOvgFhzFlwc0ccjnOaoSSD6pPbAJwBqLKMrzAUp2M2lOIT1k+C392IBJ+EEEIIIYQQbhlKMhoFjywmvUuQoWH2iMY/2jl9vCN7pCPVGAyUlpVjs7Vv3qSW7Junbfptp/Nno8Ve8wegsKgEsAcvThRW8sXXP/LOh1+ybuNWjEZTm7a1Z18GpWXleKvMpEaUOdurqw0c2/1Ni4KEjsCLsfI4pZmLnYEXi1GHQu2Dtc7Qpr51pNiYKMD1eDp4R09oFIzQhqfar+F6+6/S+KMJjHd7LTiCU84AVNZ3zv1XqDQoVF7OZXU5q1t0XBdcMosn/ngXiy6/mFtuXMSfHvkD0VERjZbrDtdsd+K4Hm3mWmwWE5a6Gucsfz6hKWhD+3Z1F0U9MuxOCCGEEEII0UjF4Z+pLT+EWnt6SJZJX0hpxhcoVRrA/uDesNYMnH5A78ihWEajicXfL2PXnv1YLBZCQ4KZN3cGQwe1fv3doU6M1WpFr692vq4fFDKZ6gA4uPVzPllTgdGsAiD9wEG27tjNvXfehI+Pd4u3ZSjJoKw0H2+V2VlPqM6qpLDKl3C/Wmqry5qtceSg0vijDU2luiANs6EEq0mPQqlGofYBQJ+/FaWXtkuHecVERzJkUCqZB/a5BNm0Wh/ifU9gMeld9rEt10LDGmc2qxksJrBZW1U7qr6I8DAiwsOafL87XLPdjeN7pjRzMYbidGwWe2DWHgTUdHHvREOS+SSEEEIIIYRwYSjJoLYsG5P+JGaDPTBRW36Ygh0vYyw/hNVicqmlog1PbfRw3dG1Zr74+kd2pO3FYrEAUFZewUeffcPxEydbva6mhk41rLFUUlrG4SM51NYaO2w/HJRKJb2SEgBcgkK1ZjXVfmNQePmTczSbodFFeKvMzs+dzC/kt+27WrwdR9Cih+YgUf7VzsATNugRpOekzhe/wMgWZ6n5hKXgE5qCzWLCWHkMi9mAzWLCZq7tNkXLr7lsBpeP1xLsr8Km8kUVPZ1BQ0aAWd9oKGhLr4X6GtY4s5j01JZloVD7eCz7ry39PF/YLCZsFhMKlQZNcE/79WmulWG/3YxkPgkhhBBCCCFcaALjUfvaszBM+pNYrXUUpr2GzWJCqfEnLPWqTq2loqvSsye98dAiq9XKpt92knBFbKvW17CwdMMC0UaLio/f+4yMg4cA0Gg0zJ4xmamTxrZ/Z+q5ZPY0Xnv7Y0J89M7A08GKBO66ai5GQyVVhs34qM2EaGsp0J8+3gezjzBlYsuKUTuCFqHo6Ben5HCBP1F+NXiprNSa1QTHDqLnyCtalaXmyCyxWUyYKo6iUGnQRgzsNgXmrTX5JMYE0yspwW2WUP19PdO10HB/mpohr6GOzv5rbT/PF7WlWfbAn0qDd1CSPRPvVNZTU0XjRdeQ4JMQQgghhBDCRf0HXau1DlPFUcAedIgccisa/6hO7U9VVTVWq9Xte5WVujats+HQKThdIPrLTxaTcfAQRpOJgoJiqqqr2Z+RhdVqZdqU8W3anjs9k+J56N7b2LhlBzUVBwkM78N94ycSEhJERYUXuwsiCfExuASeAPz9/Vq1n45zObR/EuEhBZSW1lFn8yY8eRZTJk9F5eXVogd0R+DFUVzbYtJTpz/pzDzpLlo7FLS5a6Gh1syQ19HZf63ppyeZzWbMZkurhn56gsWkx1CWgVobjpd/NMG9ZjuHOyrUPvjHjJbAUzciwSchhBBCCCFEIyqNP36xY6lOe9XZ5hUQh/7kb3j5RXXqA29UZBh+fr5UV9c0eq9nUtuGGzUcOgX2AtHePWayb/9BjEYTe/dnUld3esjbv//7DpGR4Qzq33GFjKOjIrji0tnAbJf24OBAevXu68y+clAoFIy7YHirtlE/aJHYI5bEHrGEpCxA4x/dqvU4Ai+OGk8qL1+UgYkYdTnUlmVhKDuIf/SIVq3TU9wFHZoKBjV1LbjLKOqsGmfutKafnmA0mliydCXb0/ZSV1dHQnwcl86dQa+eCR7ftjuO61Hp5Utwrzlo/KOcgUBzTSl1NYVd0i/hntR8EkIIIYQQQjRi0hdSvOft07VUgpIw1xQ7a0B1Zi0VtVrNnJlTAaitNXL46HF27T3AsZwTBAcFtHp9DYdOhaQscNbTqcj+AS+Fibz8QpfAE0BdXR0/Lms8xMpTrl10Kf1Sejtf+/pqWXT5xa0OuDUVtGjtOdSGp+IXewGAs8ZTaOqVaCMGotaGUVuaedbV2GnuWmjqOu+MGmcd0c+O9tniJWzeupO6OntB/OO5ebzx3meUlpV7fNvuaMNT7UPtFEp0OaucxeQDE6djw4ax8th5OxNgdyTBJyGEEEIIIYQLi0lPacYXWE16lBp/IoffjcY/Go1fjLMIeUcUUW6N8WNGcM2V8ykoKqamxkBYaDBxsdF8vvhH1m/a1qp1NRw6Vb9AtJfSRM9oDVVVjR/mg4ICKSouoareLHWe5O/ny523XMtfHruXB/5wC3974n7GjBrWqnV0dNBCpfZ1Bp4cxy6s30K8gxI7rLh2Z2ruWuhO+9PV/Swrr2BvembjfplMbNyy06Pbbk5A/ETU2lCXQuy6nFUoVZrzvhB7dyPD7oQQQgghhBAuTLpclCoN3iF9CEu9Co1/FBq/SMqzlwDgE9KnS2qpnMwvJCmhR6P2FWs2MG70cNTqlj3enGno1ITpKr7/9R9U1xicn/HVaomJikCj0eDj3bnTuIeGBBMaEtymz7amRlFLOJa1mE8PgXSss7Y0q019NJRkuJwLsAfNOqNYdFcOo2uNru5neXklNputifcqPLrt5kgh9rOHBJ+EEEIIIYQQLrr6QbcpuSfy3bbr9dWUlVcSGRHW4nU1VxNoYDg8/tBdvPi/dzGZ6ggM8CMyIhyVSsmYUUPx8vJq8z50Nk+dy+qT26gtzXR5wDeUZWAx6ly2eyaGkgx0x9e6BAssJj1Ht39MYUEOhaZE4pLHMmRgP3JyT6JSKemVlIBS2XGDeFpTH6ordWU/o6Mj8FKrqTObG70X3yPG49tvTncpxC6aJ8EnIYQQQgghRCPd8YE8JCQIjjZuN5rqyD50lJoaA0mJjTOj2mLyhNGoVCp+WbUOvb4atVrNBSOGMG/O9A5Zf2fq6HOpCYx3GboXmDjNOctYa4c6uVtX+qaPyTmSRa1Zze78fPJ//Ziy8gp690xAoVAQHBTITdde0WHnurW6MlOrq/j5+jJl0lhWrtng0h4aEszY0a0rgN/RuroQu2gZhclkdJ87dx4xm81MnHsTG5Z+0OJUXSGEEEKI84ncL4m26OiH9OMnTvLyq+9htVoBsFptZB0+isbLi16ninD37pnILTdeia9W2yH7YDabKSuvxN/ft8PWeS6oX0vKoa1Dneqvy2yxsGPXPmqMSnbnR1BlsJG2Ox2rzUZKn56Eh4UA4Ofny18fv6/Ts9CaytRy9D8wYco5G4AC2LZzD1u2plFjqKVfSi8unDyOoMDWF/3vKA1rmjUMhEoAqvuQguNCCCGEEEKIDud4SK9f1NrxoKg7vrZNs1Al9Ijl1hsXER0VCUBxSRk+3hqXOlCHj+bww9JVHbMT2Gfai4wIk8BTA46hTvW1dahT/XVV6auxWqxkFIditKgpKS3DeqrWUEXl6UBXdXUN+zOy27EHrWMoycBi0rtkapVmLkaXu9El+HGuF7i+YMQQ7rvrdzz+0O9ZcMmsLg08QdcXYhctJ3+2EkIIIYQQQnS4jhyaVV//fsn075dMjcHAi6+8S0lpWaNldu5OZ9HlF6NQKNq7Gx3GZrNRVl6Bt7c3/n6+Xd2dduvIoU711+XIrEyNKGN3fgQWi8W5XMOsy1qjsVXbqaurY9feA+QXFBMdFc7wIQOazJwym82sWbeF7Wl7CVaXMDzRSELPFKIHXEFI8nxKMxdjKE6npiANL/8YvIMSJcumC3TX+nSiMQk+CXGWs9ls3erGSgghhBBnD6vVyoHMQxQVlxAdFUFq3z4ddl/h6VmofLVa5/C7hiwWS7e6RzqQmc23Py6npKQMhULBoAF9WXT5xfj5tj8I1VFDG6trakjbvR9dlZ5eSfH0S+nd5PFrbqhTefaSVp3fhutKSFlA5uF/g6GSoTHFVFX7k5sHCoXCpaC8UqmkX3LvFu9fpa6KV978kJKS08HKlWs28ofbbyA4OLDR8p988T279x0AQKeyEKoup1K3E7VKRUivGdgsJmwWEwqVBpWXvxS47kLdsT6daEyCT0KchYxGE0uXr2Hrjj3U1dWR2rcP8+ZMJyoyvKu7JoQQopt48bWP2PBbGgWFJXz42j9I6Z0IQG5eAX//15tU6qrw99Pyp4fuoFdS1xTtFV2rSl/N6+98wsn8QmdbQnwcd95yTYcNMfP0LFQD+6ewftO2Ru0DUlM6dDa09igqLuW9jxdjPjVLmM1mY296JrVGE3fdel271n2m+kPQslnncnLzeOPdzzAYDM62fim9ufXGRW5rvDUc6tQw0Ngw8NVcgAxotK7Rs+9n2y//gZoKEsKVFCfE4a3R4Kv1cX5+9owpboNGTVm6/FeXwBNASWkZPy1fw3WLLnVpLygsdgaeAIwWNbvzIxhKMUUFOdhMizFWHkOh0uAdmIhC5dWirK+CwmKWrfiVg9lH8ffzZcyoYUybMq7bXKsNNTxvhpIMVD6hWGrLnOf3XC+0LjqOBJ+EOAt99Pm37M/Icr7en5FFTm4ejz/0+w75C5oQQoiz39SJF3Ddwrnc8dDTLu3P/+c9Lp0zlbkzJ7FmwzaeeeFN3nvl6SbWIs5lP/2yxiXwBHA8N49fVq7jsnkXdcg2PD0L1cxpk8g6dIyCwiJnW3BQIPPndp8Z6bZsTXMGnurLyj5CUXGpSzZPa3XU0MYvvv7RJfAEkJl1mN+272bC2JGNlm/NUKeWFOgOTJjisq6IqB5cdPVTnDy2l2HeCfwxLppDR3LYm56JSqVkxNBBTc505wiYmHS5znVaTHqKjm3HW6UgRFtLgf70tbdv/8FG62j4ewH2AFRGcSi9eugxqsqxWUxoIwYSkjy/RVlfFZU6/vvGB9TU2I+z0Whk6fI1VOqquOLS2W73pSs1PG8mXS6VR1diqs5H4x8L2M9/awOd4vzVPUOsQogmFRQWuwSeHPT6arbt2NMFPRJCCNEdDRvUr9FDbVlFJRnZR5g1bTwAUyeMorC4jNy8gq7oouhiu/ceaKK99YXA3Wk4nCokZYFLoMRRhLw9/P18+eO9t3LdokuZOmkcV152MY8/dBfhYaFn/KyjgHTDPrelEHpzKnVVbXqvqSGF9TkCPs7jmvVdq2f5Ki0rJ7+gyO17e/dnNvk5bXhqo/W7G+rUMEBm0hc0KtDd1LriU8aRlNgDtVpNv5TeXHnZXC6fP7vZwJPu+FqK9rxL5dGVp7ZXSHn2EvqGFTM+MY++4eVE+58+715ejfMxwsMbXz/eKjOpEWVoNVbncDuFStPoHDRV4Hrz1jRn4Km+LdvSqNJXu/2MOzUGA0XFpS61sDyh4XlTqDSYqvOxmvSY9CdB5X1eFVoX7SeZT0KcZYrdFNV0KCou7cSeiO5OV6VHqVSeE0VNhRAdo6i4jPDQYNQqFWCvoRIVEUZhcSnxcdGNlv/6h5V88+NKwD5USJwfOqpMUmuHZrWVWq1m5PDBNM7PaVpHDVdriZ5J8aTtSW/U7uXlRY/Yxr9323fuZcWaDRSXlBIRHsbMCycyasTgJtff3qGNKqWqyfcc3xXt4enaX/U5AiY2Qxkm/Ums1joKdryM2jeCmBAFOYVQa1ZTbjg9fG/ksEGN1pPQI5bkPj3JPnQUsAeehsYUE+irILLHQPxCe2Eoz8ZmrnVmO52pwHX97Lz6LBYLxcWlBPj7NbtvdXV1fL3kF3bu2ofZbCbA3585s6Yw9oLhzmX27s/kt227MBhq6ZvSi0njL2jzENqG563y6Ao0/rGY9CfR+MWgO7rcvpwHzqM4N3Wr4FNLahDkFxTz9AtvknUoh9joCD56/VmX93/4ZS0ff/kjNpuNEUP68/A9N7kdpyzE2So2OhKFQuH2ISA2JqoLeiS6m7yThXz17U/k5OahUCjom9yLRZddTEhIUFd3TQhxlrli3gyumDcDsM/8NHHuTV3bIdGhhg7uz9btu9y2d4TuPAuVp2bic+eCEUPYsi2t0VCuWdMmoa1XwwggbXc6n371vfN1cUkpn371PSqVkuFDB7pdf3uHNgYHB9IrKYEjx443em9EE9tsLU/X/qq/HUfABMBYeQybxUSdxUREXF+KagrZvN8Lo8X+fJiS3IvZM6a4Xdct11/JD8tWsWPXPsJ8aoiN9Kdnr75EpF6GSuOPb9TQRoHU5q7pqMgIoHEmmUqlcptp1dC3Pyx3+X2t0uv58pufCA0Jpm9yL1au2cjS5Wuc7x/NyWX33gweuPtmvL01Z1y/Ow3Pm0KpJqz/NehPbHQuI4XWRUt1q2F3jhoEX733b6678hKeeeHNRsv4+mq548aF/O2xuxq9d7KgiLc//Jo3Xvgzi99/gbKKSr5f9mtndF2IThMWGsLI4Y3/+hUWGsIoN+3i/GIw1PLaOx+Tk5sH2DMVMrMO88Z7n0nWghCCyIhQSsoqMJ8armGz2SgsLiWqHTVnxNnrktnTiGuQeZMYH8dF0yd32DZaOjSrs3XEcLWW8vbWcM8dNzJn5lR690xkYP++3HrjVUyfOr7RsqvXbXa7jqbaO2po4zVXznMJgCgUCiaMHdVkwKu1mgqQdcTQy4YcAROFUo1XQBwAXv6xqNXeTJl/H3ffdSfXXnkpD95zK3fdel2TgRkfH2+uvGwu//f0Yzz8xD+5YNrNxA25qlEgNTBhSouu5/FjRuDr2zgLafTIoQQGNH+9GY0mdqTtdfvexi07qDEYWLFmg7Mt2l+Pt8pMQWERW3fsBto2pLThebNZzZQe+Aybpc7ZpstZjb5gZ6cMYRVnt26TEuSoQfDyPx8F7DUIXnj1Q3LzClzSwIMC/RkysC9pexqPUV+zYRsTxgwnLDQYgAVzp/HhFz84/2InxLni6isuISoinG0792A0GunfL5lZ0yfh4+Pd1V0TXSxtz36qq2satRcWFXMw+wj9Ulo+JbEQ4twTGhxE3z5JLF+9ibkzJ/Hrxu1Ehoe6HXInuqfmZg1rbUDHXi/pNg5kHqKouISY6Ej6JvdC0VHj7rq5zsrGAdBqfZg5bSIzp01sdrniEvflFZpq76ihjeFhoTzx0F1kZh1GV6Wnd88EIsI7JijdMEBWP8usuQLd7dmeLmc1NquZuir7H+Pq9CdRBiaiy1lNXPJ8EnrEtmqdCoXC7XFsTSA1OCiQe+64kWUr1nIw+wh+vlrGXjDcbRCyIYOhljo3RevBXjfsZH4hdXX2gFC0v56+4eXUmu0z9B3LOcH4Uf1bPaS04Xnzjx1D0Z537DWfyCdiyG1Un9yCsTKH6oI0tBEDCeu30KNDWNuiI78zO3Pd56JuE3xqbQ0CdwqLSomOOj3VfExUBIVF7mvgSA0DcTZTKpVMnzq+Rf9YifNLeUVlk+9VVOo6sSdCiK723H/eZfO23ZSVVXL/E8/jq/Xh6w9e5NF7b+aZF97iwy9+wM9Xy5MP3d7VXRUtZCjJoCzre7z8o10e8kozF1OnLyCU1j/kKRQKBqQmMyA12TOdbidPPtx5eia+1srJzUOpVGAy1aHReLm8F9dEaYWOGtroOM79+52+DjrqOHdW7S9Hn8uzl2A+VfPJyz8Ws6EEtW8Epup8AI8EvNxxd+1GhvpyzZxBaMOvbNW6goICCAsNobSsvNF7vZMSCAwIcL4uN/hQa1bjo7bXqQoKsLapMHjD82bS5aLxi8HEqdnuLEZCkudTmrkYs6GUulOF5D05hLW1PPGdWX/dnVU37lzRbYJPnU1qGAghzkU9E5v+Bz4pwf3MMEKIc9Nj993itj0xPpa3X/5r53ZGdAhrnQGzoQSzoYRSILjXbCqO/IyhON35/rnEkw93rcnGOZlfSEFhMVGREcTFdnx9TV2Vnvc++opjx09QXqHjYPZhYqIjSYy3DxlTKBTNZky1NyPH0w/RnVn7yxEwsdmsaPxiUPuGET7gOnQ5q1AqvbDZrB0e8HKno4+pQqHgktnT+PCzb1wSJ4KDApk8cTTBQYH0S+lNZtZhjBZ7xtPQmGJ8NRb6h53AYvRu9ZDShufN8VrlE4qltsz5OqzfQmojhmAoy/B4QfnW8uR3piYwHqvFhM1Q5hJ0MxvK7NefzP7XSLcJPtWvQaBWqdpUgyAqMoy8k6dnEcgvLCYqUmoYCCHOH/379aFP7yQOHT7m0n7ByKFER0V0TaeEEEJ0CJ+wFLQRAzEUp2MoTsdcW+4cVqSNGIhPWEqX9c0TGUqeLArekmwcdVAfPvzsW9IPHHR+rn+/ZG669opGmUnt8eU3P3Hs+AkAQoLtQYS8/EIMtUaGDEpl+pTxpPTp2WHba6gziq+3N0DW2u1oAuPtmTqnrkln5s6pdm14qkez6jxxTIcO7k9wUCAbtmynokJHUmIPJk8Y7awXdcM1l/HVN0vZuz8To0VNobkXFw20oj1VlqMtQ0obHgfna//TQViVxh+/mOF4BcRSnvUd5toK55BWx/a6aihaw+9MU+UxzMZKFAqV8zuzrX0z6XJRoMCkPwlAedZ32Kxm52yAMvSusW4TfOqIGgRTJ1zAnQ/+nVuvv4zQkCC+W7qa6ZPHeLDXQgjRvSgUCu743dVs3LKDvemZqNQqRgwdyJhRw7q6a0IIcdbpbvU8VBp/wvotpMRisj9IVdinga9fa6UreCpzpmFAqH5GhTlkAh98sYzMrMP4an0Yc8EwZl44EdWpEh5n0pJsnJ9+WeMSeAI4kJnNzyvXMX/u9Fbvjzv66hoOZGa7tAUHBRIcFEhq3z7ccfM1HbKd5jR3nFuTuVJYVMLho8cJCPCjf98+LT4XHc1xbutfc/UDXY7AkyezvVp7TMvKK/hx2WrSDxxE7eXFqOGDmDvrwkbF0JMSe5CU6D6T3Ver5abrrqC6pgaDvhyK1zr3BTw7pNQxhNVcW0Gd/iRmlYaKIz8T1m8hQJcNRav/nVldkEZddSEoVfjFjGp33zSB8ah9w7Ba6zBWHsMrII66qjx7cXvfMMl8cqPbBJ+AJmsQPPvS20wcM5yJY0dQW2vkylv+SF2dGX11DfOuvYeLpk3grpsXERcTya3XX84dD/4dgGGDU1kw98Ku3CUhhOh0Xl5eTJ00lqmTxnZ1V4QQ4qwl9TxazpOZM+6KgivDx/DKm4udE2xU1tWxfNV6yit0XLNwXovXfaZsnO0797j93I60vY2CT/szstmTnoFCoWD4kAH0Te7Voj6YTKYm68+aTHVu2z2hPcXXbTYbi79bxuatO51toSHB3HnLtUR6cCbNI0ePs3rd5lNDIsOZNnkcvXsltuiznZHt1dJjajSa+N+bH1FWXgFAndnM+k3bKCgq4a5br2v1dn3UVgynAk+dVeDd8b3o5R+NWhtKbVkWhuJ0SiwmFCoNNnNtl9d/Uqi8wKICqwVTxTFM1UVUn9zS5nNuP7/TKdjxMjaLyfnHALOhhPAB13XpcMPuSmEyGc/7atuOmk8bln6AWt2t4nFCCCGEEN3C+Xa/1FxNoK6qZeIolOuoV+LlH0vdqSEfXZ39ZDHpKdrzLgoU9oc8OHXcprvUh2nLeusH/ABy8iv4ZpMBo8X1OlQqlfzl0XsJDg5s+46cknXoKA8+/gy6Kj1+vlpiY6IJ8PcFwNvbm+f//qhz2a++XeoSeAGYOmlci7OjnnvxDQoKT5cOifbXU27wYebMWUybYp9cxtMZd2k7t1OS+S0qagkJDiI6KgKNNrhF13na7nQ++vzbRu1JCT24/+6bPdLfrENHeePdT7Farc42pVLJnVeNpXfq6EbZiqX5mfyy6Th792eiUCgYNmQAF88Yiylvhcu11RG/246MSTidVWOz1GGp0+MdlNho/Zu3pvHVtz+5XdeD99za6pn5zhQ4D0yY0qHXUcPtAc7vKZvFhJd/jNv97gyNvjN9ozCUZWCzWVF7h+AT3Au1b1ib+uY4pqbqQmfGU53+JF4BcWj8orq83lV3pOzqDgghhBBCCNHdmHS5BCZOP50dkfUdZkMZVoupyx4qDGUHnQ9R2oiBhA+4Fm3EQPt7xekYyg4293GPMulysdksGHU52Cz2jB2/2LHoclahO74WQ0lGq9fZMAAYkrIAlXcgxupyhsYU461ynXrearVSVOJ+puvWSD+QxevvfILNZqO21khpWQX7Mw5Spa8GYGD/07W1jp842SjwBLB2wxaKilvWl4WXzkajsQ+vivbX0ze8nEl9DYwdaQ8QOI5DW4/jmfz4408c3Pwe1boSispq+X5rHbsP5GCurbQff5O+2c+n7dnvtv3Y8RNuZ2frCMtXr3cJPAFE+urITf/Jpc8Wk57Sg9+xc80HnDz8G2azmbq6Orbt2M1bH31PQILrKJm21EWqzxGIKc1cTGnmYixGHQq1Dwq1N2ZDKcbKnEbHtKi4pMn1FRU1/V5TtOGpBCZMcfmecgwD7OjAk7vtqTT+BPeajXdgIl7+Mah9Qtp9XNuq0XfmoBvwixoBNhtmYzlmY0Wb+2bS5WI2lGGuKcY7MBGVly/egYmYa4rtsy3qcjt6d856EnwSQgghhBCiHscDpC5nFf6x9vqhNqsZY+Uxe0HZLnqoUKl9UWvDnFlOmlPTh2sjBqLWhqFS+3ZJv8A+A5a5phibxYRRl4OlrobiPW9jNpS1ebhNw6LgGv9oQpLn4+0Xio/aTIi21mV5pVLZIcO8fl65FpvNRkKPWHxO1dyxWm2cyCsgNCSYi2edDlhkZR91uw6bzcbB7CMt2l7vXok88dBdzJk5leQBY+iT0p/hA3tSk/MzplPT13tq2vqKCh0ZezfiozZTa7bPkqYzerMu04eSSpOz+HpzLBZLM+9Zm3yvPXJP5DdqKzec7nN59hLnsSsuzKVCb6Hc4OOyfHFBLsd2f+PSpstZfcZgW3Mcw/nq9AX2oIfi1OO2zYo2YiBe/jGNjmlMdGST64uJbtvsitrw1EYBFU8UeHe3PUf9J4XKC7VPCND+49oShpIMt9tQqLxdMkOVGj/U2lCUKm8USrVL32oMBvSnhvOeiSYw3mVWxZCUBfZaT34xMttdE879nGkhhBBCCCFawTGFtrW6kKI976D2jaCuKg+rxYhJl4vKJ7RL+qUNTyWUxkWyw/ot7NKZlewPm6vsMzyRj0KlwVSVa6+toj9JaL8rAfvDYWv62FRR8L7jbua3zFcp0LsWtB49cijBQe0bcmez2cg7WQCAt7eGIYNSKSktp6bGQHhYKI8+cKdLEWit1qepVTlnGXPn2PETlJSUERcbTUx0JMHBgcycNhGAGn0ZWVvep7Y6D23uESLCQ1s8BK61juXmcVLni9Vqpdzg4xzKaLSoydYl0Xtw3zOes0ED+pFx8FCj9uioSI/VfAoPC+FkfqFLm9GiJq+un0u2IkC1Ucnu/AiXYZreKjNDY4ox6NWEhfZ2Dqs1VuZQmrnYZQhra4Y81i80DmCzmMBmdRmS1nBdw4cMYNWvmyhukLU3sH9f4mLbFnzqKs0NV/ZEvSmHpoYa1pZm2mfjixgC2IdB2sy1+EYOwSckGZMuB4tRR/6+r1i+V0n6wVxsNhu9eyaycMGcZmeKNulyUao0zc6YKXUBXUnmkxBCCCGEEPWYdLlgNVNTcgCLqcpez8M3Cqu5Fpu1jtKMLzz+V/ymdHZGQ0s4MpTUvmEEJk7DYigDFKBUodaGU1eV1+ZhY+72NygkkqtvvJdhgweg1WoJCw1hzsypLFwwp937olAoCA0Jdr5WqVRERYbTMymewQP7NZp9bOjgVOeQufp8fbUMGtCvUXuNwcArb3zIy6++xydffs/zL73BB5987cweqqjU8eJrn/HthjLy84s4cvQ4u/cewBYyyiMP7cGBAQAU6P0b1dAy2zT8sO4wr7zxIYu/W9bkMMLRI4cwsH9flzatVsvVCy854/atVisHMrPZvDWN/IKiMy7vyG65sN6kKt4qM9H+9t/HiRMmE5g4zeUzvnGTG+1biLYWH7UZv8BwZ1adT1g/zIZSDMXp1JZmAW0b8ugoNK72CUahtG/XMbTL3e+ql5cX99x5I2MvGE6Avz+hIcHMuHAiN15zWYu21500la3oCAh6Kmu0YQH5+hmD3kGJ+ISluPQtrN9CAuMnEJI8H6UmgH170yg+ccBZ/P/w0Rxee/sTjEZTk9vs7OGN5wLJfBJCCCGEEKIeTWA8KNUolF5Y66rRBCZQV12IT0gyFmMFCoVS/qpdj+M4qHxCqTiyzD67lcUESjU2ax3VxXs6fLariPAwbrz28lZ/zlEMumFB6vrnc8rEMXz7wy+NPjtl0phGbX6+vtx201V88sV3VOqqAPtMbzdcfVmjQBXAkqWrOHw0x6Vt974DxMVGM+PCCSxbsZaqiiKGxpQ53zcaTaRv/IRJlz7iNgDVkn1qSlJiD+J7xJJ74qRLu9FYx8YtO1Ao7K8PH81hx6593HPnjfSIjXZZVqVSceuNi8g6dJTDR3MI8Pdn+NAB+Gq1TW7XUJJBtTWQNz/83pnx460yM25oD+YtvA2FY8MNPuPIbhk2cD6muotZt+5XkrSHCPJTcmGfcQwf1MuZdeQQpT5Mr/gIjuQWO9sK9P7E94ih16gbnMdNG9oXbcRA+5C5sgy8AmLbNAOeY9hZfbqc1c1m/QQG+LPo8otZ1IpL2mg08dMvq9mxK526ujoGpCYzb850wkJDWr6SDtZUtmJI8nyPfmc2zDpyZL3Vz0pqqm9FDGLPib0U6F3Pja6qil179zNm1LAmt3umGTOFKwk+CSGEEEIIUY9K409Y6lWc3PIsNoWSulN/rbcYK4gYchtYjB3+cNGeAEJ3oAmMdw5p0UYMxGYxUVuWRV1VHuaaYrQRA7t89qczzQIG9ofJSeMvwGazsWbdZip1VYSHhTJr2iSGDHR/HpJ7J/HU4/eRczwPhUJBYkKc2+CJzWYjbdc+t+vYsWsfMy6cQFbmfobGFDtrMGUUh5IaUYauotDtsKWW7lNzbrnhSr785icysw5js9mIjYnCaDQ1KhZuNBr5ZeU6br1xkdv1pPTpSUqfns1uq36fd2ccR1fuBaidw+Cspfns2BjJqImXNvpcw+yWkQOmkewXRV2tD14+QQQljW1yyNfCib5sPz6MXelHUCqVDBsygBlTJ6Cql7XmGMLaXADjTDpz2Nn7nywmM+uw8/WefRnkHM/jsQd/j08zQz49rasCMo6MM8d5g8YF5N31oUJvbhR4cr5XqXPbLtpGgk9CCNHBrFYr+zOzKS+vJDEhjsT4uK7ukhBCiFZw1jAKSnSZQlvtG0H1yS3O2i0dpSMCCF2t4XAbi0lPWeZie5F2iwltSHKXTzveMHhRPzDQMLOlf78+RISHEh4WSkR4qNtgUn1KpZKeSc1nxthsNsxNFOc2m+0z90X4m1yKfxst9v+PSqxwW0emqX0yVubg5R/jsk9NBTODgwK54+ZrqNJXU1dXR2hIMH988lm3/Tx89Hiz+9gSmsB4zAotNboShsacDrA59jvtYCmjJjb+XFPZLY56WA2vQZf6S0YdM8ZewLxLmh8K2JIARnOa60NH1gE6cbLAJfDkUFGpY+fudMaPGdGm9ZaWlZOZdQRfrQ8D+6fg5eXV3q52WmC9LRlnAIkJTd+nyz18x5LgkxBCdKCy8gpee+cTSkpOp8sPHtiPG6+5HJVK1cwnhejedFV6tmxLIz+/iIiIMMaPHkFwcPsK+wrRXTWcQluh8kIZmIipOh+l0qvDH5paExTpruoPaQH7Q59CqcY7KAmLSY9Rl4OvaWiXBqBaMjTHaDTx8RffkX7gIGAPKk0cN4pLL555xgDUmSiVSvr3S2Z/Rlaj9wb2TwEgMXUi6dt/aVT8WxM7g8CEPo2uO3f7ZK4tx2woxcs/xrlcS4KZAf5+zp8DA/wpK69otEzQqRpR7aHS+KONn0Xtxq34qM0Mi7HXenIE3HokNP2I2lxwqCOGfLU1gOEIsDT8PXAU2e/oYWdFRSVNv1fc9HvN+emXNaxeu8lZ98jf34/bbrqqXQGYjg6sNxXIMpQdpLY0s00ZZz1ioxk+ZCBpe9Jd2pP79KRfSu827LVoihQcF0KIDrT4u2UugSeAvemZbNyyo4t6JET7FZeU8q+X3+LnFWvZve8AK9ds4PmX32w005AQ54rOnkLb8YBcf5auhtkTZwPHQ2T9YUeh/RbiHZTgfAjsyELtNpsNq9Xaqs84ghf11c9sWbJ0pTPwBPZs5nUbt7J5a1r7OwxcevGMRgGc6KhI50x3s6ZNJDxhhEuB7NS+fZg1c1aTD+kN90nl5Y9PaAo2cy3l2Us4dmgPmZvfpaaqtMXBzInjRrltnzB25Bk/2xKh4THo1K77k1EcitGiZvDAxoXaHZoKDjmuq/YU5G84ZC4kZYFLULipa9cRYHEsU//3wFGovK3DzvJOFnLs+IlG13lMdGSTn2nuvaZkHTrKql83OgNPAHp9NR999q1LW2s1Vwi8tYH1hscZTp+zyiMrMFbmtLnQ+XVXXcrl82fTKymBpIQezJszndtvuqrdAWfhSjKfhBCigxgMtW5ToAF27dnP5AmjO7lHQnSMn1eso0rvetNtMBj46Zc13P67q7uoV0J4TldMod3e4T7dRWcMOzIYaln83TJWr9uM1Wpl7AXDueLS2URGhJ3xs81ltijUvmxP20u0v94l8wggbed2hvfxbXffI8LDePyhu9i5O53iklJ6xEYzdHB/1Gr7try8vLj1xkWczC+koLCYyMhwlwLf7jI/TPpCSjO+RKmy1y9SqLxQqDTU1ZlJ37UDXdVawJ5BpU2Yzbz+pzOcmjJl4hhqDLWs37QNo9GIj48PUyaO7rDgk8WkZ+YwH/anq6mrsw85TI0oo8p3VJNDxjqinlJzQ8CANl271joDCrWPS+ZixZGfqdMX4B2U0KZgdX5BER9+9i0FhUXYALVKxSWzpzF5wmgUCgUx0ZEMHtiPvemZLp8LDw9l+JCBrd5e2u50t+2lZeXkHM8jKbFHq9cJLcs2bKnmMkS9gxLQhqbiE5bSpqw3R4ZjU0FX0TEk+CSEEB3Ehq3Jvw5Z2/FXIyG62sFDR9y3Z7tvF+Js1xkzNjV8CLaY9FQc+RlzbQVqn2CgZcN9upvOOHZPPfsS6zdtw2o9PS36gcxs/vXM42i1Pk1+7kzBC7/EOYR5l9M3vNyl5pK3ykyCTxa64zaXfWwpo9FE1uGjKFDQN7knPj7eZ6zJExsTRWxMlEubuyFMJn0hBTtexmrS4x3Sh7DURc59OnToGNXVOhyDXQ4UhaLL3U10bAKjRw5tdvsKhYK5s6Yyfcp4KnVVhAQHdkj9Hzh9Hnw1FkaOGssxfSyK8m0E+akIi9CgsNYCja/59gY2zzQELDBhCoEJU1p17RpKMtDnb0Wh9nEGoBy1zgCCe85q9e+v1Wrl7Q++ICc3j0NHcsjLLwSrjVVrN3HJnGncev0ikhJ7cMPVl7Hy143sSNuHyVTHwP4pXDRjMhpN68+T4z7VXeDVYtI7hw+2RUcF1tsSyJKZ57oXCT4JIUQH8dVqSe7Tk+xDRxu9N6SZFHIhujtfrZbq6hq37UKcqzw5Y1PDh2CA0szFGIrt2QeB8RMx6nI8MkOWQ0WFjt37DmC1Whk8sB/hYaEdtu6Gx6hhPRxoe8Hh3fsOsHbDVpc2s9lC2p50Nm9LY9rkcU1+9kzBC2oLCIpMptaUho/aPvuaoxh2VGhIm+pv7TtwkE+/XEJtbS3V1TV4+3jz+1uuZUBqSrOfs1qtHMw+gl5fQ+9eCYSGBLvN/CjN+BKrSY/y1AyNGv8oQpLnU5T+BWrTCbxVXtRY7cGn1IgydudHsG3nnjMGnxy8vTUtyihrjYbnIVrjj8U05oxBpPYGNltSW621AYz660ShxGI2UFeVZy+yHzEQn7Dmz7M7mVmHOX7iJOn7D1JUUuoMslZU6Pht2y4sZitPPXYvPj7ezJ4xhdkzprR6Gw0NHtCP45kbGwVeI4J9CDTuQHe8CmjbxAdtraPlztmUIXq2z2DqCRJ8EkKIDrTw0jm89vbHLlOzpiT3YtL4C7qwV0K0z9gLhvHDslWN2sdcMKwLeiPE2a/hQ7AmMNEZeNJGDMQ3aii+UUM9Nsxv647dfPnNT846Mj/+vJr5c2cwZeKYDtuGQ0cXHN7SRO0ls9nC3n0ZzQafWhK8mDs3iHfeKyY1JNdZDNvHx5v4pORWPyxX6av56NNvqNDpyDp0DIOhFoD0/Qf577/+2mRto+KSUt58/3NnDUmFQsGUiWOZP3d6o8wPpUpzKuPJHnhysNqs6I1elBm0LrPJDY0pprguxu1228Nms5GZdZiy8krie8SQ0CO2yWXbGkRq78N8Rw4Ba7hOR/DYZjEBoFBpUJwaBtla1dUGTp4sRF9dg9FowmYDpUqJSqmksKiEGoOBPekZLQ4gtsSA1P9n773D2zjPfO17gEEHQRJgp0Sqi1QvVrOsYku2LMtNtuWaOGXTNskm2ezmbDvn21P27Dm7Z0u2pWzsxCl23OUuF9mSrN5FiRIpUqRYJHYALCDqAPP9AWJIECAJNhV77uvSdYmDmXfeGcwM5v29z/N7ZlNRuhJ/88fKdXKpM5dHbrMQCfaMufDBRKRKDm5vooSs8TJSCufNXsF0MlDFJxUVFZUJJCfbwV/96DucOVuBu7OL4qJC5syarhoWqtzU3L5+De7OLg4dPUU4HEaj0bBi+WLu3rz+endNRUXhZpplHjwI9rWfQzQ50FnzcZTsUI5hoitkQVQQeWXne3EGxrIs8+a7HzG/dDbZWRMb5TLRlfzS04eusmm3Z4y4/UgRbUVTCvjh97/LiaN7MHYdwmIxk5udRdbsraMe3J45e4FAMEjFxRqCwZCy3Ovz84//9gz//o//I67CXIznX3ozrniJLMvs+fQQ06dNYUHpHFqlmQRbPsBqMWNLs+IofSxOeAp2NyIKYSRdNmdqTQTC0UiWJfntGEWJhdPSR3UcI9HV3cPPnn2e5pY2ZdlIlX5HG1k4USLmZEXOyOEgcjiIoNWjsxag1VsVw/fRCiOzZ07D6e6ks6sbSQoDEA6HCWs0aAQBSZKSRiOPB0EQ2PHII9RdXkzbhVfRa0Ksc2SiEzXKOQ92N476GTuRHnATLWSNh5GuR0vBypu+gulkoIpPKioqKhOMTqdjxfJF17sbKioThiAIPPzAVu68Yx3t7U4cjkwyhhkAqqhcayY6umY0+x2r4DV4ECwaM8mYES9wTIZfycUzu9HKfqQBwwCDViLT5OdseSWbNq6d0P1NdLTJbWtu4Xcv7qTD6Y5bbrWYeWDbnRPS5zSThiWFHsJZ05RlY4muCIVCuDu744SnGP5AgFNnyhOKkXQ4XdQ1XEna3sHDJ/joo48o0F7AKEZNutPT05gvfkRWyXalb7FrZn7mZj6teR3CAUWAmjPVwt0bH1LaPHHqLIePn6a318fc2dO5Y8OtCdX4YPhr/fX3y+OEJ4hW+t134Ch3DBOJNhoGi5iarNVcLX8LUQjgyC5MeTA/GZEzPtdF/K4qBK0eg61YMXwHxhS5aLNZ0et0aDSDCtMLoNFoEbUic2bPGFNfR2La9NkUZH8lQZwLdjfGPWOD3Y1ojXa663crz1i9bWrCsU6kB9y1KGaQKrHrMdDVgLPyFTJmbFXEJUE0IsjChEfZfRZQxScVFRUVFRWVlLClWbGlfT5fmFRubCaj2tRIjFfwuh7pI76OCgzecyzJb48z045FxBikq5Oy34mMNimaUsA3vvIEz7/0Zp8fTgRHZgZ/9IdfnhBRfCKjK+aVzEaSpKSf2TPS6fX6EpbHIl0AwpEITqcbn8+P2Wwi4Oti/WwvRlHCL4nRdDrJRVp9NVptfN9MWaVMy4K//NNvc/zkWSUae+mieYpx+Pu79/H+R/uU/bW0tnHu/EX+5Htfw2I2K8uHu9Z9Hic15VV095hJS7MyMM77VNn5CROfBgoNjfWXqD+wD5M2SFfAwMXDGp5+uoviqdZhxd/JiJwJBz34nZVK5OJgEcJSsHLUgsiVphbmzJpOT28vbncXwWAIrVaDRqvBaDSweuXSuAqIE8lQzyVb8WZF/GsrexZZDiN529Fb8hHNDrRG+5DPvonyz7sWxQxSRau3YrKX0ttyCsnXQSToQdCICGK04IGn+SganQmDrZje1tMImqjsYiveBDAuA/ebGVV8UlFRUVFRUVFRuWmZjGpTqTCedLLrlT6it03FkTMF8+WGODNtoygRCIuULEoe9VTfeJWycxVoNBqWLppPYUFu0vWGYqKFtgfvvYtlSxZw7vxFdKLI7FnT0Gg0+Hz+YavdpcJERlfk5WbzwL138U//9kzc8uKphRgMeubOmp6wTW5OFllZdq42tXL+QhX+QACIpt7NKQD93Cz8kl4RD880Z2M2eygqTN63dFsam29P/F59Pj8f7z2UsNzl7uTw0dNx2wx1rV9pqKGyppkDp5vxBsBkNFIydyYmowEgLrVzItDqrXj086mr34tZFyLdGMAT1OHp9fDr51/jL37wZbpq3h5S/J3oyJlwOIzHWUs40I0hvThpm1rRPHJDgzDo9aSlWViysJTGK810ON1I4TAZ6Wncu3UTjz9876jbTIXhnkvd9buxFW+mu343stdJsM9UPUgzGbO2KRFQk51SZsoqpbvlHLUtISLoKZk7E7PJit429ZoLOkbHHIz2Ofjaywl01aFLK0QraECOoDXYkHwuXH2CuyF9GoJGpLN2FwCy5FeO5/OEKj6pqKioqKioqKjctExGtalUGE862fVKH9HqreTOe4SZrk5qqi+wND+aKhWI6Che/hRZOYUJ27z7wR4++mS/8vfuPQd4YNud3L5+TUr7nCyhrWhKAQV5Obzyxi7e/9mviUQi6HQ6Nq5bzbYtt4+6vRijja4IhUKcKjtPbV0jtjQrq1cswWHPVD5//OF78Xp9vPXebgTA4cjEajGzbPECZs4oTti/IAg88fB9/PAv/kYRngAyM2xUN/ew77yEKaOQQDg6jAuERSo7p3JX0cZRXTMtre2EQonpgAANV+Ij4JJd6z2eXiouNXOmORu9sRdvoBuf38/F6lqWLIz2Y+H8uSn3JxXCQQ9Xz78NQCCsJSxryLN6WV7QSnlbmMsnf0e6SRhSAJmoyJlAIMhb7+3m2MkyQqEQy2aZWbdpE1kpthkKhXC6Okm3pSUVS3NzsiieGr0XB0fz7Xhw66h9RFNNDx7puRT2u5QIRoOtmEB3PaI5m67LH0aP+xqklFWe3k31iVfx+AXONGcTEYw8+sAmplsuXRcj75ixvBwOEuy8jKDVY8pegK14M52178WtB8QVlvg8+j6p4pOKioqKioqKispNy2RUmxrNvseSTnY90kcGDkDnrHycDOsruF1OCPsoWP51HHmzErZpbmmLE55ivL3rY5Yumk9GxshpbpMptL334V6OHj+t/B0Khfjok/1kZqRz66plY2oTUk8TCgSC/McvfktDY79Ys/fTw3zty48zd4Anz1e/uIM1K5dy8kw5kXCEhfPnsmTRvCH3P3NGMdOKp2Aw6AmFJGy2NDLS06i6dJmT1b2sWBY/hJs/f/Goz2FmRjqCICDLcsJn9syMhGWDr/X2DhcV7XYCYZHp06ZwvqKaYDCE1+vD0+tl4by53LF++JS70XimxURMIeJVUg4X5HZQaPNQaPNgECUiQRvajCnDCiATkQL2/MtvcLa8Uvn71CUv5Q2v86ff+xo52VnDtrl3/xE+/GQ/Xq8PnSiyasVStt93V4Ix+5eeephnfv0STc2tAOhEka13bWTWjGkp9xNGlx480nNJb5uKu/pNAAStDp21gFDPVTR9UT0TYdw+HIFAkBfePkZppqBU5Ktot1Nz7DlybynBnJZ1zQSd2DmUJT9G+xzCQQ8hT5NiPB/quYos+TFlLwCikU5yuF/sNecs+lz6Pqnik4qKioqKioqKyk3PRFabSpXRpJP5OiqIhHwYHXOUgWk46MHTchKtaMaUVTopBuOxfccGoNHUmY8RNWDT9SBrggQb3yVs/4OEPp+vqE7aXiQS4XxlNWtXLx9x35MltMmyzOFjp5N+dvjYqXGJT8kIh8OcLa+ktq6BjHQbK5Yv5sSps3HCE0BIknj1jV385Z9+Oy5CpXTuLErnJgp8Q2E2mynIj09vnFY0hdq6xrhl04unsnkMJvEZGTaWLJzH6bPn45brRDHp9zr4Wg+Hw5Rmd3OmORuMRpYumo/T5SYQCPDwA3dz9+YNyvHHvqsTp84SkiTml85m9bwMAi2HUvZMi4mY9qxC3jyhIxAWOXk1F60QodDmwWKQSU+zTroA4nS5OXf+ovK3JEnUNVylw+niUm0dd96+ju333UVebnbCtqfOlPPGOx8qf4ckiQOHj6PX67j/ns1x69ozM/gvP/gmdfVX8PR6mT5tSpwPV6qMNj14KHEuJjzFtrMUrKG97BfI4SCBrjoM6dMm3bPufEUV3b0SZ/zZrC2+ikUfVKI321x+Fi3YPCHifSqiaOx6jHk8aXVmNH3RYH5XFZb8W7AVbURvmxq9rqt2Imh1GGzFmPOWYc0b+dn5WUQVn1RUVFRUVFRUVG5qJrraVCqMJp3M11GBq+oNJF8HpuwFOEp2ACjRWqLJgZ3JSxeJDUAlr5OWEz9GNGcjedvRWQuQfB0IgiapSKfTDT1U0PeZVqfCRBkODyQSieDzJRp2A3g8vWNuNxnBYIifPvM7Ltf3Cz8ffnIAs9mUdP32DicdThfZWY4x73P5kgUcOHw8bpnBoOfbX/8CRVMK6HC6mVKYR8mcmaNOw4rxxI77MZmMHD9ZRkiSmFKYz4P33pnQ72TXekbni7R3uPrN6xHJyXZgNpvYvHFtXJ9e2fkeh46eVP5uvNJEVWU2j2+wjloUybJNZUnbHiXiLRzR0BUwMqV4NhqNZkQBZDwVKgFc7q64aLHKqlq6ezwA+P0BLlbX8B//+Vv+4k//ELMp/vo4cPhE0jYPHzvNvXffkVjdDphWPGXEPg3HeNKDB56rgWKLPr2Y3qbD6K0FBD1NiKYsZDky6Z51oT7z/kyTH40gk2kK4PQaCUc0dGunx1XeG+uzJdVIMVNWKWHJi7ftLLLkV+6LaKGLZvzOSjJnPwAQJ9oKWh3B7nrCwSVxvw+xSoGx8x27JpNVELyZUcUnFRUVFRUVFRWVm5bhqk2FQ71YC9bEvbiPZqA5HKNJJ9PbpqKz5iH5OvC1l9PRlxrod1UBoLPmT2q6SKxvbWXPEgl6CIWDGGzFiGYHWfO/oAzaBp+XJYvm8faujxMqthkMhgn38xktWq2WGdOKqK1rSPhsdhIj7/Fw4PDxOOEJIBAIUFd/hfy8xAgXQRAwGAzj2ue9d99Bc2sbNbX1yrLSubO4566NSrW68aLX63j0oW1sv+8uQpKUIJbESHatz731qzS3/h10tpFp8tPisaLRaNjx4D1x/XO63Bw+diqhzdrGdq4Eb6HIWJWyKBK7Np945D5WLJpFe+Xr6DV55OYXkzP3nhG9xMZboRKiJvKiKCJJEj2eXkV4ArBaLAD0eDwcP3WWDWtXxW3b7fGQDJ/PRygkYTBMTprwWNKDB58rU1YpkZCP3vYyeho+BQQM6UXYSx4l7HfFRUZNllhSOncWoijSG9SRZgihFSI4zH46fQaKhGNI3iJEs2Ncz9LRRIppRbMiPMWuJ0fJDuU8+FwX8Tsrh52gCHY30t2wl0g4iICAaHYoxu6Sz4UsR9D0TaR8FgQoVXxSUVFRUVFRUVG5aYkNjAdXm2ore5ZgVz29Gh2W3MVjGmgOx2jSyWKDEidRw9mY6WzMnNZRsmPS/T+0eiuO0sdoP/tLtDorglaHrXgTemvukClw6bY0vvD4dn7/ylsE+syvTSYTTz+xHaNxfOJKqgwXqXL/PZv5yTO/IxgMKp+lWa1s2bRuQvswVPqh3qDDHwhgHCQ0lc6dhS1tfN+n0Wjgj775JeoartDW7iQ/L4ephfnjanMopK5LGAYN2AeKtMmudZ3Rxsbtf0bluYOEGiPMNxlZuXwxuTlZce3UN1xN6isFUN/kYsHto/dMAyi0Q9oU+6i8xMZToTJGmtXCultXsufTQwQC/ded0WAgJ7s/YszpdCdsO2NaER0droTlUwrzJ014grFVm0x2rnyuCmTJjyG9GKOjBJN9bnR7azQ9dDI96wBsaVYe2rae+lPP0+XXk26EnoCOomwRkQDB3mbspY+N61k6mkixkX4DgBEnKGLnWfa5CHqaiERCcdGpekv+pFcQvJao4pOKioqKioqKispNy1ADAEfp4zgrXkSj1Y95oJnqvgcyMJ1soHCi1VvJmLGVkLedgPsSGq0BnbWAjBlbJ1x4SibYBD2tOCteQjT2V2IbOAAdasC4ZGEpJbNncPFSLRqNhpLZMyYs8mYkRopUyS3ayI++/3UOHjlJe4eLwoI81q5eTrotbUL7MVT6YZY9k/VrV3LwyEklOmx68VQef+S+Cdv3tKIpTCsaX+rVcAIekHKa0WB0RhsLV2xl4Yro34FAkHA4HGeency8XPnMZhi1KBJjLF5i40lBG8j992wiJ9vOx3sP0nClicx0G4UFeYhi/3FPLSxI2O6uO27jQmU1PT290dRQIVrN7r6tm1La71D4/QEiciRp5NpYq02O5VxNlmfdQJbOcVDAXFpdPno0Myi1XcFi0CmV98J+lyKGjZXRRIqlklI80jUaO88Aga66qGH5gOjUya4geC1RxScVFRUVFRUVFZWbmmQDAL01l5zFfzBur5MYo03XGyycALir38LvrFCqHoU8TXTW7prQyKdkgk3Q00rLiR8TCXowZM7CUfpYSgPQGEajgcULrn3Kx0iRKk6vgYgmwAPb7hyz7xGM/H2vXL6YyqqahO3mzp7B9vu2cOcd62hovIrNlsaUgrwx9yMZvV4vp86cx9Pby+yZ0ya82pmlYOW4o4Eu1zXyxjsfUt94Fb1ez6pbFnP/PZvR6XRMK55C0dTCBGP2DKuO2el1hAO+UYkiAxmLl9hYK1QORBAE1qxcxpqVy5hWPJUTp87GfZ6Xm8OSRYl9yHLYeeTOEv7tVx9Q3+zCaDBgz8xAJ4TwdVSMWrjp7Ozm5Z3vUnHxErIsM3vWdHY8eE9cBNZ4qk1OxLmaaExZpeQCBUZ7X7pw1ITdYCtGRp6QSYWxRIoN19/BDL5GB55nXVohwc7L6KwFSnTqZ0V4AhCCwUDyOMjPEZIksW7bl9n/7nOIoqrHqaioqKioqKgM5mZ9Xwp6WuIGT5lztqO3Di0QjDRYtxVtTGmQOHAbQTQSCfXS23ISORJGNGdjyJhBsLMWYEJT75JFOjgrXopGW+mt5N3yA/TW3DEd0/VgcBQOgDeo5bXDfq60dALR6JonH71/1MIMpP59v/HOh3x68BiRSASAwoI8vv7lx8lIt03EYSblcl0jP/vlC0rKI8DihaV86cmHk5pTJ2O4yJd4YTT+HKcaDdThdPF3//xzQqFQ3PJlSxbw9BMPAdDj6eWVne9SfqGKSCTC9OKpPHDHLEze8oTzfrXsRRrqqqnqsBMyTGXNqmUsml+S0rGO9nyM9liTEYlE+PTgMY6dLCMYCrGwdC6b71ibtDKd++oZ9r/7M3oDQtSkPSxi0ErcMtXN2hUlZE67I+X7MBKJ8H//6We0tXfELU+3pfFXP/ouer1OSXf0OyvHJKZP9LmaKFK5psfav8lse6R9Sj6XEvkUK57xWYt8UsUnbt6XKRUVFRUVFRWVa8XN+L40lsHTRA4+Ym0FuuoJdNUTkXzoLHnkr/wTtHpLfLW7OdsnTABKdtyRcBBH6ePoB6SkTJT5+mQzUECMyDK/+7iTq85+E3RZlvH6/OzYfg+zphczY3pRym0P9X13Optp7wzQa17JkqXLyXLYcbu7qGu4Qka6jenTJteDRZZl/vc//EdSj6CnHn2QFcsXpdxWKvfBaEXaGG++u5s9nx5KWC4IAn/9F9+PE+f8/gBSOIzVEhVmBkacSZLE8y+/xUsvvYzd7Ke2XcvUwnzy83K4b+smNm1cm/LxDsX1EBYGcvjwQepO/A6jKOGXRCra7ZRmuzCKEjNnz2P++q+nvP/zFdX84rnfJ/3svns2c+VKM+cuXASiguUD2+4clQ/Z9T5XwxETjAc+0wYKOMbMWWTM3Dqutsc7+ZAqA/sdqxwo+TriPJ8+SwJUapK5ioqKioqKioqKyk3E4MFT5pztcelF4WDyylOxtBRl3aqdYx5wxdIpRGNm1KQ3cza5S7+J3pqrmJA75j85ocLTwP0OxFH6WJzwBCiGtwMJBz34OipG3IevoyLhHKa67WgYnALj7uxiiqkRgzYqPoVCIc6WV1J2roJnf/0y//qz5/pMyENDNRlHsu/7UvUFjpfV8MaRAO98dJi//YefcPzkWTIz01m6eP6kC08AzS1tSYUngLPnK0fVllZvRW8rVtI9AeX6iH2PydKMhrpHBn73Tle0jwatRJ61f31ZlnG7u+K2MxoNivAE0ZSk2P305rsf8e77n9DR5aeqGSQpzOX6KzhdnXzw8X78/gDjZXAKmt6aF/fdxzywJgtvAM40Z+OXRIyixNL8NkWI6jQsG9WzxeXuTLpclmWef+lNTp89jyRJSJLEydPn+NmzzyuRUKncu9fiXPV4evl470Fefv1dDh87lfI9a8oqxZA+DQEhmnoX9PQ98zYjyxECXXVjfg6ZskqxFW2Me9bHnhGTESEaO8+yHEFvyUdvzYtGp1py0VsLkJGvybV5rbg5pq1UVFRUVFRUVFRURsGN4HUycFAfM/r2NB1BZ8lVTMitecvHeISp7TfGYM+S8ZSdn4iS9akex+Doi+ZDz2MUJZbkt3OmOZvqmqv0en0Aiul3VXUtu/ce5J67Nqa0n4Hfd4+nl+bmNiracwiEo0OlSCTCyzvfZcG8OZhMxnEdkyzLXK5rpKe3l+nFU4eMRhlo2p3w2YCUux5PL+crqtBqtMyfNzup6bSn5aRyLRtsxQhaHZ21uwCIBD1KOfdUvJcGf/dTCwu4WFHOkvx2jGL0/Ld4rOhEkdzcrIS+JCMUCnH0RBk9nkSxq6WtHYc9g+aWtnGLfmMxKZ9ISubM4u1dH1PRbmdpfpuyvKLdzqaSBSm34+uooKggM26ZQSuRafJzvkFKSMkMhSQOHzvFn//137FohpElU7zk5hcPe+9O9rm60tTCT37xW7x99y5HYd+BY/zRt55OmrI4mLSp65D8rgSvsth1PJz300geb2PxEhsrA8+zUv1uwHmOLb/Ro1NTRRWfVFRUVFRUVFRUPlPEBhe2oo3Ky3xscJHK4GkiDGfHWmVqvKS63/GUnZ+IkvUxRqrENlhALFj0KBVVf4tRlMg0+eJK2qen91e5O1VWnrL45O1x0nR2J3ptCKerE4DSbJfiywNRgeTipcssWTj2QaDL3ckvnnuR5pao8CCKInfecRtbNq1PWDc3J4vCgjyuNrUkfLZ0yXwADh87xWtvvq+Ibnq9ni8+sZ2F8+Yq64aDHrxt/YbYgmgAwNdeDoDOWoBGo0tZpB383a9cdCs9l7oQwtEIHrcvKs5tXL8mqRCWjEAwRDAYRK9PrKIYCkkIghD33Y7EaAsGXIsqbQCFBblsWL0I/5UP4pbfvcyEI92QUhsx8c9isLFs4UxOnavBoJUU8U+WM2jx9B+3JIUpv3ARnz+AXqcDyY6mp505/gAw/L07medq51sf9AtPfbS0trF7zyEe2LZ5xO3HWrnwWgnnoyG2v8Em5MmW3+yoaXcqKioqKioqKiqfGWKDC3f1m3HCk7v6Tbob9o7KZHc06XqDSSVtJVn6S2/zKTwtJxP6lGoaSarpMuNJL5yo1MSB31XsPAz8roCEFJicvKlYp23jYkcmzT1WYua1WY5MbGlW8qweDFoJOdJvazvc+du9+yPe+e3/x7mzpzh4ooL3yzR4Q1oluiqW3gcgDhONlAq//f1ORXiCaKTWrg/3cqGyOun6X3jswTjPJEEQ2HDbKhYvKMXpcvPy6+8qwhNAMBjkt7/fic/n71/W3Ygs+TFlL8CUvQDkCHI4qHyeXnQ76dPvjDvHssbI8UY7L+9p5f/755088+uXaGltBxK/e3/D+9yyaCZZuVNpDM4lv3AaT+y4n21bbk/5vFgtZnJzssnNyUajia9aaEuzMr90DvbMjJTaGumamui00Ng+U0lBDQc9rJvTy9L5xdizC4lkb2TegsXMLHKk/GyJiX+Brga2LAzwyNaVbCjxk5WhZ+rUYu7dsjau8mNbewe+vpRFs9lMICxypjmbSw0dBLzucaUVjwWny817H+xh7/4jSsTiQIa6F5KRLL14pAjVweJp0NMS97yfiGp5KkOjRj6pqKiojBFZlunx9GI2GW8a810VFRWVzzrjjcoZT7reQEZKWwESZuB7m0/RceF5pQ1r3vJRz8qPJl1mPOmFE5GamMp3lay9++69l5mz5nLyTDl+f4Berw+HI5M8q4e5WW78kkjajJVAotn2wOM/XXaeE4c/Ym5WEL8UHZS7PSHOVwXZvtrQF13lp8VjxWq1UDJnxrDHEwyG2LP/MGXnKtBoNNjSrHh9fjy9veRkZ1FRVYPRoE/Y7sSpc8wrmZ2wPD8vh//2Z3/EhYuX6OnpZfbMYrKzHACcOXtB8fCJ70OQ8xVV3LJsUdzx6m1To+eiaieCRsSQPg1L7lIs+csS2njp9Xc5frKMaJyCRPmFi9TWNfDnP/xDbGnWhO/eoNex6q5vsC4Fg/KheGDbZp79zcvMnT2Tuvor+Px+DAY927bczhceezDldiYyKi8VRhNJE3u2ZOVOZfZt8esGuhrwuS7GpeEmi9bS6q2Y7KX0tpxC8nUwK70XYcF0BDEacRaRmlk608zpGi8QTcsEsNnSSLdF76VAWOR8SwbLvD4l7XMsacWj5eiJM7z02jtIksTVphYarzZTWJBH8dQCZR1DkvtjKMYSoZosYkryu9FZ8+O2u1mKMdxsqKMlFRUVlTFw/ORZdn20F5e7E6PRyG1rbuGeuzamXPpYRUVFRWVyiA0u2sqeRfa54tIxbMWbRxxQTKTXyXBpK+GgJ2GQ3NtepqznbTuL3lo4poFzqukyww3egt2NREI+jI45cQMyn+siWtGM3jZ13KmJY02dAZhXMpt5JbO57+5N/PsvfkNHhwu3z4hfEsnKMLBwiougp0U5f5FwEK3RDkQjjq40tfDJ3k8AuNiRidtnJBAWMZtEeo02zrRkkGn00uKxYrGY+cpTjww70STLMj//1QvU1NYD0NTSRl39FeyZ6ZTMmUl9wxUuVF5i0YJSDIPSywLBYLImo+dCq41Lo4sRDkeG3CYcif8sdr0N/L4EjUigux5zcEncee7s7ObEqbMMxuv1cejoSe7evCGhLRmoPPJ7jtSn4w9qWDBvDretWZE0jW4o5pXM5gff+SoHDp/A5e4ky57JxnWryc1JzTcqxniuqbEwGrFrqGeLyV5KoKsev7MSk33uiKlgRsccjPY5+NrLCXTVoUsrRCtoQI4gGmzseOJRMvefpuxcBVlZdkwmI4UF/cKgQStRmuNCr89Rlo323h0tvV4vr+58j0gkgkajweHIpL3DxdWmFhz2DMWIfuXyxSm1N5605oHiqeR3E/I0I5ocSduGz1ba2/VGFZ9UVFRURsmFymqef/kN5W+/38/uPQcQBGFUYeYqKioqKpNDsLsROSIR9DRFqyJpRKwFq6OVkVIYUMQ+G+gdM1A0mogZ8aEGyabsqPGwLPkndeA83OCtrexZwoFuwoFOTNkLcJTsAMBZ+Qq+9nK0hnS0hvSUTapHOg8jRVAN5+GTkVXKn//xtzh3/iLtHS4KctLIFcqJBHuUNiPhoFIZ60qwhNfe2UfQ10m+WEGGEZr8UxVvJ4DsLDt//L0/5GpTC6JOZN7cWeh0wwsplVU1ivAUiUS4crUZAJe7ix5PL1aLGY1GQ3NLG9OKCuO2nV86J6VzNZBFC0p478M9CctFUUyIohrNQL2tw5k0ogqgrc2ZtK2jH/4nHa2NpEkiNc3Z1NY1UH6hiu9+8+lRTcpNLczniUfuG8VZSM5EFQxIdV+jEbuSPTeMjjkYXBWjitYStHoErR45HCTYeRlBq8eUvUDZ5/33bOb+ezbT1u7k//3454T60jNj/lAFWRYstqxr5kV3sapW6QPA9OIpBIMhurp7cLk7SbNaWLNyGWtXp1aAYTwRqgPFU63OiqTV43dV4ax8hYwZWyc1Uu7zjio+qaioqIySTw8eS7r8wOETbL1zgxr9pKKicsOz/ekfoNeJGPTRFIenH7ufzRtXX+deTRxaox3J14EcDiqRAW1lz6C35COaHSkNKK6FMW2yQXLGjK0ASQfOyUQYT8tJBFmIS59KRSAbbvAme51odCbCgU587eV09HkE+V1VAGh0ZgRBM+7UxFhfh4ugSvV7WLp4fv+xefLizl/mrHvxNB2hp7ONytMHEEJ2luS7CPl1tLp8HK1sZtacDGX9gvxc8nKzycvNTukYABqvNCv/9weCSFKYUCiEp9fLsRNlZKTbsFrNeL3euO1mz5rOyuWLUt5PjLzcbDasXcXLb7yHqNWSmZGOKGp55MGtpFktceuOZqCek+1Ao9EQiSRGVuXn5SS01ekJ8c7JMEvyxLg0xdq6Bs6evzisQfvF6lqOnzpLKBRiQelcli9dMCHvUBNRMGA0jFfsGo2AFbv2ZcmP0T6HcNBDyNOEHA7GeXnFyMl28PWvPMHrb31AS2sbWdYQU/NszJozX2nbZC+lt70s7loYr8ju9wcQBEFJo9MM8ksTRZH5pbPp9Xq5Zekidmy/J2VfLxh7hGoy8bSzdhe+9nJ87eVEgh4EjXjNPLA+b9xQ4lPj1Rb+5//7OV3dPVgtJv7rn3yTGdOmJKz31vt7+e1LbyPLMssXz+NHf/RlRFHkVNkF/vi//j+Kp+Qr6/7nj/970txqFRUVlbHi7uxOutzn8+EPBFKu7qKiojJxnDpTzsGjJ+nt9TF7ZjGbNq6NM+pVSeR//eUfMWdm8fXuxoQTHXjuRm8tIOhpQjRlEXDXABAE7KWPxVW/G2qQci28Y5INkjtrdyWs113/MSZ7KZ7mo3GDIk/LSZznX1DWs+QvS1kgG2nwprdNVSKdYpXRYtEVjpIdcWXBB287GuFppIic0X4Pyc6pp+kItuLNVF36BQatpJS5D+vT+LhcwtXjI8/Ti9VqQSeKPHjvXSn1fyD2zHTl/3qdjkgkjLuzG1mWMZtMBIJBAq4gq1Ys5e47N+DxeJk1cxqL5s8dk+Cy66O9fHroGGlWMy5XFxpB4E++/3UWzS9JWHc0A/WMdBurblnC4WOn4tpIs1pZvXIppj5hK9bW5bqzBCQtZ5qzFeEpRl1945Di0/u79/H+R/uUv8vOVVBWXsHTT2zn8PHTnK+oRq/TccuyRaOqMHg9qkxOhNg1nIA1UHSOiX+yoCES7EGrs6CxFRPorsfvqkrwjQKYM2s6f/7Db9Hd48Gg1xPpqVHa83VU4Gk+iiAasRSsVISnsYrs7R1OXn3jfaou1SIIAvNKZvPIA1uZN3cWJpMJny/eZNxqsfDQ/VtGJTzFGEs1vmRCrKNkBx3hIL72csJBD6Ix45p4YH0euaHEp7/7l1/y4D23s+2u9Xyy/xh/848/55f/9r/i1mlqaeMXv36V5/7jb7BnpvNf/vs/8cZ7e3jk/jsBKJ6Sz29++rfXo/sqKiqfE4qLCmlta09Ynp3lwGQ0Xoceqah8vvnokwO8+8Enyt8trW2UX6jiT773dcVHQuXzQ2xwIZrspM/chqviReRwkEgkhD69mLDfRVhvwVn5CiFPC3aGHsRMpnfMcDPwEE2/G5gC0ttehiAa40QYb1u/N09vexm6tIJRCWQjDd4yZmxF8rsJdl4GQGctIGPG1iEHeKMtw55qRM5oo0KSCQ/d9btpDRYA9cr6Vc4sZs3Jo7WtgxnTiyidO4u1q28hJ9sxuKsjsnhhKe9+sAd3ZxeiqEWrFZFlGa2oxWCIpuxpNALhcJiNt63GaDSMeh8xLtXW8cHuTwEwm0yYC6OTXm+9u5uF8+bGVTuLMdL3NVDg2LH9Hhz2TE6eOoYh4sJeuJR7tmxUIqrixKqMqMgfCItxwhMw5ARAV3cPH368P2H5ufMX+Z9/9294+kyyAcovXGTTxrXct3VTwvrJmKiCAakyUWLXUAKW0VFCb9Mx5XhMWaWE/E7c1W8hh3wYMmfhKH2MztpdhDzNcb5Rg4mZi2PoP/6B4q7fWTlmnzmAUCjEf/znb+nsiopWsixTfiGaDvtnf/xNvvzUwzz3/GuKACWKIg/df7dion8tSCbEQlRY11kLEI0ZwOR7YH1euWFyQ1ydXVRU17Jl01oAbr9tBa3tLhqvtsSt98n+Y9y2ehkOewaCILB92yY+2nv4enRZRUXlc8rmjWsxDYpuEgSBbXffkfSFT0VFZfLw+wN8tOdAwnJ3ZxeHjpxMsoVKjP/5/37GU9/8c/73P/1iyIjOmxFTVim2oo3YijfT23QYOSIRCQdAjhDsbkTQ6pWIHsnXQSSUWO47xlhKeafK4EGy3pqHJbvfbNecswi9NU8pay9Lfsw5i/qjgKp2Ikt+TNkLMGUvUDyiJqpsejjoiQ5oe64qy0KeJjprd6VUEj4VYt/VwL7GxAJb0UZloJjq95DsnMbOn+RzMdvWELd+abYLi0Fm+rSp/PC7X2P7fVvGJDwB6HQ6vvONL1IyZyaCIJCWZiUnx0G2ww4IWMwmSubMRK8TlcH5WDl5upw8qweDVopb3tPZxuULnyasX994lYvVtQQCyY3NY6mN7uo3CQc9aDQabl+7mD+4O4cnNhWw465Sshz2pNvOnF5MQX5uwnKTyaRU3BtMTW190rQ+l7uT8vMXE5bv+fRwyucs1WsqFY4cP83/+cef8qd/9bf8y09+xcXq2oR1hrvmYmLXSAwWsDLnbFe297adjROdg54WeptPIod8aPRWHKWPo7fm4SjZgSG9OOV9xoidm4HPlbE+Q86crUj6PbW2tVNxsYa5s2fw3//i+3z5qUd46tEH+e9/+QNuXdWfLjyU19ho8HVUJDyfwkEPvo4K5W9TVmlcEYVYGqMhvSju3MfuB5WJ44aJfGprd5Flz0DsywcVBIHcbAet7U6mFva787e2OcnL7a96kJ+bTWubU/n7anMbX/rOX6HVaNh213oevu/Oa3cQKioqnwtysh388Xe+ysf7DtF4pQl7ZgYb161i1oxp17trKiqfO1rbOwgOUSmqvvFq0uUq8NN/+K/k5WQhSRI/f+5V/tc//Jx/+psfJaz36lsf8drbHwETMzC4VuhtU5XBnN42FdFop7f1JOFgNy0n/gUEDYKgxZS9AKNjaLPnwdEIkr+TztpdOEp2jLskd7IZ+JhvkyzISurM4PQovbUwZY+osRIOehSBDvpN0P2uKnzt5Tgh7hyMh1QiqFJNaxoqvcxWvBlnxYtkZeVwpd3PoSotpdkujGLUfHnKkidSikSSJIl9B45SVl4JwLLF87ltzS1KFbwsh51v/cFTBAJB3njnQw4fO4UsRw3ItdronL/BYCAzoz9Fr6W1nYuXarGYTCycX5JSqXlTpJm5WW78ksiZ5mwCYVExkg53HMPXkYMpq5S2die//O0rtLS2Kft+6P4trLplSVx740kxFQSBb3z5cX7+qxepvVyPyWRkSmE+O7bfM2TkqWWI5d3dPUmN3SORCHX1V1iyaB5Ol5tTZ84TjoRZOK+EwoIkwtcEROXtP3Sc197sT4G9XN/Iz3/5At/9xtPMmF6UsK/xpKCOFK1lzV+FL2ZIXrUTjVbfF/H0OHpr7pj2OZCJMmh3dXYO/Zk7+pnBoGfJonlxn9VebuCd9z/hcn0jaVYr625dwebb1456QncsPn3XOlLu884NIz5NBHNnTePN5/8Vq8VMW7uTH/63fyDdlsbmDYkGmjfry5SKisqNQU62Y0IqsqioqIyPjHTbkOa4Az1YVOLJ6ytfLooij23fwmN/kCg8ATxy/52KtYEkSazb9uVr1cUx4+uoIBLyxQ0owsFepEAnfnc1ks+FRmfBkrdsWAFlcDSCwVaMq2onkq9DEV+AcRmQJ1t/oHF4jIGV9lL1iBro3ZSsStxIvighTzT7IFm1u5Cn+ZoNykab1pSsT2G/S6nMd/tDf07G+ctUXbxAgamagiwLU0pTMxZ/5tcvUVlVo/zd0HiVS7X1fO1Lj8WtZzDouWPDGk6VnScQCCjCE8DGdasUgWnn2x+w78BR5TOz2cQ3vvwE04oTPW8HMr1kBeV7TyniWUW7ndJsF2kmgXR7HnrbVGRZ5tnfvBxnExAIBHjx1beZUpAfJ9oMHHB7ezpwn/wtJpMRo8U+YvRLTW09v3/1bTqcLkRRxJ6ZwVOPPkB+Xs6Q28yZNZ0sh50OpytuudFkxJ6R/LltS7Ny9MQZXnrtHeV5//5H+7jzjnUTXmlYlmXOHn0PgzYcVwVRJwQ5fuBNZkz/o7j1xyt2pSJg6dIK4sQhR+ljivA0ln0OZKIM2oumFAz92dTknzW3tPHTZ36nVMLr7unh3Q8+wevz88C2zSnvG8Ymok6EeKiSOjdM2l1Otp0OVydSOAxEb/rWdie5g0Jfc3MctLR2KH83t7aTmxNdx2IxKwp7TraDOzeuoaw8MXQToi9Tv//F3/P7X/w9v/vZ/5mMQ1JRUVFRUVGZZNJtaQmzqAA6UeS2NSuuQ49ufHx+Pz0DPFU+2nv4M2M8Hpv59rkqsOavInP2AwB01r5HyNuGRqNHo7Og0Y4cXTJ4Rtycu0SJAPK1l9PbeiZOFBk4sEkl9WO0JEvNEUSjYgguiMa4lJG2smfpuvxRXOpIrI3uhr3D9sWUVYp9zoNkzXtKEehixryO+U9in7NdaW8ijzEZE5HWNDAVy2DOYPWKpTz9hae47b4/Zcr8e1MaYFbX1MUJTzHKL1ykrv5KwvLsLAff+9aXWTi/BLPZRF5uDg8/sJWtd24E4EJldZzwBOD1+vjtiztHnBhfsGAxmrw78EvRCnNL89sw68PMmjMPx9ztaPVW6huvJvWnlGWZ46fKEpZHBAO7y0KcPFNO+YUqTp4+x/HLxmHFh16vl/987kVFRNLpRFzuTn7+yxeQJGnI7QRB4BtfeYLCgv7slox0G3/yR18jLS1xfwX5ueTmZvHKzvcSJho++mQ/V5paErYZD13NZykwXmFJfruS2hiLLLNJFyf8Gof4VLAYw4nO3fUfT0hK2HApf6NNOyuZM5OZMxJ/SxbOL6F4amHSbfYdOKoITwM5eOTEkGmiQzHWFMLhzv1nlc7Obppb2q55EM4NE/lkz0hn7qxpfPDxQbbdtZ49B46Tk2WPS7kDuP22lXzrh/+Tr33xIeyZ6ex892MlsqnD6caemY5Go6HX6+Pg0dPcd/fG63A0KioqKioqKteKxx++D5PRyLGTZYRCIQoL8nhg253k5mSNvPHnEJe7m7/4X/9CJBJBlmUK83L4/370revdLYWBxscxUk1tGzjz7XNVoEsroLN2F962s9HKUPp09LaphDxN+F1VOCtfGTL6KdmMuDlnEZFwkLDPia/9HACCaMToKFHWGUvqRyokSw8x5yxSUuMs2YsVYcZd/Sayz4UsR8ZcrW+oaA5r3vJJO8bh+jHeyITxRqfUNwydxlvfeDVptFJhQS5/8PSjSbc5ffZC0uVOl5uGK01DDtYhKt48tmMHtRdn4by4E1HUkuWwkzv/EeUc+f2BAW12cqWpBZ/Xh8lkJC83MSrpnXffRXYeVUaHkYhM5+XdHDmSx+rVa5P24+TpcgKBQMLyzq5uLly8lLTyXoycbAc/+v43aG5pIyRJTCnIQ6PRYM9M56XX3qW9I2qrMnvWdJ7a8QAVlZeGFLTOna9kygAhq7OrG6fTTU5OlmKSPhqsjhkgWjDSGxdZZhQldCbHqEy4g8EQTpebjHQbJtPoi9JMdvW+iUw7EwSBb37lSfbuP0JZeQUajYali+ax4bZVSdf3dVTgdsYLhwatpFRN7OruGbUH20SlEH5W6e7x8PxLb3KxOiqk2zMzeOj+u1kwb+j084nkhhGfAP7se1/lb/7xP/n1i29hMZv4qz/5BgB/+8+/YN3qZaxbs5zC/By+9sWH+eYP/ycASxeVsn3bHQDsOXCcne98jFarJRwOc8f6ldx71/rrdjwqKioqKioqk49er2PH9nvYft9dBIJBLGa1wt1wFObn8Juf/O/r3Y2kjFfUSFahLuTtIBzoRKtPx5y7mIwZW5Wqcr72cnw5ixJKk8cYuC9fRwW9TccQBA1aXXQgI4dDIGii1ahEc9STaRz+OcnOR0x0Ucy3jXZlQBjrtyALSrreQGFmoPfVRFbrm8hjTIWJ8PAZL8Ol8cYqvo2G4SIOUolGCAc92KQLWKbkK9uU7XuOTyoN9Ppg9sxpgIDL7Y4zyu71+jhfUcWpM+UsWxKN5Av5uwk1fYRRlPBLYpzY4rz4BuFli5NeL71e75D983iG/mwgg9PzZs2Yxl/96Du0tTvR63TKudX2+QInQ6uJfhYOh3np9Xc5cepsn7gOOTkOFs0vYca0IuaVzErJR0g0pJG3YAd1J36nRJYBBCI6lqx4OuV758OP9/PxvkMEAgF0osiaVct48N670GhSTz6abE+iiU470+t1rFm1jNb2DsrOVdDU3ErjlWYe2HZn3H0Se9YvzGmnrl6K8y0zilKfN9ro76uJSiG82Rk8iVNdU8fujz/i7Im9VF6JkJ+fQ06WA5e7k+eef5X/8oNvjrnYwmi4ocSn4qkF/OLH/z1h+V/+8dfj/n7gntt54J7E3N4dD9zFjgfumqzuqaioqKioqNzAiKKoGP+q3JxMhKgxeOZb0Ijo04owZExTopwcJTtwAiFPM1oxNbFSb5uqpLkB6KwFhDxNQNQXKda3ZAIYpCb6DBwwxAZngmjE3CeQDRSTgDgBavA5iA0aJyoKYGDfYsforHyFQFf9hApb15uhIu/m5ENmRjruzi5luSRJBAIhys5eoKm5lTUrlikDbEmS+ODj/Rw5dhqf38/c2TO49+47FLFlycJ5nDh1NmH/9swMJeopHA4jy3LCcy1ZNMzBXT+hs6OZIr3IGWc2J8/04PcHEiqHZ2akY8/MYPfeg4r45OusRycE4wzMzzRnR4WAiG9IEWL2zGl8sDuxup4gCMweZyrv4IHwvJLZGI1G/H5/wr6WLZkPwPu7P+XYiTNAv8gmSWHOnL2AyWjEaDRw/9ZNrFm1DIc9c9j9r1mzFr0mhPPi6/gDQawWM0VLn2L6zNQiRI4cP817H+5R/g5JEp8ePIbJZFRSL1PhWngSjUXcHeo+CXQ18NPn99PU3KosP332PI1Nzfz5H39LuZZjz/qpeQFumXqJc83piuDpl0QWLduQ1Hx+OCY7SuxmwNdRQVjyRidE+p7H9Q1XeeeVZ7Ab3Cwu9NLTI1NVU48ckcntKzxy5Php7r9ndB5bY0F9Q1NRUfnMM54UDhUVFRWVa8d4hJsYg2e+RWMGkXCQjBn3xA3eHCU7xvc7MEwExVhSPwZHfQ0Uu3zt5URCPoLd9aMS4iYqCiBZRBqAHA4S8jQDIBozb/r0lpEi777x6Cp27rlE9aXLBIJBmprbyM/N5lTZeQD2HTjGt7/+BYqmFPDia+/EiUvnK6q4XN/In//wD7GlWVkwbw63rlrOoaMnlXWMRiNPPfoAnl4vO9/+gLPllYTDYeaVzOah+7cogsngaBh3T5D3zwgsyYt6QMXSloxGA1mOTGxpVsLhsCI8CQK0tfdXC7flLaRHnMv5xm7FYDsmQG1YMX3Ie2TWjGksXTSf02fPEwiGqK+/QluHk8KCPPbuP8qdd9xGRnp89Irb3cWx/W9Q2xIkw57H+ltXUliQO+J7mcGg50tPPsRzz7+mpPqJosjDD9xNlsMOwOFjp5T1ay43IElhQOZCZbUSEdvZ2c2e/Ud4+smHWLxg6Hs/HPRQbL3KlAHraCOVhINzU7rGDx45Gfd3SJLw+wJ8su/wqMQnuDEi/wYy3H3S0XaFSI8TiD9HHR0uzp2/yNLFUaEw9qyn+k1uWSSTa2+iu0cAMZ2C0odYterWUffr8165buCEhSAaCQe6cVa+QsP5C6wsbMLtFahohyt9Pv9XmlrIyXYgCALdPeP3D0sFVXxSUVH5THMtfSlUVFRUVMbPUMJNssptvc2nkAVZif4JBz04K18h5GnBkF6kzHwT6Ka7fneccJJs8DbcZAWALPkx2ucgh4MIGhFN+jQErR5Z8scNbMYi+iSL+hqIt+UUglaXshA3kVEAyfrWWbsLv6sKQatX0hBv9vSWkSLv7IXz+PbXVuP1+dj51ocJxt1+v5833/2ILzz6ICdPn0to3+v1cejoSe7evAGARx/axq2rl3Oxqgaz2cSShfMwGg384789w5Wrzcp25yuqaGpu5S/+5Nvo9bqEaJjm1osEJC1nmrMV4SlGmtWKLS3R92iwJ976ux6n/JnfQThq8tze4cLl7sQX0nKp5Tm2bF7P3NkzEtr54hPbcTgy+ad/f5arTS0YDXqaW9r4t5/9muOnzvIXP/xDMvtSFts7nLzw3L9QZGkmQxI5cyqbk6fP8Y0vPoA9fGbE97LSubP4n3/1x1yorEaSwpTMnRnn6eT1+oCoz5Knr6iC3x/E6/Up4pPT5aYgP4dXd+5ifsnspNGyE3HvdHf3KP+vb7hKc2sbkYiMIAj8+oXXeOrRB27aSN3h7hNvUIPbl9zbqr0jvrph7FkfDuxUrq3MOdvRW/OSbT4iY40SG+m5P5oJ7M6aXZhzl8VVIgx6WvG2niJj5tYxHVeqDPxeBNEIggZfezkWWgkK4AmZ2XXGizcQTesNBIJEIhG0Wm1fmu7kc8NUu1NRUVEZCre7i737j7Dn08M4Xe5RbTv4BzLoaRmyOpGKioqKyvUnmXDjrHiJrrrdcdWXeptP0XHheZznX8DTcjLqydRahq+9HMnXgckerWBkspemVBktNlkxVHU4AGv+KgStHkETHTQKGhFBq8eavypOeBpL9ahklZpiYpfBVoygjaagpBpdNBFV4obqm6vyFSX90JS9AHvpo2OukHUjkew7kALdNLZ087M3r/Kjv/4x//Tvz9J4pZnauoakbdTU1tPc2j6kb1NbmzPu7ykFeWzauJY1K5dhMhmprqmLE55iuDu7OH32vPL3wApd2X2RP4GwGCc8AWy8bVWCz5EgCNy1aV3csmnFU/jzH36LTRvXkma1EgwGmTG9CI1GoLaugZ//8gVqausT+qXRaKi93EBYkki3pWEwGJBlmc6ubj49eIx//o9ncbujqYoffnKAJhdKlb4l+e2Ytb1UH30u5fcyg0HP0sXzWbF8UYKZ+JxZ06PHp+k/3kAwiF7fX90y5rfU4/HQ0NiUdB8Tce/MnB5NO2xp6+BqcyuRSPR6sFrMnC47zzvvfzJiG8mYjEqaoyXY3YiteHP8feJzEQkHsU2/V4meG8zA6oYwtEg/nufHaCvXDffcd1XtxFX1RsoVQztrdtFZ8x4tJ35M0BNNOwx6Wmk58WM6a96js2bXmI8rFQY+v2TJTzjoQQ4H0Yh6nF4jF9qzcWTnK+sb9Ho0Gg1FUwtZ1heRNtmo4pOKisoNzaGjp/hff/9vvPHOh7z57kf8zd//O58ePJby9mMtu6qioqKicu0ZSrgREAh6mpB8LmUiobe9P+qku34P7pr36Kx5B60hA1P2AnRphbir38TTfBSNaMFSsDJuADJ4wDbSZIXWaMfnqkCW/HF9kyU/Plf/gHA8A9dYJEAMORKt7hUTnqLHmtrgzJRViq1oY0K0V+bsB7AVbRx11O/AvsUGNUb7HBwlO8YlbE0E5y5c5Jlfv8S//fzX7N5zMK7S22gZ/B00Xm3mrSPddHRFvYYaGq/yn7/6PcFgKOn2er2e/LzsIU2lBxtsD6bDOfQkm9PVmXR5Xm42C+bNTVieZrXy8IN38/UvP86MaUWYTCamFU3hq08/mjTlzJ6ZwX1bN6HTiRQW5KEbEJkTiUT4eN+hhG1CoRAXq2vw9Z1zWZZxd3bR1d1DW7uT/YeO87f/+BOqLl2m9nKDks4XE6CW5rcR9HaCaMWpXcxr7+zjlZ3vJRW6RuK+rZsxmUzoRFFJ9xNFMa4IRZaj3+tJb9AntAETc+9s2bwes9lEW1uHskyjESieWgDAkeNnRl3mfiSB/FoIULE+dNfvxloQrTgvRyQCXXXIEYlCe7RC4WCKpxYyr2SW8vdYRfqJZuBzv63sWfzuGtrKnkXyudBZ8xFNDgJdDTgrX6Gn8cCwE9jm3GVo9FYiQQ8tJ35Mb9s5Wk78mEjQg0ZvxZy7bNKPJ/b8ksMhxZPQ5phORNZSmu1ixhQHpXNnkWXPZM2qZTx0/9185+tfHLW/1li5OWP9VFRUPhd0dnXz6hvvEYlElGWyLLPz7Q9YMG8O9syMlNpRy66qqKio3BwM59kBIMuROC8oU3bUMDkS9EQHP30pQ2lTN9Bdv5twoJtIOEiwp5GI1IvJHvVr6W0+RW97GbIUFRRig0mjowS/szKp31SqfiLjMQgeGAkQG9DFjjNjxtZRp/5MpFfMwL6JxgwABG3/4H2iTZBT5cOP98cZO9fU1lNWXsH3vvWlMQ2oBh5nRJZpaWmjNFtQjLghagQ+FCuXLyYzI51VtyyJ8yACsKWlsXrl0mH3P6UwecpRJCLjcnfy6+dfw2g0sGrFEqYVTVE+f/qJh3j3g084dvIsgUCAkjmzmFZUyM+f/T3dPT2sXrmUbVtu53L9Ffy+AMFgCL0+8fwEAkFc7s6kfWgdIKTE0Gq1mEwmtFoN4XAEn89PKBQVTTWCgE6nIxQK8crO90izWnC63ATC0Yp6sSpyoqjlaI2eT468rrR78MgJ7rxjHdu2JBaZGorCglz+7Aff5ODRk8yeMY1TZy8QDAaVan852Q5y+tIN8/NymFIwdHrXeO+d3Jws/vR7X+d7P/oftLZ1YDQayM/Nxmw2AdEUzXA4PKrUu2tdZXK4PkheJ21lzyCaswn1XEUOB5F8HWiNdr7+pcf5aM8BTp89jxyRWbywlLvuWBcXgXej+DPF9ttW9iwB9yWaj/1TdLnBRvbir9PTuI+Qt41gzxW8LafQWQswpBclff7qrbnk3fIDRXBqP/OfAGj0VvJu+UFcKt5oSdW/Nhz00Fm7i0B3fTQ9XKvHlpnDvHlpNDbWs0zr5EpOCbd9/QusumXJmPszVlTxSUVF5YblfEV1nPAUQ5ZlzpZXsnHd6qTb+f0Buns82DPTEUVRLbuqoqKicpMwknCjNWbGTSRkzIh6aLirdmJIn0agqw7RlEX35Q+i2xpsZBZvVoQod/WbGGzFuAaIV3rb1LhZeFP2Qnzt/X49scmK0YhKYxm4Do4EMNiKFfFp4L6u1eBs4GAn1jfJ50KWIzhKH0sqhE2ECXIoFCIQDGG1jFyFsNfr5aNP9sf32+fnfEUVR46fYd2tK0a178HfgTZzFR7/OSVFbKAAlZ6exrIl8zlw+AThcBhBEFi0oESpGLVj+z1kOTI5fOw0dfVXQIDCgnzOlleyesUStFpt0j4UTSlgwby5lF+4SCQSQQpHELUamlraOX6yTBnAHzl+mh3bt3Hrqmg0hV6vY/t9W9h+3xYAnv3NS/z4J7+kw+lGlmVeeOUtrBYzt65ajihqeXvXx3z7618gLzc7bv8Ggx57ZkZSAWqwTxRE09hW3bKE6kuXaW5tJxAMKp8ZTUal/fYOJ6vuvoO6hisYtBKl2f3+PxnpNq5c2YtBmx2XsrV7zwFWr1gyYmW6gWRk2BTBSpZlLlbX8sm+Q5RfqFLEtqwsO19+6pGU2xwr9swMtt19B0ePn074bOb04lF7Pk1EQYbxEp3Q3awILKFwEJ21AMnXgd5aoHjrbdty+7DC4bWo4pcqWr2VzFn30nrqJ8jhAJFQLzprPu1lv0A0Z4MQjWIUtHplQntI7z5rLo55TyrCE4Bj3pPjFp5S8a+NLYsVgTBlL1A8CTMy0snIWIQs+ccU+TpRqOKTiorKDctgj4IeTy8tre0EQyEKC/NYsXxRXCh1OBxm59sfcvT4aUKShMVi5u7bV1Jq759duZnKrqpV+lRUVD6PDCXc6G1TowKIvxOt3oqgEemsjXpoyBGJcNCDLq2QUM9VNH0eSbbiTYT9LmwDBKje1tPIchg5HE2bik1QxExa/e7quH3HJiuSGZ5P5DM5WSSARjQrEVqx/UzU4Gwkk92Bg51gdyOSz0XQ04Tekk/Y755wISwQCLLz7Q84efocIUmiID+XB++9S/HxSUZDYxMhKRplEwyFqLp0me7uaLrOP/37s2g0GtauXp5yHwZ/B4Jo5lL3dGbZLsdVkQPIyEjHYDAwZ9YMjAYdd2y4lalTCpS2NBoNmzaupcPppsMZFVquNjXzys53uVBZzde//PiQ/Xj6ie3833/+OfsOHEUKhUhLs5JmtcS9F8myzFvv7Wb5kgUYBqWPXaqtY9dH+5QUPp/PTyAQJBAIUl1TR+ncmXT39PDSa+/w/W9/JWH/d95xGy+99k7cMo1Gw6YN/RXIZFmmq6sHo9HA/fdsprOrm7d3fUxXn9F2WpqF+aWzsfeZjQPcsmwRWoJ0VO5ERCIY0WEu3IBBV4tRdCcIfLIsU3GxhtvW3DLMtzY0giBQMmcmJXNm4un1UnO5HovZxMzpxQnvmJPFlk3rqLx4STkvEE3NvO+eTcNslZzYPTswml+OSBhsxdf0XTbsd0UjnsJBxY8uc87X6W06PKrnwY1SxS8c9OBpOqJMYAAEOmvQiGYikg9k0IhGDOnTEDTisBPYQU8rzgsvxC1zXngBnXnskU+pRrzFnl+G9GLSZ2zBZI+m4sae09b8VWh0pus6hlDFJ5UbhkNHT7HvwFFcLjdFRYVsvXMDs2ZMu97dUrmOLJw/l9ffeh9Jkuhwuqm6dBmIVreurW3gn//jl/zg219VZkffeX8PBw4fV7bv7fVycO87OG7NJCev6LqG9Y4WtUqfioqKSj+x51+gqwHJ14FosgMCvvZyZDmCoNGhMzkIdjWAHCbQXY/BVoyz4iUEQYNosmMtWE3X5Q+VNgVBS8jTogziBDFaoSnm6TTwBb+t7FmlnaGeyTC6qkiDSRYJYMlfhtExJ66NiRicjfQbYylYmTDYkeUIeks+otlBJBStKjZQCBuvEPf7V97izLkLyt9Nza384le/50c/+CY52Y6k29jS0pT/V9fUKcJTjFffeI+CvBymT0stHSnZd7Bh40befdcTJzwFgyHOlVdSfr5S2fZy/RW+/4dfUSq7AbS0tiek3kG0el3VpctDCmu7PtyHy+Vm4bw5yrHVXG7AaDSQbus/Zr/fT8OVpoRKVdU1dXQMqCzmD/RHIzW3tFEydyYCcLm+ka7uHno8vQT8AYqLChFFkTUrlyFqRfbsP0yH083Uwny2bF7PzBlRE+3yC1W8+e5HtHc4EUWRZUsW8PQTD7H9vrvYd+Ao73ywhzSLOc73as7sGWSk21i9MJ/OtBICET3ZJdux2LI4cOBT/FJDgsAHYDIakp6jZNTVX8HpcjN1SkHCNWO1mJP6XE029swM/vT73+Dw0VNcaWrGYbdz25rlo4rmgv57Nvacgv7U3EBXHRrRjCV/8j2FIHp/CII2rhBCb9NhbMWbCftdN9X76cDnnmiyk1a0kY5zzyHLEaRAN6IhHUEQMNqjwv9wE9gxc/GYx5Nj3pM4L7ygeECNNfUu1Yi3ZM8v4LpEkw2FKj6p3BDsO3CUnW9/oPxdU1vPT595nu/94Zcpnlp4HXumcj1Js1p44pH7eOGVt6hruAJEZ7FmTi/CYNDT0eFi/6FjbL1zI+FwmENHTya00eKxcrrByOPrEk0jx/IgvlbRSDdCXr+KiorKjUJsRldnzUNnzYtW8gn1EomEiAR70KcVIQU6MaQXEfK2I5qzkZEVo3I5EorzJyESRtAa0Ij9A1tT5my87eeSepDIXme/31SSZ3Ik5MPTfHTcEwbXKhJgpN8Yk30uJvvcuMGORqvv+6wUT/NRfK4KMmc/EJfuMdbJkc7ObsrKE82SQ5LEwSMnlFSywRQW5DJzejHnK6vo6uqPLNFoBHKzHciyzOHjp4cVn3w+P8dOldHU3EZutoOVtyxBq++Pql6/diUWs4lPDx4jXdPNtOIpNLe009Ye73/U2dXN+x9/yhOP3Kcsq60b2ni95nJ9UvEpFApxcND7TCxFr7mlLU58AjCbTAltWC0WpboawMAYH0EjgCyDIODzB/jXnz6nVBK2WMw8fP/dLFuygBXLF7Fi+aKEtq80tfDL376s2CJIksSxE2eQQhJPP/kQDz+wFaPRyCf7DineWFOnFPDkjvuB5APkJUtvYddH+0gTe+KFJ5MpqZH6YHq9Xp557iUu1zciyzKCIHDLskU8ueP+pMbvLa3tyvW2eEGpkhrY0trO7r0HabzSjMOewcZ1q4eNvEuVNKslobrgaNHbpiKIRqXKpNE+J85zrbe9DKNjzqRHQMXu9djzYOCzI5ZydzMxMNoxFiGrTyvE234OZIlwoBPRlKWk3A03ge1tPaUITzGhSWfu94Dytp5Cb906pn6m6l97o0STDYUqPk0ikUiEsnMVVFTVYDToWbF8MVML80fe8HNGJBJh956DCcvD4TB79h3my1+Y/JxslRuX5UsXkpmZQVNztGRpZmY6+gEGopf6qqEEgkECgeTVbRpcxoR0iVgKh6+jYsQHsizLHD52mktnPyFX34AtM5cFG75BWnr2pEUj3Qh5/SoqKio3CgMHrIBiQK4zZYEpC60hXYlMcsx7irDfpaTpyZEQgb6IqJg/ScjbihwOEuy5qszeB7rrseavihvADZysiIR89LaXxT2TBdGIyV6K0TEnWvHuJpkwSPU3ZqjBzkQfa2dX95CVv9ydXcNu++UvPMJPn/kdp86cB8BoNLB+STYhjZZAOBoFDcknijo7u/nXnz0X52/0yaeH+aNvfinO32j50oUsX7oQiL63/slf/u+kfblYVRP3d1qaBUmSaGt30uv1YTIayMnJQq/TxUVtDaTX6yM4wDcJIDfbQUtrO4FA/PKiqYUUFvRHUoRCIS7V1pNmtZCT7aDhSrTalcGoJxgKIQgCNqsVfyCI0Wigta0jLrKot9fL7156g8KCvKT+TgAHD59I6sd55twFHuy5C1ualW1bbmfdrSuoa7iCLc0aZ4wOie9KVouZLz71JM+/9CYQjV7LSLfx9BMPJaQUJuP1tz7g4NGTNDW3EgyGMJtNOF2dTCnIS/AH3bv/CG+++5Fyve36cC8PbLuT0rmz+Of/+KXyLtna1k7FxUt86cmHWbJo3oh9GInxTl5q9VbMOYsU8UmWAghanVJ0QZb81FefxEMu04qmYEuz0tnVTdm5CkIhiQXz5iT4e42FG8UofDhGc66ViFKjne763Ug+FyFvO0Z7CSFPE1pjBpGQj0igR4l2GmoCO2NmVFgy5y5TIpxiJuTe1lPK52Phs+Jfq4pPk0QkEuGZX7/Ehcp+34D9h47z6EPbWLPy2oRE3iz0en30eJKX02xJUlVD5fNHTraD/LycpC87Nmv0gWs2mcjLzaGltS1hnYXTdCOmsAnWGQiCkPCSE4lE+Md/e4ZP9h3CYoBtyzR4vT7aX/u/bH7oB3ivfjppgwu1Sp+KiopKP6as0gTPE505Gzkcwpy3DJ3J0T/g6Hvxz5z9AD2N+5EFgVCf0ARgSCsi0BVN5TbnLSPYXU840I3PVYHRMSduv7Fnrqf5KAga5IiEoBGRIxKEg3iaj6LRma7bhMFYB7Uj/cYMN9iZ6GPNzclCr9cniC4ARVOGj4BPs1r4kz/6Gk5XJ11d3UzLCjE3y41fivoHzZ09Y8iJol279yUYa3s8vby962O+9qXHku5Po9FgMBjw+/0Jn8UqmcUozMul6lIdnV39qZlNLW2sWLaIZYvnJ20/3ZaWYPhtsZiZPXNanPg0rWhKnGl2+YUqXnjlTbzeaEpkdpYdl7sTT68Xk9GA1+cnEg7T6/Vxuuw8DnsmhUmqvUUiEY6dLOO+rck9iQYey+Dturp7sKX1CZdpVhbNL0m6bjJK5szkv//l96m53IBGIzBjWlHSqKXBSJLEux/siZq69+H1+rhYXcv7u/fFiU8ud2ec8BTjrfd2c7G6NmESU5Zl3v1gz7jFp4myUrDmLScS8uFtOaWku2XM2IrHJ/Pm67+lrHYfEI2Um1Y8hfqGq0h9nmjvvP8xWzavZ+udG4GoL1hrm5P83GxEnciBwydwd3YxrWgK625doXyPg7mRjMKTMZZzHfttCQe649KLsxd+WRGkYpGvg6uaDiYmMA18LuutueitW8ecKTG4EMLN5F87GFV8miTOllfGCU8QfYC98c5HLFucaAz4ecZiNmFLS6O7pyfhs/wJUOhVbn6sFjNLFs7jVFl53HJBEFi7pt9IdNOGW/n3//wNoiiSZo2GzFutFlbddjdCx76ks7R+SeSNNw5RUf0KGo2G+aVzuG/DDBz5JWj1Vn76zPO8+sYuTHoZhwF+vRseX28l1w6Xjj5HbnbWpA0uPiuzHCoqKioTQVLPk3CIQHc9ge56HPOfTHg2avVW0qauw+eqUiKcJL8byefEYJ+DNf8WrHnLCQeXDDtzPzjlJWZsDv0V867HhMF4BrXD/cYAIw52RnussXSoZJhMRjZvXMt7H+6JW+6wZ7Jm1dIRz4NWq+Wxh7bxu5fewO3T4pdEjKLE+rk+lpXmxR3LwImiCxXVSdsb/A4/mNUrlrB3/xHl70hEpqu7hyWL5hEIBJX3/A8+2c/0aVO5VFNHj6cXiJpN5+XlYDIZk7YtCAL33LWR519+M04kKS4q5Hvf+jI+nx/TgCpyAN09Hn79/KuK+TpEq9bdsmwR625dweGjp6ipa1BEKINeBwi0tXdgtSROnMUErBjVNXVUXbqM1WImNyeLiouXErYxmUzkDREtlSparXbUaW7hcIQrV5uTflZdUxf397kLF5NG2MmyzKkz5zGbE7+T9g6ncs7HykRZKUTFi3pFeILoPfv6kQAXavvF0EAwyAsvv8msmdNwZGYoyz/Y/SkzZxSz64O9XK6PpoQ6XZ20tnUwd/YMtFoN1Zcuc/xkGT/4zlfJSLcl7ceNnNo11nM9UFQbmC0xsNhEqsLRRPu23gzRZqmiik+TRMWgsNsYgUCA2roGSufOusY9unHRaDRsvn0tr7/1ftxyURS5fcOa69QrlRuNRx/ahizLtNYdx9mrR2dK596772DWjGmEgx72fPQGuw7WoxO1NLe20dVt4un7FnPb7fdhz84nnBl9SEteJ+1nf4lozETWWvjdBy20d0Z/sCORCO31Jzi99wDLlq/ELS7l6IkzmPQy25YJpBnhQKXMu8c8fOkuPR6PF3umxIEamSO//zcAFi8s5f57Ng85Y5QqN+IsR9Wly+z59DDtTheF+bls2riWogGVfVRUVFQmk6SeJwM8m7xtZzHZ5yZEACXzJwHQaPVKNaBRz9wnGcAOJ+ZMVqW8sQ60RvqNMdlLhx3s+FwX8Tsr49ocanKk/EIV7+/ex5WrzTjsmWy4bRXr165M6NNdm9aR5cjk8PHT9PZ6mTNrOndsuDWuqu1wLFuygJycLI4cO43X46Ikp5lcu4neuneB5JFZRqMhafS9wTC8yfW2LbfT3e3h9NnzdHX3UH2pDrs9ndNnL1BxsYbHHtrGsiULqLx4CZPRwML5cwkEgsiyjNFooK2tY1gx7pZli0hLs/LpwWN0dnVTPLWQTRtvHdKk+szZC3HCU4xgMEhBXg7p6WnKhG5MUJCkMLV1jUwvnpLQj7lzZgLR96LfvPB6nBG8IAgIgpAg4tx953p0A2wRQqEQHo8Xm82qeFZNBrIsYzIZCQZDCZ9ZLPHXjqgdeuibkW4jGEqMvLNYzAlBAy53J+UXqtBoNCxeWEqa1TJsHyfCSmGoe9bb4yTNW4FB218lsLOzm3A4Qnu7K058Anjut6/i9UXFRVmWqW+8it8foPFqM9OKolGGnV3d7N1/hAfvvWvEfk02ob50UVFMTbYYz7mOPYsHPpMHimqpPqsn2rf1Ro82Gw2q+DRJDFeZYTzK+WeV9WtXotfrotXu3F0UTS3g7k3r1YGtioLRaODRLfPoqGlCwkDBwkcxmDMIBz1UHnyWnoZysk2ZhMNWZlst5Fk9pIUvEmmzEE6P/thYCtbQduo/kMPRl4urwhzaO+sAyLN60GvDuH1GXN09uNqv4uysZnZGG2vyQwQl6PRocHm03L5AIBgMYTDoOV9RTbA7hCYS/dE/ceosV6628F9+8I24cHH31TM4ew1kOHLJctiB4Qce13qWIxKJcODwCaXE9cL5c7l93RqMfc+ycxcu8svfvKy8bHZ0uLhQUc13v/UltSiAiorKNUGrt2LJXtzveRIOImjEOM+Twc/Gnsb9+N216G1TlGeprXgzHed/S8jTkrSKXLI0Nr+zipCnGaN9juK1orEVI4gGZMmviDHJxJxUKuUNtd+RBKqxDrRG+o3R6EzYijYmHewMd6yDJ0eqa+p49jcvKb8dTpeb1996H1mW2XDbqoR+LVuygGVLFox0KQzJlII8HnkwmvYSHFDJEJJHZq1cvph3P/gkoZ1Vtywedj86nY6nn3yIuzdv4H/83b+ycEEJel10WBUIBPjdS28wrXgKRqORru4egqGoMBL7TTUajUMKTzHmzp7B3NkzRjhi+vaZKJrE8AeC+P2JnpiiqKUwP4/2Dhderw+D0UC2w87C+XNZND8qyp46Ux4nPEFUsNCJIjZbGvWNTUwrKuSBe+9kYZ8xuCzLvL97H/sOHMPv92O1Wti8cW2C99JEYTQaWL50IfsPHiMU6hfgLGYTa1ctj1t30YISdr79gZKKFkMURXY8dA8vDIo2A1h364q497l9B47yxjsfKuu98c6HPPXoAywdIo0yxngjI4e6Z3vLXkyoEhi7tsKRcEI7tfWNSoRaMBRSro0Op0sRnwBqLjek1K/JormljZ3vfEj1pctotVqWLp7P9vvuSmqwP5ior2txXHqirTiaRpqK1+t4mQzf1hs52mw0qOLTJLFi2WL2HTia8ADLzclOMN1TibJ6xVJWrxg5tFrl5iYYDHHm3AUlr3zOrOkjvoDF0NumYrLaCQe66bn8LkLxJlw1H9LWXI9fEnH7+oVdt89Ii7Od2b0u2sqexVa8CeeF55HDQWQ5AhoRTdvHGLQimSY/ywtasRkD1Hemc64li7kBGbmninsW+fEGRM7WBSmv87Fxnpk0ow4JI3LWRrrqX8RuCrAkP+otEQiLtLS2UV5RpXgd7PvgRTrr9uANajjTnM3M2aU89fCd+Bo/GDb/PHbM12KW44VX3uLEqbPK303NrVRcrOEH3/4KGo2GXR/uTXiehSSJjz45MKQvhoqKispEY8lfRkTy0tt6GkETfY3NmLEVrd6a8Gz0dVTQ23qGUG8zoiUHiIo5nbXvEey5gkZMHMQMlS7hc1Ugy2HCwR60umiUg6DVIWj1mHMWoRXNQ4o5AyvltZU9S+ase/E0HVHW1xrtdNbsItBVN6Y0jZEGtclELb1tKob0aaRNXTeq3xit3jrssQ6eHPlk36GkaU6f7DvE+rUrU/79Hy2ppq1v2ngrbR1OTpw6q/Rz0YIStm25PaX9NLe1oRMTo3oikQgnT5czb+5MPtl3iK7uqLWE1WJmxvSiCRdiSktmJRXRBEGgdO5Mrlxt5vCxU3GfSZKELEcwmcx4er10dLhwOt3k5GTx5rsfsfG21Zy7cDGhTbe7i4vVtSxcUILVYqLD6eLEqbPML5mNRqPhk32H+GD3p8r6Hk8vb7zzIRazOWn1vFTw+wN0dnWTmZGe1LrkkfvvxtPjob3DRSAQxGIxk5eTxda7Nsatl2a18IXHt/PCy28q/mI6nY6nHn2AJYvmIQC7PtqH0+XGbDaxfu1Ktmxar2zf1u6ME55i5/GFl99k7pwZwwoj47VSGOq9sHDRY+w60kqLp19oykhPQxS12DPS49oQBAGHPUP5W9Rq0WgEIhE5IZhzpGiuyaTX6+U/fvFbPH2pqpIkcfxkGS5XJ3/0rS+NuL2n5aTyPIylW3fW7gKikxQwcQWChkL1bU2OKj5NEoUFuTyx435ef+sDxZAwLzeHr35xx3XumYrK9aO1rYOf/OK3yksYwIpSGw/v+BJGS38o+VAzvQNfcL09Tsrf+QfcnV00Nnfz8QU9+QV2jMa+mcewyPEGO6sWexG8DXjbzqDRWRAEEY3Oit91EaugZW1xmLLmbEw6CbNOojijC60QQfB5sBpCSBL4QnDucoiH15rJsGi42Gpg3QM/oqGxCa0QIdMUvccHzjq1tzsBOHOugvf2XWBJvgajKLEkv52Ky2GOvV/BvFn5KeWfDz4HE/2D2dLaHic8xWhovMrZ8xdZNH+uUm1wMI19lXRUVFRUrgXhoIdAd70iPEH/AC6ZT5POmofk6yDgqqLj/PMA+F1VCIIWo31OwvN3qHQJyeciEvIi6tMTon38zsqoB9IQkUKxlLu2smcJuC/ReuonGNKnIZrsSmlvyetERoYxpGmMlO43nPeIIX3akCkmQzGayZH2DlfSNrq6ewj0VVybaEaTtq7RaHjq0QfYeucGmprbyMl2kJPtSGk/rW0dHDpykuaWNhwOuxL5FMMfCHCuogqj0UB3jwdZlvH0emnvcHH7+om1lZhSkMeG21ax78BRZZksy5TMmcX+QydIs1pIs1oU3ymAdqebjAwbFrOJYDBEa18q4IFDx3C53Jw8U55gSB6JyNRcbiAiywj0C4dl5yooO1fB0sXz+fTgsaR9/PTQsaTiU2tbB8dPncXvD1AyZybzS2croqQsy7y962P2HThGa1s7oqjl/nvu5KH7t8S1ccuyRRiMBvbuP4LT6WbqlALuvOO2pBkUSxaWMnfWdMW3qnTuLCUr5ZZli1g4v4SQJGE2GRMMz8vOVSQVU0OSxPkL1UOKaxNlpZDs3hQNadyx9UlqnnuRUF+EnVar5fZ1awgEg0qUl0aj4YFtd9J4tVl559NqtWRnOWht64gTpQBuW3PLiP2ZLE6cOqcITwOpuVxPw5WmYTNjwkEP3rb+d9pYanYsYjbm0TfZqL6tyVHFp0lk5fLFLFk4j8v1jRiNBjU1ReVzz8s7340TnvKsHsy+Rso//U+WbvoOwe5GpdRpbKZ3sMGfVm8lreh2jr/6t0qo8PnWDJrbW3F1V7N00Tw8vV7q6q8QjkR4yTaTe0o9mHVh5Eg3acWb8DTuJRzyoRWNZKYZuVV3FY0gE46ASQwxJ8eLGHER0ch0B2Qa2oNsWmqg0K7FoNOwOncKDl0bQeEMYpoXKayhJ6CLi7zKz8shHPRQcfJDAmGRM83ZLMlvxyhKLM1vo8stENHOJOsG+BFqGEZAami8ypKFpWRmpCctdz2U/4SKiorKeEgWrRP0tOKseDHBv2moAZxWb8VRsgMn0YFHbPAhaPWYshfgKNmR1KA8WbrEwApIo0mFHijmZM66l9ZTP0EOBwl01ZFWtFH5vRPNDkWIGk2axkiDWlvx5gnzHhn4ncSOaeBkUTLhqiA/hw5nogDlsGdOWvGdsaSt2zMzsA/yxhmO93fv4/2P9hEMhahvuEp941Vmz5weN4A3GY10dLiYOb2IKYX5eL0+TEYDWlHLJ/sOc+/dt8d5JEE01aisvAKAxQtKlUq/Xp8fi9k0bKTY9vu2sHD+XMrOVRAORzhfUUVl1SUqq6Iii16v57Y1KwiGQmTZMzl49CRdXd3IQH3jVUVU6ezqIRKJ4PH0JqTreXp7CYZCmM2mBHPucxcusmTRPLp7klew7uxMrJJ34tRZXnjlLaWa8YHDx1m8sJQvP/UIgiCw59PDvP7WB1ysriUcjkb2VFbV0NrWwR9+7am4thbOm6uk/o2EyWRMSO+srKrh7V0fc7WpBbPZxNrVt7D1zg0pVdwbicm2Upgzazp/9aff4ejJM3R3e5g5vYjFC0sJBkOUX6giJEnMK5lFRrqNru4e6huu0t4RnSSdXjwFW5pVeZ+zWMzcvXkD80pmj/u4x4rT5R76M6d7WPEp2N2ILPnj0rHlyIB0zOzFk/7efSP6tt4oqOLTJKPX61LO11ZR+SzT6/VSU1sft8ztM0bT5TqaaSt7hubmFkKeKzi9RgzWPKYYPGgu/gaj1k9aVx0ZM6NlSmtOvRr3QrR+nkBTm5lWl5fmlnauNDUjy7B0/nSmpbuo7xAptoewGKCnYU/0BSsiIUs+8tO1hIxhfCHQajTotBE0BJCRiciAILCw2IhOC3oRZECUrtJx4QV0RO/xti4NejGipN7lFRQzZ3q0uk6W9jJ5Vj0tHisV7XaW5rcBIEdkdNlrb4gfH3tm+pCfZfaFbG9ct5qdb3+Q8PnG9ZPj4aCiovL5ZajUN2fFiwTclzBkzlKE+5EGcFq9lYwZW5H8boKdlwHQWQuUVL1kJEuXcJQ+RtjvHnMqdDjowdN0BEP6NAJddcjhIB3nnsNgK44TtEabpjHSoDbsd02I98hYqzdt2riWC5WXEjx2tmxeP2kpd5Odtt54tZn3P4qWtNfrdBQXFXK5/grVNXVkpC9Cq9Vw25oVWK39ZtcGvQ6dqKW2rpH2DhdNza0cO1nGlk3rlBS8T/Yd4q33divb7PpoH1Py83B3deH1+rBnZnDXpnVJLSoikQgXKi/R1NzKjGlFtLS1x032QdR8/GpTC9//9lcAKCuvoKurG6/Xj9sdnVwyGPSIA8zBQ6EQd2y4lT2fHkaWZTQaDXq9jjkzp+P1+ejt9WE0GkizWtDpdAiCQPHUQuoariT0cVpxvO1IMBjitbc+UISnGGXnKjh34SKL5pew78Axqi71C0/RY5V56fV3eOj+LeSOs7pejIYrTfziuReV/Xi9Pj76ZD+hUCjOdHvxwlLe+3APuZYe3D6jYvCtE0VKZxcM6Sd0LawUMjJscSmCACaTNiEaK92Wxp/98Tc5W15Ja3sH+Xm5LJw3B38gQHe3h+wse8rm3skYi3fdYKYU5iddLggCUwrzkn4WY+C5Dgc9uKt2ImhEDOnTsOQuxZK/bNKP4bNUnW6iUcUnFZVRcrrsPJ98epj2DheFBbls2bR+1GVhP48kq44SiwjakO6nrqEJja8Bk04i1xLkV3sDTD/3f5hRmI7DKmFq7aDUkoPfWUnA68YviVS02ynNdmHWSTyx0cbH57PwSxrWLLCjMzlYUtiNUZToDerZX5PGXaUdaCISCBoE0YQs+YgEOhG1Rmw6EVnWIEux8sICMhEcVghJUeHJHwIZAVEXIezvQmuwUThzFW3NOUjNn5JhCfDQigDz1m2m89JbhAPdpGXk4Pb5MWglSrP7Z39NJiO4jxLOyrvuAtTM6cVMKcxPKFVstVq4ZelCADbctgpZlvlk32G6e3pw2DO5e/MGxddKRUVFZaIYKvVNo9VjyJyFo/TxlAdwUY+nXYR6rirLQp4mOmt3JY18im2TarpEKmlqA0Ua0WQnrWgjHeeei0ZAddeTPuteRcwZbZpGqoPa8XqPjLV6U/HUQr77jaf58JP9NF5pJsuRycb1q1m8YHIHXhOdti7LMhUXL1HXcJWKi5eQwmFFpMnPy8Fms9LR4WbG9CLu2xqtxNva1hHXRn3DVdr6UvJtaVZ8Ph9vvPMhmZnpTCnI4+1d8d99U1Mrh4+eYtmS+RgNBlzuTl589W1MJmPc+fP6fPz0mefj0uAv1dYzdUoBBn18ZNXl+kY8vV6sFjMrli3iVNl5amrr6fH0RqvveaBoaqES7ZNuS+P+ezazdvVyqmvqsFosvPnuhxw8cpIOZ390ii3Novg/brv7Dn7+yxfiBEe9Xs+WzesT+uLz+fD6fFy52kJXdw86nY7cnCzOV1SzaH4J9Y1XkaRE0+xgMMjps+e5e/OGEb651Pj04LE4gSvGwSMn2XrnRiVKLyfbwY675tFa9T7+UPQdNiwYefLhu/AP4+M51LLRXJPNLW0EQyGmFuaPOxpLFMWEyC+L2ZxyZcmhGKtIPZili+ax59PDNLe0xS2/ZdkisrNGTos1ZZUmPE8FjUigux5zcMmwz76JOIbPUnW6iUYVn1RURsHxk2d5/uU3lL9rauv5Wd3zfOfrX2TmjOLr17GbALPJxNzZM6isqolbHgiLiDm30lr3HlrBxJT0XgLBCPcv8dDsBovopdNnpdPXilBZwVRbD2npuZxpjr4kNHWbKLa1Y9UGyU+3MXXqdHLEXmyGJgRBxuUROFhpZPuyTmR0RCIhNBr6ZtqiQpgcCSDLoeifWiMCMrIsIggeBAEMumgUVEQWiKAlGhAlEAl5iASc3L3li4SDt9Fy4sdEgh66q55HNGaiNdiYv/4R9le/wBRdJUZRwi+JVLuyeeyO3Bsm/FYQBL7xlSd47c1dnDt/kUgkwqyZ03j4/rvjqnNuXLeaDbetIhAIYjDoJ23WWkVF5fPNcJWCkqUqDzWACwc9OCtfifP6gKjnk6+9HCckCFCTkS4xcBY8llpnsBUT6K5HDgfpvPQ22tLHlZS70e53pEHtRHiPjKd607TiKXzjK0+ktJ8bkVAoxH8+9yLVl6KRcw1Xmmht62B+yWzM5qjBtMVsxlJk5o4Na5g1YxoAuTlZrFm5jMPHThGJRGhtdxIOhzEY9WgHRBcdPHKS+SWz4ybnZFlWvBbd7i7y83KUzz49cCxOfPrw4/0J/ot+v5/6hisJk6MajUYRzRbOm0trWwedXd2Ew2GkcBiT0Yjf5ycUktDpRNbdugKIpknG0rIqqy7xfp+huCzLBIJB9Dodv3/1bf7LDwqYPXMa3//2V9i7/wjtHS4K8nLYuG41ebnZcX0x6PX4AwHKL1QpAlMoJHG5rpELldUAFOTlKP8fiC3NmlSUSoVQKJqK1uPpZfbMaeTn5eBydQ65bndPD9mGfsFj1W1baErvpKPtCiVzLBQv2U6k48iYUllToa3dya9feI2rTS1AVBB87OF7r2ta3FCMVaQejE6n47vffJrdew5RfuEi+WkeZs1bzW23rVPWGS4SaTzP8Yk6hs9KdbqJRhWfVFRGwQeffJqwLBKJ8NGeA5958amzq5vqS3UYjHrmzZ01ppDcRx/axk+feV7JMwdYtnAm06xXOS/LSLJIc4+ZLGM3Bh0UZUFLt4zFCrIMbQ1nKbnrMbLscympeh1D9yH0ESdGQcJugSJbJ+XVLeSVBMiy+NAIEeSQzIOLejBqQQqG0JsyiEi9EAkRVZuEaOOyhIyeSFhGq9OjiUSroGgEkGUBjQBmA8hyGFFrQWuwIvk68LaX03z8x+jTChHN2YTCQbS66A+arXgTems2X398I5fPdOLukdCaV/Ktx2/FkW64ocJvbWlWvvKFHQSDISKRyJAmsIIgTIpBrIqKispARpOCNlSKRE/jfkKe6IAt5vEEKIJUyNOc8PydjHQJxbNwgKehaHaQPuteOi+9jUarx1nxIoKgQTTZR9zvaFJCJlJM+7xWb9p/6LgiPEFUiLlytYWayw0snN/vMWQymSidMytu20cf2sa04ikcPnqKg0dOEggG0Wg0lF+owmIxUzJnBt3dHkWMCgSD1DdcpcPpprW9A6PBkBCR43J3xv19trwyoc9ZDjt1DVdibzkK80pmK7/hb777ER5PL1aLGYNBj98XQAqHkcIR/IEgTz32YFKBo7aukWWLF+B0uamuuYxepyMYCvHx3oOEQhLf/toXmFY8hS8+vn3Y81pcVIjX608qIjldbvz+AE8/+RB7DxyNi6LSajVMnZLPogWjj7y+2tTKz3/5At09/SmJt65aTmFBLrV1DQnrWyxmxX5A2b/eSsHCxzD13Vehqx8i+d3orPlx99NoU82SIcsyr7/4Uzpag8SG7V3dPTz/wu/5/le2kjN9xZjbngzGI1IPxmI288C2zdy1qrAvEukKsuSFFCKRxvMcn8hjGMhEpCN+FlDFJxWVFAmHw3QMUbWlpbX9Gvfm2vLx3oO8+8EeJS8/zWrla19+bNQm+vbMDP7iT/6QC5WXcHd2UVSQiaX3KD5PmEBY5JIznduKr+APytjM0OOFvAw41mSnKMODPeyhp/EAkZCPrUtCnDtnpKoWpjoEzAZYNVsm2+6BsA+BCKJGJi8Dur0RBDR0esJYwm6MehnkvhcyQUZAQzgioxGCBCUBORjCoIsgaqO6VFgGjSBHhShBICzoMJmzkYI9COEAflcVciREJORDa0hH0EbD3GMzyxkFiynR6xN+dG7E8Fv9oBB9FRUVletBqtE6I6VIWHKXoDPnYnTMUbZzlOzAl7MIrWhOeP5OVrqEKasUX0dFwoBIv/gPcFe/CX2V59Kmrht2v76OClxVb6Cz5ilRW7EIr5CnBTtMmph2o1RvkmX5mkbgDhZ3LGYT04qmUN94lVAohE6nQ6/X88XHH0z4DRUEgVW3LEGv0/HGOx+h9fmUz3p7vdTUNrB541oWLyzl9bfe50JFNb4+T0tRqyUQCNDW7qSwIE851qKpBQn7GExuThY+vx/NALuDwoI8Hn1om7LO4eOno/sRRURRxGQ0IssydnsGX/nCw9y6Krk3TigkIQjR6CpZRkkBi0Rk/P4Ar7zxHj/6/jdGPK+CIDC/dDZ1DVcUH0+NRsO0oinodTo6XG5K5szkr//8e/zkF7+l1+fDbDKRn5fNjrvmUZhji2tvuIF8w5UmGq8088Y7HxIMheIEuUNHT7L9vi2YzSa8Xl/cdgtK5/Dc869Fv4P8XO7YeCtTC/PjhFjJ7ybkaUY09UdHjSXVLBnVZz8hV6wnPT+a4hcIixi0Eguy22k8/w5paTdeJM1Ei9RjiUQa73N8oo9hotIRPwuo4pOKSopotVoc9sykFRhyJsjw8Eakrv5Kgg9Bj8fDr59/jf/2Z380qhe/mOq/YN4c5W9XazvdXV20R2YzPeMinqCOPIuAgIzVKBCUNawuauFwQx4lhR58Hefxtp5Ba0wngohHU0BEcKLTehGECLOzupDCMjptn8AkgMWgocsHRp1MKCih12rQ9qXLy7JApzdCmhFkQcaglWnrDmG3igiiBuQI4TCgjRYVloGA140WPzqTHVkOgxQg1NuGoNGBHMZWfAfB7vq4mWU1/FZFRUUlNUYTrTPSwGSgmBNDq7dizVs+qj75XBcR5Pjfu9HOWo9lQDT4dyIS8iH5OpB8HTiBjBlb6azdpaQWRkLxg+eJEtNulOpNR0+c4cOP9+N0ubGlpXHHhjWsXb2cCxcvEQqGmDtnJmlWy4Tuc2CKXIyC/ByyHHa237eFtDQLC+fPxWwyDdnG8VNnKS4qpLKqJi69zuvzsXL5EtKsFlatWMqho6cIh8P4/AEEjQaDKOLzB+js6iYzIx29Xs+dd6yLa3vp4vns3nNA+VuWZQRB4NGH7uXO22+jofEqmRnpcRH6vV5v0uMSBIFgIMiiYTy5FpTOYc/+w7gHVa/LSLeh0QhcbWqhs6ubjHTbEC30M2tGMUsXz6e720M4HMaWZkUUtehEUSmGsmXzelYsX8TpsvPRqm1TRAzes3HX3VADeUmS+M3vX+dseSVen48zZyswm03MK5mNXtc/DK6ta+AH3/4qH36yn8t1jaTb0phSkMeBIyeUide29g7KL1zku9/6ElNy0xUhVquzImn1+F1VOCtfIWPG1jFXlBxMbzgdvyRiFCWW5LcrXqcxG4eJTvGbCCZapB5rJNJo3rsHRyXFvAIlvxvRmDnuYxjpdyoseQkHPZ+LqChVfFJRGQWbb1/LS6+9E7dMEAQ2b1x7nXo0+Zw8cy7pcpe7k9rLDUnTDXs8vezdf4RLNXVYLGZuXbWcmTlhuhv2EgkHcZQ+jt6aS3WTn8ZzF5AjIaYbwecPYtULtHjSSNP14gmZKMzwoxHgjpktZGaVEPE2IsthJJ+L3h4tJ852cktfIRKNRiAckjHo5L6opuhLmE4nkKWDQDBCJALygED0cCSCXisTkQU0CMgyZJo1+ANhAt4IZr2AVgNdvQKyxoBFLxEMa9GFQoh6PxrRDAYjkrcNQaPFnL0GS+4SLLlLbqi0OhUVFZWbhdFE60x0ikSyGWpPy0mc519Q1rHkLxvzrPV4JyKMjjmYshfgay/H114ejbroM1M3ZS/A6Jgz4fuE61e9KeZJVFiQx6Xaen7/ylvKZ909Pfzm96/zq+dfJbNP6BBFkfvv2cz6tSuBiUl1WbZkATWX6xOWL1owl823p/b+FwgGycywsWhBCS2t7fgDAdKsFvJyczCbo96K9ox05s6eQdm5CkStFqPRgLfXi9vdhSSFuWXpQqZOKeB02XlqLtdzy9KFWMxm7rpjHdW1dXyy9xAdTheRiMyMaUWsW7OCnGwHOdmJBs1Gg4HMjHSmF0/hcn1/ZbpYpJYtbej7ZvMda7l4qZbzFf1eTDqdjul91ewEQUg5inrdrSs5eqKMdFv8/tasWhYn5tkzM9jU964dvffqUoqE+fTgsf7ItT7Nz+v1cbmukbmz+/2wZFkmJ9vBFx57UFn2j//2TEIlvpAk8fHu3dy/UowTYmMCsK+9nEjQg6ARx52mBTBzdgkvvprN/OxmrMaQUjHZL4lkzLr/hkt5nSyRejJTfjtrduF3VStVRwHazz+Pr+0sGtFI2tT1CRPKo93vcL9TRkcJvU3H8DsrPxdRUar4pKIyCtasXIaoFfl4X/QHvjA/ly2bP9vV7oYzdJSSVAbp9Xr5l5/8ig5nf4rihcpqtt+znjnmIAH3JVpO/Ji0uV+gYv8vMGrDZFsCaDURAvJfOJwAAQAASURBVHotPSEDS1fehta+lJZTPwdJwGoGo8lKpLcR0eQgotHj9zqZluHlweVhjKKARtAAMnpdNFVO6LNy0ggCseAsk0GDPxgBWSASAUEDokZAq4+KTgMPJxSW8XjD6DVaQn1tWPQBQrJIQNJijIQJB71oRBP6tALCfjcyAlcba3n3zGukZWRz69IlWIwNn6kfDZXhCYfDRCIRdDo1fVElEVmWaWntIDfHkVCtqOx8FYvnJwoHn0dGG60zkQOTZDPU3razyue97WXo0gpSimyYDI8Prd6Ko2QHHeEgvvZygp1RL6KYp9VkDUavdfUmr8/Hr59/nYvV0SIloijS2+tVqo5BNM3rYvVlNBqB5X2VuyRJYufbHzBzejF2vWtCUl3WrFxKbV0DJ0/3T8blZGfx2EP3pnw880tmU1Nbj8VsYub0ImV5bk62YuRdWJBLfeNV9HodgUAQl8tNLEjK6XJz+NhpTgzow64P9/Gtrz7JtOIpiBot2Q47ZpMJk8lAerqNX/3uFf78h38Yd85iaLVabl21DJ/PR2ZGOk53J4Ig4MjM4N6td3D2fCU5WY4Ek3CIevH88Lt/QF5ONq+//T4Gg4Fshx1RjEZSLZiXPApMlmXOV1RzvqIKvV7H8qULKZpSwLe//gXe/WAPl2rqSLNauXX1MqYU5PGTZ36H0+lmamE+m++4jSkFedG+j0JwPnmmXPm/2WzCZDLi8/lxuTuJRCLKc3hw5V5ZlhNM3GN4XLWEA3lx+xt4T4aDHkRjxrjFkUgkwp59h6lrcnP+fDv33aLpq0hnQrItZ978RWNue7JIRaTWZczm9NnzOJ1ucrKzWLSgZEQf2clK+fV1VOB3XyLYG6327K5+E505l96mI8hyBEPG9DFPKA9+/mv1ViwFa3BVvKhEU8WuEb+zctwG5zcLqvikojJKVixfxIrlN94Df7JYOH8uh4+dSlhuHvQCFePQkVNxwlOMXZ8cY+pTd9F9pRKNpxXnoX/GoguTZpBw+YykGUJ0+vUIAjR0ZTFbf56svBm4Wqpp7gJ70IVOJ6L1ddEp5xLxyljECIumaxE1gNBvqhkTm5JlBBr1GgKhCGEE9EJ0I6HvnwaIyNFZO6tJg9WkQdQI0TQ+HSCDJhJGFgUM2hCyFCEi+Qj7uwiHepGkIHKom7Crlz1nDLgvXGVmoQlteQVLNnyNjIyRQ9CbmlsJRyJMGeDvoHLj4/P52fnOh5w+U05Ikpg9azoPbruLwoLc6901lRuEuoar/PFf/T2t7S7SrBa+980n2XZnf+nxH/7Xv+fjnc9cxx7eWIwmWmesA5Nk4hCAyV6Kz1URN7CNVcqTJX+fz0snOmvekAbDN7rHx1iEsWuZPv7qzl2K8ARRUenc+YtMK55CliM6cOvq7iEUCgH9qWax/58qK+eeTasmpGpVZ1c3W+/cwB3rb6W+8SqZGTbmzp4xqnL3t61ZwbnzF7lc36gs0+v1PLr9HuXv/LwctBoNsizT1dOjCE96vQ6XqxOXu4v5pbNJt6UBUc+ll15/h8ceupeay/XYbFZsAyKI3J1dnCorZ83K5N5N27bcTiQic/DICYxGAzZbGlqNhtfefF9JDVwwby5PP/FQQiSTKIo8/eRD2GxWPvx4P4FgAK3WxPTiqXG+UgN5/uU3OXGqX8jdd+Ao2+/bwobbVvGdr39RWX7qTDnP/uZlpQ9Ol5vzldV8/9tfiROgUhGcw+H4yKVZM4qpuHgJSQrHAqFYtmQBy5cujFtPEATsmRkJ5u4AYUMRtqJbE+4fQatHZy1ANGYA4xdHPvxkP3v2H2ZaQSYlC3oRwl5CUogsRwG3zTMSCfWOqu1rYXg9kkh9+IKbf/73b9Ha1oFWqyE3J5slC0v5zje+SHZWYoRerI+TlfKrt01FNNkBCHqaiERC9LacQtDoEY3pZM3/otL2aIT2ZM//oKeV9rJfEAl6ABCNmco1MlCgaz/7K7R6a1zxic9SCp4qPqmoqAxL6dxZrFqxlKN9xpQQfel4/JH7ks5UXK5vxOXuosfTi16vU2bD6huu8Lf/+iLTstO5fXojshzCpovQHbARkQU+vVzILEcXRlHC6D1FsNdKS0M5V90iRjFMICgTDnnRasEkXKUzGEarjyBqo5XoACIRSOVd0KDTEAxGkoo7GgEQQKsRkPvaFLVC30ugzJUOifxMGQEtECYc7CUs1SFHgmgF0Gkk6lslVuVdYaoD0sQAnR1l/ObZf+TJL34Di6Y76Y/H1aZWfvviTlpaoyHVDnsmT+y4TynZrHJj88vfvRJXDan60mX+4xe/4S//9DtYLebr2DOVG4V/+fnzbN28jqd2bOPU2Qr+/l9/hcvVxRcfuy+6gjz89irJGevAZCRxyJy9EG97f5RJxoytAIrwFPI0KYOWwf2AiSvXnex4nZWv4HdVKYPdkKdJ8ZtJJfrpRhfG/P4AZeUVCcstFjOt7R2K+BQTJyxmU8LveTgcHnVK5uDBeVu7k5dfeQWPs5YWj5WC/Fwee/jeURdbgaiA9N1vPs3Z8kqOHD/NlaZW8nKyuNrcSpapC6tjBgaDnvnz5lBReYmAt4viHC01rRosFjP+QBCj0YDT1amITwDNLW1U1Vwecr/tQxTKgai59wPbNnPPXRvo7fVx5Php3t+9L26d8gsXee/DPTx4710J27vdXTReaUYG/P4gaVYrd96xLqnnVtWly3HCU4y3d33M8qUL434n3/tob5wvFoDX6+Xv//nn2DPTEXUiK5fMZkleW9w6ycSeaJpj/3ppVgvLlizAajazaeNaZs+cxrS+dMHBbFy3mtffej/p8oH3R+zekSU/hvSiCRNHDhw+gUEb9Xoyilr8UiYV7XZyfUEiwZ64tkcSlibqnj9y/DSfHjyGu7Ob6cVT2LJ5fcL9kKydYFjkJ78/yLO/eYmeHg9Go5F0WxpNza1oNQKvvLGLb3/tC0n3OZkpvwPbikRCSiSpVm8lZ8m30Ftz49ZNdT+Dn//WgtW0lT1DJOhBo7diL32c3qbDcdeIwVZMV8MepN42JK2ejJlRYbq3+ZQyGQI3fwpe6pK9iorK5xJBEHjikfv43re+zObbb+O+rZv4rz/6bkKIMkAoFOLYyTIqq2q42tTC5bpGTpWdp6vbw+X6K1hNAjPtXbj9BrQaDWEpTK6lB4M2xCxHFzWudBDApvPiuXqSiBQgEglzvsVKfUeYxg6ZcERGpw2Tmx4VmiKyrMwOxlLtUkGvT+HxJ/eLUQLQ0xtG1ECPTyYohQEBORxCDoeRo/7maDWwfXErC4sgyxZdt6NXT3unxIUDz9DdsBdfR/9Ltdfn46NP9vPdP/1rDh87idfnB6Izff/5qxfx9HpTOyCV68bVptY44SmG1+vj2Ikz175DKjckFVW1fPUL27GYTaxbvYxf/PivefuDffzqhTeud9duagYPTPSxaKS+F/9gd2PS7QYPDoKeFmUwI4hGfO7quPU7a3fRWbsLiA5ChAEGwwO3jQlLsUFNbB/tZ3+J5HUmDP4G/h6kgs91UTEXN2UvIGv+U0pUlq+9HJ/r4ohtDHfsN0KaRyAYJJwkrX9KYT5ypP9HPt2WhihqmTqlIGHdhfPmAv0RMgNJFiETG5y7q98kHPQQiUT45XO/xhEuY26Wmzyrh6bmVn7+yxfw+uJN3VNFq9XS0tbOoaMnaW1rp+ZyPYc+eZVju5+lveJ1DNowSxfNY2ZRFl/cnMm2lTaWzbYgCAJpfR5MySbNpuTnDbnPgrycEful0+nIyLBxYgiPz2Mny5Iuf+Y3L1FzuR6DXofDnkEkEuFXv3slafR7xcWaJC1EI9oG/n76/YGEytKRSITyimrOnLtAZ1c3Pe42nJVvUHHhbPR+mrM97noO90WWANyxfg3TiqLiUigk0dTcitPpZvPta7l9/eohhSeA9WtX8sC2O0mzRs+9PTODxx6+l6WL58etN9Zn0HCEw2E8nl4yTX7FXPxMczbdAQOnm3Pi2h587UK/IBZ755yIe37Pp4d58dW3aWpuxefzcaGymn//+W9obmkbdjtZlvnXnz3Hs79+CZerk0AgRFdXD80tbUQiEZpb26mqrh3yvjJllWIr2hgn4sWer7aijeMWY7R6K9aC1Yp3HoDOWhAVhwZcS6Ntc+A14Kx4WRGe8m75AabMGXGfd13ejatqJ5LPCRotcjhI65mf0372OTouPE+gq+GGeDZPBGrkk4qKSkrMmF7EjCRpdgPZf+h4Xwpbf2lfSZKorKqhICeNVUWdmHVB0gwhXD4zmYYwem2YQpsXGYF10zykZ6QT7r2MLEsYtHC6yk+RowuTViLDokWQQdsnCGlkgV5/GLNBm1LE02gRhKigJMjgD8oYDVpEEcJ9UVAyEQRZBllDRNYQiUSr7FkMMhEZAtFsAFx+IwtzO/D1GNAaZqG3TSUc9NBe8zG/ee8y5yobuNrUAkBPVwdrFhXSSw7BYJBTZ8oV41SVG5MO19Azyx3OxOqYKp9PBEEg4A8g9s3w5+Vk8e9/9xd898/+D6GQBGqW7ZgYqxfRUFExghg1f5Ylf4KZMEQFn8wBFeZ87eVEAj0IWl1CRE1M+Gg/+0tCnmYkrZ70WfeOK9JIK5oRTQ501nwlyslRsgMnEPI0oxVHjrScaJP2iSbdlkZBfi5Nza1xy9OsZr74xHbkiMyV5hZyshxsvn0te/cfocPpxuPpxWA08MA9m5ViKKmmZA4enLeEZlBsrFYG/m5f9Lrwen2cLrvA2tWjq5YYDof59Quv88yvX1S8NNNtaWhLi+no7KGlqQ5RfJOH7r6V3MhxrtQLdHllmjoFCgtysJjNVF26TJY9I67dmdOLmVc6mwXz5lJ+IV54zMvNYcmieUBUwLlQeQlPby8zpxclTXEK+INJ+x4MhhKW1TdeVd5bBiJJEkdPlLFty+1xy43GRN+pGAM9qQwGPWlWKz2e/kG/09WJz+dXqufFBJmmNg8z09aQ1Sf2JIuEMRoNfP/bX+HjfYd49jcvIYoiDkcmJw68xYXz5/jG176K0WgAkqeg3b5+DRtuW4U/EMBkNCYV/ybDD02r1VI0tZCGxqgg4vYZCYSjw/apRTPJnH2X0nY46BkxynK897wkSezeezBheSgUYs/+Izy54/4ht62uqWPPp0fwBwKDtpXo7vGQkW5LiHQbzGSm/AY9rbSVPYMcDkajSdMKkbztaDQ6JSop2N04qpTFmOAXSw0VjRnIskRawRolmir2nfidVfS2RwVeQdCiz5hOsLOOUG8rIU8LosmekOJ9M6OKTyoqKmNmcJjvufMXsafp2byigP1nO/D7A2gEAaNRz5LZdsy6bjJMQbr8enqDespbHdwzrxmtIDHN7kOjs6ERPMgRPWFJQquRuXeZgLtXRiMI+INh7Fat4uUkRWT0Og29gTAGnSYuBW+iEICwLGPoMyXXiRAM9flJxZSpSACtJiqKyYCMQEiSETWQZhK41dqCL6SjV5uuVNJoPf0zXFfPUWIVqRCixpxmA2xdAhnWFpwRMy0eK71eb8K5jv0fiHv5+Kzkg99sxPy5kr08TZ2Sfx16pHIjMnfWNA6fOMvmDauVZTnZDv797/6C7/yXv8XvDwyztcpwDDUwAYYtX53MN8aYORtf+7m4QZk5Z5EiPlmyF6O35sUbDIc8iNrMhIiamPARK8Uuh4O0l/2CnMVfw9N0ZEyRRqasUuwkDnQdJTtG9RswmdWjUuFKUwt7Pj1Ma1sHuTlZ3LH+1jiPvIfu38LPf/l7xdMJoubcjzy4FYu5X2Dz+wOcu3CRxivNCIKAxWSk4mINrW0dZGUYU07JHDw4D7ftjos4iQ38IVplb7Ts+fQIB4+cjCvi0tXdQ9XlFkRtAdnZQfID3YQb3mf5gunk5+bwqw/aKSk1oNeJCILAjOlF9PR4lN+a7CwHTz76/7P3nvFtXGf69jUFHQQJsFNiEUVKonq1JEuybEvucYtrYidx4pRN2WSzSXaz2ZL/tndrNtnNpq3Ti+PYjnu3XGXLkmX1wiaxiL0AJEAQdcr7AcSQIEGKlKgSB1d+sYgBZubMYDBzzn2e534Sg/777rmNHa+/zXsHjqDEFZYtWcg127ciyzK9fV5+9NMH8foSkyGCILB541puu/m6lDYurqk2LBY0TcfrGyAwFKR6fgVe34BhjA4wNDRs/B1XFNrbuwzTcqvVylVXbE7xiVq7ahkvvbJzQkRbTraLRQvmG68FQeDyLet5+vlXEtvt6KaltY3gcAiPJydRtCGY+M4Gwlaq+oLMLZta7BEEgaPH6inMzwOgyBlkYd4AEWWIN3e+ztVXXTOlGCyKYloD9STnykvpxuu28aOfPmgcL4DNZuPqbVtShJfpCktn85sPBkMMTxKJf7rIp7b2TgYGBzGZTEhSDE1TjPfCkShzSuwsWlA15Tk+V6ixIN7ah4yopOS9WRRNxIIJw/mhtp1E/S3TTllMRqIlJzIAdE1BjQwy2PQ8stWDozjhwyaZnYgmG7oSSfEVlO35xEO9IAjI1lxyKq97XwhPkBGfMmTIcIaE+2vxNTyeMvtqlhXWlPSQbY2SlVVGq9eCJEvIkkTM7KB3uAVBgOGYmYNd+cQ0E56VdxNt/DFqLAB6GMniJhyJ09QvUO4O4rCCJOp4h3QKcyQkKaEuxRSduKJjNgnYLRLhmIokigizrT6RqJin64mUuriiIwiJKnkwJs1PGP1H1UGSEuvZpURqYJYUw+3JITbcy0DDY8QC7WhqFJsMV9ZE6euBrYsFsqwwGFTpVmQQYcH8eSm5+jZPDcGuPSmz86oSMiplwB9+PvgfGrkeN+vXrmT3GF80SAyU1qxcNslaGf7Y+NRHbyMwZsCWpCA/lx/859/w5POvn/9GvY+ZjseJ2VU6ISomFmjFUXIJNs9Co7PvLEpEuQi6YAwaIGkwXGxULhobUTN2X7I9l+yqDxhmsz37v48luyLFUHYmzEYUwLmqHjUdmlva+N4Dv0JREoPQ9o4uDh+p5fOf/qiRBlVVWcHX//xPeOfdAwwM+qkom8sla1ZMqNz2yhu76O/3MXfOaOpZYGiIx556kftuWT0jr5ixg3PXSJpbbZ8nRXiCRCrb6zt309XTR3FhPpesXXHagfPe/YcxmSYOu/q9PiIVpQTkhUDAWL5k0738x+UejhyrIxKNsXhRFR53Dv1eH00tbbiynCysrjQicWRZ5trtW7l2+9aU7YfDEX7528cM4QkSaVA7d+1l/rxyIzIK4LrtW2k80Uy/18fxuhMEhoKYTCaGQ2G+9tf/QnnZHGw2K5UVZWxYtxKTLBOLxzlW20goNJoy1d7RyU9++TCf/eQ9xrJcj5t777qFhx9/jvBIelWux819996OIAgcr2ukt89LcVEBV1y2EU3T+I//foBBfwBZNuFw2PH5Bmk80cKC6nmGIJOfP+q7NtlvYDgUouVUu/F6IGwloshYZYVYxw5iwRVn7Md2Lv3TqudX8OXP388bb+/B6x2gpLiAy7dsSBEBk0xHWDqb37zTacdmsxnf3VgK8tMbhSfJz89F1xNVEqPRGKqmoY0YwYuCwPIli7jzg9dPuY1zRSzQhiiZsbiryK25G7OzEJOjkIHGJwGwuqvIKt2CEvFN27/P7CpNpG6PTFhYPQsQpNH71nDfIay5C4xzPjZyTo0F8dU+jBLqRTJnIdvzMNnzZnRvPh/G8mdDRnzKkCHDGaEqIZSwFyXsxUvCjHVjhY/h7iAIYJY0o5NYs7CKG669kp/+8mG8ve2J8HXZwd03XcOc+SsJZZnpfPdbSCY7uhojppkJh4YJWCDbnpi1ynMJmGUBQQBFhaGwSjSukZtlQhZ1nDbpnGWtJKvhaRqEIipZjuStM+E3JRhmU4kWSCIouowoJGb4BF1DEiXE0HG63/0v1PgQkslJf3wuA+EhnFaNOzfbCYXCDEXguQOwdKmZtcuXML+yPCWkerjvEAhiykMt1HvYSBF5P+SD/yFy120foLi4gPf2HyESibKkpprt42Z+M/xxs2RR1aTv5eW6uf/eW89ja97/nM7wW7J6Jo2KiXjrsHkWpmwvKUDBeIPh8rQRNWN9YGyeGsyOAgpWfJKe/d9HV2Mo0UEchasuyGz2uaweNR2ee/l1Q3hKElcUnnv59RTT4VyPmw9ce+WU2zpyLL3HVcOJJsSsO3CVTT8dauzg3G6zUliQR43iS4l8qiify2NPvYg/MBr99Nqb7/DFz96XVhQwji8ex2oxk+3KSllX0zSsssqKOanRVMnB5trVqdWV83I95OV6OB219Sd46PfP0NbeSW39CQoL8iibW5JSoe/1t/Zw+FgdjSdacDjsXLp+NV/70qf52a8fobm1nVxPDnl5Hny+QU42n6KxqYXVK5bQ2dXDoSO1bNq4jt8/+XyK8ORyOfG43dQ3nqSltd0QEzVNo7iogL/40qfo6fNikmUq55UxHArzre/+mPaOLmMb88pLWbtqGTWLqkZECp0Dh48Ti8Xp9w0wN1SM3W5lfmW54ec0FZIoIYoimpYQPKJqIpptZXEfLlk5q7TTc1VYIMmcksIpU9qSnE5YOtvfvCzLXLFlA8+99NrE5ZdtnLJtyxYvpKJsLi2nEtdUKBQmEo0hyxL33XMbX//zzyJJ0mmPcSZMV4CZbsrkTFIWx0fL6koUQTKlRDala4caCzLY9DzRQCu6GkOUrZidJQiSedrf08VeTAIy4lOGDBnOEJtnIbb8pYbnhRL2kS10EzVJHG/TaemXEU0Jz4G7PvgBcnJc/M1ffIGmllNEIjGqKsuxWi0judY/AU1BjQURRBmr5mNhsY4ggqaDzSIgCokHk6rpaBqEIxouu5Qor3wuDJ/SoOug6IlwdHGkAl5i0jEZ/pQQoFQNZElFFM2gxRElM4Iooqsx4qFeRNmKLbeGvOJNvP7oo6wq7iXL6cBsNvH2fp2Fi0r5+L23G6aW42dr9TEhy7oSBV27aLw6/lgRBIGtm9azddP6C92UDBkycPpUlLOpoDSTdbV4mGDXHsPTw5JdgRIdRI0M4mt4HFG2p0RTjeVczWCfy+pR06GltT3t8tZTHWmXT4UsS0SjMYZDYaxWC3ZbIipYFEVEUcAyzSixdIPzavMOXK5W8vJCtIbnUbN4OQ2NTRPa7w8M8ewLr/HRD39w0nYuqalm5669LKiaR8PJZvz+hNhUMSeXe67MwWnRZ00IfGfPfr7x9//BcCiMoqgEhoLEYnE0VTO8O2OxOK+8/jbzR14PBYM89tQLDAwGkCTJWK7rOqfaE8JQNBojFI7gsNtGPJl0Ltt0CV7fAKqqYbWYcTrtxGKJynyd3b1UlM9l/8GjPPnsy/gDQ4iiyPIli7jrtg8gCALPPP9KivAEiarJPX39CIAkJfp3S2sW0NrWgW/ATyweY/v6Tdx4XaqR/GRYrRaWLVnIoSOj5v5RVaa2z8Pi6tHUqDNJO5XMTqy5iwj1Hp7gHWfNXXRe+mTTEZZm4zd/9bYtWKxm3nzrXQb9ASrK53L9VZczt2Ryw3tI9I/+7R//kq//3b/T1tGJ2Wwm15PDlkvX8Wef+8Q5EZ5mIsBMJ5J0pimLzqI1aPEwoe79CFJiEjKZOje2AmHy/p5sXyzQhqZEsBcsR5DM6EqiCJEgW6f1PZ1rMXQ2yIhPGTJkOCMksxNH/gq0+DARbz1B7wkGB/10BWx0hez0D0bYsnEZf/LJe4ywcFEUqaqsMLahxoL0Hv0NihJFkK1IJgexkBcBbcRjSUeHFB8nSRTQNZ1Ct4lYXMckiaPazzk27JUkcDtlBBJRUOk0L20kPQ898SHR7ESUrYCAEuonYVQuoMaDLFpUwk0bXHS096OqKoIgcOfl+Vx525dxuPJS9z3mwSeIMqasOaDrxkPtfHp1ZMiQIcMfAlMNGM7GJHi66yZns/2nXiXmrUcQJayeBZjsBSghL7oWn5CCkeRczGAnBztj259cbsurmdD+cyV+uXOy6e3rn7A8aSg9XXRdJzgcYt/BoynbWFhdyarlSzCZph95mm5w7llwC4LwJPm5ATaXbcCWV8NTz+1Iu/7x+hNTbv/qbZdR39hMb18/SxZVE4nGsFrMfOUT2zAPHz6tKKBpGoeO1FLbcBK7zcq61StSPLKSDAWH+bfv/IjhkWgkWZbQdI2BwQCyLFNWWoIsy3T39uF0TDSnf2vXu1RXzTNeD4fCBIPDiKIw4tkz2vFpPdXBpg1rOHjkOPUNTXT19MGIR7w7x8XefQd56rkdHDh0jOKifHKyXWiaxsEjx9HR+fi9d3Dg8PG056unt58sp8N4bbVaWFhdia7rfPkL908r4mkst99yPYODAVpHDLwtssr2lWYKC0b7WmeSdhrur2W4810QRHQ1jiCZ0NU4CCLDne8iyfZzHmkyE2HpbE3Rz3SSbV55Kb/40X+y/9AxfAODlJWWsLRmQUok3mxxLgSYmaYsJu6TrUYffeznk8LT2Pt7LNCGEvKiKWFMziIcxWuxeRYa36GzeD2iyXba7+liLyYBGfEpQ4YMZ0i4v5ah9reIBFrRdQ2/fwhd18izR1hSBKKUS/2JJhpONLOwunLC+oqi8OYzP2S4vx5NU+kdUFhfOYAggHlkEkQQhBE9KdXIWZYFNA3sNoF4TEOWRUaT3s6e0Yimia/FMRl2mj762qiMJ8BIlDiiWUCLhdG1OLqqgCgimbLQdZVw/zE6dv0zcwuqsdoW89ArPSzM68cc8fPyI//Mqm2fp7xiNE1n7INP1xSjJKzoKkeQTLQeeow9LTn0D0aYV17K1s3rycmZWUc+Q4YMGd5PnG7AcDbeSdNdNxZoQ4uHUWN+JHN2IlqVhF8UgkA82J128DfbA6jxg52kMDZezBorPJ2r9I3LNl3Co088N2H51s0zG9Tu3LWXgYFB8nLdRmXRQX+AwFCQD9587Yy2Nd3BudViJjqmapemaei6jtUyeTU3gCyng499+IP84sHf03iyhaLCfO6562ZKFy4k3O+Zcr+apvHAzx+idozA9cZbe/jQHTdxyZoVKfvZd+AIweFUb7nsLCcDgwGi0RixeBxZlsl2uQxfq7HEFYXKilKO1zXS1d1Lc2s7/kDA8OwZ2x1zu7NZsayG/v6EMXkSRVVpbm1n17sHGB4OMTDoZ2DQT/X8CvLzEimDh4/WMegPjN+9QWFBHiZZTklRhIQX2EyFJ0ic/y9/4X5aWtsZ8HaTpx3CLMbOOtpsvL+PyVlCfMSo2pa/9LxEmsxEDB/PbFWMmw5Wq4VL16eP8pxNZluAmWnK4nQ+n+7+rqNjdpYg2zyG5+CZVEy80MUkTsf5yVXJkCHD+w7J6iE61I4S6iMeDeILyZgljSxLDLctykAQ6hub+Iu//Vf+8u/+jYcfe5ZwOGKsv/Pp7zDUW4uua/x2Z4yDJwPEFQ2TqKPp4+WmkSpyYxaKI9FFsklE0/VZrXI3vpKuYek07v1kBJTRtpHPSAIoom1koZIIm9UVBDFRilsyJ0QhNTpINDLMb1/zE4xbOdpbRESR0WNBnvz9qBnr2AeZIFtTjAsF2UKfL8jRIweQfDvp6mjhtZ3v8K3v/pjBwck7dhkyZMjwfiPcX4saSwyCk/dNJexDU2O4F9ya0tlPfu5cY3aVYnVXYnaVJSYttDjRQCsmZxH2guVkV1496aDQXX3zaJsbHp8Q3TDTdow9/liwO2WAlM40d/znew/9BCXkTfm8GgsS7q9Nt8tJ2bxxLTdetw27PWHS7XDYuen67TMemO557yCCILCgah4rl9VQVVnOkppqCvJzcdhnXjkrWQFxLOMH55esXQkkJtAaT7bw7nuH2PPeIfq8A3R09ky67Z7efv73/35Fb5+XbFcW4XCEn/ziYQ4fqzvtfg8dqU0RniAR9fX40y8Ri8VTlvsDQZwOR8oyk8lEXq6bsrkl3HDNNj73qY9wyweuntDXAZAkifXrVrFm5TIjDdJqtSLLMjabhYYTzUAikv2ySy8BEobTebluRFFAkiRMsowrK4u+fl9K+061d46aFIxEra1Yln5QvXbVMr7wmY+ypGYBgiBgkmXWr1vF/R+7c9JzPB0qyueyqMxhCE/u6psxJ8vYj1zvsUDbme8g3Uk9D0zn2j2XKIrCG2/t4bs//AXfe+BX7N57IG0F4PNFUoAZy5kKMOMjy053vUzn8+nu76JknlCE4ky+w8kmXc7XM+90XFSRT20d3fzDf/wIf2AIp8PG33zlM1RWTFS3n3rhdX71u6fRdZ01KxbztT+9D1mWT/tehgwZZo/4UAe6FkcQJRDsxNUw4biMzaQgCyoBbxNen0JBfi7RaJRde/bR09fPn37mYwx1HyHYX0++M0yP38SasmEaOkGWACG9Kj4+GglGRCENZOncP+zH7zspdmkjolPSlDzxYbAJCpIlGzXiNdaRzFmgKYnQWc8GYsEO2iNlDEcTncqkCabbFqE7KHGsrpEVS2smGNcGu/YYxoWaEmHX8SBZQqJyS2JdJ0PBIK/tfIdbb7zmHJ+ZDBkyTIe/+of/5l/+7kvG65/+5nE+cc+oyfgnv/RNfvzff38hmva+IG0aQ9hHLNiJ2VGMGhk4r75GSRKDoO2E+o6CphLztwCghPrIW3wPZufE1KnUdc9+BjuZPjf2+H11j6DGgliyyyY1zR37+b7DPyUe7EI0O/HU3JU2EsriWUhzSxuarjOvfO6U/e9tl29i6+b1BIdDOB32tJ/t6OzhnXf3MxQMUjmvjA1rV2GxmAmFw7y3/wiHj9Whaxoedw52u80QsxRFQVXVWUnpGZ96eM22LQz0d/HWa8/Q158oPZ+X60YSRb73wC/5xlc/nzad7eXX3ppQKSwSifLdH/yCm27YzuKFVcyvLE/bhtqGkwD0ewfo7OohEo3isNuZO6eYllPtLBiTJjd/XhnFhfn09XuJx0f9IUVR5JP33cVN12832rxrz76UKC6AS9asIMvpwGSSWbNqGUPBYWRJxOsbpKOrB/9QEJvNyt2330hF+VwikSiSJKa04dCRWoZDYTRVw5WTbUQ4RaMx4nEFs0nG6XRQVJDHjddto629k67uXmP90rklXHvVVoLBEDabFY87h5xsF4sXVaWtKnjkeD2v79zN4GCA8tI5XHXlZoqLCtKeS5h+pNB0iAXa0JUIVs8CdDWGIMqI2RWGX8/FUmHsXKLrOg/8/HfUN540ljWeaKappW1aZukzYbqpwLNdzdNZvH5CerQ1d1HatMrpXl8zub/P5Lh7D/0EQRCRbR4j6ioy2ETPgR9RuOozxjYuVAW8i0qV+bf//im3XH8FN1x9Ga/ufJd/+taP+Ol3/zHlM53dvTzwi0f5+ff+CY87m7/4f//FE8+9xu03XTXlexkyZJhdRJMNk70Ai7sKNBWhq46+YQEdAVELMTSc6PSMrf5ysqmVltZ2LKYsouEIglVhbk6MSAi2LhZGo4fSaEmTRTadJ6/xSRFFUFUdYVwDdVTi4YTwlPCBGolWEiRMjiKcczcgyXaO7u8BRmc0o6pslBCORBKdwvEPMtFkM2advZ11HO94DouUbwhPSZpazmL2LkOGDLPKuweOpLx+6LEXUsSn5taZmy1nGCVtGoOuYXYUI9tzjfvnmQwwz4bEIGgHsi2PuBrDlDWH+FAHsi2PQOuOKQdDszGAGi/Kucq34at7hKi/BV2N4SrdMum2xg6OJJMTRTIj2/MJtO6YkAbYNSjyy//7H0NoyHI6+fCdN1GzcPIqj7IsT+rzdOhoLb/4ze+NCmWHjtSyZ+9B7vrgB3jgFw8RDA4TCAzR1d2Hw25jSU21IWBVza+Ykd/TZKRLPRS0CBsrA1gGrOR63PRHc7BZLQB0dvXwL//5fcpKS1i0cD4b1602Kp6ON1P3+gZpONGMrutYrGZeef1t1q9bxYduv3FCO2xWCz19Xk6cbDE8NP2BIQJDQfr6fSnCz5KaapYtXYSu63R09RAYCmI2yVx/9RWG8ATgcefw+U9/hGdeeJXGE804HHYuWb2c/Pw8nnpuB/UNTUiSiMedTb93AK9v0Kjs67DbqKxIGJJbrRbmV5ZzsqnV2LYgCIRCYRx2G7m5OfT1ewlHokiShCxJCILAB669ElmWyXLKfO1Ln+ZobQN9fV6KCguonl9O48kWfv6b3xOPx0fO1wAnm1u5/Zbr2bxxrbGv3XsP8NCjT485rwMcq2vky5//BEWF+Wiaxr4DRzl8rA5RFFi9cikrltbMWgqaLa8GLR5OFBTQE9eqIMoIkhlH/or3vfAEiSyHscJTknffO8gVWzZMKQTOhOmmAs9mNc+x+7TmLgBSsxFcZZenXW8619d07+8zSYEeattJdOBEYpJg0Z2YnYWYs8vxt+xIWKQ078Cz8JYLWgFv0mGboqr86OeP8Lmv/RP/9f1f4hvwp7x/z2e+PqsN8Q36qW1s4pptmwC4YvM6evp8tHV0p3zu1Z3vsnnDanI9OQiCwK03bOPl19857XsZMmSYXWx5NWRXXo1ksiNKJqqrKhBEEzFF4u0GMw1diZK7voFBTja1MjCYuIe8+sY7fOt/f8Ov3wjTNxAjFFFYNAfMciLySZsQpXvhwnZPR1InkyUhkSqoJ9qf8HxK5OPpOoRjMsGoRlwZWUHXEtUC82pYlMYPCxKzlAurEu+Nn/FIPiRigTayi5djkuUU0SpJtivr3Bx4hgwZZoFx97bznK3R1tHNp/7s77nzE1/lE3/6tzS1pK9A9odAMu1rbBqDr+4RtHgI2Z571mkMZ0PSSFYJ92PJrkCSbYmKd+F+lJB30hSf8QOoM00bHCvKeeseYaDxSUN4EiQz4YHGSbc1dnAkSCYsrnKUUB9K2JeSBugov54f/+apFA+foWCQn/7qEYaCw2m3PRWapvH4Uy8awlOSzq4evv29nxAc2ebckmKsVgvDoTAdXYmUN7vdxlVXbGb/waPU1p+YsI2ZMFmqYmR4AEW3oMi5hvDU0dXD8boTHDleT239CR5/6kW+/8Bo+vxYD0ZN0zjZ3Iqu68iyhDQyi7Zn74EJ6XWqqjLoD7B33yF6+7z4BgaNVDabzTrh84Ig8On77uaeu27h+quv4N67buHf//Gv+PIX7p9wfGVzS/jcJ+/l2//6t3z1i5/i8PF6fvf7p3n1jV00nGzm0JFaBv0BGk+2EIvHEQQBi8VMb5+XX/72MWM7H7zxWiPq7FR7J76BQcKRKEPBYQ4dqWXunBLK5hazavli1q5axuc//RE2rFtlrJ+sgLft8k309Pbxzf/vv/mrb/47O3e9S1NLW0r61os73jS+U13XeeHlNyYcVzQa5dU3E+O/Xz74GL95+AmOHKvj0JFafvarR/j9k89P+b3PBDUWJOyrRVciKb9TXYkQ9tWet1SnsSnHKW2bYUrsmWy3uXXyic7mSSpbngnTTR2eaZrcbOxzpszk/m52laKpMZSQN6UNSsiLpsZS2pBVugWLuwqzo5hA6w5iwW4ivkYkS3ZiAiE6MGvHcKZMKj794Ke/Y8++w1yx+RL6vQN87HN/ndIx6e6ZWKXibOjt85HnyUEeKbcoCAKF+bn09HlTPtfT66WocLQyQXFhPj293tO+lyFDhtlFjQWJeOuMB27Fmo+ybv1GFlbN5cp1xZQV5xAIBOnp7aenz0tt/UmO15/gnb37GQwM0dYb5YEXBrHI+khHQjfS1zQ9nQh1caHrEAgqJPu1AkkxSidZDCZhPi4giTr9QRttg3YEQUJXY8bDpXJeGevGGYZCoqRtTo7LmPEY+zBKPrQCp15HGzrJ+jGduCSCILDl0nXn6OgzZMhw9lwYb5AkyWjzh3/6n9x7543807d+dEHbc6aMvUdCIm1B1xSi/hbiwU4srvILarRqdpWmGMm6F9yKbPNgdpYklk/S8Z+tAVQy2itpipw0RrblL8WWvxRdiaQVs9INjmR7LrItLyFeaQlRxVW+jdoTHYRC4Qn7jsfjHDh0bCanC0j4I6Uzo1ZVleN1o2KLySSzYmkNlfPKsNtsfODabWxav4b/+9lv+eVvH+NHP32Qf/r3/53Ui+l0A+vJfLdcOQUc7C4gqiYirRRFpa09YTI9NuWu5VQ7+0eO/7JNlxjL/YEgiqICUFSYb0QzARw5Vp/SnseeepH9B49hs1oRBIF4XGHQ78ckyyysnpeSrpZElmW2XLqOz37yHj51392sWrEk7fG3tnWw/+BRenr7efbF1+jv9xnv5eW6sVgs1NafNMQfQRCorChFFEUaTzTT158YY80pKeQbX/0869euQlFUFi+qZv3aFdjtNjRNp62jk3vvvpX//dbf89EPfzCl6vFY3tt/mKee20E4HCY4HELTdLp7+jg1cm4hIWomr42h4PCkpuXtHV2cbGrl4JGJ1fTeeuc9evtmZ3w4m0LHmXK6fuKZClBTbbe38UVefOoX/OinD3L0eAPRaCztNmZzEnS6Pni2vBpcZZdPmHRwV9+Mq+zyGZt3z6b3XpKZXDexQBuCIBIb7koIUA2PJyYuhrsSy8d8VjI7KVhxP7I912ivrkSwF6zAUbw2cb+fpWM4UyZNu3vljT088N//j/xcN3fcfDVPPvcaX/yrf+E7//yXVFWWXeg+01nz6FMv8/unXwa4oIZoGTKcS4ZDIeJxZcbli6dDutKuBTW3YZKfJNvVS46liVPj1hkY8FM6p5jGEy0U5kh86uocwjEV0SoiaDqySSCugo5AKApZFo2L1bJNEMBpl1Nej94WEzKUpoEgJHIJPfYoh/pEVq/bjKYMp3iOfPiOm1i+dBFHjtUjiSJrVi01OmfTqXh0yweq0XWdd987SFxRyHZlcf01V6StMpghQ4YLg67pnGhuM6oXaJqW8lo/j4p7Mtr8O//yl0Ai2vxb3/sFbR3dlM4pOm/tmA3GR/boamxCZI+9cOUFE6BigTZEyTytMuhjmU1fGsnsxOquJtS9H0EyY3GVG22ZrB3pnvGu8u10v/cddDWGGgsiW3MSfiKh9F5FMJo+PhPsdhuCIEzonwuCYKSxGccmiRQV5FE6t4T588r47x/8LOV938AgP//No3zjq59LEXmmm8qSzpelcNENrFy1lz17DwAJEUTTdOw2K3m57pT9n2hq5ZI1K1i2eCF3334jL7z8Bv5AMNHuwnxK5xSnfN5kGu1XhMJh9rx3EFEUyMl2YbFaiMfjiCMRSB2dPSiKyvf+75f09nmRZZnVK5dy9ZWbp0w7DIXD/OQXD3OyeTRVrrm1nYqyOcY5EgSBmoVVHDleh8NuR5YlCvJysdmsxjrBYIj8vFwgcT+rrT9BOBJB0zSKCvNZtXwxiqIiSSJVlRWn9eB6e/c+42+r1UJwOOGp1dPbT1npHATAbDYbhup2mxWbzTbBSwsS4lljU0va/ei6TuPJFgryc1OWT+arE+g7geSqSlsdcDZ/p1PR3NLGj372II0nW8jOdnHF5g3ceP02crJds14ZM8lk2/X7utl3pJl97R6iqoyqahyvP8HC6nkpnlx5eR5qFs6fleNPMl2fpKnS3qbrnzTTfc6E5H5UJZSyH3f1zUS8DSmfNbtKkW0eo1iFaHKixvyYXWWJiQxXaUr707U3p/I6gIuiAt6kw7rgcAhPTrbx+ubrr8BqNfPFv/pXvv1PX5v1hhTke+j3DaKoKrIkoes6PX1eCsfdGAoLcunoHFX5u3r6KCzIPe1747n9pqsMLyhFUdhyw32zfEQZMlw4/IEhHn7sWY7XNaLrOnPnFHP7LdedUXnayZjqgdu0ewcWzwBF4S58A4OIgkhengdFidPV3YtIlNs2CDitOiZZpLVXZdFcE7ouIIkjaWvo6Be5yJ2McEpnhg6AAPG4iEnWkcQ4Swq8hL3HyZl3reHbFO6vxZZXw7LFC1m2eOHEfUyzZOwdt17PjddtYzgUwp2TPStGqxkyZJg9ItEYH/vcN1IqZ370s98w/j6fRZKmijYfLz5d7JN1yXukt+4Rwn1HDdHJlr/UMP2dqc/HbHI2g9PZ8qVJDExaMTmLkUxOBMlkeItM1o7x7U56V5mdJei6Rm7NXcYAt9Rcj0VWiSrShH3XLJrc82kysl1ZLF5UzbHa1EGYKIpctmk9HZ1dtHV009vbj6IqZLtcXLbpEt4b56uWpK/fS2tbR0ofaLoD9sl8We648SbmFBeyd/9hrDYrPb195GRnE43FsFlHBZqxkVAb1q3ikjUr8PuH+Pb3fkpgaGhCW9esWmb8PTgYQFEUBEFgTkkhTS1tWMxm/IEhgsOJZ32/18eRY/WUzS1h7pwiXtzxBo0nmvnCZz46qeH7k8/uSBGeAPq9PswmE3NKRg3wBQHKS+eS7UojuthszClJ3CsG/QG+872fcvhYLX7/0Mg597GwuhKPO3tkW6e/wQWGRqPQiosK8PoG0PVEZJmmaUiiyKYNazGbTQwFh+nq7mXl8hre2bM/ZTuiKLJ103q6evom3ZfTmWoKn06MHPB2895L3yXo76W+z43ZvZDbb7mOsrklqedikt8pJK6f6Yock7F332H+6u//nUAgcX7a2rvo6uqhta2Dr//5nyBPs584Uybrf5441W8IT4BhOD8cihjiU3XVPO6+7QNGX3Smgs9knK0P3kz8k6azz1igLeW4wv21KGEfgsmKs2gN4f5aJKuHWLDdMCdPHjfAcOe7RLx1Ke0P+2rTCOAJ4V+NDREf7gFBRBDN5C2+ByCl/WZX6YT2DjZNTDU9GwP2s2FS8am4MJ/GplYWVY+a2F1zZcKP6Uvf+LeUCgqzgScnm4VVFbz4ytvccPVlvPbWXgryPBM6QVdsvoQ/+fN/4JMf+SAedzaPP/sK27duOO17GTL8MfHjX/zOCAGHRPjxD3/yIH/9tc+T5XRMsebMmOyBa8quxmQ6TGVFKZUVozMug/4A/f0DzPUkLJEURaehPUaOQ0ykr0mgqgJxTUQS45hGqthN4kF+UZDwedIRBWFMGxNJeAICVrPGSHQ9HrtCPNSLLiRSLaZr9jfdWRer1YJ1xH8iQ4YMFxe7XvjVhW7CGfGHMFl3JpE955OzEZHOdtA2dmBlyS5Pa7472XbSRUKNLQWeHJiaowGu3byQJ19P9R/atGHthMie6fLhO2/ilw8+bhgZW61Wrr/6clavXMoXvvJN2ju6jM/KssTe/YcnrRYHGN5LSaYzsTOVcfHgiafYtO5m1qxayo9//jvejSs0tSTivQvyc5k/rwyTyZTibQQJYcTtzub+j93Jj3/+O4aCCUFBkiRuvG5birCRl+vBarUSiUSM9Lzm1nYikSi5nhyKCvPpHhFYTrV3Eo3H6O8fYPfegzS1tnHbTdeydfP6lP1rmsb+NCJdfq6Hvn5fivgE8IFrr6S5tS3lfCeXJ6PQXt+5m0F/gFyPm86uRBCAruu0nGrH7c7GYbexoKpi0u8myfzKcry+AQCynA4WVc/nVHsngiCQ5XSweeM6rt1+GU8+u4Odu95NfKeCgCzLmEwy4XCE4qICrr/mCuZXllNcXMAzL7xKJBJJ2U9Otosli6pTlqUTI99++lvEw4NEFJmBsJVosIMf/Pg3/PXX0lc0HMuZiBzpiEZj/ORXDxvCUxKvb5D6E00cPlrH6pVLz0l0DqTvf+4+YTKEpyQOu42C/Fy+9qVPI4piSpTYbJ2LsetoaoycqhsZ7nxnzHe2HTXim3JbM40Sm+oeMLaqXFKI8tY9StTfhGTJJuZvIxY4RTTYjq7GMNkLyFZCRLx1qNEAzuL1MxDAE0UrNCWMIFtRo340JcJA4xPGJItkcSFZPYl2ISDbc3GVb2Ow6XlCPYfQtTiO4rXkVF53xgbss8Gk4tN12zfz3sFjKeITJAQoQRD40c8fmfXG/OUXP8E/fev/+MVDT+Gw2/jrr3wagP/v2w+wZcNqtmxcw5ziAj75kdv4zJ//AwCrltdw6w1XAkz5XoYMfyw0t7SlCE9JIpEIe/cd4sqtl56T/Y7tIC9dvACHw44SCaRUYMvPy2XzxnX89FcPU9qjElNUegc0jrUJzC8GkwySqBNTBZxWkaQhr6Ylqt2dz8iAqdBHjMR1EpXuJElA12DEuQpR0EEXEMSEDCWKAqJkQ5JFBMmMbMmZkdnfbJeMzZAhwx830402/0PhTCJ7/hCYjUFbuvS56aT9jed0EVxXrK6hbGErBw4fR1VVli9dNGWlu9PhsNv57Cfvod/rY2homJLiQiwWM4P+AB53NksXLyAajeFw2LHbrITDEcOIezxOpyNt5PfpBuzTOXe/f7mW5tY2FlZXcuJkC/7AEL19XooK8vmzz39oQmpXkvLSOXzzr75IbcNJopEoC6orJ6R1mc0mtl++iWdeSDz/CwvyiERjSFLCoLu7dzSyZ3g4RGtrhzEB1d/v4/GnX8Rht7F29XLjc7quo6jqhPaUzi2eEJ20oLqSD1ybGEft3nuAuoaTOBx2Nq5blSL0nWxOiG5ZTgelc4ppGxGqIpEoAvCRu2810gB9A4PsO3CUWDzOkprqlO/l6is3c7yu0TCUd7uzyc/P5b57bmdJTTWCIPDOu/t57c1do43UdRRFYd2a5XzwxmtS0g3tNhuf+fiHePCRp8b4UxVx7123TIgKG//dtu77lSE8HezKN8SWcDjMe/sPc/mWqYMbZisVrrm1jcFBf9r3Bgb89PQlPJhnIyIoncgd9tUT8dalfHZZsZ89LdkTBCinw5HW5mO2zkXy96ipMQRBZLjzHVzl2wm07kAJefHWPoQoJSpLT3ZPm242wfh9prsH6CEvuq4Zx+Us2YAaHxpZ5ifibyY+1EE83IcgiFhzqgj1HjaEImvuAqy5C07blrFFK6w580HXifibUaODhHoPY86aiyW7DHf1zanV7mruwuwsxOpewFDbm+iaimxxG/5SF2piZlLx6cO3Xz/pSldfcSlXXzH7A9jy0hIe+M7/m7D8G1/+VMrrm6+/gpuvvyLtNqZ6L0OGPwYmM1883Xtnw/gOshJo4/67tnPozZ+DkuhExMU8br9uFTWrr6KoMJ+f/uoRnI5B3jkRpDRPos0ns6hERRDAblYQRUAXUDUdQYDhiIrTNjGk/0KSFJ4gITppms5QWEMSBRw2ETQBWQJZFJHMFmSrG0tWKf7ml4DphUTPZsnYDBkyXDh+8uvHT/uZ+++99Ty0ZPrR5n8InE1kz8XO+EGb2VVOZKDRGLyM9/pIx2x60pwugmt+ZfmU0UdnQl6uh7xcj/G63+tD07SEUDPOy9hus7J6xVL2HzpqLJNlmTtuvT5tCtrpBuy2vBr6fQN4Q1Y0f5SCfGfKuZOzqzh0JFH1zWI2saSmmmg0hqKqLKquZOniBVMemyzLLK1ZwL4DR/jd758BYMWyGtatXm4IQduv2ES2K4u3d7/HUHAYp8NBYX4uFos5Jb0vFI6kiFdJEerNXXtTxCdJkli0YD619SeIx+Nouo7FbEaSJD52z22sWrGE7p4+igrzKS+dQygcpvVUB/MryydEUUGi/5fjlElaHpfOLWZuUTZmzUt30Mk3/+rPcI9U+tt34Ai/efhJNE1D0zSeem4HV162kbtu+wCQ+K6/+qef4o2336W9s4s8j5vNG9elRGPt2Xsw7bl8b/8Rbr/5ugnL51WU8o2vfo7unj5EUaSwIC/N2iPnZowYmRQya/s8E0SW6fSlZypyTIbVYjEqCU7YhyRRXFR41v3EyUTuZCqzbMtNubeWlWQTjfaliHIAl25YfU7PRfI+I1k9BFp3oEYDBFp34CzZwMCJZwxvvdOJWWO/ZyUyaLwemyqZvDee7v45NovB3/wSlqzEvnU1RmywCR0dQRCRLNmga6NVEUeOO9xfi7NkgzEuAHCWbEi5N48vWuEq38ZA45OE+o6Arqe0P6t0C2FfA4IgEmjdgat8G7FAC5bsSkSTjex52yccw/l+Pk4qPtWfaOGNt9/j0x+7fcJ7D/zy92zdtJYF82f3AZMhQ4azp3zEMDKdN8e88nNTTnNsB7n30E+IRsMM9Z+i0O1Al8qo2XI1Hr0BPX6CtgYHA4N+rlxTxFyHysoqJ9G4goBIXPVjNWlGhFMiikhA03UcVglNT0RAJUn684rnOSJKGInCkgWBuAKylEi9k0SBbIdENK4TjYMs6MiSALqKpkSQHQVoWhx9xKh1OiHRszVrnSFDhgtLw8mWtMsFQeDQ0XoCQ8HzJj7B5NHmf2hcLPfI2fI0GcvYY4n6Wwl27Db8rNzVNwNMKwpqtryjLgYK8vMQRXGkSm4qJUWFXHXlZtavW0ld/UlsNitrVy/D486Z8NnTDdgd5dfz60dfTPGdWrV8CffcdTPyyLmLRKKo46OIRvpf0Vj66l9jCffX8tRrR3nnvdFqZCcajtPR8A63fuhPjGXr1ixn3ZqEgDToD/DP//6/xBWFXI+bto4uotGEqXsyDc6V5TSEKH8aoeTyLRt45vlXjagZp8PO6pXLKCstwWI2s37tSgBefWMXz7/0OvGRlMXy0jl8/CN3GNEtSdHi0kqFhkaFqCpjkRRWzvVjlcFUsMQQniKRKL977Fk0TaO9s5vOroRZ+uGjdQSHQ3ziI3cgCAI5OS5uvmH7pOdsOE1VRUhUVkyYm0+cpBQEgeKigkm3mWSsGOl0OkCAmnzfBJFluv6ps5EKV142h/nzymlr78IfSPUIW7p4AcsWLyA20HBW98DJIpPiwW4ATM7ilO3q+hPE4wrtQ3E6/DImWWbLpku47NJLJt3HbKUFJo9j7PH5m1+aUNRhKpLfsxIZJB7sRJHMDDY9T+6iO4DRe2rU30JW6ZaUc5dOmBp7XIIok7fkI/hbXiY22IyAgMVdhShZESRTynGH+2vxN79MbLgLs7MEQZTRNYXeQz/G7Cg2jnd80QoAQTIjShaj+MNY0bxgxf3GM8NX+zCCZMLqrjTWTXrNXqhnwKTi0y8feorLN6cv011ZMZdf/PZJ/vlvvnjOGpYhQ4Yzw+POYfPGtezctTdleUXZXJYvXXRO9jn2QefvbaW//RiioBFXh9jZIrHS/yirl1QQVU385MFXCYY1LJJKrklCMMdYVOZA0iMI8fRmtuKIuiSiJ8x6RyrLCWC8HoumwWz7bSf7uTqjRuPRWMJpPBDWkAWwmkGWBMyyQExhJCpKAEFC12KEeg8hiCbjwTidkGhbXg1Dw8McPukn2LqbhdWVLKia9wedSpIhwx8j//bNL09YdrT2BN/7yUPIksRXPn/feW3PZNHmf2icr2pTUzFbnibpSA7afLUPo0hmdDWWMuA424pWf2i4spxcun4Nb72T2sfJcjrZcMkqBEFgYXXlaau9nk60fPPVpzhWm2rKfeDwMQoL87h2+1YgEV00v7Kck02tqKrGiaYWvL5BIFEJ7s233+WyTekH5OH+WrrqnkfrbsQi5Y8KN8V9MNhF07E3qFyydcJ6Odku7rv3Dh569GmGgkGW1izAOzCI1WolHo+Tl+uhvHSO8fmK8lShRNM0Hn3iecrL5mB32IhEogwFh3nn3X1Eo9GEeXR1JZdespqnntuRsm5rWwe/efhJPv+pjwCjooWHAHdtdfHi/jAVzj5sZpWc3BI2XHubsW79iSZisRg9vf2cahu1hlBVlaef38G88rnTsoVYtKDSSKEby7zyUiwW82nXn4zxYmTxgm3kdg3i7WlnZfFolM9M+tKGyKGqnGrrpN/rI7LnOHr+Vq679rppebAKgsAnPnIHcUXh0OHj9Hl9iKLItdu38sXP3ockSWd9D5wsMsmSXUbOvGuw5i5I2a5nwS2sLFrD8mvK8foGyct1p1S5m+pcxOIKsVgMm816VvYRZypmjf2eTc4iZJuHiK+BcN9R+keKVehKBE2NERk4gRLxTXpPh2QU1mj0pKqE6T34AMLIIERHJzpwAsmSjTW7MiUlXLJ6iA13ocWCxIKd5C7+MN7jDyZe04VkTUR8jv1+ISGO6UoEW/5SbO5qooHWCVFuZlc5wY7dKCMeiK7ybca6Z/tMOlsmFZ+O1Z3kb7/2J2nf27xhFd/9vwfPWaMyZMhwdnzwpmspmzuHvQcOE4/FWVKzgC2Xrks7IzRbJB8E7+39Z2JRK7n2CENRE0sKvMTD0NE7RIO/nGA4MZMSVRN59CuL+/ANDFPgGCLp8ZSCkBCYxkY3KUoiFU+ShJTlyWAvQSAlSmrSanRp0PQRYWvc5wUBQmEVDXBYJVRdB3SGQhpxFRq7VBaUSORmSUiSgNk0ookJlkRD1Bi6EkUXVazuKkzOIiLDPva9/L8c759DtqeATRvWTvCHqK0/wU9/+ZIx8/jqG7tYsayGj334tozwlCHDHygtpzr54c8fZv+h43z49hv49j99LVMs4Cy40JE9yUF41H8Kb90jKYaugmxFi6eP1pgOyUGbIJmwuMqJBlqJ+Brw1T2CIMpnXdFqppyLCK+Z0N3Tx/q1K8nJdrH7vQOEQmEWVc/nuqu3TvBMmorTDdjfePqZtOvt3XfYEJ8APnjjtXz/x7/i4OHjhvDksNvIy83hsadeID/Pk9b7yuwqZTCoYpUTglNtn4eafB9WWSGiyLT0KFQuSV1H13UaTjRzqr2Da7ZvwZOTg91ho7x0Do0nW/i/n/3WMFZXVY1Bf4BIJMaO195m/bqVZDkd1DWcpK/fiyxLFBfm09fvo6s7YRLe5/VRVJBHQ2MTR47WpRVzGk804xsYxOPOSREt5uTDfVfZiESdWBxuChffnnKNyFJiyJmuAp0gCOzctTdFfBoOhWjv6CbblUVRYb6xfPsVmzled4J+34Ax72ixWLj5hqvSfl/jOXS0ljd27sHrG6B0bglXb9tC2dyStGLkpTf8Ocfe/D/6eztYbHFRUrWByzZdMq1qwkmhQokGOFR7it0nTNTkC1jlCJGuV/i/H3fxZ3/6uWn1y4uLCvjm179IU/MpwpEoVZXlE54XZ3sPnEzMMTsnpmKP3a7DPrXxOiTORX/d4zTUH6GjZ4jjvR6WFfupKs8Hzsw+4kw9rsZ/z4CRXhjuO4rJWYwluxz3iJfUZD5VWjzMUPtbRtSSbPNgy19B93vfRo0FESUz9sLVxIc6UGNDqFF/okKdbDW2ac1dhNlZQizYidlRTLD9LcyOYmIktqlGfOBMpJwmz3e4v3bCdWqPrUyJcjO7SokMNCaENDVGNNCaeD82hChZkG2eFDPzi8bzaSg4jGmSEp2yJKeUw8wwe8RicfbuP0xzaxvZriw2XrIqJdc9Q4bpIAhCSpj2+UCNBempf5ZwOAKIDIYtuK1R+kNWNF3kYEcWPYOps1WDQYXXj6ncsXoYHAmRSNUgokjIkopFSgo4icgjQYC4moh+isU1skY8oJKv44qOwyYl0uKS7RrjyzQdJkvhUzQd2SQiSwKqBuG4gKLoSJJIU6+IyyHRMQBOm4RN0hAFCMctZNsSsyg6gCghmhy4K68nKhfz1uP/iqCG6O73cfC4k1279/GZ+z9MVWUFkJihfOjRpw3hKcmhI7UcOlLLqhVLJrQzQ4YMFy+9/T4e+OXveeWN3dx8/RU88rNvke3KOv2KGS5qJLMTm6eG4e79KOF+tFgQQZQR5IQnT7BrD6LJNuMOfrrUsIHGJwn3HSXqb8GSXTErFa2mS7i/Fl/DE5icReQuuiPFGyYe7MbDzGbTBwb8HDh8DLc7mxVLa6Yc1Hd09vDr3z1uCCV5eR7uueNm5lWcecTXVAP2eCy9R9v4dLo5JYV8+fOf4E+/+v+QJQmn00GuJ8c4lrd370srPklmJ+RfTqSxCaussKo4cVxJg+vbLk31JlIUhZ/88mFq60crCjqdDj57/70IgsCCqnl86XMf54239tDW0cWRY3VkORzUN56kvvEkr765i89/6qMTUrd6R1LvAGJjjq25tZ3q+eVpxZFIJJpyHEnRQhRF7DYr7qprJ1yTC6vn4XQ6UvaRJD/Xk9Kul17Zycuv7jT6PvMry7nvntvJcjrw+Qax2ax0dfUQV1RWrVjM5z5577TGSu/uO8SDDz9pvPYfr6e+4SRf+vwnmFsyUYyULVks3fqZMxqcJ0WOQEjjjTpryoSrVVaID7Zx5HgDK5dNb7uCIMy6n9pYzmVhm1igjZONx2nvHuKdlmzCMY2I4kFRurFYzDgKZ3Z+z8bjKp3onFN5HVp0CDUeRLa6R0S3wil9qgAC7TuNqKXs+TcweOJpQEh4PFndWLPnIQoSiOJItGoA59yNRrU7SbaTXbEdJAsdRx+nveMUwWCIXrWKtauqKEpzTtK1PxZoMyr9JT2otOhQogCHOYuIr4Hh7n1oahSToxDPojtnNTp3pkx6l8/15NDU0pb2vZMtbeR6cs5Vm/5oCYcjfOf7P+WRx5/lvf2HeeX1t/m3//oh9Y1NF7ppGTJMSfIGJqghIqrMsV4PTkscSdTIs0cQBY1yZwf5ntGw3PbOburrj1Pl7kuE4SoaOgmBSRIT1eKS6DooukBMERAEAVEQsJolVC0RqRSOwcCwjtkkMt7qaibC01SYRtLpdB3a+lWOt4PFJKLrAsVuHbPZQZ5LIBTVAAEdkSyrloh2AkSrC5OjkPzFH8Kau4Ddb73AnpZs6vvdRjXAuKLw1HOjD/9T7Z0TOopJjhyrn5XjypAhw/nhf3/8Wz70yb9A0zQefODf+NJn7s0IT+8jrLkLsHoWJGaa/S2oSjiRIjfGHHympIvGECSzMaNtpN7Fzs+EsBYPo4T7CfcdxVv3CLFgtxE1oIT7px3hpWka3/3hz7n5Q5/m69/8dz7zxb/mzo99gbqGk2k/rygK//ezBw3hCRLV3H70s98SCp95VNlULK6pTrt8Sc1EE/GkkfW8ilLy8zwpIlqycls6li1fSVso9bqo7fNgsiXEuLHs2rM/RXhKbvt3j41GaJXOKebeu26hurKCgrxcbLYxhuShME8+9/IE709FHfXOynKODthdWU56evs5VttAfWMT4UgESFg7jPVPSl6DQ8Fhunv78QeG8LfsmHBNyrLMx++5PUUkEgWBeRWl2O02Q0Q8crye5156LWXS7WRTK7/7/TN09/Tx/Qd+RXtHFyXFhZSXluDzDfLGW3vSnd4UdF3nxR1vTlgeVxReee1tAMMHZyxnGkFpy6vBVXY5vSwz/KKSAlSy39fTOzEK7EIwXsxxL7g1xQPqbO8vYtZ8XjsS4zev+dm9r45DR2rZva+eNxvsHGqzn7Gwl7wvJqu3JdscC6TXL5KM/Z7HRpbKVjeAcU9NCqtjSYr9SW8li7sKs6OYQPOLiWin/GXkr7if/GUfw7PwFrLnXUXxuq+Qt/QjeBbcirNoDe7qm3GVXY4trwazq5TOupc4dryegQE/8XgcW6yR3zzxDnveO3ja9idTvgOtOzC7So3KeLHhLkTJgj1vCZbsCkRTIkJNlGwEWncQC3bPqOr2bDJp5NO2y9bzX9//Jf/5j1/FMcZlfzgU5js//BXbL5+6xGSGmfPmrnfp7OpJWRZXFB576kX+6iufvUCtypDh9CQfBFaHB6++kHzTProGNEo9EIiaEQSYU5BFqSlEe2sUmSAHezu5YbVAeR4U5oiEohpWc8K02yRqhq+STqJjJ4oORGKIetRYnvzXahWwWwT0EU8oTQdV0zHNgvA0mr4nIAg6igqK5KE0b4hgVCMYgUgMbDYJl82K1RTHbHMjaiE0JYKux0GQkWUH1pwqokNtROvaEAb34bbZDeEpyam2DiKRKFarBZNsStsmANk06e07Q4YMFyEPPvocWU4HjU2n+Mu//3baz/zie/98nluVYTYZKwzFBptTzMHPJHrgTLw+ziXW3AXY8pcaKSpKZID4UEeirflLseZOXd0tyQsvv8FvH30aRRk16z7V1sk3//k7/OyH/z7BP+ZobWPaiZhIJMKBQ8fZtGHNtI9hummDN163jZbW9pT95nrcXHdVqg/T3n2H2fH6Wxw4fBxJFJk7pxiPO9t4v7pq3qRtEfUod1yeR2N9N8PDCRFt3bwoK7fdOiHl7dDR2nSb4FRbB4ODAXJyRkvcjxepkjSeaCYv180la1fy7sjANic7i+HhENnZWcY24orCgD9g9EcA6hubWL60hk9+7G6jEl8yner4sYN094dGUgc7ycvpZImiTEi9m19Zzn//xzf523/4FpFojMJ8DyaTCZMs84FrrgQmr2Z3rLYBi9k8IRIcYPe7B7ju6sun9B2KRKJ4fQNp32vv7J50vbPBlldDUUlqm6KqbPT7Cgvy06123jnXRRvCkQgv7m5PiZiLRKMcONbMnNLJfx+TMR2Pq7G/8+TfyWO15dWgxoKEffWjUUhpIqhc5dunjAaTzE5ya+5KSVXMrbkrJVXRaKtztGJjUtBMin4dp5oIxaSU1NuVxX28+uorXLJmhfF7S0c6s/hkZTzR7CQ80Iggylhz5qNEBtGUEErIe8YVB2eDSUcv9334Zr709VruuO8rbFi3nPxcN33eAXa/d5jy0hLu+9DN562RfyzU1aef8enp7TPyqzNkuBhJ3lxfeKsev3c/RZYow8NxGsOQlWUmq+oWPNktoMe5e2OMU21etCHwOAWcdhNRTUCNR+kZBI9LJNcxEgUFxBQZm82C3eJEiQXQ46MPL33En0lERxdAGFGkFFXHJE9+s07jUT4pSeEpgYDFrFPsHETTwR8SONCks6jUjC5IIMSJqVLC7FxL7gnQNURLNoJsIdyXKAEtylYGwtYJ+zObzZhGhKU5JYUUFRbQ3dM74XNrVy2b5hFkyJDhYuCvv/KpC92EDOeI5CBCVyJYPQtQY0HiwU7DHPxsmInXx7lOm5DMTuwFy9Fiw0R89cQGmwGwehZiL1g+7QHMMy+8miI8JWnv7ObQkVo2XpJasn14ODTptoLDk0cWjWcmxvB5uR6+/uefZe/+w/T2eSkuKmDtqmUpotCe9w7y20eeAqCstISGxmbqGk5Ss7AKd46LvDwPWzevT9uW5H7tJpXVa9ajuFYT7X4bqxRHGnoHNVaQcj6n8gYSpdRElsmMt00mE5Ik8aHbb6SyopR9B49SOqeExpPNmM1mo6fT1d1LIDCEK8uJxWwmFoshiiLB4DBW6+i2Y4E2TrU20t0fMgy5D3bls5I+mk7U4Z6bek22dXTxyGPP4rDbCASG6OnzctUVm7np+u1GNFUywmo8uq7TlaYvBCNi2UBgSvHJYjGT5XQyFJwYxZOfl5qyN5u+ZlWVFcwrL6WppY1+rw/vwCCiIFCzsIrlSxbOaFvninNdtKG7py+tgJKuYuV0mSplduzv3Jq7iOHOd40UaF2JoMXDhH21RP2nAD3h8TROdFNCXry1DxlV5tKl9gFnlaqYFP0Ggkrq72ckNVOI9hCJRFMiGCH1+ky22Vv3CFF/ayL9VTKnHO/Y9seGe4gGWrG4yhEk03lN204yqfhkMZv53//4a156dRd79h2hrrGZbFcWn//k3Vxz5Sbkc2hc/MfK+IsriSiKZ1XBIUOG88GpAQtvvHMYyEaSJPoCKksLBpDNJpZ6mtGVKBF/CzYUyvMhFHdjNYOGQFSJkWWOIQk6FpOODmiaQEQRES3ZWN1ziA93oSuJmcG4JhCNS5glDVnUUmzKdUgRnsabjSeFp/H/To4w6jlFom12i8BwRKWuNcKRzgIqSh2IqkYoJlKYrYEaQdNiIEok5bF4sB3JMppik1+xgZ63dzM8HMJiseBxZyOKIpesWZHSyfzYhz/IAz9/CN/AIJC4H1y9bQsLpphNzZAhw8XHDVdddqGbkOEckRxEJDv8ksmOOMYcPOyrx1k0/eicdFwsVf2C7buI+lvRdQ1BENF1jai/BdCR5Oml0AQnEZNUVSEUmphGVz2/YtJtLZg//WfhZCXlJ0s9sdmsk1arA3j5tbeMv3PdOSxfuojunj4UVeGGa65k08Y1kwoi6aJN1MLSScXE1SuW0JDGhmN+ZTmuLCeKoqCqGhaLmXnlpew7cASLxYw7J9sY+K9dtcz4e8O6VWxYtwqAaDTG3v2HaW3rwJ2TzTPPv0J7RzeCIGC1Wgxz6+DwMPsOHDV8KW15NexrMdM8MnCG0dSyjqE4m24brQoXiUT54U9+w/BwCJvNSuW8MiAhdI0VfxZVz+dkU2qVQUh4fC2YP29ChggkDMfzct1pz3MSURS54rINEyr4CYLAlVs3Gq/PReXKT3/8Q/zZ1/+RA4ePE4lEsFqtgMDjT7/EbTdfO6NtnSvOZdGGYHCYirI51DU0oY/xxrBYzMa1NJuM/Z2Heg+DIBoTv1bPAob7DqErESzZZdg8NROq+bmrb2aobSdRf8uk0WCni5qajgCVPLe+ww6iagsw+vtx2yKEhOIJxvLprk8AXY0RD3YBIFvdWN3VhPuOpHzOVb6d7ve+k0jZjgeRJfes+XrNhEnFp/rGZv7y779jVD7497//c+afhalfhtOzft1Kjtc1Tli+dPGCaVUTyJDhQnLoyGhIeHfQCSLUDThYae0jGAxh0b0IooSmxTBZs6jM8xJRJFRNRBR0YqqM2aQgSwKKJtA9ZCfLqpNvN6NGBzE7itGUKNG4QF8wjscWQRRGH2K6rqPpMG4CEB2IxzTM5sQbAqNV8eIayEZqXppKeyPCk64lfBGiMbBbBSQRnFaJVVV2jvqyONyTz4a5HczJiSAJOpqiYXLkAyImZzHRgRNoSpSYvxWrez7WgjU88/IphoLDdHT2oGkaNquVD91xIzffsD2lBcVFBfzNXyS8MEKhMFXzK8jJdqVpa4YMGS5mkjP6NuvIjKSu8/QLr3OypY1lixewfWvGzuBCc6YRD7a8GlQlRKj3cMpM82DT88SDXUS8ddg8C8+6g3+hq/pJVg/RoTbi4T4EQcSSM5/o4EnUeBBE0SgNfjrWrVme1t8pJzubRQsmmnMX5Oey5dJ17Ny1N2X56hVLZ2Q4PllJ+dOlnqS7LpToEHLkFDC6zGG3MX9eGa6sLK66cvOUbZmpmLh+7UpONp9i775DxrK8XA+33HA1Dz7yFPsPHiUejzMwEMBsMREKR2hubcdqsbB4URWrVy6dtCKcxWJm88a1bN64FoBXXn97klYLKGpqxFpfKJuomprOFlVlOodSJ80PHjmeNoJt0B/g1TffwZXlJC/XzZZL13HwyHE6RlLhevt99Pb2s7C6End2dtrqxVdctmFak/RXbr0UURR5feduBv0B5pQUce1VW1MEkJkKlNOhubWN2voTOB12nI7EeK7f6+MHP/41V1y24aLNbJnJ/VBVVRpPthBXFKorK1JEk/nzysn1uFm+dCHdPf3EYjGcTgdFBfksX3r66K+Z3pfH/851NW68Fx/uRbbmINs8KeJiuL/W2JZkdpIz/7oJ+x37+wRmJVXRllfDlstsHKk/ZUSCJVMzb75hw4SIsXTX52DT80R8DQiSGcmUNCFvxVFyifHcSXhb7cDsKEZ36OTW3DVjsWy2mFR8+u4Dv+WKLeu48ZrLefL51/jejx/iv/7pa+elUX+srFhaw7VXbeWV1942cpqrq+Zx120fuMAty5Dh9KSrUhNVZWr7PKyRZCyOCrydx+gaELGLfhxmHZscJ6rIqJqIJEs4bCI6EFas5BSUU5ifQ6z/EEpMQbblYvMspKO1H5PYhUnSRivaaaDpOrIkTOiYKIqGySSORjrpif+LYsJE3Ah/Gv0DIxYq+Y+Q2IeOTn9Aw2UXkUWBHIfA/VuHcC++DToeRY+KgI5kzcGevxx39S0MNj1HfLgbBAlBENGUCI3H3uFUc4jSOcUUFxYQjkQwm81IkoTJNNHnSRRFFi9Kb36aIUOGPwz+7l++x+Wb1nHD1YkIqB/87GGeePYV1q5cwn/+788ZGAxwx81XX+BW/vFythEPynAvWiyYMqjJqbweb+1D5y0t7lwTC7ajq7FENSdLNqLJjmTJRo36Ez5XwfYUb5PJ+PAdN7Fr935ONo9GuJhMMrffch1zStKvf9vN11E1v4KDh46jaiorltaweuXSGR/DZCXlpxKe0l0XgyeeYmVpmINtTPBuLCqanpfPTMREQRC4586buWLLBppb28l2ZbF4URU//OmDRkRUd08fza3tyLLEyuVLmFtSxHAoTGVFGZ/++Iem1SaAKy7byO69B4jHU/2V8vM8LBuXKrZsyUJe37l7wjaWLVmYMnAOBFLT3XRdp987wMnmVhpPtlA1vwIBKCudwyc/dhe19Sd57Knn6ezsJsedja7rNJ5sRtM1aqrn09PbT1aWk00b1kxI05yKy7ds4PItG9A0LW2/9UwFyql4+dW3UvyOkgSGgryxcw+33nTNjLd5rpnJ/bCltZ2f/foRwx/NbDZz563Xs3Z1ouK2253NFZdt5JXX32b+SMQbwOJF1WkrQZ5pO8ano7nKt+GrewQ1PozJWYIWC6CEelGjg+TMv/609/jxv8Pk9scul6we496eTjyejnBWVVnB/R+9ixd2vEFbeycedw6Xb9mQNupy/PXpq3uEqL8FTYlgK1iOp/pmQ1QK9R5G0AUcxauNSEvZnntOfL1mwqTi04nmU3z7n/8Ck0nmsx+/kzs+/pXz0qA/dq7dvpUtl66jvaObbFcWRYUXhxldhgynY/WKJbz1TuqspEVSWDk3iN3koLO7h67uGDnWKP6IGZOoIgoaNrsVWTZjszlRw73I9nw8uYtQhntQgol8bB0dq3sBatSP21yHKCUqyOkk/qPrY8zFx6XYmccJT3Flcj8oTRdGPJ7GoCeEJ4uciHjyh3R6BlXcDhGbRcRmUpB7nkDVwgiyBXveUnRdRZBGZ+HMjiJMzmKs7mpigVYGGvewsjjOwa58QCbL6QDgeF0jsVgcs3lyo/EMGTL8YVLX2MJf//mnAVBVjSeefZW///rn2bhuBYeONfDv//3TjPh0AZks4iHqb8XkLE6JeBg/eAj31xL1t6DrGq7y7SkzzQICluyKP3jhCUCS7ZjsBVjcVQi6bhjZ6oKANlI6fDrketw88L//wiOPP8u+g0dxZTm59QPXsG7N8inXW7G0ZkIVuJky3ZLySXFi7HXRW/t72kOlyEP7yXWZmVs2n9cafCnbEkWRq66YOuoJwB8Yor/fR35+Lq6s6QsaJcWFlBQnBLrOrp6UVLykobaiqPT29TO3pAiLxUxXTy9e3yCn2hLm8DULqyak84zlmm1b2H/wKM+99BqalpiUy3I6uPXGa1i2OFV8unrbFk40tdLe0WUsy/W4J0RZjRUdYvE4x2sb6e33MTwcIhaLE45EWbywilNtHTz74muUzinmnXcTAlgoHKGrq5f5lWUU5OVSkJ/H5z71kbRt13WdI8frjWj81SuWsiRN5cJ0wlOSmQqUpyMam9z3LTSJx9VUzKYn1WRMNwJMURR+8suHU7y0YrEYDz7yFBXlc43qhjdet81IB40rCksXL2Dd6uUTInt6evt54eU3qD/RhNPhYNO6xSzOzTptO8aLVACDTc8bwgyAIJlAlNDVGL2Hfkz+ik8x3PnOtKLa0olgZlfpBOFqrHg8E+FsSU01S2qq0XV9SoNxSL0+1VgQNRZElCwJf6oxHlDhvqPEg12IJttFkbadZFLxKa4ohumt1WohGotP9tEMs4zDbmdhdeWFbkaGDDOicl4Z1161lRd3vImu61gkhbVlAyyb70YJe+nsHsZljSEKcQrsYcyyjiBALBrEbrIjiFnY8pchmexosSFiQx2oUT+SNRuruxpX2eX0H/sVVjGCpuuoOoTjMjaTgiiOjVpKiEyKqiOPiExjb+OKriMjJP438oZGwtNJFFK9oERAFxJRUroGkiggixBXYHBYI65J5LtF1MggAPaiNeQt/pDxYPTWPpRITRhjZqjGVhJ79wBWOYzbFkmZMRUEYUI4eYYMGd4fhMJhcrITvm+NTa2omsr6NYnCASuWLKDP65tq9QznmHQRD0pkACXsxeQsNj6XbvCQHKgRDRBo3ZEyQJLtuWSVbrlQhzWr2PJqyImHDc8UAEGUEWUrrpJLZzSAcTrsfPzeO/j4vXecq+ZOYHxJ+XQ+Ld39wzz53Ms0nmjGYrGwfu0Krrvyepre+xWNdbuM1Jh6VWbOig/x4Q/n8+rru+jt9zGnpJCrr9wypUeVpmk8/NizvLvvEJqmEYnGMMkyOdkucnJcbN64lvVrV07reLwjXpBJxtjpEI3GjP2dbG7jo5/+c2RZpiDfgzsnhw/feTMrl6X/vkwmE3/39S9y2y3X8frO3VjNZi7fsoH5leUTPmu32fjzL9zPkeP1tJ7qYGgoyNw5xRM+N7+ynBXLajh0pJbWUx2EwpGEmbLVgizLBIPD1DU2YbdZOdnUSk6OKyXyStd1mppO4c52GSl5w6EQ7753iO7efkqKCli3ZjlPPbuD3XsPGOvtO3CErZvXc+uN048umkqgjAXaZiz8bNqwhseffhFVTTXYtlrMbJjmd53kXHhSpWO6EWB1DU1pTdw1TWPfwaNcs23U63Dp4gUU5OfyzAuv8PsnX+D5l15n/bpVXLNtC5IkMegP8D8//LmRnhkKhXn8uTfxb17FpZXClO0YK5Z56x5BV2NEfA3ouooo29B1BUGQMOfMIxY4hRYL0rv/e1hc5SnRQJNxJumYZ7KOIAiTiotDbTuNZ0ny+pTMTmRrDjo6WnTI2A+AbMvFkl1u7OdCp20nmVR8UuIqDz/xovE6Fo+nvAa485aLL0wwQ4YMF45rt2/lkjUrqK0/iYNuCuQWZLOdaCSE27wLUJCJY5EBASIx8Ic0REkjx6YiWXMQJTPEQDTZ0OJhRNmJvWAVg03PIUoWJJsHNexnOJwQuDRdQBZHJSOdRKSSKAqoaiIdT9USqXOaloyQEkb8oWRkRwFKdBg95kfXEyJURDHhtEkgCAhqeCTKSkDTdDxZIo1dcXKcFvJcMkJS9BIl4sNdgGA8sIkGsGRXkFW6JWWmwV56LXvfeWlCqP7iRdVp0+4yZMjwh092Vhbdvf0UFeRx6Gg9SxZWGbPvk1V4ynB+GR/xIJmcyLZcdCUy5eDhXKTqXIwkypPXTqiglDDgrU0x7r0YOV1JeW9XHf/7s9cIhxOm55FIhDfe2oPXO0hXm4+leaPiwfFeD3uefYO/+srn+PIX7p92G156dachjkQiUQ4drUNVVcpKS5hbUsRvH3mKYHCYbZdvOu22SkuKEUXREMRyPTkMBRPV/5wOO5qmceRYAx2dXeTkZAPQ3dNL9fwKfv3Q48yfV2ZEXqdjyaJq5hQXUlt/goFBP5FING3ElCiKuJxOdu89QCgUZu/+wzzxzEtcs/0yrt2+1fjcxz58G+/uO8Q//Ov/4M7JTkl9CwSG6Ov3kZ/nIRQOY+1JBD7YxuxP03V8A37y83Pp6/fyPz/4RYrw8eSzLxMORyb4P73x1h4uXb+GwoK8Kc/noD/A3r27sfh34c6SKSgqx115lXGN9x76CQJCilgxHeFnw7pVbN64jj17DxCJxlhQDAMRG5u3XGYIetONXDoXnlSTMZ0IsNhIVFeRM8hA2GoYzwPEI4EUL6Wh4DD/88OfExy5Rv3xOC+98ia+gUHuvesW3t69L60v2Ju7j7B144dQT70waTvG/paj/lbiwS4EyYw5a24iKihpzK1EyK64Gn/T84bxtqf8ztPet87kHn8m60wmLvYe+gnRgROE+o8ZzyRBtoIaQ9cUJEsOseFE9GFyP2Mnvi8mJhWfltTM54233xt9vTD1tSAIGfEpQ4YME/C4c9i0IVHVJ6neh/rr6GjcjaqoKCooEnQPwokeyHGaiGDGma2jhr04Si8j1HcE2ZqLEhlAjQcZaHgMUbYi2/MRAIs9B1kOoqsC6Cro8YRApCc8nxQNzFIisqo/oDIU1ijMkbGZxYToBCAI2DwLyFv+CfqP/ILI4Al0JY4k28nKL8ZetI6h1ldQ42ZQ48RVGYFhJHQWl1uRBB1BkjE5irHn1hAeqEePh+l+7zsUrf2zKUNZL9u6lYaWHhoPHKazq4dwJEphfi6fuu/u8/hNZciQ4Xxy1eUb+MY//g+XbVzD759+mU/fNxrxcbT2BKVpogUynF/GRzwIkslIoT7d4GG2U3UuRk4n3lzsvlanSz1542CvITyNZffut9m6KFUgrsn3cbArn4OHj3P1tulHtr2zZ7/xd1d3L+qIgXdvr5e5JUUA7Hh9F5dtuoRBf4DDR+sQBIGVyxdPMKbOyXFx6fo1huVBUWE+g/4A0WicvFwPff0+Bv0BnGMEJl2H5tZ2PO4cDh+tM/pr6Xjj7T08+czLhrhlsVi4757bqFlYRTwep7bhJKqiUl1VwS8e/H1KpUJd13nh5TdYMH+eUdlOFEU2rFvFkpoFxONxWk510NnVY6TdJYUop8OBpmmoqjrBm0mSJLZuWs8zL7w6IeLmVHsn4XAkbeRZfWPTlOJTc0sbP/jJb/CYfSzMG6CjU6a33sOnq3OMa1wP+9B1bcbCjyRJ/N3Xv8hzL75G8/E3qSkYJL+wjJXb7wLSR1NOxvkUuqeTorqgupI52WGq3ANElESltqgqY5EUFuacInCqzTimd/bsN4Snsew7cIQbrr6Cru7e9MesR+hreJGsMUXh06XKpps88NTciRoZGE3PG6lSZ3GVJyq+Wadf8e1M7vEzXWcycVEQRESzEy0eIhzswupZkFhB17DlJ7zvkmMmWXJPq20XiknFp+//x9+cz3ZkyJDhfUjyAZpVshap+Coa9z2LjIokwiO7dbxBgcvWFBEeyKZ67QZyiosI+xJ5+rIlG8mSjRIZRB8p56wF2tDVOLoWQ5LMqFocJBt6PI7AiDE4ArI06jouiQI2i4h1xPtJUUEUdSRJIOA9SXj3t3HmzsOcNQdRso6k+rkJNL+EIIqYHPmYs8qwBjtRooOJzoGmoGsagmjCkl2BrsVwV9/E4Iln0WJBQj37yZl/3aQdCJPJxJVbN/LegcNkOR3k5XrwuLP5+a8f5Qt/8jHKS+ecj68nQ4YM55FPf+wOrFYLR+tOcMctV3PjNaMRASeb27np2ssvXOMyTJmShSCiawqCmOg2u8q3TUi/UWNBBpueR4kMIltzgPQDJDg/ni3ngovJN+RMmSr1pK+/bsJ7FklhWWUAswARxUZtn4eafB9WWWFlcR9oEyM1pmKsQDMcChONRhkeDqNqGpIsMbekkFwPPP38K+zctdcoTf/MC69y5wdvYMO6VSnbu+3maykqzOfdfYeIRmNs27qJ8rI59PZ5eXv3PhRVpfVUB6CjaTqiKBCPKwwPh1CUVEPxsXT39PHE0y8Z+weIRqP88reP85G7b+HXv3vCOJZQOEJweJiCvNwJ29l/6JghPiVZtWIJ7753kLklhfgDQwSDfUDC5sVsNlE1v5zjtYnqcNmuLELhCPF4nGxXFl/90qeYU1LI8boTE/ZlkmU6B/0pywYG/XR29RCJxqhvbOKqKzdTUTZ3wrqPPPEcsViM7ljiuk5E8fTzxtt7uHb71pSUuzMRfrKcDu667QOoN15urB9qfR75DCKXzofQPZ0UVcnsxOmwc+nW6zl14EHjN1Hb52H7ChMum5ByTN29fWn3pes6vf1eCvInXj8WSWHVnH6scj6SxTNpO5JtToplsjUhwIy9B6uxIBFvnZEO7Sm/c0YV36brF3c260wmLso2D55FdxJo3UHU34quREHXjGtPjQ3jrX0IaeS4Aby1vyO35u6UIhAXw3NmUvEpQ4YMGWYLNRZkXqFEqLSUxqZWNFXlsiUWagdKGYxnYZJlSqouIdz2ovGgc5ZsIHrgRyNG4YmZJkEUQZAQTXaUyCCiyY4WHwbRBFocEBF0DVUFUYZoHMJxEbcDwjENHZFIXEcLgTtLxCxHGfZ3MxxWqFx1O9GhU2jRIYa9taixIUTJhK1oDYIaw5Q1h9hwd0J0kixIsoSmqygRH9acShwFK7HmVBvC0+l4/qXXyXI6UkLe44rCS6/szERAZcjwPkSSRD7+4VvSvnf3B689v43JMIHJonqSxq2yLc8Qlby1v0tJvwGMzwFYcypRlVDKoCa5D+C8eLacKy6kb8i5Fu3mFBey78CRlGVuWwR3lkxEkTk0EtVxsCuflcV9WGWFmvKsGbVxQXUle/cforfXS09vP76BQSRZwma1Mjwcor6xmapKjdd37hnxs0ygaRqPPP4cixdVpxiUC4LA5o1r2bxx7YTj8Q0M0tPXT31DE8HhEKqqIooCdpsNSZZZunjBpOfiwOFjKcJTkmBwmG9/76c47DZjWTwe52RTKy6nc0JaXjJqaiw3XreNjs5uOjq7Wb5kIbWyjNc3wPzKcgrz85AkkcqKUppb2yguKsDjzkYURT50+02sXLYYAKvFQjye6kec68mhu7ffeN3vHaDhRDMmk4zJJHOstoH6xia+8JmPpghQ/sAQnV09xuuxlgi1dSe4dvvWlGv8bISf2YhcSidoeE++SMNgOae6BsjLdbPxktXkZLum1aZ0zCTKcdOmLVSUzeHUgd8iaCHWLMsm25U14ZiKCwvS7ksURQrz88jP87Brz34iY9LQ3bYIZcXZWB2eKdsxHbHsbCI3pyvGne06MLm4aHYWpl0OEGjdkTAcH9mPt/Z3RAdOGNkYZmfhRfOcmdzqP0OGDBlmgbE3u4U1ywk7N5KVU8jcwiy2VAWxSApXXbkFKdY35ga9nWDnbixZc5Ht+UiSGdDR1BiibEWLBREEES0+jCBIgACCGVXTCMUgENYZGNIJRzVMEvhDAg3dAu1+Bye6BX7wio2Tvhz6h+30Ddto7x7gyNu/4tjRQxyrP0Fde4xgKE5oOIjv5Msgmon5WxElC4IAksmGikQootB0spEfPH6Cx559E93knpbwBNA6UnVmusszZMjw/uMrf/sfF7oJGUaw5dXgKrs87YBAtuVhyS7DveBWJIsLQRCJDDYRH+5hoPFJhnsOEu47iq5riCYH8VAifUSQrajRABFvAwONTxI49TqqEkpJq4gFu1MGKLPp2fJ+IumFMtD4JGoskW6V7F8ETr1OuL/2rPexft1K3CPeSEm6g06sRZtYuPETKCSElagqc7C7gNz52yldcOmM2lg9v4LjdSfo7O5lOBQirijEYnHsttGcIn9gKEV4SqKqKsdqG6Z/PGtXEolEjRQ2SHhfRmMxqudXkOtxT7pussrdeAb9fsLh1BTELKcTWZbp9w5M+PyKNKbmWU4HX/3ip/j0xz/ErTdewz//3VfYunk9JUUFaLpGW3snff0+cj0eLt+8nttuvo5v/tWXUqohpquMKMsyH/3QrUZ6YltHF1arhUULqpBGUvcURWHHa2+nrGcyyROq34VCEbp6+uj3DaZEiE0WyZL8vqdDUlwYy3QFrPGChnvBrcSxcHD/XjoP/47ek2/x+muv8q//9UNOtXca68z095HufpgUa1xll08QLuaUVrLmyvuZVzaXbFdW2mPacMkqspwTj3HdmhXk5LjwuHP4wmc+Ss3CKiRJItuVxcqNN7Biy72nbUdSWNLUGK7y7ZidRYl1LC6UkJehtp0zPqaxjBeuxm4/KVzNxjow+TUWC/akXR721U/YT27N3Yk0vVgQb+1DF9VzJhP5lCFDhnPK+Jvv/Qsk3nrrTZSuV8gxqdy7YSPL1o36JZhdpcY6oiULm2chsWAHmr8VXQmjRHyIsh1dCSNbctB1DcmSTWigmbgq4A2KdPQrlOUnPKCCEY3Xj0Z57WCEBaVxmntiLF4yn3fb7eRYw+w62MHq8ghXLtNwWYcYGNYIhi3ErCI2k4KuxvA2v47N5kTX4tjylxIb7qOpM0iWCYaiJsyEeHv3ewz6A0bUkqIodHb3oms6eXluHPbUEtTunGx84yrVJJdnyJDhj4ODR+svdBMyjGH84CMWaENXIgnhacxMee+hn6BrcaL+VgTRNPKMy0GNDSFbstF1zTCEtecvI+yrNZ6DNs9CbJ6F04p8mO1In3B/LaoSwuZZmJKqEvE2pJTjvhg5H0bLDrudL33247yw4w3qGk5is1nZsG4Vl226BEEQmF+9iENHalEUhWVLFk1IEzpdG2XnHJ5/6UkWVM1j0D/E0NAwsiyjqiqxeBy73U5xYT6yLE3axvEiyVTMqyilIC+XrCwHZrOZSDRRXW7Z4kX4A0MT/JTGsnzpIl5+deeE5ZIk4XSkmpSLokBVZUVKSiHA5o3rJq3eLQgCixdVs3hRNZCwI/jNw09x4PDxhGm42UxFeSmHj9VRVjrHEDQCQ0EaTzRTNreEFUtrOHyszihPv3rlUj58x02Iokhzazu/eOA/CevZxMaYYFskhagvNb3SbrOxdPECDh+tQweamk/RMxJBJUsS//hv3+VP7r+HAo/9jCJZxnMm6VtJ0kXv7G31EAmqeGxhPLYwvnCIg10JA/bPffy2M452mUmU43SOKcvp4Iuf/RjPv/QGtQ0nsdusrF+7ku1XjJrrzy0p4jOf+PBp2za+Hba8GqL+FiIDJwi07jD2a84uJxbsJupvMczPk2Jccv3pRG6eScrxmawzWbSUEvbR/d53MDuKke25KddexFuHo+SSlPu62VlI0do/w1v7EKJkvqiKYGTEpwwZMpxTkjdXLZ7olNhtNq6+6hrU2CbCvnokOVGR5b0DRzh8tA5JOsrqlUupKl5vlHPW1Rj2vBrCvgbQdFRlGEGU0bUYFnc10cEmInGJmCrxXlcRK/JbEdCJKPDUnij7WyQKS8oYFiSQvDQ1n8KVtZiOXoUeX4i9cbhsiQC6jssO/cEoZlkmropIok44qmK1qlhyKtFiQwxGHYSjQWLxxEzlQDjx77HaBrp7+mg51c5vfvckR2sbGB4O4XZnc9N127j37luNDtTWLet5/KkXJ5yvrZsuOR9fS4YMGS4G0gcXZLhImGzwkFtzN97ah9A1hai/BVPWHNToICZnCSZnEa7y7QRadyR8XfoSaVzjO/2nS92Z7ZLq4f5afA2Po4S92PKXkrsoYXg/Nq3QM8NtnilnIqqdL6PlnBwXd99+Y/r3sl1s3bx+0nWnauOgaTXf+tcf8trO3QA4nQ7y8z1EIlEAsl1ZLKlJCDEedw6D/sCElDVTmlS5aDSGb2AQjztnQpU3AFe2k7WrljEUHMZkkrFarAwM+mk91YE/MGRMeL3z7n52v3uAUDjCogWVbL9iM1dvu4wnn32Jvj4f/sAQiqqybtVyOrq6yct1I0mjIpk7x8XX//xP6Ov3EYlEqVlUNaMiCiuXL8brG6C3rz9RPS/LmbTu5KVXd7Jpwxp27z3AMy+8akRxZbuyuO+e25AkmeKi/JRIrmJnkPXzFfzDfSkm2CuL+yjMtaWIDwB33HoDwWCId/cdoqe3H0GA4qJC8vM8+AND/ObhJ/n8hzefteH+maZiJUl3Tzra0EbQn0+hc5jirNCo91KHirf+cfR48JxGu8zkmPLzcvnohz94TtqRVboFJeIz9mtxlY96JuUvxewqPat76JmkHM90nclSA3sP/QQtFkR36GmvPUm2I5mdKfdWs7OQ3Jq78NU+bJirXwwm5BnxKUOGDOeFYNcewr7alIdQ0vhv54uv8tZhr/HZQ0dque7SMlaWRtDRMTtLkG0essqupPfA9xEEGUEQsbirEiawkgUEgf6QjUUFQWLkIOvD9PgVgjGJNSuX4nQ6UFWNgUE/0ViMPq/PmKVbWSEQVGwogRhWExS6IBYHyQyhmIwoJszFowON2Dw1BL0hDnQVpJiOJjs3Bw8f5+kXXmH/wWNGB2lgwM/Tz79CJBrjK3/6SQC2blqPEld4bedugsFhsl1ZXHXlFtaunhhKniFDhvcnH737pgvdhAynITlImNipv5uuvd9CjQ2hDzYDoIT7yVtyb1pvDrOr3Pg7GSWgq/FJKy7NdqSP2VWKyVmMEvYS7jtKv5ookR7xJdK4TM6i85KKcTai2h9CRcF0bZTyN/J/332Y4PBopa/xVb/GRjvddMNVRCIRHn3ieSPlS1VVPnDdNuy2hNeSrus888Kr7Ny1l1gshtlsZuvm9Vx/9eUIwmjKXtmcEvr7feRku/D6BjlyrAFN07BZLfzrf/2Qe+++habmNl57c5exTl+/l+N1J7hq22ZAwDfop7/fh8Vq4cjxenJyXByvP8GSRdVG5NQVWzZSs7CKmoXTO0+DgwH2HTxCLB5n8aJqykvn0N7RjTtnok9RLBbj3f2HePLZl1OW+wNDPPXcK/zNX3wh5Zghcb0Xzakg2tRgmGDX5PuwmhRK5s6bcK0nonLu4z//58eIooDDYcdiHhXz2ju6CFKAq+zyszLcn42KkePft9mseH0yp/zZ9AQdhh/Zmrn9aPEg8jmOdrlYqmCO3+9wz4GU95P33Ysh9WwyJpvwKFhxP0NtO8kq3TLptTf+3gow2PQ80UAr+sj9froRdueSjPiUIUOGc85UnWh/SGfvsR7G345eeKeN5UuvoqB0KbFAG5LVkwgfle2Ish1rTiWIsjED7QtZMYle4qqEjsS7XRUEh/zILqtRZliSRBZVV1J/oplwOILZbGLRXJFl853IokjPoESpW0WSIMuqoWo6CBIRsQiIo6txwgMNyLk3E4geSDEdddsi9Axn0dLWQW+v1xCekvgG/JxoaqG5pY15FYkH3rbLN3HFZRsZDoVx2G0zCqfPkCHDHyb/8d2f8bU//TgAHxsjPv3n//6Cr37hYxeqWRmmIH2n/jl0JYKmRBDNWaCryPZ8Aq07cJVvT6l8p0QGGGh4nMhAIzmV1xNo3YES8hIb7kK256OpMRgXJTDbkT6S2UnuojvwAuG+o4Y5uiCZjUio8zEgORtR7WzSlWaCqqqEI1EcdtsEUeO066ZpY9P+RxHUEGaTCXdONgNjKrLNKS4kMBRk7pxiykrnsO3yS1mxNDEAXVKzgD3vHeClV98iNBzmmedfYdfufdx+y3X09nl55fVR76JYLMbLr+7E6bCnRGddtmkdh4/VEQ5HaDzZbHg5lc4tIRqN8uOfPwQw4Tj7+r384IFf43Q6iEQi2EcMxgcG/eTneaiaV878ynIqK8pYtmRh2gpyY2luaeOZF1+luaWNSCSKd2CQosJ8BODFHW+y5dJ1ZGdnTb5+86m0y72+AZpb2iZU1JPMTlZc8VngB3R3tLBK7sVms1JavojqDR+f9HrJdjkNz6jxaJqOraBmQuSeZHZidpVOiKZKx7moGLl+7UraO7qAhB9ZbZ+HVcW9FOTlIpBeoJ3NlN7zXQXzdG1Pir+CKGPJrkCQzOhKJOUeas1dNGG7yW2MP5bx25/Ntqb7zGSpgek8ZcdGUY29t3rrHkFXY8bkgi1/KYJknnGK6LkgIz5lyJDhnDNVJ7q5P5eo2j9hHV3XafGaKZ7vNErbipIZq2cBOVU3Mtz5Dmo0QKB1BzlVN6LrTxHTzIT7veiazoL8YYTF24m8dxxIVFzxDSSMMivK5nLDtVdSWVHKU08/jSR34LYOUVwqMBgUkCQQ0DFJEFEkSsqqKV50Ld3vfRddi1NmbaQ4P4uuviEOduXjtkXoDjpZv24lfn+AWDyW9jwoisKgP5CyTBTFlIp3GTJkeH/zwqtvG+LTWF56bdcfvfh0qr2TuvqE186qFUtwOuynX+k8MF4wMbvKCfUeSlRFNdkT/k6iiXiwGwHoP/Zr4sGE2a+rdAvhgUYj4kgJ+4BElFQyqndsmt7YQclsR/pIZic5ldehhH3E/C0AmLLmkFN53TkbiKQbqLvKt+OtfSghuE1TVDvbdKXpoOs6L+x4g5279hIKhfG4c7juqsvTmlvPpI1KtIOVxV4OduVTVVlGw4kW/IEhIJE69qd/8jGu3HrphO1lOR0cPlpPLBozIqN8A4P89FePYDKlH8K99c57bN28nlPtnTzx9Es0tZwiFovT2+/DJMvYbDZKiguMSmiBoSCRSJTCgryU7QSDw/T1+wBhgvn4wKCfvFw3+Xm53HhdqnF2Orq6e/n+A78iriioqsqBw8dQVY14PE7Z3BIAdu7ay4duvwlZllMMvgFqFlZhs9nSbRpIVApOh2R2snTTR5jT8DiariMKQqJowBTXyfKlizh6vJ6Orh66e/pQ4gpZLidrVi6jsCBvVtJhZ7ti5OaNa+nr9/L27n3IRKkp8OHx5FBeNgeYKNDOdkrvmRzTmYpfp2u7Fg8T9o2aqwti4neiq3EEyQQkolCHO98l4q2bsI2o/xSgY8kuP+tzM53zDLNX/XTsWCvqP0U82JkyuQCc92i0dGTEpwwZ3oe0dXQRiUQpL52D2Wy60M0BJu9E2wYnzmYVOYMMhEcjlmKBNpSwD13XKFhxf2KWyVHAQOOTKCEvgyeeRpItLFi0lMUFl+I7+RJWWcFsixIKV7LnQD3H6hqNCi1Wq4WW1nZuun47hfl57Hzhx9jMA4gCZGU5iYu5xPRezJKK22Umb+5KrO75lGz8K8O87xN3XsIb+3s5XteIYDFzy+UruWzTJbz25ju89Y6TLvpSjslkkrHb7EZnIEOGDH9c7HxnHwCaqrNz934YU8a8vauXLOfFIbRcKH7+60d5+bW3kGWZnGwXTz//Cvd/9M5JDYvPJ+MnUEI9B9DVOJIlG2tuDQKJ9DVdjREPe41KrLb8pdgLV2IvXGlEHCXTH5LC09g0lfEDgtmO9FFjQQabnjeEMYD4UAeDTc+fUeTT6QaQkw2+Aq070DUFJR5CtuYApxfVziS1Z2z7kqbUUw1wX3p1Jy/ueNN47RsY5DcPP4HdbjP8mKZisjZme31YO3tw2yJEVSdLaqoJhSMocYV//uZXKC4qTLu9U+2dnEpTAVdRFFrbOhBFkf5+H5qu43FnU1KUiKIa9Af4/gO/ZjgUQlFUzGYToihgt9uoWViVsi2L2Tyheh0kjMUtFgvmNCJXMkrblTW96+X1t/YYApE/EERVE15WXd19zCkuQpIS2+vs7uH+j97J40+/ZHg/rVy2mNtvvY7mlnbeeXf/hG3b7TYqK9JHyw14uzmx++fEo36ynA7yPBPTW8ezbvVyHnrkaU61jf5GhodD+AND9PZ5yc0+98b3M0UQBD5407VcuXkVPccexSp7cLjyJxVoz4d5/1Scjfg1VdsF2Wp4xSbF38Gm540oT4urHEEyERloNCqRjt+GyVkEkDLREBloNLaZ9I2ajngz3fM8m9/F+LFWcsIheb2fq2i0mZARnzJkeB/h9Q3w0189QkdnNwA2m41bPnAV69euvLANY/JO9KqlV/Psixai0YTpZpEzyMK8AZAdLJyXeAiYXaXouoaAYNw0k526obadRP0txkMsFmhj7tJbjFnk267ZQmNTC5UFKg2dJjzubErnFNPb188LL7/B7bdcR9Htd9N9+NfoyhB2dxWCZMI1708ZaHgMLR5GtnkS7XAWUrDifqMNt5cBpIbBXrp+De/uO0RXdy9DYzwdKsrmsnXzJZOGc2fIkOH9zbd/8GsAYvEY3/7+r4zloijgzsnmy5/96IVq2gXnhz95kJ/9+hH0EUHOYjFTs7CK3/zuSf7fN750TlOSpzsDP7ZTr8aCCIKILbeGvCX3AKPG3Vo0gGjOwuqpThF0chfdQchdbRiQQ6rgMj5KYLYjfdRY0GgjJNIwlMgA8eFuwn1H8Y60EZjW4GQ6A8jJBl9KyGtEfiU5nSgw09SecH8t/tbXaOkc5KXDEgNDMRZVlbB9SRynRU/ZJiSio+sO7MAiJVKXklgkhYN7nmVJzZ9NeT4GBwPsO+pDCGczb8Fa8sa0sWr9fRxq+TndLQPG5+02K1dcc+mkwhPA0NDwpO/5/UMpkdTDwyEGBwPcddsHeOfd/dTWNdLT50XTNKwWC/l5Hjq7E6lnRYX5WC2WxDmwWVmyeAEnm1pTtu/xuJlTUsSgP0CW05HSnynIy0WWZTasWznlOUnS2zsa3T42vS9Z5c8mJdoiSdKIb1QVAwN+LFaz4XG1eFEV69asYO++Q8b6kiRx5wc/gMk0cZK1uamBvS/8DyYhSkSRqe0TuaRykFVLKqb8/URHosyWLl5AIBDEZJLJy/UgSSI7d73LbTdfd16M788Ei+YjxyEiWfKnFGjPl3n/ZJyN+DVV2625ixjufDclPXos9qLVxAKthlBlCFDjjh8YiR5qJdix24geGvvedKKSUiORWvHVPowgmVL2Ewu0zep3MXaslRT2x95bzybCbrbIiE8ZMryPGCs8AYTDYR569GnmlBQxt6TogrVrqk50tP0lPvPRW/jt4zvo6/cyELZisuWwpLqEQNMzxmdFyTzhoZTMgU4OHmKBNgKnXkcRrJwYKCIW0phr1Vhe0MWawhzq+910B0dv5IeO1HL7LddhdpVizSpAoNAIyx3ufIe8JR9BjfgmDECSr1ta29nx+tt0dfdSkJ/Ltssvpaqygi9//n7WrFzGi6+8iW9gkPmV5Vxz5WWsXb3sPJ3xDBkyXGw89stvA/DX//Q//PPffPECt+biobb+BM+88IohPEFiANhwohm7zUpzSxvzK8un2MKZM5MZ+HSdekFKGBInZ5e1WBA1FkS25qRNZYsGUgf4Uwkus23iGwu0EQ8mfGFs+UuxFywn2L4LRTInIraC3US8DYR9tROOPZ1AJ1k9KV5V6QaQ6QaKuhonNtw1JuVw+qLaTFJ7zK5SGlp68fZ2UO2SqY16cIXf49gRnVWr100Y4Aa6j1Dm6KLAIk+ojpaj+6b08zl0tJZf/fbx0XSx109w6fo13PnBGwCQLVncfs/nKN9/mCPH6pFliTWrlrF8yUTfmbFUlM/BJMsT0srC4QgOu42h4GgUEUAoHKaifC4vvPwGXT2j0df+oSFOtXdisZhpbmmns6uXstIS5pYUceN129i0YS1PPvsye/cfJh6PM6+8lJtvuAq73cYDP3+IhdWVnGhqITA0THlpCRXlc7nt5mvJy/VM2f4kxUUFNLcmvHSyXU5MJpl4XEGWJSxjovNXr1xq/O12Z6dsQxAE7rnzZtavXcHxuhPYrFbWrFo66YTe6y8/Qc6I8JT8Pnc2SOTnDVFRIk36+xkY9BNXFFxZzgmRXT19icI4F6vx/UwE2gt5DGcrfk3WdrOzCEm2G8cf7q9FVyLG/c5ZtAY1ttLYrz1/2aSTAa7ybfhqHzbuj7oaOyPD8kSkWTnBjt0okhmLqxxXeSJVNdkOV9nls/JdnI/U5NkgIz5lyPA+oa2jK0V4SqLrOnv3Hbqg4tPpOtHFhRrf+Orn6O7pQ5IkcrMtM3oojX3g+kM6tcd3E4qKieomzbuIhQdRbDkMhK0p64miYNysk+LW2Jt1oHVH2tkTgKbmU3z/x782Opte3wD1jU188mN3sXhRNVdv28LV27bM5mnMkCHD+4CM8JTKvoNH05o6h0JhhkPhlHLus810Z+BP16lPeDa9giDKaWebz2RQMNsmvra8GjyAqoSweRIlySLeOgA0NUZWyXpDeBp77FOlzgmCOCpATfKsHj9QVONBZHv+hJTD2fYiCcUEnj8osLxAxiorrCruBWA4KnOwu5DStannOytvfiJijaHU6miygtWZP+lAMx6P89Cjz0zwKdq1Zx8rly9mQdW8xHmQJDasW8WGdauMz+i6TmAoiN1mTRu947DbuXrbZTz74qupbc1yYrGYKcjPpaunj3A4jM1mo7gon0BgmJ4xkUa6ruMPDBni7srlixkKDqNrOh/90AcNwefOD97AbTdfi6KoWCyjVd6+8dXP0dR8ilAkQq4nB0mU8AeGqK0/yan2TtatXk6uxz3ld3H5lg3sO3iUaDSKKIosrJpHXWMTJcWFiKKIJEnceN22afVTqyorqKqsmPIzQ8FhDjVFKHK6GQhbjUi2qCqzpyWH5Rsun/Qa87hzMJvNxGITvTuLC/OB82d8fyZMV6C90MdwNuLXVG0fe5xT3UOTQnu6bST/FiQTFlc50UArEV8DvrpHEER5RlFJaiyYSPMbEbGigVYGGp80TNAliwvJ6iHQuiNtW2byXVwsVQdPR0Z8ypDhfUK6nP0kw6HweWzJRKbbiS4uKjDWOZOHkmhy8Ng7EcqtYkpncyAq8N5JE66c1Fve6pXLzvhm/eIrOyd0NjVN44Udb7J40em9ITJkyJAhA6iKSq4nJ2XAnMTlyjqnPnnTnYGf6jmhhLyGF+BkwtKZPmdm25h4/HopPlYjEQDjj30qgU62eXCUbCTQ/KKxzfHP6vEDRdnqRlNjuMq3z4qoNhm+gUHCMcGo/JWkts9DaVZwwucls5PyNfdS+9YDKf2HuG5h4cbJq6OdaGolHE7fxzp8tM4Qn8Zz8PBxnnnhVfq9PnRdZ+6cYq6/5goWL6xKEWOvunIzc+cU8e57hwhHIixeVE1Bfi4//MlvsNmsE/yOrFYzHnc2NquFcCRKLBZHHzEMt9usuHNcuHMSZuPtnd0p0UaSJE0QewVBMCIPdV3nwUeeSkl9e+mVnXzk7ltZtWJJ2uMEKMjP5Yt/ch8v7HiDppZTlBQXct+9d+ByZRGNRlm0YD7Zrskr3c0UkywjimJKpLtxPCb7lNeY1Wph6+b1vPzqznHLrVy26ZKzii6ZzQpzZ8PFECFzpuLXTNs+qeH5GKF97Da8dY8ApPhG9R35ObGhdqL+FizZFUbk0umqGybbmoy+SlafC/cdNVL5xhabONvv4nxXHTxTMuJThgzvE8pL52C1WolEJopQi8cZTF4IZtKJPtOHUmdXL939QUKW1M5mR6SUqBJN+ez8ynKu3X6ZMcM305t1W0dn2uXJcrcZMmTIkOH0LKmp5sDhY8wpKaSjs8dY7nTY+fwn751xqfuZMp0Z+Kk69eN9B5PLew/9BD3sS61eZ/VM8F25kIOC6Rz7VAJdcuA0lulEfDEmsngyz6uzJT/Pg8MCNfm+lOU1+T5yirLTrrNy5Wrs5vvpPPwQ0WgMh8NO5dqPUjI3vYAETBmZl6xQN56TTa384sHfJ6r6nuqgs6sHOMDOXXtZv3Yln/nEh1JS2pI+SEmSYtX4/obT6eDyLevZu+8QSxcvoKOrh86uXkwmGavVYlSWG93OpE1PS239iRThCRKTbg8//hxLahZMWeBmTkkh93/0zpnt8AyxWi0sW7KQQ0dqJ7x3yZoVp13/+qsvJ8vp4O3d+wgMBamqLOe6q7aS63ET7q89IyH5XFSYO1MudITM2Yhfs9H2qbYxvtpdxNtAPNiNFh822j7Y9DyQEKhg8u9t/H7UWBBf3SNE/S3oagybuxo14ktU6VNjuMu3Y3YWpkxsDLXtJGf+dWm3n47ZnrA4F2TEpwwZ3idYLGZu+cBV/O73z6R4Zyysns+KZRfPTed0nM1DyWRKeDSM72wuLQqweMllrN+wke6efkqKCqicV2a8fyY367xcD6dCE6vQnC78PEOGDBkyjLJ65VKOHK8HoDA/j0F/AKvVwle/+KmU+/S5YrqTHZM9J8b6Do6NlBIQ0HXNSNcyu0onDDQv9KBguseeTqRylmw47Yz9hRzkWiSVG9aI9HQqI4bTiTS6LBssK+hCjQUn9CXUWBCPXo+pdA66rmO325CHD6DGyibtd1RVlpPtysIfGJrw3tioorHs3LUXXdfx+gZHhKcEvX39dPf08uuHnuDPPv+JSY9NEAQ+/fEP8egTz3H0eAORaJSSwkLuufsWnA4H119zBb995CkqyuZSOqeE9w4cQRQFSopTzc1XLk+ce13X6ezqxWIxTenjdPR4Q9rl4XCYk82thkCm6zr7Dx5l38GjKKrKssUL2XjJKmQ5MezUNI0jx+tp7+gmL9fNquVLEEWBg4eP03Kqg2xXFpesXXFW0VB33HoDQ0PDNLUkKiqLosiWS9dNS3yKeOvYtG4xl226xFimxoIpkS4znbC80BXmkiTvVa6yy41jSEZfnS8x/GzuC7MR3XO6bSTfAxjuO4QgSkjmbEzOImSrO6Vow1Tf29j9AEZqtiW7AjUWJBpoxV64Ekt2BZGBEymCvKt8O97ah4j6W04bYfWHRkZ8ypDhfcSGdauYU1LEu+8dJByJUrNgPiuXLz6nnhmzzdk8lHKzLVy2MEwslNrZtMoKi+b0M29u3mm9AqbLFVs28IsHfz9x+WUbZ2X7GTJkyPDHgCiK3HfP7TScaKa+sQmnw86aVctmNQ1nMmYr/WT8M8nsKkW2504YaEb9rZicxSkDlvOZdjNWJEseuxL2oesauTV3TXrs6USqgRPPICAg23NP+6yeSYW6maQlHT3ewO69BxgeDrGgeh6XbboEh90OJPoSlaW5WK0WXq+zopvi6PnLWD4/iInohG2qsSCn9j9IQ/1RfIE4tX0e1pSHWVJdAkx+LYiiyMfvvYMf//J3BEcqwomiyAeuvXJCpFGSZJW6Pm/qRJmm6cQVhZZT7Xh9A1NOZrmynNxz5y385ndPcPDIcfp9Pv7vZ7/l2u2XsXXzenKyXezctRd/YIjKeWU0Nbem9AWvunIL5aVzOF7XyKNPPI9vYBCAeeWlfORDt6Y18jaZZPyBIQYG/AiiQH6uB7s9UY3OPMaz6tEnnmfnrncZ9A+haRrHahs4XtfIpz/+IcKRCD/48W9oax+NHn/mhVexmM30jzkfO15/m8/efw8V5XMntGM614nTYeeLn72PU+2dDPoDlM0pIWck3XAqzjRC6XRC8tmabM8GgyefJ+JrTPnNxoI9Rtqwq2xyL6zZ5GwFpNmI7pnONsYalkMi0kkfk0VhL1h+2u/NllczLc9AZSQCauzzIpnKLTlKaGo+hSiKlJfNOefRwOeajPiUIcP7jNI5xZTOKb7QzThjzuahFAu0saR6DgePKxxsyyKqyhzqKeSmdTKeLHlWO/irViwhFo/z0is78foGyMl2se3yTVy6fvWsbD9DhgwZ/lgQBIGF1ZUsrK48r/s9F5E5yYHx2O34ah8mNtyFrsYwOUefz+cz7Wb8oDoWaEMJ+4gFOzE7ilEjA2mPfarUuel4N013oHi6QX80FmPX0QEOH61DkkRkSaK5tR1RTAzEmlvbOHi4li9//hNYrRZj+/lLS1l9zenFrPBAC8ePHSQQ0o3qaLtOSsRirVy61jrltVBRPpdvfv2LHKs7MS0Po/KyObScakfTtJTlZrMJizlhBRCPK+lWTeGxp1/k8LE6RFFMHEM4zONPv4jHk8OyxQtTfk9DwWEOHj5OLBZjSc0CTCaZ5196nYcefRqXy2kIU82tbTzw84f4yy//yYT9dff0cay20Xjd0dlDZUUpixdVG1GKvX1eXnj5DeoaTxrHIIoCvgE/Wzevp+FEc4rwBIl0vmBwmEUL5hvLotEojz75PF/94qdSPhvqO85Ay6uYbTl4FtxyWnGobG7JpCJgOs5lhNKFrDAX7q8lMnCC2HAiVXOg8UkcJRvpO/QAWiyIxV113qKvYOYC0oXwyxo7HlFjQQYaHjdMyO1Fq3EWrZly/cBQEEEQkCNTP2fUiG9SYbJPWMb/z959BzZ934n/f2pLtizviRfgjQGzdyCBDLIISchq0qZJ0z3Sa3t3be/u22uv9737XttLf3cdaZs0bdo0CUnT7AEkhBH2NrYxGLCN95Zla1jj94ewsCx5ywtej3+CPx/po/dHw/Hnpdf4/376LOZOb2ZlXGwMn37o7hG9p6caCT4JIaac0X6rYYjLJxG4beEMcisbsVi6yZqVgSl84LG6Y7FscRHLFhdhtzvQajXT/tsIIYS4loS6QWv/AErvuG67uRK304bWlIbHaZuUsptgF9UejxtteDLqsFjfc9D/3AcL0HH5wgnjlXKu0ZYSDnbRr9RG8Ke/7eNcpbeXo8vl5tDRk0RHmfwCLA2NTew/dIx1a5YDI/tb4lyDklO1xoDpaIero8mYk8qMIc5Jo9FQNMwWB+tWL+fo8dPERkfR3m72bU9PTUGhUBAfF0tiQtygxzh/sYq33t2BWq0mKtJE3z8/Ptl/hLkFuX63jzCGs2blEgDe3baTD3bspqq6luqaOtRqFXnZszBdDpjV1Tdy4WI1M/s0Mz9/oYozZ88zMyOVyuoa3JebmNfWN/KD73/T9/fP2YoLfoEn8GZ0VZyv5Ojx05yvrAo4l7a2DixdXbjdHl8wEbw9NNs7zERFejOWDh05yY4Pt5OhP4vRAImVl5h/3WforPoQe0cVGmPSqLMKPR4PpWfOcf5iNVHGNGZHXBh0iuNoTOaEOa0pDbXBW1LpsNTidvfQdfQXeFwOlFojsfkPhHwNoQoYTWa/rN4AfN/XTaHS4DBX4nIUBX3O6huaePmvb/tKPrNmZ3LPhoXEpc4Z9P8z/QOTyrjlPPv/vURPn+FGzS2t/Pb3L/Iv//i1oBMypwMJPgkhriq9v8Tzc01Bt4+HvmOJhRBCTB+hbNDaN4DSUrbVN1rb43KgVOtR6aNRqLSTUnbTv6lua9lWX1lH73jx3t4iwx1XHsovdQYrS6p1FXCu8k3fbbu7rbjdblpa2+nq6iY8PMy370JltS/4NBKdlq6g09HsLjXtPbGjOKOBRUdH8s2vPsaOnZ/w8mtvYzZbSE6KJyrShFar5YF7bh/wyyy3282fXvwb+w4d5WzFRcDbYHtOXrbvbxFLV7d37XYH+w8d42zFRYzhYaxYthBnj5P3t+8C8F3UOp0uyisusqio0Pe4Zov/RMDiUm+/p+SkBOLiYujo6ESlUhEVGUH35ccDaG5pGzBrq66hAZUysA2EUqVEgYL+p6xQKHxNzE+XnuXPL/8NgDZVPEXJTTirKlB9/DRJMTqc1mY0xiTffUcSnHA6nfzmuRcpP3vety0xSsF9q8MwXn5vjTVDKZQT5kYT1On7+XK7e3C0XwBAodISP/8JtMbEoPcb7VqCDWIYbcBoMvtljfR16+np4Ve/+5NfD7hzFRf5XVsH3/v2Yr9j9/3/TLDA5MVjf0XpsdE/XNNpsXC67Nywg91TjQSfhBABHI4edu09SHHJGVQqFYsWzGXF0gWS2SOEEEIMoPcCr6VsK9amYjwuh2+ktkKl9fYMcTnwuJ0olN4/wSeq7KZ3fTpTBpaa/ThVWr+x4YNdFPbvg9J7sdn3wikUgaiBypIO7PWfWqbpM1XN0i/4NNpeYdmzMwfcl5M18LS7gbhcLmx2O2EGQ9C/nWKio9iy+VbuvWsjJWXnOHe+ElNEOIsWzMUUMfD7Ye/+Ixw9UYxWo8Gg12G12bHZ7FRcrPJNNs6alYnd7uD/+/Vz1NTW++574PBxkhITfD9HmiKob2gCvH/3mTstRJoiUKvVzMr0b/bfd5KdRq0mLja6z74rX8ClpiSh1+uw2fwnDKvVKtJTZxBpiuCdDz7y25cQF4tWE5g9Pic/hzCDt6fUzj37fdvtLm9PzwXJjdTVN5KSlIfeEDPqrMK9+4/4BZ50KieZxibOX9Qyb04eMPYMpVCV+I4lC0ilNWJMWU7D0V/6tmmMKXTV7kMbnjDicxtsLb295AhBwGgy+2WN9HU7frI06PCBltY2ikvLmV8YfMJ30ABXTx1FyU2+MuC+gk02ny4k+CSE8ON2u3n62ReouFDp21ZxoZKLVZd4aMudk7gyIYQQYmpTaY0YorPprj+K4nKAp/eipTcopTbEodZHARNXdgOXJ3a1nfUGwlwO7B0XaTv7ui8wNtRF4XiXvwxUlhQX5R/80eu0REdF0tbe4Zd5rFarWbF0dH0XkxLjWb1iCXv2HfLbvrCoMGjT64F4PB7e2/4xu/Yewmq1Ehcbw603rRtw8p1CoWBOfjZz8rOHdfyjx4t998tIT+XM2fN4PB7a2804nS4S4mO5/rrl7Dt41C/w1Lu2w8dOkpyYgFKpICY6ym9SX++k5JtuWEOEMdzvvouKCvlgx+6APlURRiN5OVdKH/Pzspg3J5ey8vOYO73ZU+FhBmbNTGfxgrmkpSZTdamW4ssTLsH7HKckJXDo6Enf8dPTZnDf3bf5btPa2u77d9+pxj09PXhQoVB53wejCU6cPF3md+yi5Cb0aifN7aBKuhHaDowqQ6mvUGUQjiULyGFpoPHE73xBcU3EDJzdTSiVmlGdm9aUhtvlwGNt9VvLlSEGD/gmYvZ9TdTJG3jvo0NUX6olNiaa1SsWk5yUMOhjTVa/rJG+br3DBILpW2Lb10ABrvb259FfqiPaYPPLylQoFKMKiE8VEnwSQvg5XXbWL/DU6+Dh49xw3QqSEuMnYVVCCCHE1Nc7QltjTEGlNaJQqjFX7sCUsQEAtSEOXWT6mMpuRruutrOv+6Y3eVwObK3lWJuKfdlZQ61hPMtfBitvydCfIz5KT1P7lW/7s2dn0trWTnRUJB6Ph7i4GDbffvOQF7GDufeujWTNzuT4idO4PW4WZoVRuHBtwDoHCxa8t/1jX1kbeHu0PP/ia4SHh4WkoX5vgAggJjqSeYW51Dc043A42HD9KjZcvxpjeBjl5y4EvX94mIFuqxVjeBgKBeTnZtHU3EJXt5W1q5azfGkR+ZczqPqKj4vlwXvvZOvf3sHhcABgiojgsUe2oFZfuZwMMxh4+IHNvPTqW9hsdjweD3q9jjUrl/h6SH3qtnk0rl1ETYOZ+LgYsmZl4O7pYv2SFGo7w4mKjCAz3T/gl56aQktrm19wyOZU0+TOYo0hyhv8VCjxuHpQqLxZWsMNTvSdAhhtsPmOfbwuntsikgiPG9sQgl6hKPEdbRaQy2GhpfRF3A4LSq2RhPmfw1K7H6VSg8PibQA/0nNzmKtRoPDdv638NTxuZ58hBq0BASOil/KzX77ga6AN3oy8z3/2QV9AJVgpn3cq30soVVeCzSMJ3I+l/9RIXreZmQMHqgcKYg8U4Jq55BFKatXUX6zzu/1N69cEnUY5XUyJ4JPb7ea/f/U8+w6dAOD+zbewZdNNQW9bXVPPD//raTrMnRjDDfzTt77ArMsv9OZPP4lWo/ZNivj0/XeyYd3I676FuJZdrLw04L7K6hoJPgkhhBBB9A2g9A8wtZS+iEKhRBeZHrLJeiPR/9t1l8Pia4bucTnQR2cPeRE3nuUvQ5W3fHbLdbzxcQVnzp5HoVCweEEBd2+6BZVKic1mJyY6KiStAYrm5lM0N9+X5dV+7o1hZ3m53W52f3Io4Jgej4edu/eHJPg0rzCPi1VX/k4LDwtj9sx0Zs/M4K7br1w7GftlLvWKjopk4fw5HLmcQaVUKkidkcxjj2wJGnTqa8miecydk0t5xQU0ajU5WTP9Aje9li9ZwKzMdI6eKMbh6GFuQa4v8NT7vIbrTCxb4P+8euxmctPXYYgLvEjfcP1qTpedJVpruRIcqk/goQdvJTo7rU9WYSxqlbckcLjBiUVFhb6yu94MkzarnvSZ2b4SyFD3OBuL0WQBOczVKFVadNFZxOY/gNaYiCY80Ts4ANBHZ4343LSmNNRhsbjdPdg7LqKJmEFPZw0aYwrqsFhU+hjMldv97lOy/wXs3S76hiCcTidvvLOdb3/9iaDZlQ5LA/WHn/JN5YvNv39EgfuJbFieNSuT/NwsSs+c89teNLeAjLQZA95voADXHfd+jplzyjl1+gwqpZKFRYWDlghPB1Mi+PTejr1cqKrhpWd+gqWrm0e/8n0WzS/wBZX6+s+fP8tdt17PbTddx4e7D/JvP32aZ//nR779P/re18iZnTGRyxfiqhIdFTmqfUIIIcS1bKjJcLrITCLS1oxr4+6B9P12HbwX5r1jw109lkGnN/U1XuUvwylv+WLWcux2B0qlwm/SU3hYWNBjjsVosrysNhvd3dagx2tt6wjJutasXEJpeQVn+2Q2mSIi2LL5VjweDyVl56iprccUYcTjIaCJd17ObB558G6uX7uS0rJz6PU6Fsyf42usPRi73cGp02fotlrJzZ4VNPDUKyE+lls2rA3YPtrsuRkpiXzzK4+xfedeGptK0Rhn8PhnryMnayYuh7e8T22IRReZMeKswqWL5nOx8hL7Dh4FvAEob6bXHb7bjHYIwXgYzdS88Rgc4P1dsIH6w0/hcTl8Tcyd1maiczb7Su76ZjJ2dZymKNkV0MfoUk0dNps96PujpfQlX8ZWb+BsJIH7iW5Y/vin72Pv/iOcLC4DhTfwtGr5olEdS6FQMLcgN2B65XSmcDjsnqFvNr6+9c//xS3rV3PjuhUA/O/v/oJGreYLj27xu11rewdbPvst3n/ladQqFR6Ph9sf/Cq//uk/kzYjic2ffpL//D/fHHHwyel0sua2R9n99nN+qaNCXIusVhv/9l//S1ef6SUASYkJ/MM3vyBNx4UQ4holfy8NLVTjxcfLYOVtw8lg6p8xABM3sW+iDXSu4Rm3su9IKcUl5ag1apYsmMfihXMB+PFPfkFzc2vAsZYsms+n7tsUknV5PB7KyiuorK4hKtLEgnlz8Hg8/PqZP/tlRdntDsIMBlxuF+ANPD38wOZhBZr6u3Cxmt889yJms5lLtfVYurpZt2oZ//itL4145Huo30NDZbaY0tcN67PX0NjMhcpqIk0R5OXMHtbfuxP9eR/r53c81uLoavBlPPVYatFEzEChUKFAgTos1u81+eClH2GztHKmOdqvj5FOp+Pf/8+3UalUQd8fbpeD2PwHcNlafc937/OsNaUN+XxP5u+tqf7/hIk2Jf5yqG9sISkhzvdzcmIcxaUVAbdrbGolLiYK9eVIu0KhIDE+loamFtJmeMdr/vC/fo3H46EgdzZffux+oqNMAccBeOWNbbz65jbAv35aiGudwaDny597hK2vvc3FqksoFArycmZz/90Dj/4VQgghRGj6uoynYNlZ+tg8uhtP+mURBLs46nsB53Y5iMq6g67afX0yCjbgsrVOmXMdq2BZXhHp1/ObP73h1xuz/Ox5KqtruPeujdx64zqef/E1v2sLnU7H+rUrQ7YuhUJBfm4WmbE9ly9qtbzxznYuVl1Cp3L6GhTrdFry87K48YbVRBjDR90nxuPx8PyLr1F9qZbDx05htXp7b52rqOTYyRL+96f/SkpyYtD72u0OausaMBrDiI+LBUKfPReqrJ7EhDgS+1yPDmUiy7l6hWpq3lhZm0tx91hxWltxdjehM3kTP1S6SJzdTWiNKeijswIyPRML7uWDd7b6BZ4Ali8p8mXSBXt/xObfj8vWGvB8a01pw570F8r3XN+AUu+/gYDfn8CEv0emugkJPj3x5A+orqkPuu8Pv/hxyB7nVz/5J5IS4nA6nTz93Cv86CdP87N/+07Q2957543ce+eNwJVv8oQQXjNSEnnyK4/RaelCpVL6Rt0KIYSY3n73/Ku8+uZ24i+PSp+Zkcq//uOXJ3lVYqL0v1C3NpfScf4D1IZYjMnLfBdOLWVb6bHUE9PnPr0Xvm6XA4VCSVftPkwZGzBXbsfZ3UJL6Yu+hsBXwwVVsPKmc0e2cqmqhf6XUHv3H+b665azsKiQ8PAwdu7eT2tbB+lpKaxfuzLk/TL7Bz5Oni7za8YN3vKxkrKzPPbIFpRK5agfq7KqhsamZk6dPuMLPPUqKTvLb557kR989xsB99v9ySHefv8j31j47KyZfPrBuwnTeoZdNuZ2uykpO0eHuZOZGakDBrkmI+g70eVcMD7lcyPV+95TqPW4XXa04ckodRF4XA5c5g5Uukg8Hrdf4KnXwkVLaDU72PbRHux2OyqVisUL53HHxvW+2wxUVmjK2DDq53s0pYpDnb9KZ8IQk4+l7gAKtR4Aj9OGy9mNraUMl91MeMrSCX+PTHUTEnz67VM/GHR/UkIs9Y3NzC3wjhmta2gmKSE24HYJ8TE0t7bjdLl8ZXcNTS0kxsdePo43Wq1Wq7l/883c/3jwwJMQYnj6j9oVQggx/d10/Uq++aVHJnsZYpL0vUD1Zi8047Q2o1Bp0ESk0H7+XaxNxb79/e/X20jYZTdjrtyOMWU5befeQqnSTsoFVbCyFqe9k0sXjkPYTNJTU0YcfBmovMnSXkJRcmtAzxqPx8PFykvExkSTmz0rJM3FB9M/8GFUW5ndZwpcm9V7MaxUKgOy1jstXZSdqUCr01CQmzWssrm2djPd1uD9rE6dLqOltY3YmGjftnPnL/Lq6+/63e7suQu8vHUrm5ZqgpaN9e/P1NLaxq+e+bNfGePihfP41H2bpkQm/ng24K+6VMvJ4jJUKiUL5s3xC15OdnZl73vP3lGF09qMPibHu8PjxhBf6P2n0zZgMGzD9atYs3IJTS2tRJoi/K41BisrNFdu9wW7Rzrpb6BjjmbSaN/PXlfTCVAofb8v9TE5dDeexOO0XQ5O5WKIyR2X98h0NSXK7m5Ys4zX3/2IG9Ysw9LVzY6P9/NfP/x2wO1ioiLJzcrk/R17ue2m6/hozyES4mJIm5GE1WbD6XT53sDbdu6TxuNCiJBpbWtnz77DNDa1kJQYz+rli4kaoKxXCCGEmA70sTkY4guxNhVjbSrGaWujp7MGAEN8IfrYHL/b915M9r3o7rjwgS/wNNEXVMFKn6qrznN02//idnRypjkamyqFB7fcOaKA0EDlTerqGvTqRl9pW1+RkRHDPn5ldQ2fHDhKR4eZWZnprFqxCLPZwvade6mqriU6OpJ1q5dRkJcd9P79Ax+rsjq5dOnyFLg+gbF5hXl+gZpdew/yxjvbcTq92VHh4WE89vAWZs8Kfs3U0NjM/kPHqKmtx2534AGUfY6n0+mC3u+TA0eDbm+uKaW7cwZhEbFDlo29+OpbAf2zDh89yazMdFYuWxj0+P11mDs5dOQkHWbv8zx/bv6YssD6G20512B9gD48Us+2D3f7tn+wYzeb77iZ61YtDdm6x8JvkALgcTnA4/Z9VoAhs7B0Oi2pKUkB24cqK3TZWkc16S+UpYr97+9xO337PE6733PRu67xGNIwXU2J4NMt61dTUn6e+x77FgqFggfu3kjWTO+3Jrv3HWH3/qN875tPAPAPX3+Mf/vpb/jDi28QHmbg+9/6PACtbWa++6Of43a78Xg8zEhK4F++88VJOychxNWjpraB/3n6D77U8eKSM3xy4Ahf/+KjIU+lF0KI8fbh7oMcPVFKpMnIZx+6i0VFBQPeVnpkXt1UWiOxeVtodjmwNhX7JlYZ4guJzdsy4AXSeE29G6n+GUDhqWs59N7PUbi6fRlAdpeZZ/74Mv/8D18bdkb3QOVN2cs/y66T/x/1Fv9soZTkRGbPHN6X3sdPlfLHF17F7XYDUFZewYe79uFwOHzbmppbKD97nofvv4vFC+cFPU7f1yA1ORFLZxcflel9gaeU5EQ233Gz7/a1dQ289ub7fp/jrq5unvvzq/yf7349YIhAZXUNv/jN8zgcDtJTU6i4UIXd4UCn1aJUKjEY9Oh0WuYW5PplPQEDTv2r6wyHqEVEz5o/aNlYp6XLb6JfX0dPFA8r+FRWfp5//+kvMJsthIUZiIuJInv2TL74+EMjbpI+kN5yrubWdiqrLmG3O1AdP0dy4X0sWbos6H0G6xVlbmvg1MF64MrnyOPx8Prb2yiaV4ApYmyfr1A1vx7q8z/aLKyhygp7ezz1NZJJfx6lno927eP4yRLUih4W5MayfF7wAO9g+p6/QqlGEzEDPB4UKu/7qu9zEcqSv6vBlAg+qVRKvvPVR4PuW7NiEWtWXBlPmJGWErSMb0ZyAn/8Zej6RwkhRK+33tvhCzz16u628u62nXz24S0D3EuI6aGuvpHqmjpio6MG/PZbTB9D9dncfNt6Hn1wE2q1mhOny/nuvz7FM//zQ5ITgzfalR6ZIpipckHVPwvhwuHnfYGnvhlADoeDo8eLWbs6eEAgGENcPk6nE6fT6QvMmKLiufvBL7L1tXe4VFOHQqEgN3sWD9xzx7BKwTweD2+8vc0XZOp16nQZOp2W9NQUv+3vbvuYRQvmBj1239dAqVRSkJdFWgbUuecQl5ASMLHtyPHioAHkTouF8nMXArKs3n7/IxwOBwAxMVGsXrGEA4eP4Xa7iYmOQq1Wk5+bxROPPhhwzNzsWZSVBw6PijRFkJqzIiD7qH/ZWP/npy+Pe+ggeHu7mW/+449oaW3zbbtUo8PpdLF3/xHWrVk+5DGG0hswOlN6in1Hz7DtuIMN87UkRHdi2f97PLhZunRFwP0G6xXV3OHwlUz6PZbLRUnZWZYvWTDq9YayQfp4fv4HKivs21x8pOVzvcf8ze//QknZWd/2iqpGTle9zBcee2hEa+x7/h6305ctqjRloFBpfM8FMKo1X80T8qZE8EkIIaaycxUXg24vPxd8uxDTgdvt5k8v/o2jJ4p929LTZvDEow9Iv7dpbKg+m33Nn5NDTlYGZeXnBww+ielvsAsZrSmNlrKt2FrLUai0vnHpttZyWsq2Bs1+CnUPlbHqm4XgdLoAKG2K8evJBNA1QDZOsOenqeESH37wGvtPt6NQKJg3J4/Nd95MpCmCjLQZfPvrT9DebkalVo3o9+WlmjrKyitwu91ER0ei02rp6XFSU9dAj6MHi6WbxIQ4YmOiAG/fo65uK8bwML/jDPQaRGAmSldN9MyFAQErl8s14Lqc/fZdqq3nvW0f09XVTViYgZTkRBITYtl441oam1pZvWIRhQW5XLdqadAsopXLFnHkeDGXaup825RKJZvvuNkv8DTQe1PruERmeioXqy4FHHteYd6A59HrhVfe8As8Adhsdqov1VJcWh6S4JPDXE1DbSW7DpzmTzvasdg81DQoeeh6EzMSFRw98GHQ4NNgvaK6wnKxu/YEfTyNemyX7e12A42tVgzqTjyevxGZuWFUza8n6/M/1vK58xeq/AJPvUrPnKPifOWwv3zre/4KtR5cDt8+hVqHQqX1PReGmPwRr3kypihOJAk+CSHEEMLDw2jvMAfZLlMAxfS1a+9Bv8ATQFV1Da+98T6ffujuSVqVGG+NTS0kXB7UUl1TT3lFJbNnXnsTd64VQ13IqI3Jvma5hvhComZt9DUctzYVY02YhzFpkd8xx3vce1NzC5dq6omNjQ7IBAqmbxZCVGQEKCA/PrApeG7WzGE9P7auNva88RO0zi4Sw6Optxg5fqqEhqZm/v7JL+B0OulxOkfc9/F06Vme+eNLVFyowuPxoKisJm1GMk3NrXR3WUEB7R1m2jvMpKelkJqSRFiYAYM+sK/SaF6DuQW57Ny9P+BYOp2OnNlXnptz5y/y62deoL29A5vdQaeli+aWNgoLcggPM7CwqJBvfPmxQc9Vp9PytS98hoNHTnDu/EUijEZWLF3IjJQr0+qGem/evX4ev9nahsXS5btPbvZsVi1fFOwh/ZSfPR+wze32UFldy959R9Co1SxfuoD5haN/nxri8nnpo1/x2kdmLDZvNlan1c2fP+xgbhYkZPR4X+cgWWsDla0tmBHGmx98EpD5pdPpmJOf0/8ww+J2u3lh6xscPnrSNxExxnSJfEsrOq1mxL3axvvzP5CxTvqrrK4ZdN9wg099z7932l3fRuthCfN80+6UGgOm9HUjWvNkTFGcSBJ8EkKIIaxYtpB3P9gZsH3VsqH/ABJiqjp87FTQ7SeKS/3KTMTV5dfPbeXM2YuoVEqUSiXf/uqjpKcmT/ayxDgZ6kJGH5GOzRCLxpjsy3KKzdtCC9BjqUOlDgs4ZqjGvV+susSxE6dxudzMK8wja1YGf3nlTQ4fPekrD5s9K4PHP30fYYbgX/b0z8JIyllPUqOF+pqLFCU3+QJQixbMDXpxGez5OXvgL+Ds8psaB3Cppp6f/s/vaGhoosfpJD1tBptvv4mZmUNfDNrtDv74l7/idruJi42mqbkVjwdKy8+j02oIC9MDV4IUl2rqSU5M4LpVS1GpVAHHG81rMHtWBmtWLmH3J4d825RKJfdtvhV9nwDX2+99hNPpJDExnsoq7wW7y+XiUk0dudmzWL1i8ZDnC94A1JqVS1izcknQ/UO9N2fMms8/fWcxx06epr3DzMyMNHKzZw1a3njhYjVvvfchB4+cxNxpQalUEh7mfe+0tbfjdntQqZSUnjlH6Zlz3HrT9dy0fs2wzqe/bquVPSeasTmVwJXMMYvNw9Hybh4qih5wrYOVrX3qvk289Ne3fSWPBoOBzzx0t99rNBJ79x/h8NGTANhdakqbYligbuTCxSrycmaPuFdbqD7/ozGWSX/R0VED74uKHPEaes9fqTH4AkK952+IyR30uRhszeM5RXEqkL8shRBiCDdev5rOzi72HzrmuyhftXxRSNK2hZgsPT3OoNvdbjfuYfTUENOTDGO5tgznQqanu4GwxIV+F5JRs26lu+HogBdIYx33vu3DPbz9/oe+n/fsO0RMdBStbe2+bW63m7PnLvC3t7bx0JY7gx4nWBbG0lu+wbkDf6C5oRp9fAoz56xl/gClWsGeH3t3W0DPKIDycxdobG4h+fKgkarqGn71zJ/5x7/7IjGDXNgClJZXYLfbAZiVmY7H7aGlrZ2enh6USgVz8nPQqFVUVdditdlRKZUsXjCXm9dfN+AxR/Ma3LNpI4sWzOV06Vm0Gg2LFhT6rd3j8XChshqAlKQEnE4n9Q1NuFxuurtt3LzhugGDSSM1nPemClixdHiT7eobmvjlb5+nx+kkPi4Gu91Oe4cZj8eDSqnE6XSROiOJuNgrzdG3fbSH1SsXDxjcHIzd5iDCGE54WBgOR4f/uamU3LHxhqD3G6psrWjOJgrynuTM2fMolAryc7LQakffIP3o8SsZzjqVk/x47wTB1vYOXC73NdP8ujA/m9iY6IByzJjoKObOyR3Rsfp+xoL9eyS/C4OZKgMdxoMEn4QQYghKpZJ779rIzRuuo6W1jfi4GMLDAr8NFmI6mTsnl4bGpoDtOVmzxvSHrhBXg6up4etgFzLW5lLsHRdx2lr9Sp/Mldtx2c1Ym0tDfr7t7Wbe3bYzYPuHu/aRNTMdjUbDhcpq2tq9F/TnL1SxccNaoqMDsxMGysLIWvYZ0of5WvV/fsLDwwJ6RlltdtraO0jtUzYG3kbm+w4e47abrx/0MTx9yqhUKiU52TNx9Dg5f6EKg15HUoK351psTDROpwu1WsUdt64fVhPz/rqtVqov1WEMD/crc+uVmZ5KZnpq0PsqFAoiTRF0mDtRKBRkpM0gNSUJu8PB7JmZbLxxnd/ta2ob6LRYSE9LCRrAGepzFMqL7F17D9Lj9H6pkp6aQrfVikqlwulyEWbQk5gQG1C61tPTQ21dA1mzMkf8eFFRJuYV5tPS1g54sHRbvV9QqlTcc5c3yBfMcMvWiuYNPIV0JHr7efWW3OnVTmxObwbUsuXhI+7VNF17EqnVar78xMNsfe0dzlwuy8zJmsWWzRunXKb3VBnoMB6m1jMthBBTWIQxXBoxi6vG+rUrOXP2PNWXan3bIoxG7r7z5kHuJcTVb7peXA1ksAuZyegvUna2Iug0M7fbTUtbBy2tbdhsdt/2xuZWfvm7P/Hdb30pYEoajD0Lq//zEx0VyfKsVj4uU/kCUHa7g+goExFBRt33zdYaSF7ObLRara+UCkCrUZOelhJwTmq1isKC3CGzqYL58ONPeHfbx/T09AAwMyONRx++l0hThN/tenp6qG9sJtIUganfOa1escQvK02lUhFmMHD9dVeyvc2dFn7//FZflpRGream9ddx4w2rfbcZzudIa0oL2UV2Y3NLnzUrmZOXjaWrG6vVxpKF84I2LwcCnpvhUigUPHTfnXSYzVRfqqOzqwudVkthfi7f/daXBrzfRJetzS3IpfpSLdEGmy/wdLwuntT02SQW3DPiXk3TuSdRbEw0X3z8U1itNjx4RpXxNt6m2kCHUJPgkxBCCHENMhj0PPnlz3Lq9BmqLtUSGxPNoqLCUfeVEOJq0ffiqvHEM0Rn3Y6ldr/vYkCljxmXjKDxMJwLmYnuL6LTaoNuj4mOxNpt9Qs8AURFmmhqbuF02VnmFoysPGYoAz0/RQUZxER3sOecERdabli7kp279/sFzTweD41NLRw/WUJ7h5n5hfmsXLYwaBaFwaDngXtu588vv+43de7uO29mZkYar735AfUNjajVaormFXDvpo0jPpey8greeGe737YLldX8+eXX+fLnHvZt+3jvAd7btgur1YpSqaRobgH333M7Op33ddlw/SrsDge7PzmE3W4nLMzAhnWrWLpovu8YL7z8ui/wBNDjdPL2+x8yIyWRgrxsYOgghUofE9KL7BnJiQHTiY3hYcTGRLPl7lt56n+f9WVG9crLmU18XOywH6O//Nwsvvftr7Dv4FFfX6pli4swGPSD3m+sAdORWLt6GaVnzvlerzarHo0hkns23TKqoNfV0JNoqNdnMk1WQ/eJIsEnIYQQ4hqlUqkomlcQsvR+Ia4GvX/sN554BnvbORqO/hJdZCZqQwymjA2+kjSY+hlQw72Qmcj+IoUFOYSFGejutvptz0ibQYfZQmv7lf45Go3aVyLW2tqO0+nkZHEZzS1tzEhJoiAva1Slab0Ge35mzlAyf8U6v9f4w48/8f274kIV5k4LcbExVJyvpOJ8JSVlZ/nCYw8FXdPCokJmzUznxCnvUIfCglwSL5fb/ePffZH2DjN6nW7UXwAcOHQ86Pbys+dpbzcTFWWiuKSc195437fP7XZz9EQxao3a11dLoVBw+y03cNMNa+i0WIg0RfgF1No7zL6ypYA1HD7uCz4NFaQI9UX2dauWcvDISaxW//fVDWtXMCM5icc/cz9/feN9GpuaUSqVzC/MZ8vdtw77+ANJSoxn8x1TN2NYp9PytS9+huLSciqraoiOimRRUaEvADOaoNdoyiWvplLm8TSZDd0nggSfhBBCCCGE6EOlNRKddTsNR3+Jx+XA3nGRiPR1vsDTVC8v6TWcC5mJ7i+i0Wh44jMP8NyfX6HD3AlAWJiBB7fciQL495/8kq6ubvR6HXFxMWguBz4ijOH8x89+TXNLq+9YmempfPHxT406YDOSC707b91AYkIcBw4dp6GxGbVazdyCXFSqK2VzZeUVlJ+7QG72rKCPFxVpYu3qZQPuGwurzTbovihMfHLgSND9R48Xc/cdN/s9j1qthtiY6IDbOhw9vmmEgY/jn7U2WJAi1BfZsTHRfP2Ln+H97buouFCFyWRk1fLFrFzmbVielzOb7337y7S1daDTa6dkydV4USqVzJuTx7w5wRvvj9RIf2dcbaXM420iM+MmmgSfhBBCCCGE6MPlsGCp3Y8uMhN7x0U8LgfNp55DZ8pAHRY7bcpLYPALmcnqLzIzM43/891vcO58JS6Xi6xZGWg0GjweDyuWLqTiQqXf7XOyZ3H8VIlf4AngYtUltu/cy+23BJ8sNhwjudBbtriIZYuL2Lv/CF3d3UGPd6GyesDg01h5PB7Kyiuoq28kPj6WOXnZvp5RudmzKCuvCLhPTHQUSZcn9Fm6gq/Z6XRis9mHFcSLj4sJOjUMID9ntt/PQwUpQn2RnZyUwKMP3zvobYI1rp9uJjOLaDS/M6ZznygRWhJ8EkIIIYQQ4rK+F1dqQwwR6etoPvWcNwPKXElk1u3TJvA0lMnsL6JUKsnJmum3TaFQ8PnPPsiHuz7h+MlSFAoFRfPyWbd6Gd/7158EPc6JU6VjCj6NRm+T6iSjhTar3m8yXpRRPS49waxWG08/+4Jf4+ykxHhys2dx5HgxHR1mausbMUVEEGm6nEmkUrH5zpt9ZYBZszKoqq4JOHZcXAyRkcNrvK1QKLhn0y08+/xWnH16KKWnzWDlskW+n0cb2Ozp6aGquha9Xh90Wt+1brKziEbzO+Nq6BMlQkOCT0IIIYQQQlzW9+Kqt8eTzpSB3VyJx+Wg/dybaOc/PuwLpqnc62QkpU8TdR46nZaNN65j443rfNvcbveAvZ2CTcAbbwV5WeTNUJKoafNND7O71MSYdGQaKjBXnQRCGwR4f8eugIlt+w8dZ/cnh8ienYlKpWJGciKt7R1kzcokLTWFFUsXkBB/paH2ujXLOXGq1C9rSalUctdtN42od1ZBXjZ//+QX2P7RHqw2O3Pys1lUVIhGo/HdZjRBiqPHi3nl9Xd9/cBSZyTz6KfuIS42ZkTP1VQ3ls/SZGcRjbZccjR9osTVR4JPQgghhBBCXNZ78aTSx/h6PKnDYonMup32c2+iVGmHXZI22VkKwzGc0qfJPo/eBtFHTxQH7Fswf+IHJiiVSu65/1EOf/A/YGmlKLmJDnU+Ny/Qo3B1BQ0CjDV4d/xkid/PLpeb+oYmALJnX1lXXEw0sbHRbLptQ8AxTBFG/u5rj7N33xEuVFYTFWli1YrFpKYkjej8S8+c49U33qO5uRWFQoHH46GwINcv+DTSIEVdfSPPPP8ydfVNWK1WDAYDdkcPzz6/lb9/8gsjWt9UNtbP0lTIIhpNueRE95aDwM/cgcPH+fjjnbi7a1EYZ3Hz+utk4MoEk+CTECHS3m6m8pJ3ikR6aspkL0cIIYQQo2SIy8faXBqQuaGd//iIStImO0shVKbCedx1x03UNzZRW9fg25aXM5v1a1eO+2MHE5+Yyk33/zP1xVtxOTrRaTWAK2gQIJTBu95Sv3abC7fbjUqlRKdyEm2wUW/xPmZzc+uA9w8PC+Om9WtGfd5NzS0888eXfSV3Ho+H4pIzPPdnO1/9/Kf9bjuSIMUHH+7m2InTOJ2uy1vaqW9owuVyUVldQ0bajFGveSQOHTnJvoNH6eq2kpOVyfp1q8bcDL6vUHyWplsW0WT0luv9zCk0Rqrtuew6UMzpU8e5ocCBIc7FmebzPPfnRh5X38/cgtyQPrYYmASfhAiBv731Abv2HsTtdgMwMyONxz59HxHG8ElemRBCCCFGIxTTuKZClkIoTIXzMEUY+c43Pk9ZeQVNza2kzUhmZubkBu9UWiPxORuHDAKEIuBQNK+AsmPbyY3zlvodrYlFrVYxI8FEUXITerU3GFRvMY5rr6R9B4/59Xrqda7iIvUNTb7m5iO1/+CxPoEnL6fTSfWlWl8Z3nj76N0XeG93ma+HV0NjE2fKTvPFh9cTM2NBSB4jFJ+lycgiGovJ6C2nNaXhRM/Jwwdot3zC3l2tLMty0GVWYVGHcfpCJ1qDgg93fiLBpwk08UXSQlxlDh05yc7d+32BJ/BOOtn62tuTuCohhBBCjJUhLj/gYm6k07h6sxT66hugsDaX4nJY/Pa7HBaszaWjXHXo9a6l73l4XD1oTRkTerGrUCjIz83iulVLJz3wBAMHAfq/nr0X274AVPlrARfjQ7llw1qMsbOwOdXo1U4WzmhhzaJZbFqmQa92YnOqabPqMRgMrF29bMznNtD70tUZOFGvV4e5c9SPZ7Pbg27vtHQxM2P8X+u2muOYqz6mKLkJncobXNOpnMwMq+DCsb+F9PM41O+EwfTPIorO2ewX2Oz/mk0Fhrh8TOnr/N7rvZ8JU/q6cSnZVWmN7Kkw0m5xoVc72VDoIFznprapk6ffrOXM+TpOnT7Dzj378Xg8IX98EZwEn4QYo4NHTwTdXlxSjtVqm+DVCCGEEGIqGSxA0Vsa0veisffi0ly1c0oEoHrX2FK2lfbz7wLewJPdXElb+WtY6o9M8gonx0iDAGMJOADo9Tq+8qUvsPDGr5CemU3RnAy+ef8s5uenow2L4rxlJnMKi/jGlx4lNiZ6WMccKMDUXvHugO/LWZGNJBkDAxwatZq01ORhPW4wGWkzgpa35edmodfrRn3c4WqzGuh2KNGrnRQlN2HS2X0ZZe0WV0hLS4cbtAymfxaR1pjkF9h0mKtDts5QCkUgf6SOFV+gtMnbrF6tVuFw9PDuERst7VcCnRZLN4ePnhq3NQh/UnYnxBjZ7Y6g291uNz1OJ4YJXo8QQgghQmcsjaKH6nViytgw6b2UhqI1paFQ67E2eZt962NyUKivBAO6G09iiMkddhClq7ubvfuOcP5iFZGRJlYuWzhh/XxCaaSlRKEolVIoFOTmFeJIjfOVayXExZK78nPcZhxZ0/DB+lA5u1vw4IEg78uklEz0kWawtPgd76b11xFmGP1fvSuXLaK1rYOW1jba2jtQqVQkxMVw3+bbRn3MkYiJS+Z4fQJFSY3o1U4WJDcCYHOqUYUvC1mG31j7H4WiHHiiTdbET4PGTU6kt/+ZwaDHbOniloV6Xv3EmxygUEDqjCSOnyphyaJ547YOcYUEn4QYo4K8LKqqawK2p6WmYIqYenXXQgghhBiesTaKHipA4bK1TnovpaGotEbC4+f7gk8elwOFUo0hvtD7s9M27ItIS1c3T/3yWb+G2IeOnODTD91D0dypd9E8mJEEAULZcHk0QaxgF/8qfQxulyNogEkdFospY4Nv2mP/9+UXczRs+2gP2z7cg6Onh2WLi1i1YtHwn7wgNly/ivYOM/sPHSM+LgaVSsXyJQtYvy54Q/n2incJS1yI1nilv5XD0kB3w1GiZm8c8eNHRZmYM2c+pWVHfIEngLOt8XzpwdA1tQ9F/6PRTJubLL2/Q92qMMy6RURExpMYGz5ukzJLys6yY+cntDbXMTO8AqXbgc0ZxomGZGLcZqIiXDy4LoL9lUZi45KJNEWE7LHF0CT4JMQYrVu9nOKSci7V1Pm26XQ67rnzlklclRBCCCHGaqyNogcLUHRW7/Zt751c5XE7cTks3hKuKRB46hWevBC3s5uuhmMolN7Lh6hZG1FpjSPKXvh4z4GASWxut5s33t7G/MI8FApFyNc+lLFkZQw3CBCqhsujCWINFEA1V25HoVBeCUAFCXwONFGtpamFA4ePo1Qq0Ou0nDhVQmXVJb7+pUeJiY4a8jyCUSqV3Hf3bdxy41qam1uJi4sZ8Evc9op3aa94B3P1LpIWP4nWmIjD0kD94adwXy5dG00Aasud13NYf4rWJiVutxujMYwH5yURYwpd2V+oM5cmK6touLSmNKobzNRUHafbsZ/SphiWZ/UwNzcNgzEmpNmdxSXlPPPHl/B4PCQZLYTrPNQ02DjdHIPBGMHus2HcPN9D3owIlJExvumQC+YVhGwNYnASfBJijPR6Hd/40qMcP1nKxapLREWaWLp4vkTShRBCiGkuFJOpgl0AOszV2Dsu4rS1Xs4w2YHH7cTecRGPy0FL6UskzH98ygSgXA4LdnOlL/AEV7JtRnKBe/5Cle/fSUYLbVY9dpea1rZ2mltaiTHpJvSieayZbcPVN+Cg1IRz5ux5ys5UYNDFMW929rAfYzRBrMECqGpDDOEpKzBfeN93+94A02AZVm+8sx2LpctvX3uHmXe3fcyn7ts0mqfoyuNHGIesHAhLXIi5ehduh4X6w08RW/AQLSUv4HZYUGqNhCUuxO12o1QOv72xy2Ghq/IdcjITUGTPwpB8HfaGPaPKThtKqDKXJur9OxZnLzawdXcXRclKXzmjvQtKzqnYsOWzIf0dt+3D3b7m4b2BpbaeFJJSorl5w1oUijvYvm0b3S3Nvv1LFs1n0YK5IVuDGJwEn4QIAY1Gw5JF86ReWAghhLjKDJYBMlq9AQFndwv1h59CHRaPs7sJjTEFp7UZhUIZ8gve0QplyZjRGAZ4A0+5cW3YnGqO18XT49Gi17gn/KJ5rJltI2GIy8fj8fCnl/7GkWNXGhy/96GSRx4IZ8H8OcM6hm/dw8yaGSyA2lta15e5codfyV2w1/x8xZWJdz09TlwuFzqdltIz54Kuu73djN3hICE+NiTZbVpjIkmLn/RlOjUd/w0ASq2RM/blPP2z5+nq6iYzPZXbN95A1qzMIY8ZLLAXZoodcXZar8amFmpq64mLiyFtxugbsQ9mIt+/o3Xg0HHsLjWlTTF+5Yx7zyhZ1u0mShu6x6qta/D7uTfAhN3BymULAVi8oJBTp89g7rSQNTuT1JSR9UoTYyPBJyGEEEIIIQYQikbR/fUGBBpPPIPbYaHH5UBnykAdFkvcnId9F/5ToXRmrCVjfcuCVi5bxIlTpXQ5NKiUbt9kMVX8CuyXPvC7aL5wsZoz584TZjCwsKgQY3hYyM9tqMy2c5WNfPDhq9TVN5IQF8v661cxtyA34Ly6urtpbesg1qRFYW8c8Pk4XXrWL/AE3rLDrX97h8KCHDQazZBrHk3WTLAAqjFl+YABppbSF1EolKgNMUFf80STk/JaqLhQSWtbB+CtBFhYNAe73cHp0nJsdgcpSfG8/cFOzp67AEBcXAxb7rqV3OxZQ57nULTGRGILHvIFngBKO7J495MrU6gvVl3i6Wde4O++9jmSkxIGPd5YyuH6vhfcbjd/fvl1ik8dJ1pvpd5iJDtrJo89vAWDQT+WUw4QiszM8Wa12dCpnOTH+5fb5se30m1pISoqcMLhaCUkxFFTWx+4PT7W92+NRsPCosKQPaYYGQk+CSGEEEIIEUQos376U2mNxObfT9PJZ1FpjChUGkwZ69EaE6fU5KqxXpT3LQvSabUkx0dgspdjsynQG7SkxkYwe2YnLrsSlc5EVNad/OW17Rw+etJ3nLfe+5DPfeZ+crJmhvz8BspsO1fZyNPPvoDb7Qa8gYxn/vASjz68hdxkMFftRKmNYM9ZI3sPl6Dy2FiY2sKs1BgKVz4U9Hk5XVoedA3d3VbOX6wOSVAmmGAB1LZzb6FAgTosNiDAhN2MLjKTiLQ1QV/zzPwa3tr1BzrMnb7j2Wx2zlVU8vf/8p8o8JY+nTxdRkxUFKkzvNklzc2t/O65F/nut7886t5QvRyWBlpKXvD97PZ4MHTsJEKbTqfjSo+mHqeTXXsPcv89tw95zNEE9vq/x3d+coLik8coSm5Cr3YCcPbcBV5/exsP3HvHSE5xWEaTmTmRfaIKslOJ7N6DXu3E5vRmQOXHtxIZrkTTvhdXQnzIgmQb1q3iDy+8GrB9/bpVITm+GLvhF8IKIYQQQghxDemf9aM1Jnkv1C+XujjM1aM+dm9AQK2PRqHyZryYK3fgclh8F7zW5lJclxso972ftbl0TOc1Uoa4/IALxOH0qOlbFnTiw1/x7DO/Il1XTmKMHqtLR6tyHtlZM1GpvJckpoz1nD57yS/wBOBwOHjh5dd9gaBQGiiz7aOdHwZ9vA927PKd18UL5dguvU+Yqoui5Ca0CgfnKps5eqY56GNptQNnNmnU45MT0D+AGp2zGZXOhFKlxYMHU8aGgACTKX0dUbM3DviaLyyag7rfeuPjYmhrN1NVdQmATksXFks3VZdq6ezTH6rH6eTgkROMRd/m4kqtkfiiz+NWGtAqnWzIqiJCa/e7fVNL6wBHGrv+pW8lJw/4Ak82p5o2qzfb6cixU75+RH31/4xbm0txWBr8PuODfeYHev/2/73R9/it5a/RUrbVdxuXw0JL2VZay18L+e+W+dnRxEXpfCW2ZruOU03JZM7Oxe3oHNPv0P4WzJ/DIw9sJikxAYVCQVJiAp+ZhpM0r2aS+SSEEEIIIUQQY51MNVCGgbX1DLaWskEzqhzm6infTHgovvLC0leprjzGgmTvxbfNqeFcWwwFmjLa2jOIiYoEvBfNp4tVQY/V3mGmqrqWzIzUkK1vsMy2OFcJOlU0dpf/5VJdfaPvvHbu+Z6vibL3vLwX2NWHSli1ak3A4y0sKuTjPQcCtsfGRDMzc/j9eerqG9m7/wjtHWZmZaaxfOkCwgyGoLcdrGwSuxmXrRWMib7bDyeo2N1tJTd7JlZbCjabnTCDHpvdTlNzK46eHgB6Lv8XoLmljQhjuO9nszl4YGS4uhuO+gJPvdPukhd/g7o3v4dO1UN6lJnTjfG+249nX5/+pW9ZEeex2Zy+90Lv+8fpcuHxePx6XvXPmnKYq+m4sA1HVx1aYwrg/d0z0Gd+NJmZLmc3TmsLTmsLLXinVraffxdrU7FvfyhFJs9j+Y2PU1LZietiIxFGIyuWLiA2cnyGCyxaMFcaiE9hEnwSQgghhBDXtJ6eHt55fycHjhzH4eihIC+LOzauJz4udtSTqQabRGXvqAI86CIzBuyjNB2aCQ+HSmuk2ZON273Ht62i1URBQit6tZPmdjuzl272ndsMdQM6lTsg6AOgVIW2aGOwwEy0UU20wXalafFlCfFx3rVowjlVa6IoyerbV9oUg92lprPTfxJcr4y0GWy+42befHcHTqe3JCsq0sRnH9ky7Ebcp0vP8uzzL+NyuQAoLjnDvoPH+MaXPxu0L9ZoA6iDlWalpWZ6M7X0YNB7S9ysNhuAL8hkNIajVChwezwBGT9ZszKGda4DiZq9EfBOvdNeDpyFR6fSk3wfZSff9ws8hYUZuG7V0jE93lD6lr5FRZmor2/yvRd65eXMDpi+1/8zbkxZjqOrDrfDgsNSCyqdX3Cp/2e+9/3b092EaebNvszMtrOv4+hqoKV0KwnzP+t3H0NMLob4QqxNxVibinFaW+mx1Hr3xRdiiMkN+fMTkTSXZUmwbJn/9qkePBehJ8EnIYQQQghxTfvjX17j1Oky388ni8u4WFnDP37riwNmlAxFa0rD7XLg6W7xCx45u1tQasIwpa5BH5szaEBgqjcTHg6Xw4K+67jftnlJzXg80N2jxRK21O+iOSnWTLShJSDoEx8XG/KpYYMFZpKs8TQUHwD8Ayc3Xu/tH+Pu6WJ5Vg+2Pkk8+fGtHK+LJ3t2pt99+gZy1q5exqIFczl7pgQ9beQWrQ8ISgzE4/Hw2lvv+wJPvZqaW/h4zwFuu/n6Qc+zr8ECqIMFTl12M6b0ddy0/jrefv9D331MERFER5mIj/M2d9ZqNMxISaS6pp7YmCjf7WZlpjM/BGVQvQGovm648XbCo1Pp3H8Es7mTWTPTuWXD2iH7S1XX1LH7k0O0tLSROiOJtauXjagnVXdnC1XHtqLGTmpKMh0dnb73gt2lJsJo5K7bbwq4X/+gc8eFD9AaU3BYatGGJ2O+8L73dgN85g1x+XTW7Kenq4GmE7/1ZYGFp6zAcvjnODwXaK9I8nuuVFojsXlbaHY5sDYV4+i46D1WfCGxeVumze8VMT1J8EkIIYQQQlyz6hua/AJPvcydnRw8fIJ1a5aP6rgOczUKhdKbwQC0lb+Gx9XjK6lRagxD9lEabjPhiWwgPBK9AYsIgwKlNoIjlQby470ZTyqlm7KWOL76oDcrpfdCPDyxmkJ7HQ079/oyZiKMRj790N3Dzg4aiYECM3MW38Sj+jQ+2LHbO+0uPo4N61ayeOE833llZ8RztNhCcX2U77wWp7exZt1i37GCBXIMGjdpujO47GbsranDfo3a282obVXoVHq/rBqdyklz5REgePBppIbKumu3ewOyc+fk0dXVjUKhIC9nNo8+fC9/efl1ei5ndaWlprBm1VJMEUbs9h4K8rNYuXQhKlXw0spQWLa4iGWLi4Z9+7LyCn773Iu+gF7FhUoOHT3JN7/ymC+QNpgDBz6h8sgLaBR2bE41Dc5Z3LmikC5zE2mpbtyxa1m8ZOmAQez+n3GFUk1swUNYLl3JFBysgXjU7FuxtZXjdlioP/wUsQUPeRuxe9wotUbCEhcO+7kQYrxJ8EkIIYQQQlyzGptbBtzX0Bi8cfRwaE1pqA0xuN092M2VaIwp9Fhq0UTMQG2IGVbZ3EDNhPtmQQyVpQKTV97SWxak1plYtvFJzrz8LsfrVBQlNxFhgHtuXel3gd8bfLv9lnxWLF3AmbMXCAszUJifHdDgeiLML8xnfmHgc9d7XlGxyazdfB+aw6W0NteREXOJ1EQTkdorZXchLZ/svkBeQjtWh8qXVaNTOSlKbiLZ1I21uTQkr3X/jJy+WXfH6xJ4649/9N1WqVTy4L13smTRPMBbUnfs5GnsNgcF+dnj2m8pFN58d0dAJll3t5VtH+3loS13DnrfmtoGdm57nZxYe58eT1Ze2avhszdnkuDoxJSegGGQ7Mn+n3GP20lLyQtow5P9BhEMlO2oNSaStPhJXwP2puO/AfDrh9X/8VrKtmJrLUeh0qKJmEFPZw221nJayrZK9tMw1Dc00dXVTXpaChrNwEMERCAJPgkhhBBCiGuGy+XixKlSzlZcxGgMZ/bMNBQKRdBJVMlJCaN+HG9GwwbqDz+Fx+Xwlbc4u5uIK/jUkBd4w20mPJV7Q/Uva/v7J7/Apdp6bJY2kiJdGBMLB7xvbEw0K5dFT9RSR6T/eW2+Ix0Inm02WCBnpOWTpvgsImOS8DTXUZTc5Btbr1c7SUzOCOlrHSzrzh5exFvb/EfZu91uXv7rWxTkZxEeFoYpwsjaVcv6H25Kstsd1NTWB913/mLVkPc/dPQEdZ3heDwe2qxXstFqG8106jeSnOQeNBjY/zNuTFlO44nfeXs+UUf8/Cfoqt03aANx8AagYgse8gWeAGILHgoIPAHYWsp9zcUN8YV+DcetTcXY4ucTnizZUsG0d5j5/fNbOXryNG6Xm+SkBO69ayPLlyyY7KVNG6Ht2ieEEEIIIcQU5XA4+NF//g8//9Vz7Nl3mG0f7uZ3f3jZ10S6r5joKJYumj/qx/JmNGxHbYhDodKijZqJQqVFbYjDXLl9wFHovrX2a4bd2xepN9DUO6K8N7jhC0CVvxbQRHsyGeLy/daQmpJEVk7+oIGn6aD/ecHAfZR6Azl9DVZKNRCV1siyW75BRHQieo13yl64zkN6ZjY5Kz4b0te6NyPH0t1NS1s7doeDmtNvoFM5A27b43RSUnpuwGNZm0sD3u8uhwVrc2nI1jsaGo16wKykyIiIIe9vs9kBqLcYAxrk212qIbPQ+n/GPS4H2vBklFqjd9qdyx70Mx9wHEuDt9Suj5aSF3BYGgJuq9QYUBvifD2etMYkYvO2YIgvRG2IQ6kZXY+7a8HPf/F7Xnn9Xc6Un+dsxUX27DvEU7/8PRerLk320qYNyXwSQgghhBBXvZraBv71P37O8ZMlAKhUKmZlphEfF4PNbufmDddx5Fgxdru3XGjjhrXoL0/yGg2HuRpndwtOazO6yEwUSjXKyEwcllqUSs2Q/ZhGMqVsuL2hrmW1dQ3UNzSRkT6D2JiJzagaTvnkcIVFxHLd7V+lvvglHI4ewsIMxOXdE/LAU0PJK5SePkFjq82bYZVQhcmgYH6ymxN1CQHBFtUAkwinclmoUqlk1fJFbP9oT8C+1SsWB7mHv/y8LPYfOhawXafTMSszfcj79/+M9/6s0sfgsrX6DR4Y6PeFw9LgK7lTao2+nk+9PaD6l94Z4vKJIfD3SmzelknvETeV1dY18MFHe+jpuRJ8dbs9nL9Qxfvbd/GFxx6axNVNHxJ8EkIIIYQQVzW3280zf3yJyqoa3zaXy8W585UYjeGAmblz8th447qQPabWlIYHD1pjCmpDjK8cDsDjcQ+rRGq4U8pCGdy42thsdv7pRz9l/6FjOBw9GPQ6brlxHf/4d18clwbm/Q23fHIkxzNX7kCn1aLTaoHQv9YOczUV5SU0ttp8vaWO18YzJ74ONTaiDTa/aYQ6nY6CvOygx5rKZaEAt960DofDwb4DR+lxOgkLM3Dz+utYMH/OkPedNyePeYV5nCy+MrBAoVBw9503o9Nph/X4/T/Lvp/7BIwGm0zY3XDUF3jqDTRpwq70gOpuOIrW6D8ZcKTTDwWUlJ2lp6cn6L6ys+cneDXTlwSfhBBiijhz9jzFJWfQaDQsWThvTL1GhBBCXHG24iKtbe0B2Rkej4em5hbSU1MwjCHLKRiHuRqlSuuX8dG390+osgxCHdy42vy/p55m196Dvp+tNjuvvfk+iQmxPPbIfSM+3kgnC/YvrRrL+2Asr/VI1q0yZbH/nIpmS7wvw8nuUnO6KZmYMAdN1gjA26Rbq9XyyAObB8wSDGXPq/GgVCq5+85buPWm6zF3WoiJjhx2c3uFQsFnH95CcWk5JaVn0ev1LFk4l5TkwF5L4yVqtjewFJa40Jfh1NuEvLvhqG+/GJuY2Gg0arVvkmNfSQmBZdsiuCkRfHK73fz3r55n36ETANy/+Ra2bLop6G1/9ss/snv/UeobmvnDL39MzuwM377qmnp++F9P02HuxBhu4J++9QVmZaZOyDkIIcRYvLD1DQ4ePu77+aNd+9iy+TZWLpOmj0IIMVY2u7c3S0J8LM0tbX77XC43s2dmhLwUayRlc2MRyuDGcNjtDlxuV9DR8eXnLnC24iIRxnAWLSgkPCwsZI87Gm63mx0ffxJ031vvfjji4NNoSshC+T4YyWvd09PDoaOnOHvuAvFhZvLi2zBFJw5r3U6Xk5qOwNfX7lKjMM7g/3zjU5SUnUWlVFFYkIPBoB903dOhLFSv142qzFahUDC3IJe5BbnjsKrhCRZg0hoTAzKeBlNT28CefYdobesgPS2FNSuXYIqYOq/PZCvIySJrdialZ/x7m0UYw9l0242TtKrpZ0oEn97bsZcLVTW89MxPsHR18+hXvs+i+QVBA0fXr1nKw1tu4wvf+lHAvv/8+bPcdev13HbTdXy4+yD/9tOnefZ/Am8nhBBTSfm5C36BJ/B+G//am+9TNC8/6B/4Qgghhi97diYatZqoSBMZaSlU19Thdnun2+XlzOaRBzaPy+NORHnLRAW5Oi1dPP3MC+w9cBiPx8OCeXP43GceYEZKIm63m+f+/Ipf+dHb73/EFz77ILNmDt37Zry4XC6sVlvQfZ2WrhEfb7QlZKF6Hwz3te7p6eEXv3ne1whZp3LSkdLE3FwrMPS6wwwGZmakcaEysMF1QV42pgjjiCZ8jaYstG+mltPpZNfeg5wuPkmEppOMvNWsWblk2BlKYnClZ87xuz+8hMvlzWY7c7aCg4eP882vPE5UlGmSVzc16PU6vvaFT/Ob516kvqEJp9NJpCmCTbffRF7O7Mle3rQxJabd7di1n00br0elUhJpMrJ+7XK27dwX9LYL5uaREB8bsL21vYPSs+e5ef0qAK5fvYSGplaqa4KPzxRCiKmipOxs0O09PT2crbg4sYsRQoirUJjBwKbbb0KhUDAjJYlFC+aSlzObB++9k//417+f9hdYI5m8Nlrf+t6Pefm1t6mpbaC2rpG33/+I7/zzv9NttXL42Cm/wBOA3W7nhVfewOPxhGwNI6XRaEhPSwm6Lydr5oiPNxUmCw7ntd5/6LjfBK7enk1l5xvosXUMa92b77w5YBJcSnIiN6xdMaL19i8VjM7Z7BfACzb1sTfDrHf/M398mffff4849wkiXeUc+Pg1fv+nV0a0DjGw19/e7gs89eowd7Lj472TtKKpafHCefzfH3yHr37+03z5iUf4vz/4ex645/bJXta0MiXCxfWNLX61ksmJcRSXVozoGI1NrcTFRKFWqQBvCmRifCwNTS2kzUgK6XqFECKUNBrNgPt6m4kKIYQYm9UrFpOZkcrhoydxOHqYk59DQV7WhDSdnu4OHD7umxLY17mKSnbs3EtDY0vQ+zU3t1Lf0DSpPQy//LlH+Ocf/dSvV0t4mIEvf+7hUR1vqpaQ9c0UKj93AfBmPPU2B7e71JysiaCowEpEuLcccrB1p6em8L1vf5mDh4/7SrEWzp8z6N8swbTUlNLd0YQxMn7YZaF9M8zOH/ojNRdrKEpuRa92YnOqabPqqS8t58LFamZmBmabjbQv17Wsq7ub+obGoPvOVlRO8GqmvtiYaG5av2aylzFtTUjw6YknfzBgBtIffvHjiVhCgFfe2Marb24DmNRvZIQQYsnCeWz/aE/A76KoSNOovpkVQggRXGpKEqkp8qXkSJ04VTrgvlOny0kZJLikuvzF8GS5Ye0KoqN/yPN/+Sv1DU3MzEzj0U/dS/bszFEdbypOFuzfiyoszIBO5aQouQm92ht0a7PqyY9vRa2KH/a6I4zhrF+3alRramlt488vvc75i1UkGS1oIlTcY2wnM8MYtFSwb8Cod39L2Va6WktYkNwOgM2p9k3fA6isrgkIPo2mL9d0NtZAm06rRavV4nA4AvaZTN5jejweTpee5cSpEuoamshMT+W6VUuDViOFmgQSry4TEnz67VM/GHR/UkIs9Y3NzC3wjuisa2gmKWFkb+aE+BiaW9txulyoVSo8Hg8NTS0kDvChuPfOG7n3Tm9zMKfTyZrbHh3R4wkhRKgkxMfywL138Orr7/n+5x8VaeKxT9+HUjklqqOFEEJcwwab3pU6I4n5c/M5fiowMyp1RvKEXKAOZcG8AhbMKxjzcabqZEGXsxuFWu9bx7K5Baga3iJSZ6fVaqDLoaEouYmEGD3GyPhxX7fH4+HXz7xAU7M3I67eYgRLG79+9gX+6e+/ijE8zK9UMFjACMDjcqCjgzBND909GkqbYnyBJ4CoqMiAx9aa0ujodlNdepjWXftpIZs1eR7iIrWD9uWaTGcrLnLo6Ekcdgdz8nNYtKBwWH//hSLQplarWbpoPnv2HQrYt2r5IgD+/PLrfLRrH6VnzuF0esvzXn7tbR7acid33R58SFgoXGuBxGvBlLiquWHNMl5/9yNcLjcdZgs7Pt7P+rXLR3SMmKhIcrMyeX+Htzb1oz2HSIiLkZI7IcS0sGxxET/8/jd5/DP38+UnHuFf/vHrpKcG71MhpiebzU5tXQM2m32ylyKEECNy/XXLg5bORUZGsPmOm5hfmM/a1cv8ShhjoqN4+P67xnVd1ubSgJ5BLocFa/PAmVpj0X/anNaY5NcDymEObNA93qzNpXTVHgTwBaDCzLuZk6bEpHdS1xlGuLaHhGg9efnzJmTdZeUVvsBTXzabjSPHTwVs79/I3WGpp6VsK7bWcgwGIx6Vt/dUfnwrOpU3kys2JprC/OyAY1241Myz7zXR2GpDjY1ETlFeVkxjm21Ss9MGsv2jvfziN3/k4OHjHD9Vwp9f/hu/+8NLw6rMCfa89Q2ODjfQtum2DSxdXOTLUjQYDNx1u/dzXX7uAoeOnuTsuYu+wBNAVVUN2z7cHTD9LRR6P9d9z6+lbCvm6j2jOj8xdUyJnk+3rF9NSfl57nvsWygUCh64eyNZM71vpt37jrB7/1G+980nAPiPnz/DJweP09rawZPf+0/CDHpeee5nAPzD1x/j3376G/7w4huEhxn4/rc+P2nnJIQQI6XX6yZ1VK8YHx6Phzfe2cGefYfo6elBq9Vy3aql3Hbz9dJrRoyLvQeO8ds/vsr5yktsvm093/zSI759breb//7V8+w7dAKA+zffwpZN4/fNtbg6hIeF8X9/8B3+879/zaWaetxuN8nJCfz9k18gwui9mN98x82sWbmEc+crMYaHU5CXNa7Zu5ORFTFRkwVHou8FOgolrp5ueiy1hOvVRM9YRfzSdRgj44hUNU/YugebJGg2BzYY798HqrVsK/aOiwCEJcxlQd4Kjn/8B+hspiilCYthKXdvvivotLttH+7F1qOktCmGBclXehl9eMrF3PVTK/DUaenirfd2UFVdS1t7BwqFgvi4GDweD8Wl5UP+Tdj/eevtQzbSBvgajYaHttzJnbduoKOjk4T4GF9vr9IzFVgs3djs/l+cuT0eOsydHDtZQn5u1ijOPrj+n+ve8ktrUzHd9UfRGJPRRWZMyUCiGNqUCD6pVEq+89VHg+5bs2IRa1Ys8v38j994fMDjZKSlDFniJ4QQQkykDz/+hI92feL72eFwsP2jPUREhLN21bJJXJm4WqXNSOL7f/cEH+4+SHe/MfPv7djLhaoaXnrmJ1i6unn0K99n0fwCZmWmTtJqxXRRkJfN7/73P3xTWHOyZgZc/MfFxhAXGzOu6+jp6aG4pJzOjmZStUrC8WZ99C0lG8+siGCBmlBPFhyJvv2RrE3FeFze8n2FSotao2dmZm/AKT7ofcdj3bMy01AoFEGzd2bPTA96n76N3F0OCx6XA0N8IbF5W1BpjWzY8n0aS1/F7egkZuZyDDHRQY9T39CITuUkP77Vb3uCqoIemxmNfvSTLTstXbR3mEmIi0WnG/tAmPJzFzlZXIalq9u3zdLVjaWrm/Kz54f1hWQoG+Abw8MwXm5G30uv1wLBs7BUKhVut3vEjzOY/tlcpoz1eFwOPC4HCpUWlcY4JRr8i9GZEsEnIYQQ4mq1d/+R4Nv3HZnw4NOBw8fZs+8wZnMns2dmcPOG60jsM21WXB3SU5MB+PiTwwH7duzaz6aN16NSKYk0GVm/djnbdu7jC49umehlimlIo9FQkBdY6jRR6hua+NXv/kSHuRPwTnO7ZQHkz2bUWR+ToW8T5bLyCo4cO0VslI45mSbSckfX4LvvBbomYgYqTTgep21SelHFxcawZuUSdu096Lc9L2f2gFkyfRu5q/VRgDeA1kulNZKQf8+QmVopiSZM1nLfZLzSphjy41uJMqroqHhzVM+F0+nk5dfe4fDRk7jdbnQ6HRvWreLGG1aP6Dj91dTW+gWeejU1t9LT4wxyj0Dj3QB/8YK5vL99FzqdFrv9SlNyrVZDpMlE0dyx91Lra6AsOIVKi86UgUKlmfQG/2L0JPgkhBBCjKPOzsASAxi8LGE8bPtwD2+//6Hv56Mniik7W8G3v/4EMdFRE7oWMXnqG1tI6hNwTE6Mo7i0YhJXJMTwvbD1DV/gCcDuUvPRKTtxkS0kxHkbm0/1rIjesqIej45fvn6JHbuPYtDCHYuVtJaEk1hxkfW3fmpEx7S1lGNrLfdeoEdmolCqfYGb3p5OE52ZdfedtzAzI40jx4vp6XEyd04uy5cUBS03H24j9+Fkaq1bnEnJ/v3Yeq5MxjteF89D10eO+rl4890dHDh8nLr6RurqG7HbHRw+epKu7u4xNdx2uz3o9bp+vSA92B09HDtZgtVmo2huAUXzCsb0vI1FbEw0n37wbmw2O8dOluByudDrtORkz2LV8kUUFuSM6fjB9GZz9QaeerPgorM3TYkG/2L0JPgkhBBCjKPZszIoKw+8uM+alTFha3A4etjx8ScB27u7rezae3Bcp9WI0HviyR9QXVMfdN8ffvFjEkc4MXgwr7yxjVff3AYwrAa4QoyX1rZ2qqpr/Lb1llc1t4T5gk9TPSuit6zo+OEDxDgvEW+CtQUKwrUeahot7K08Tva864c9dMTlsGBtLUVtiENjTCJq1kbfBbpCrceYvGzSSgIXzJ/Dgvlzhrxd/0bu/bNfRhIwyi1aj8vlYsf+87gUrSQmRLN+3UoK584eVeDJ5XKx/9BxLtXUU32p1re909LFL3/7JxYWFY56QExMdBT5uVmUn7tA1+UMqO5uGyqViqbmFux2OydOlVJaXsFDW+4MuH8on7fBFM0r4Of/7184XXqWsxUXiYo0kZc7m7QZyWM+djC92Vy95ZcKldZbcjdO5ycmjgSfhBBCiHF0683Xc/5iNQ7HlXR1nU7HLRvWTtgaWtvasdlsQfddqg0exBBT11j6WyYlxFLf2MzcAm/pVF1DM0mDBKvuvfNG7r3zRsBberLmtkdH/dhCjEX/4KdO5aQouQm92kmPW0t0zuZpkRWh0hpxRK7kUsM2jDoPty/0ZrR02uDtox5MUe2cLC4bdkCjNwChi0wPGoBQagzjeTohEepG7gWLbqJgUeD20QQqenqcWK026uobAvbZ7XZ27t7Ppx+8e8THBVi8cC4ffLgbg15Hd7eNzq4uKs5XYgwPwxRx5b178PBxrlu1lNQU/ynuE9kAX6PRUDTPm4UVKn3LT3s5LA20lL6IUqVFF5mOKW0N1razfiWkk9ngX4zN+I2gEEKIq1RTcwu7PznEkWOn/OrfhQgmPTWF73zjCdasXEJO9izWrl7G3z/5eWakJE7YGiJNEWiCTAUCiB/n5sBiarlhzTJef/cjXC43HWYLOz7ez/q1yyd7WUIMKTYmmtQ+mRbRBpuvr09k1h1ojUne4MvlZsUOc/UkrnZwVoeSEzURfts+LvHQbfcGeVWq4V+iGeLyMaWv8wu29QYgTOnrps0FuiEuPyBYOJmN3Hvp9ToS4mJxOl0B+0ymCJqaW4Pca3jCDAa+8sQjZM3OJCxMj8ftJi42mvzc2QG3PXvuQtBjTNXnbSi95adtZ1/H5fC2J3A5LLSUvoi97Rxul4Po7E1EpK32Np3v87meDucngpPMJyGEGIE33tnOR7v2+b6BDQsz8PlHHyQzQyZFiYHFx8Vyz6aNk/b4BoOelcsX8fGeA37bNWo1161aOkmrEuPl0LFifvSTp+nqtoIHPtpzkO989VHWrFjELetXU1J+nvse+xYKhYIH7t5I1szxmQomJkenxcJ723Zh7rSQlzObxQvmhmQy11Tw4L138utn/kynxUK9xXvBPWNmEUuWeAOoocr6CJaR4XJYQpZtMSPRxOpcN20tCjxu798TawsUvH3UQ1RUJAvmDV2q1leoJ/Bdqq2npPQsWq032yUqcvQT4q4G99x1C+/v2IXd4d9we0ZKIslJCWM6dnJSAl/9/Ke9PZVOnOalv74V9HbG8PAxPc5UE2yqnblyhzfjKTqL2PwHxj2bS0w8hcNhv+YL+HvTyHe//VzAyFghhOh15ux5fvW7PwVsj4mO4p//4WtBm0EKMVW43W7e37GLvfuPYLF0MTMjjdtvuYHZE9h7Skxv8vfS1Ldz9z5++B//i6XLO9DAFBHOssUL+M6Tn79qAggORw8nTpVi7vRO7Qz1lz/W5lJay/+GxpjkzbjQGr0ZGWVb6bHUE5Nz15gugHubRNfVXODwyfP8bW87q/M8ROihx6Nl7rovsm7d9SE8o5F5/e3tfLTrSo9AtVrNIw9uZn5haC76xzuwN15ee/N9fv+nV7DZ7BjDw0hKiscYHs43v/LYmANQvex2Bz/6f/+Dpd9AEqMxnH/++69dNUHkXn0bpveaDtMqxejJXw5CCDFMx06cDrq9ta2di5WXmJkp2QNi6lIqlWy8cR0bb1yH2+1GqZTKeyGuJvUNTfy/p37jCzwBmDu72HfwKO9t+5gH7r1jElcXOlqthiWL5o3b8d09VpzWZpzWZlqAqFkbaT//LtamYt/+sejt0ZQ8YyY35DyALv00lRfPkZXeRk5mPPFZSUMfZJxcuFjtF3gCb9D5L1vfJC979piDH72lVn0DDP0DEFM1ALX5jptJTUnm470HaO8wk5meys0brgtZ4AlAp9Py+c8+yJ9fep2GxiYAEhPi+dT9m666wBNcmWrXVv6ab9tUn1YpxkaCT0IIMUyDpYnKFCgxnUjgSYirz4HDJ2htaw/YbunqZv+h41dN8Gm86WNzMMQXYm0qxtpUjNPWRk+nd8qeIb4QfezYRsv3bRIdpzWSkZkFTI3snxPFpUG322w2yisuMLcgd0zHH6jUqndim9YU+CXeVMqUWrJo3rgGPsHbJ/K73/oS9Q3e4FNSYvy4Pt5k6p1q19dUn1Ypxkb++hRCiGGaX5gXdHukKUJ6PgkhhJhUlq6ugQPLUhU+bCqtkdi8LehjcvC4HDjaL+BxOdDH5PjK8MZqqjaJVigGvjRU9Xlv1dU3cqrkDG1tHSM6fm/vHl8Aqvw1X+ApWMChf1Nqc6eFd997l+1bf8zRj/9AZdmekZ3gNJKUGH/VB556M95UOhPROZv9ApO9TcjF1UUyn4QQYpgK8rJZtXwxe/cf9m3TarV86v67JJNECCHEpMqZPZP4uFjq6hv9tqtUSjasXTVJqxLTyaIFhQFldwAul5vSM+coOXOOivOVvveYQqFgxdKFbNl867D7Xo6k1KpvplTdqZf5y456Ug3V6NVOWlvVvHV8B498KpGCvOxRnrGYLL3lp30Dj9HZm3wBqcnOAhTjQ4JPQggxAls238qKZQspLTtHWJiBonn5hIeFTfayhBBCXOMWzC9g3eplvLd9F23tVzJS1q9dxcab1k7iygbWXvEuYYkL0RoTfdsclga6G44SNXtyJoT2Nhe3tZajUGnRGFPosdRiay2npWxryLKfpqLUlCTuvHUDb733IW63G4DaukbCww3s/uQQ585X0tjUwsyMVJKTEvB4PHxy4AhJifEBk1MHKpeztp7B1lLmd9uBSq36BiTOnzlNVoS3FM3mVHO8Lh67S8Eb7+yQ4NM01Lf8VKbaXTsk+CSEECOUmpJEasrkNQQVQggh+lOr1Xz1C59m6eIiDhw+Tk9PDzdct4Ili+ZP9tKCaq94l/aKdzBX7yJp8ZNojYk4LA3UH34K9+WSm8kIQFlbz/iaixviC/0ajlubirEmzMOYtGjC1zUYt9vNydNnqDh/kQijkaWL5hMVNfR0Q4/Hw4WL1dgdDmZlpqPTablh7UoWzi+k5Mw52to7eG/bTlQqFW63h+aWVgAuVtUQGxONVquhq9vK7/74EieKS0mfkcLa1cvQOWuCNhZvKduKtakYtSEWXWSGX8+ntrOvDxiAMmWsx3z4gG9baVMMdpf3Mra+oZGu7m7fF4EHDh9n9yeH6OjoZGZmKjevX8uMlETE1BMswDQVyk/F+JHgkxBCCCGEEFcBjUbDquWLWLV8agVHgglLXIi5ehduh4X6w08RW/AQLSUv4HZYUGqNhCUunJR1qdRhKFRatJGZviyn2LwtNLkcODouolJPrWxnp9PJ07//C2fPXfBt2/bRHj73mfvJzZ414P1q6xp49k9baW72BpR0Oh33btrIkkXziIoysXLZQt7b/jEqleryPTy43d7hKh6Ph/YOM1qthtIzFeh0WmKiIqk4X8mhYyf5+hP3BW0s3mOpB0BjTB52qVVvU2qN5spla3586+XMJzVarRad1jsJ7qNd+3j5tXeorLp0ufm+NzPqv//jn5k5TXtzTqWG60KMlTQpEUIIIYQQQkworTGRpMVPotQacTssNB3/jS/w1JsJNVlUWhNKlf9oe6VKi0o7dDbRRNt38Jhf4Amgp6eHra+9M+AkXo/HwzN/fNkXeAKw2+385ZU3fFPWALQaje/fSqUSk8no9/PFqho8Hg/RUZG+7RZLFzv2HA/aWFwXmU5cwaf8Shd7A1Cm9HVBA0+9ganElAyO1SVgc6rRq50UJTehUzlZtng+arUap9PJ+zt2c7q0nOaWNtxuD263m9q6Bn7w709NylTiru5umppbRv3Y/Ruuw5XnxFy1k/aKdwMac7scFqzNwacWThXW5tJpuW4xdpL5JIQQQgghhJhwWmMisQUP0XT8N75tsQUPTWrgSWtKQxeZHpC143Ha0EWmozWlhfTxPB4PO3Z+wu5PDmLutDAzI43bbr6e2bMyhnX/krKzQbc3t7RS39BEclJCwL6KC5W0tLYFbHe73Rw6epI7Nq4HYOH8Qt5+/yNcLhcAM9NTOX358UymCLrPXcCg15Har6yt4kLlgI3FtcbAtgUDlVr1bUqdt3ITN6iK+WjnR+RFVWHQuli1MJ2Nt24AoMPcSU1tPTabPeA4l2rqKCuvID83Cxj/bCKr1cbLf32bE8WluN1uYqKjuPO2GymaO7Jj92243ve96LKbcbsc2NrO4bS1+pU29gbrIHhZ22TrDaj1L8mc6usWoSHBJzEs3VYrAGEGwySvRAghhBBCXA0clgZaSl7w29ZS8gKasMnLfOpfCtYbPOl7sRxKf3trG6+//QFqtRpjeBjnL1bxq2f+zDe/8viwehVptZoB9+l02qDbrUECNL3sdofv31FRJh6+/y5efPUt7HY74eFhrF6xhLlzctCoNVgsXUSajH1K87wiIyJ85XJ9DdRYfCD9m1KvW7OcVcsX0VBXjUHZQWzqAt9tTRFGnE5X0OPo9TqamlvJzx2/4EdXdzdajQaNRsMLW9/g1OkrTdVb29r54wuvEvvVx0mbkTzsYw76XszYgLlyuy8wpTNlYG07i8dpQ6UzoTWlTcnyvMECar3rFlcvCT6JQTU1t/Dya+9w9twFFAoFOVmzuO/uW4mNiZ7spQkhhBBCiGmqb3Nxpdbo1/Op/vBTk1p6N1DWzlgDT52WLs6UV6DWaCjIzWLfwaP89P/7LY6eHgDCwwzk5sxCD3y89wAPbblzyGMuWTifE6cCy5Vmz8wgJjoq6H2yZmag0Wjoufy4ffVmB/VaMH8OBXnZlFdcQKVUkpM1E7XaewlpMBj4aNcnAcdYvXSOL2Ci0pmG1Vh8IP0DJxqNhtT0wF5WGo2G61YtpeJCpd92pUJBSlICM5K976VQBz8uXKzmtbc+oKq6BrVaTV7ObE4Ul6JS+ne3cbvdfLL/CPffc/uIjj9wBlmiLzBl76jCUrMfhUqLIb6Q6OxNAFMym2iig7tiapGeT2JATqeTX/72T746co/Hw5mzFfzyt3/ypd8KIYQQQggxUt0NR/16PIUnzPXrAdXdcHTS1jZQ1k7/PjUjsXf/Ef71//6cP730N57701b+7rv/xtPPvuALPAF0dVspP+v9u7upqWVYxy0syOGm9df5ZR8lJyXwqfs3DXgfg0HPXbffhEKh8NteNLeAgrysgNvrdFrmFuRSkJftCzwB3H7L9axZuQTN5W1hYQbuuv0m8tLDfcGc6OxNaI1Jfj2gHObqYZ3bSH3+sw+weOE8XzZYhDGc/Lws5s8r8JUx9gY/+vejGk3wo7WtnV8982eqqmsA77XT/kPHKD97PujtzZ0jf/8M9l7sDUyptEYUKi0elwOPy+GXyTUVs4l6191XKIK7YuqTzCcxoJPFZbS1dwRsb2lto7i0nPmFUyOCLoQQQgghppeo2RsB79S73gyn3ibk3Q1HffsnWv8L97Fk7fRqaGzmlb/5NwCvulRLU3MLKpUSl8vt227p6qar20pK8vCzvm69aR2rli/iQuUlTMZwZmamBQSW+lu1fBGZ6akcOnoCu91BQX42hfk5Q96vL5VKxT2bNnLrTdfTYe4kNiYKTZ8m5X37KvUGfcazDEylUvHTH3+P93fs4vjJElQqFQvmz+HG61f73y5EmW37Dh7D4XD4bQsz6Om0dGG12TDo9X77ZmWOLAhUXXUea9V7GPUetIaogPeiKWMD5sodKJRqdJGZ2DsuYmstp7X0ZRQqzZTNJgpFSaaYniT4JAYULPDUq73dPIErEUIIIYQQV5tgASatMRGtcXICT+Df5Lr3YrhvmdBogidHTxQHTDxzOp24XG5ioiNpbfP/m1ulVLJuzbIRPUakKWLEDa1npCQyI+WmEd0nGINBj8HgH2gJ9hwN1Fh8KCNpEN6b1XXX7QOfV6iCH61t7QHblEolaakp2O0Ov+BTfFwsK5YtHNZxW1rbePb5rbjM58iNa8OJnoxFD7H0cgZZ29nXcXa30FL6oncK4+UgadvZ17E2FWM3V6IzZUzJbKLxCO6GQld3NyqlCr1eN+GPfS2R4JMYUEb6jIH3pQ28TwghhBBCXD3GezrYVNK/yTWMPWsnWCPsqKhImlvaiIuNISY6irr6Rhw9TuJio/nut79MfFzs2E5kBKby6xvqBuGhDH5kpM3gyLFTAdvTZiTz0H2bOHGqBIulm+ysTNasXDLswU3PPr+Vmtp6wLuONqueTyq3k5I6k9QUbwCqs3o39o6LvucFQKHS+srvXD2WKZlNNB7B3bGoqW3glb+9w4XKapRKJYUFOdx7162YIqbOc3Y1keCTGFDWrEzycmZTVl7ht31Ofg6ZGamTtCohhBBCCDFRrsXR6KHM2gGYW5DLjp17/bbFxUTT2tpOVGQEarWahPhYlEolD225M+Rf8g4WXAJC9vrabHYuVFZj0OtDdq0Q6gbhoQx+LF00n92fHKKp2b8/13WrlrJw/hwWzp8zorUBVNfUXQ48edVbel8zD4eOnCA1JQmV1kjU7I2+1xW8zcU9ThuG+EL00dk4zJWTnk0UzHgEd0er22rll797nq6ubsDbFN7bdsbMt772uQlbx7VEgk9iUJ/7zP3s2nuQE6dKUSgUFM0rYPWKxZO9LCGEEEIIMQFkNPrYZWakcv11K/0mw6nVKn7wvW+gUaspLa8gPMzA0kVFzEgJ7YS/oYKH4SlLQ/L6fnLgKK+/vQ273Q5AUmICjz2yhYT4sWVwDTQdze1yEJ2xYcTZWqEMfuj1Or7+pUf58ON9lJWfI8xgYNmSIpYumj+aUwXAarUNuK+r2+r3c+9arc2lAQEaRjUpAAA19klEQVQ1l6No0rKJhhLq4O5oHT56yhd46qv6Ui0V5yt9TepF6EjwSQxKrVZzw9qV3LB25WQvRQghhBBCTLDpMhrd3Gnh0JETtLWbyUyfQdG8Ar/JbJNt020bWDC/gOKScjRqNQuL5hAbEw3AogVzx+1xhwoeGmJyMcTkjun1rbpUy9bX3vbra1Xf0Mizz2/lH//ui2M+h/4Nwp22dhQqLebK7aPK1gpl8CPCGM6m2zaw6bYNI75vMBlpM9Dr9dhsgUGogtzASYQwtbKJppO2QXoYD9b7WIze1PmNLIQQQgghhJhyQjUdbLxcrLrEr595wXfBvmffIXZ9cogvf+7hKdVAOD01hfTUlAl9zOEGD8fy+h46ciKgoTp4A1BVl2rHfM79G4SrtEYcllqUSs1Vl42n02m56/YbeenVt/ye09zs2cwfpKH8VMkmmk4y0oK/LxUKBenS33hcSPBJCCGEEEIIMaCpPhr91dffC8gUqaquYdfeg9y0fg2Xaus5fPQkDkcPBXnZzMnPRqFQTNJqJ95QwcOxvr42u2PgfTb7KFeNb23BGoQDOCy1AFM2G2+0li9ZwIyUJA4ePk631UZBbhZF8wpQqVSTuq633/+IHTs/weV2MScvm1tuXDvhwdRQmleYR0baDCqra/y2L1k0f8zloiI4CT4JIYQQQgghgMDm1C6HhZayrfRY6tFFpk+Z0ei9zJ0Wqi/VBt13urSc8PAwXvnbO74skk8OHGHh/EIeeXDzlAlAOZ3OcS0RHCy4BIx5+ltBbhaHjpwI2G4wGMhMH1vj8cEahLvdPbh6LKhV3vLFqZSNN1ZpM5JJm5E82csAoKenh+//60/YueeAb9u5iouUn7vAt7/+BEmJ8ZO4utFTKpV86XMP8/HeAxSXlKNWqVi0YC6rli+a7KVdtST4JKa9qku1WK02MtNT0em0k70cIYQQQohpydpcSmv5a2iMycTmbUGlNWJtPUN34wk8rh6iZt6M1pg0qaPR+1OrVSgUiqBlXygV/O2tDwL2HT1RzJJF88gfoIfORDlw+DjbPtxDc0srcbEx3HjDapYtLgrpYwyUOdQbXDLE5I95+tv8ufkU5GVTUnbWt02pVHL3HTej1WrGtP6B+hmZMjbQUvoiKn2077ZTKRvvavLmuzvYs/+w37am5lYM+mp27t7PA/feMUkrGzu9XsfN66/j5vXXTfZSrgkSfBLTVktrG8/88WVq6xoA0Ol0bLrtRlYuWzjJKxNCCCGEmH5czm6c1hac1hZagKhZG+luPInH5UCh0uJReIM4U6mZcZjBwJz8HIpLzgTsS0qIp7LyUtD7lZSdndTg05Fjp/jL1jd8Pze3tPKXrW+gUipZvHBeyB5nsMwhl92MUmPAlL5uTM2qlUoln/vM/Zw8fYbSM+cI0+tZsmgeKcmhmdzXfw3eTK7tKFXaUWdrieHbu/8ITqcrYHtjUyv1DU2TsCIxXUnwSUxbzz6/1Rd4ArDb7Wx97W1mpCSSIU3ihBBCCCGGzdpcitaYiiG+EGtTMdamYpxdjdjNlSjVBgzxhRhicn23n6xmxv3LAgHuvWMdansVxyu8o+iVSiWrVyxmTn4OBw4dC3ocnW5yG5Hv+PiToNs/3LUvpMGnUU9CUxlQGGcN+3GUSiVFc/MpGqQpdqgMFVCbCkHRq4nb7UGlUuFy+QegXG4XCQlxk7QqMR1J8ElMSzW1DdTU1gds93g8HDpyUoJPQgghhBDDZG0uxVy1E5XORNSsW/G4HHQ3nqS7pQTcLrSmdF8Z3lRZZ2/QweWw4KzbzsbFRq5btRSzO4GM1BSioyNxu93ExkTT0trmdxylUsmSEAZ4RqO5pW1E28diJJPQnE4nb767g/2HjmO320mdkcydt24gJ2tmyNc1WqMOqIlRKSzIoepSDdWX6vy2x8VEs271sklalZiOlJO9ACFGo/9Ek76sg+wTQgghhBD+tKY0VDoTLruZ9vPv4OrpxtVjAbcLlCqUU6SEqe86286+jsNS79fPKCN7EUVz84mOjgS8QabHHrmPuNgY3zF0Oh0P3nsniZOcsTFjgJK0gbZPlK1/e5eP9xzAbvdOqbtUU8dvfv8X6uobJ3Vd/Rni8gOCoZOVjXe1u2PjevJzs5iZkYper0OpVJCUGM8P/+nvQlZaKa4NkvkkpqX0tBTCwgx0d1sD9uXnzJ6EFQkhhBBCTB8ul4ueHid6vc6XNdJStpXuxhM4ra2gUKA2xIBCiaP9Ai1lWyc9+6l/eVVb+Wve7X0yofqbkZLI97/zFc5fqMLR08OszPQpMaDmlhvX8vSzL+B2u33blEolN28Ye+PjhsZmTp4uIz42hsKCnGFP0rN0dXP46MmA7U6nk2OfvEXC7ff5Pccuh0Uyja4BCfGxfPdbX+LAoeM0NLWQkpTAkkXzCDMYJntpYpqR4JOYljQaDXffcTMvbH3D73/audmzWTB/ziSuTAghhBBi6urp6eH1t7dz8MgJHA4H6Wkz2HTrBmbPykAXkYa58iNv4EkfQ/y8x+iqP+zrAWWLn0948uQOdvFOOlvvCzwBmDLWDxoUUygUzJ6VMRHLG7bc7Fl8+YmH2f7RXuobmkhKjGfD9avImpU56mPabHZ+9r/P8Pb7H9HT04NSqWD2zAz++R++Rt4wvpxt7zAH9PUBSDJaMNhP03jiGWLzH0BrTPRN0XN2t2DvuEjU7I2jXreY+sLDwrhh7crJXoaY5iT4JKatxQu9UzQOHjlBt9VGXvYsiuYVoFRKNakQQgghRDB/2fomR08U+36uqq7h18++wLe/8inczadRqg0oNWHoI2dibTpJ1KxbAeix1KPUTH6mg3fS2Q6/bebKHdNywlnWrMwxBZv6e+mvb/Hmuzt8ASS328PZiov8x89+xS9+9kPCw8IGvX98bAw6nc5XcterzaonIqwHe9s56g8/Rfz8J+iq3YezuwVHVx0oFFibSyUDqp/mllaqL9URHR1JZnrqZC9HiEk3JYJPbreb//7V8+w7dAKA+zffwpZNNwW97c9++Ud27z9KfUMzf/jlj8mZfeVbjM2ffhKtRo1O602l/fT9d7Jh3fLxPwExaVKSE7nr9uDvFSGEEEIIcUV7h5ljJ08HbFe6rZw78Aey0uMwxBUQlXUHXbX7sHdU0n7+HaJm3YrL1oohLn9SS616s216ezyZMtZjrtzh6wE1HQNQoeJw9LD9o71BM5dqaus5caqMlcsGz1rT6bTcdMNq3nzXP7inC4tm1vJNWEp/i9thofHoL9AYU3Bam9EaU1AbYtCa0kJ6PsF4PN5gWk1tPXFxMczJy56SXzp7PB5eevUtDhw+jsfjAWBmRhqfe/T+IQOAQlzNpkTw6b0de7lQVcNLz/wES1c3j37l+yyaX8CszMAI8fVrlvLwltv4wrd+FPRYP/re1/wCUkIIIYQQQghobzf7Lob7ijbYcNqtqHSziOudImdvo6v+KE5rC+GXy+36Bn8g+BS18eQwV/sCT72Bpr49oEIVFGusPEyjWU1MXLKvofJU6W9U39DEyeIylEolC+YXEBsTDYCjpweHwxH0Pk6na9BhPX2tX7eK6KhI9u4/grnTQvbsTDZcv4qY6CgM2idoPPoLPC4Hjo6LaKNmojbETEjQz2538Jvf/4WKC5W+bYkJ8Xz5iYeJNEWM+rgejweFQhGKJfrs2XeY/YeO+W27UFnNq397j08/dHdIH0uI6WRKBJ927NrPpo3Xo1IpiTQZWb92Odt27uMLj24JuO2CuXmTsEIhhBBCCCGmt4SEWDQaDT09PX7b6y1GiMohOvt2XxDBEJOLIb6QHks91tZSNBEpviwjlc40IZku/fUGfrSmNN86ewNQoQgMeTwePnjjD1jrPsHao+J4XTzpM7P5zP0bsVW/P2lBt17bPtzD2+9/6Pv57fc/ZMvm21i5bCHG8DAK8rK5UHkp4H5RkSYK8rKH/TgLiwpZWFTot83lsNBVuw+NMQVHx0UAejpriJmgbLNtH+3xCzwBNDQ28fpb20YV0Dl3/iI/+59nKC4tR6fVsm7NMr71tSfQ63VjXmuwpu0AJ4pL6enpQaPRjPkxhJiOpkTwqb6xhaQ+406TE+MoLq0Y1bF++F+/xuPxUJA7my8/dj/RUaagt3vljW28+uY2gKDfAAkhhBBCCHE1CTMYuGHtCt7fvstve3RUJItW34lKe6UkSKU1Epu3ZUST5SZCsMCPSmsMSUDowOHjfHjgPEXJKvRqJ0XJTZRWuzj0/lkKZidPWtANvBlPfQNP4L2G+esb7zF3Ti4RxnAevv8uikvLqayq8d1Gr9dx/713kJQYP+rH7ttc3GltRhs1k57OGjwuB40nfkfS4ifRGhNHffzhOH6qJOj2k6fLRpy91Nzaxhe/8U+YOy2At1H7629vp6a2gV/+d/DqmpFw9Avu9nK5XLhcbiT2JK5VExJ8euLJH1BdUx903x9+8eOQPc6vfvJPJCXE4XQ6efq5V/jRT57mZ//2naC3vffOG7n3zhsB7/jQNbc9GrJ1CCGEEEIIMRVtvHEdcTEx7Dt4FEtXN7nZM1m/bhXG8MBeNKOZLDedHTh0HLtLzfG6eIqSm9CrnSxIbsTcqoDc7EkNup06fSbodqfTyenScpYvWcDsWRn8+ql/46+vv0dxaTmxMdHcvelmCvNzx/TYDnO1r7l4b4+nmOxNNJ74HW6HhZbSF0mY//ikPDejSSL404uv+QJPfR05XsyZ8vPk5swa05rm5OdQV98YsH32rIyQZFYJMV1NSPDpt0/9YND9SQmx1Dc2M7fAmw5a19BMUkLsiB+nN3tKrVZz/+abuf/x4IEnIYQQQgghrlVLFs1jyaJ5Q97uaposNxz2yz2T7C41pU0xLEj2BhDcbg+GlDWTes4q1cCNtVVKle/fsTHRPPHZB0P62Ia4fOwdF0Gh8OvxlLT4SVpKX0Sp0o57P6z5hfns2Lk36PaR9my6cDGwNBG8gayyc2MPPt2wdgWlZ85RU3sl+SIszMDdd9wypuMKMd1NifEAN6xZxuvvfoTL5abDbGHHx/tZv3ZkU+qsNhudli7fz9t27pPG40IIIYQQQoxC/8ly0TmbUelMvslyLkdg5sh0l5+bBYBO5SQ/vtW3PcIYjqNh76Sec9G8gqCT3bRaLYUFOeP++FGzNxKZucEv8Kg1JpIw/3FM6etCGniyNpcGPNfr1yxgUY5/8C8xIZ5Nt9844uNnpM8YcF9e9tgCT+Atb/3mVx7j4fvv4rpVS7nr9pv4/ne+woyU8S1NFGKqmxI9n25Zv5qS8vPc99i3UCgUPHD3RrJmeuupd+87wu79R/neN58A4D9+/gyfHDxOa2sHT37vPwkz6HnluZ/R2mbmuz/6OW63G4/Hw4ykBP7lO1+czNMSQgghhBBiWpqoyXJTyQ1rV1B+poQUVQl6tRObU815czIPFCX4gm6TlfUVEx3F/ffcztbX3sHpdALewNMjD27GYNBPyBrGs99WL2tzKeaqnX7vO5fDQnflu9y4wMDSRfOp6TAQFxdDYX5O0IDcUD513ybeeu9DLH0SFwAWzJtDbgiCT+CtxFm8cB6LFw6dYSjEtULhcNiv+W7bvT2fdr/9HGr1lIjHCSGEEEJMKfL30rXH2lzqN1kOvBlRV2PgqZe5/hQXjv+Nji4XjshVLF6yjAiD0hd0C3WWz0hZuropKTuLUqGgsCD3qush1D/jzpSx3m/KYqiCf6dLz/LUL56ltPwcWq2GNSuW8N1vfQm9fmICeUJciyT4hPwxJYQQQggxFPl7SUy2puYW6hubSYyPIyF+5P1hh+taDLpNJX0DUL3Ga8qi0+lEpVKNuG+UEGLk5C8HIYQQQoiryN4Dx/jtH1/lfOUlNt+2nm9+6RHfvt89/yqvvrmd+NhoAGZmpPKv//jlyVqqEAOyWm2cLi3H7fGQkzWTN97ezrGTp33TzeYV5vHIA5vRjMPc+okoLxMDm8gpixJIF2LiyKdNCCGEEOIqkjYjie//3RN8uPsg3VZbwP6brl/pF5ASYqroMHdy8PAJTpWc4cSpUiJNRpRKJbV1jajVKr9sp5PFZbzzwcdsum3DJK5YjIdrbcqiENeKKTHtTgghhBBChEZ6ajLZszMGHc0uxFRzsfIS//6TX/L629t4+a9vU3rmHMUl5ThdLuoaGqm4UIXd7vC7z6GjJyZptWK8XItTFkejveJdHJYGv20OSwPtFe9O0oqEGJpkPgkhhBBCXEM+3H2Qoye8WSWffeguFhUVDHjbV97YxqtvbgPwlTsJMR5eef1d7HY7be0duN1uwNtcu66+EbfLO826ubWNGclXxtXbbfbJWq4YJ9filMWRaq94l/aKdzBX7yJp8ZNojYk4LA3UH34K9+XgXNTsjZO8SiECSfBJCCGEEGIaeeLJH1BdUx903x9+8WMSEwZuxLz5tvU8+uAm1Go1J06X891/fYpn/ueHJCfGBb39vXfeyL133ghcaTguRKiZOy1cqqkDwN0vyNnWbiYqykRTc2tAALQgP3vC1igmRm9gqW/D994AlASevMISF2Ku3oXbYaH+8FPEFjxES8kLuB0WlFojYYkLJ3uJQgQlwSchhBBCiGnkt0/9YNT3jY2J8v17/pwccrIyKCs/P2DwSYiJoFarUCqVuN1uYqIiuaBU4HZ7A00qpZL01BTMZgsx0VG++5giIrj9lht8P7vdbkrKzlFX30h8fCxzC3JQqVQTfSoiBKTh++C0xkSSFj/py3RqOv4bAJRaoy8TSoipSIJPQgghhBDXiMamFl/T5uqaesorKpk9M22SVyWudWEGA4UFOZwsLkOjUTMzI42KC1UAxMfHoNNp+YdvfgGTKYK6+kYSE+JYung+YQYDAN1WK7/63Z+pvlTrO2ZiQjxf+fwjmCKkQbW4+miNicQWPOQLPAHEFjwkgScxpUnwSQghhBDiKnLoWDE/+snTdHVbwQMf7TnId776KGtWLOLXz23lzNmLqFRKlEol3/7qo6SnJk/2koVgy+bbMJstXKy6RGJCHNFRkSQnJbBscRHzCvNISowf8L7vb9/lF3gCaGhs4o13tvPw/XcB3p5lbe0dhBkM6PW68TwVcY3zeDycKjlDadk59Ho9SxbOJSU5tEEhh6WBlpIX/La1lLyAJkwyn8TUpXA47Nd898jeHga7334OtVricUIIIYQQ/cnfS2IiVF2qpa29g4zUGURFmYZ1n//z4/+mw9wZsF2jVvNfP/4eR48X89Z7H9La1o5arWbxgrncs+kWNBpNqJcvLrM2l/r1bQLvJLurvW+Tx+Ph93/aysniMt82hULBA/fewbLFRSF5jL7NxZVaY0DPJym9E1OVzOAVQgghhBBCTAnpqSnML8wfduAJvBf3QbcrlVy4WM3zL75Ga1s74A2i7j90jFdefy8UyxVBWJtLMVft9E6ouzx9zeWw0Hb2dcxVO7E2l07yCsfPydNlfoEn8Aak/vrG+9jtjpA8RnfDUb9AU3jCXJIWP4lSa8TtsNDdcDQkjyNEqEnwSQghhBBCCDFtLZhfOMD2Oez+5FDAlDyAI8dO0W21jvfSrklaUxoqnQmX3Uzb2ddxWOq9gSi7GZXOhNZ09faZKztTEXS73W7n/MWqkDxG1OyNRM2+1S/DqbcJedTsW4mavTEkjyNEqEnwSQghhBBCCDFt3bLhOmbPzPDblpaawp23rg9ajgfeDKiuru5xX5u1udSX/dPL5bBc1dk/Kq2R6OxNVwJQ5a/5Ak/R2Zv8SvGCaa94F4elwW+bw9JAe8W747nskNDptAPu0+tC12ssavbGgNI6rTFRAk9iSpOCfSGEEEIIIcS0pdNp+doXP8O58xepqWsgIS6WvJzZKBQKMjNSqbhQ6btte0cnnZ0WIiNNGAz6cV1Xb/lZ36BLb/mZy24GuGr7H6m0RpRxy2k4+RJqjZpIUwSmjPXDCjy1V7yDuXqXL7Onb48jYEoHWJYsnM/Hew4EZNvFx8WSmZE6SasSYmqQzCchhBBCCCHEtJc1K5O1q5aRn5vl6wO1dvUyoiJNeDweSs9UUFJ2luqaOmw2Gz/+r19SWV0zbuu5lsvPtm/fxgev/JTycxcoKT3LsROnqSt5MyALrL+wxIW+3kX1h5+iq/GUX3PtsMSFE3QGozMjJZH777kdXZ8sp/i4WB7/9H0D9iYT4loh0+6Q6S1CCCGEEEORv5fEdNVh7uTpZ//C+9s/RqPRkJQYR1Skt6F5clIC//DNL47bY7scFhpPPINCoUSh9H5uVDoTpowNuGytIcl8Kj1zjr37j2CxdDF7Vgbr1iwnwhg+5uOOVnnZaY5s+1/0aic2p5rSphjy41uJMWlYtHjZkKV3/TOdgGk3xc1m8/Z4Muj1ZGakSuBJCKTsTgghhBBCCHEVizRFYIoIZ+6c3IB9dfWNNDW3EB8XOy6P7TBX43E7cVhq0UVmolCqMaYsx1y5PSSld7s/OcSrr1/phXSx6hLHT5bwza8+jjE8bMzrH42yU5/4Ak/H6+Kxu7z/LaIJc1sD4ebqQc9Za0wktuAhmo7/xrcttuChaRN4AtDrdRTkZU/2MoSYUqTsTgghhBBCCHFVU6lUo9o35sfVx+C0NuNxObB3XMTltNJ44nc4u1vGXHrX09PDu9t2BmxvaW1jz75DY1j12LQ5YjjTHO0LPAG+ABTRi4YMtjksDbSUvOC3raXkhYAm5EKI6UWCT0IIIYQQQoir2sL5c4Juz0xPJSY6alwe0+WwYK7cjlKfgNnqprKxh4bqEmzd7Ti66jBlbBiyAfdgGppa6O62Bt13sfJSwLbhTN4LxXS+OfnZ1FuMvsBTL314NJn5awa9b9+SO6XWSHzR5/16QEkASojpS4JPQgghhBBCiKvawqJCVi5b5LctNiaah+7bNG6P6TBXY+tq5XBxFW8cj6Cl3UpVk4emFjMWhxaXrXVMxzdFGFEqg1/ORV7uadWrd/Je29nXfcElS/0RWsq2Yq7aibW5FGtzKR0XttF44hnfbXqn8/XeZjgWFhVSWOBf4qhWq7nv7tsHXG+v7oajvsBT0uInCU+YS9LiJ30BqO6Go8NagxBi6pGeT0IIIYQQQoirmkKh4L67b2Pt6mVUXKjCFGGkIC9ryGDIWBji8jlz8ChHKuspSPAGmtweJc1deiw1DeSsTR3T8U0RRormFnD0RLHfdqVSyarl/oG2/pP3tKYM2spf864zvhCtKQ2XowtHVx3uy03SY/Pvx1y5Y8TT+ZRKJY9/+j5Kz5zjVMkZbFY7161ayszMoe8fNXsj4J1619vjSWtMJGnxk3Q3HPXtF0JMPxJ8EkIIIYQQQlwTEhPiSEyIm7DHO1PVRUFCa8DkN7fHSe2pl5gx74Exld7df8/tqNQqjp04jdPpJDYmmjtv20B6aorf7VRaI9HZm7yZT3Yz3fX+GUTeEsEdaI0pOCy1KFD4glMqnWnICXX9KRQK6s4f5OTxCrrscPxUCfMK87j/rvUobA2D9n0KFmDSGhPRGiXwJMR0JsEnIYQQQgghhBgHCSYnqo7AyW8LZjSj8thwDDH5bSg6nZZP3beJu++4GavNRnRUJAqFIuhtVVojpoz1tJW/hkKlQWfKQKHW4XHafIEmtSGGyNm3Yb7wvu9+poz1Iw6QHd37JubKnRTGXznvspJTHA4rITczARjblD8hxPQjPZ+EEEIIIYQQYhzMXXJL0Mlv7tjriJm5PmQBGINBT0x01ICBJ7iS3dRLodIA4HE7fduMKcvpqt3ndz9z5Y6AJuRDOVBcj82pRq92UpTchElnpyi5ibamGhQa45im/AkhpicJPgkhhBBCCCHEOMjNnsV1Nz2AWu9tAK5UKlk4v5B7Nt89oZk/vY3De/s3RedsRqHWY2stx95xEY/bicftpPHE73B2t/hu07dP1EgCUB1dLo7XxfsCUAuSG9GrnXQ7VIRlbBxTqaEQYnqSsjshhBBCCCGEGCcrly1k6aJ5NDS2EBERjili4gMvDnP1lcBTduCEv/DEBVjqj+B2WHBQR0z+/WiNiX59okZSIpibPZOPGxopbYphQXKjb3u7Kg+jaeJ6bg2mp6eHYydLuFh5ieioSJYunk+kKWKylyXEVUuCT0IIIYQQQggxjtRqNTNSEift8XuDRlpTGiqtEWtzKR6nDUN8IeHx8wlPXkhYYhGNJ55BoVDisrWCMdHXqHykvaluWLuSM2WnydDX+bYplUo2FGpwOSyTnvlks9n5n6f/QE1tvW/b9p17+dLjnyIzY2xTCIUQwUnwSQghhBBCCCGucn2DR/2DUeBtSJ4w//GAQJNKaxxxiaBRr+AzN8VSd8lMm8WJPXwBhUlt6JQ9tJ19fcTT80Jt5579foEnt9tNU1Mzv//zK/zr956ctHUJcTWT4JMQQgghhBBCXGOCBZRGE2gKxmGuRunqJj0zm/mXA019+06NdcrfWJWUnfP9u7G5hYuVl3A6XRSXnkWn0/LFxx4iJjpq0tYnxNVIGo4LIYQQQgghhAgZQ1w+pvR1fhlOvSV8pvR1kxp4AtBqvZP+urq6OVdRidPpAkChUFBX18jv//TKZC5PiKuSBJ+EEEIIIYQQQoSUIS4/oLQuVJlVY7V00XwAGpta/LbHREeiUimpvlTLpT5leUKIsZOyOyGEEEIIIYQQ14wlC+dRU1vPufOVvm0RxnBmzUz3/dzdbZ2MpQlx1ZLgkxBCCCGEEEKIa4ZCoWDzHTcTaTLx+z9vRafVYgwP8+3X6/VkpM2YxBUKcfWRsjshhBBCCCGEENec61YtYfmSBX6BJ4VCwabbNqDTaSdxZUJcfSTzSQghhBBCCCHENUetVvOFzz7I0ROnKTtTgcGgZ+ni+ZL1JMQ4mBLBJ7fbzX//6nn2HToBwP2bb2HLppsCbmd3OPiXf/8FF6pq0Gm1REeZ+M7XHiVtRhIAre0d/PD//Zqauka0Gg3f/tqjLJibN6HnIoQQQgghhBBielCr1SxdNN/XhFwIMT6mRPDpvR17uVBVw0vP/ARLVzePfuX7LJpfwKzM1IDbbrr1elYsmY9CoWDr6x/wf5/6Hb/8r38C4JfPvERhfhZP/fs/UHKmgn/84VP89Q//jVo9JU5TCCGEEEIIIYQQ4pozJXo+7di1n00br0elUhJpMrJ+7XK27dwXcDudVsvKpUUoFAoACvOzqGto9u3/cNcBNt+2HoCC3NnExURz9GTZxJyEEEKEWHuHmb+99QH/9fPf8Otn/sypkjOTvSQhhBBCTBPW5lJcDovfNpfDgrW5dJJWJIS4lk2JlKD6xhaSEuJ8PycnxlFcWjHk/V7+2/tct2IhAB3mTpwuF7ExUX7HaWhqCfl6hRBivHVauvj5L39PW3uHb1tZeQWb77iZtauXTeLKhBBCCDHVWZtLMVftRKUzEZ29CZXWiMthoe3s67jsZv7/9u48zsa6/+P468yZzWyW4cwgu8mSpQhjvxmhFKlEpRI3Ki2Wn7KTLXfWlD0hRSUxSSVkyz1SKTe3UkrSMDsmy8wxZ+b3x+TcplmcYc6c65x5P//xmPO9zuXzmc+Zme/3c67zvQBKla/n4ihFpCQplubTwKGTOBkbl+fYqgXTruucK9dG88epeF6bMfq6nv/BR1tZv2krAFlZWdd1DhERZ/ky5uscjacrPtu2m1YtmuDj4+OCqERERMQd+IZUwewXgi09lTM/RxNSLYrUE9uxpadi9gvBN6SKq0MUkRKmWJpPy+ZNKnA83BJKXEISDetHAHA6PolwS2i+x7+zbjO79n7D/Bmj8Pf3A6B0SDBmLzPJKWftVz+djk8irELe53mg+x080P0OADIyMmjbrV/hkhIRcaITv8fm+filS5dISEyhcqWwYo5IRERE3IXZN4iyET3sVzqd+WlD9uNXXQklIlKcDLHnU8e2LYj+dAc2WybnUs+zfdc+otpH5nns2vWfsHVnDK++PIrgoMCc52nXnA2btwNw5OgvJCafoUkj3e1ORNxP2TKl83zcy8uLkBBNGEVERKRgZt8gQqpF5XgspFqUGk8i4hKG2POpa1Qbjvz0Kw/2H4HJZKLPfXdSu0b2paB7Yr5lz74DjBk2kITEZOYvXUPlihaeeSH743o+Pj4sn/8SAEMG9OGlVxbR64kR+Hh7M+mFp3SnOxFxS21aNmP/twex2Ww5Hm9ya4NcjXcRERGRv7NZz5N6YnuOx1JPbNeVTyLiEiarNb3Eb3h05WN3ezavVLNKRAzjyI8/E715G/EJifh4e9OsaWN63tNZ+z2JiEtoviTiPq7eXNzsF5Jrzyc1oESkuGnmICJiUPXrRlC/bgSpf57H388PX181nUREROTarKknczWart4Dypp6Une7E5FipeaTiIjBhQTrnUkRERFx3JXGkm9IFfsVTlcaUGo8iYgrqPkkIiIiIiLiYfJqMJl9g9R4EhGXMMTd7kRERERERERExDPpyicRERERD/L+xi1Ef7IDTGAymejbqxtdo9rYx1es2cjmz3cD0Kl9JE8+8aCrQhUREZESQs0nEREREQ9So1pllsydQFBgAPEJyTw+ZCwN6kVwU6Uwvjv0I1t3xLB68XTMXmYGD59Mw/oRtG5xm6vDFhEREQ+mj92JiIiIeJBmtzUgKDAAgDBLKKFly5CQmAzAtl376NqpNaX8/fH19eHuLu3YujPGleGKiIhICaDmk4iIiIiH2n/gMKnnL1CvTk0A4hOSCbeUt49XDKtAfEKyq8ITERGREkIfuxMRERFxIwOHTuJkbFyeY6sWTCPMEgrAseMnmTZ7KVPHPEMpf//r+r8++Ggr6zdtBSArK+v6AhYREZEST80nERERETeybN6kax5z/EQsIyfMYuzwgTRuUMf+eJgllLiEJPvXp+MT7c2qvDzQ/Q4e6H4HABkZGbTt1u+64xYREZGSSx+7ExEREfEgv/0ey/DxM3nx+QE0b9owx1hU2+Z8tm0vl9LSsFov8/GW3XRq39JFkYqIiEhJoSufRERERDzInIWruXDhIguXv8vC5e8C8PSAPkTe3ogmjesT1T6SvoNHAxDVPpI2kbrTnYiIiDiXmk8iIiIiHmT+jFEFjg/o25MBfXsWUzQiIiIiaj4B/9tAMyMjw8WRiIiIiNGZzWZMJpOrwyh2mi+JiIiIo/4+X1LzCbDZbAB06PFPF0ciIiIiRrdn80q8vUveFErzJREREXHU3+dLJqs1vcTfNzczMxOr1VqodzL7Pjmatxe/7OTIjEP5ejbl69mUr2dTvsWvpF75dD3zJXdmhNeaZFMtjEO1MA7VwjhUi7zpyqc8eHl54e/vX6jnmEymEvWup/L1bMrXsylfz6Z8pbhcz3zJnem1ZhyqhXGoFsahWhiHauEYL1cHICIiIiIiIiIinkvNp+t0/z13uDqEYqV8PZvy9WzK17MpXxHn0GvNOFQL41AtjEO1MA7VwjHa80lERERERERERJxGVz6JiIiIiIiIiIjTqPkkIiIiIiIiIiJOo+aTiIiIiIiIiIg4je4HmI/MzEzmLlpNzNcHAejdsyu9enTOdVy61cqE6Qs4/nssfr6+lC0Twshn+1GlcjgAKWfPMfmVxcSeTsDXx4f/e7YftzWsW6y5OMLRfAHmLHyLPfsOEBefxKqF07i5VjX7WM/HhuLr442fry8Aj/XuTqd/RDo/gUIqqnxPxsYxeeYSzqX+SVBgKcaNGEzN6jcVSw6FUZh8C8rJyPV1tBYffbaT1e9tIisri6aN6zPy2X72W6MWNGY0N5rvgYNHGDZuJtVuqmg/dum8Sfj7+RZnGg5zJN/TcYlMmb2En46doFJ4Bd5aND3HuKfVt6B8PbG+33z/XxYuf49LaWmYMNGqxa083b83Xl7Z76N9ue87Xlu2hszMTGrVqML4EYMIDAxwRTri5t7fuIXoT3aAKfv22X17daNrVBv7+Io1G9n8+W4AOrWP5MknHnRVqB5v71ffseyt9fx64g96doti2FOP2scKM7eRG+cuc15PlN9aRDUpfgWt/d1l3e9Kxpx1G8Bn2/dy/PdY3ls+i/MXLtJvyFiaNq6f5w90j7s60LJZY0wmE+uiP+fleW+wcOY4ABYuf48G9Wozb/qLHDn6C6Mmz+PDVXMNt+ApTL4d2janb69uDB4xJc9zTRnzbI4GjREVVb7/evVN7r2rA906t+OLPfuZOnsJb76W9/fFlQqT77VyMmp9HanFqbgElq36gJULplKubGlemDSHjZ/s4IHudxQ4ZkQ3mi9AtZsq5mrQGJUj+QYElGLw4704f+EiS1auyzHmifUtKF/wvPoGBwUyZcwzVK5oId1q5blRM/h025d069yOi5fSmD53GQtnjqN61UrMen0Vb67ZyLMDH3ZRRuLOalSrzJK5EwgKDCA+IZnHh4ylQb0IbqoUxneHfmTrjhhWL56O2cvM4OGTaVg/gtYtbnN12B6pSuVwxg4fyBd79nPxUlqOscLMbeTGucuc1xPltxZRTVwjv7W/u6z7XUkfu8vH9t376HFnB8xmL0qHBBHVPpKtO2NyHefn60ur5rdiMpkAaFCvNqfjk+zjX+z+ip7dogCoX6cW5cuV5cB/fiyeJArB0XwBbmtYF0uF0GKOsGgVRb4pZ8/xw8+/0iWqNQAd2jQjPjGFk7FxTo39ejiarzvldDVH4/5iz37aRDYhtFwZTCYTPbtF2b8PBY0ZTVHk604czbd0SBCNG9ShlL9frnO40/eiKPJ1J47mW6d2dSpXtADZf3sjalbjdHwiADFfH+TmWtWoXrUSAPff08mw9RXja3ZbA4L+umouzBJKaNkyJCQmA7Bt1z66dmpNKX9/fH19uLtLO73WnKjqTRWJqFUNszn3kqUwczm5Me46P/QUea1FVBPXKGjt7y7rfldS8ykfcQnJhFvK27+uGFaeuITkaz7v/Y1baNeyCQDnUv8kw2YjtFyZHOeJT7z2eYrb9eabl8kzF/PI4FFMm7OMM2dTiyrEIlUU+SYkplC+XBm8zWYg+9L8sAqhbl1fR3IyYn0drUV8QjLhYVd/HyoQ/9f3oaAxoymKfAFiTyfw+JCx9H92POs3bS2e4K9DUfyseWJ9r8WT65uccpYdX+63X22Su77lSU45S4bN5tzgxePtP3CY1PMXqFenJvDXa83iHr9LPF1Rzl2lYO405y0pVBNjuLL2d6d1vyuV2GvABg6dlG9neNWCadd1zpVro/njVDyvzRh9I6E5hTPyzcuiWeMIt5QnIyODJSs/YMqsJcyZOrLIzu+o4srXKEpafeXG1aldneh35hMUGEBCYjLDx8+idEgwndobYw8vuTGeXN8LFy4ycuJs+vbqRr2ba7o6HHFD1/qbGWbJvsLg2PGTTJu9lKljnqGUv39xhlhiOFoLERGjuXrtn261ujoct1Bim0/L5k0qcDzcEkpcQhIN60cAcDo+ifAC/gC+s24zu/Z+w/wZo/D/6yMQpUOCMXuZSU45a++Cno5PIswFH1kr6nzzP0/2O1De3t707tmF3gNc05gojnwtFcqR9Nc7695mM1lZWcQnJrt1fa+Vk1Hq+3eO1iLMEkrsqQT716fjE+0T24LGjKYo8r16I2ZLhVDu+EdLDh4+asjmRFH8rHlifQviqfW9cPESQ8fOpG3Lpjx0/132x8Msoew/cMj+9en4JEKvekdY5GrX+psJcPxELCMnzGLs8IE0blDH/njYX39PrzDy7xJ34Egt8lNUc1e5NiPNeSWbauJaf1/7+/v7GWbdb2T62F0+OrZtQfSnO7DZMjmXep7tu/YRlc+kfe36T9i6M4ZXXx5FcFBgzvO0a86GzdsBOHL0FxKTz9CkkfF2vS9Mvvm5lJbGn+cv2L/eujPGkBtTQ9HkW65MaerUrs6W7XsB2PHl11jKl7Pf6dBIHM23oJyMXF9Ha9GhTXO+3HeA5JSzZGVlsWHzdvtivKAxoymKfJOSz5CZmQlkL+j3fvUdN9euXqx5OKooftY8sb4F8cT6XryUxrCxrxB5eyOeePjeHGORtzfip2Mn+O33UwCs37SNTu1bFkv84nl++z2W4eNn8uLzA2jetGGOsai2zfls214upaVhtV7m4y279VpzkaKYy4lj3GnOW1KoJq6T39rfXdb9rmSyWtOzXB2EEdlsmcxZ9Bb7vj6IyWSiV4/O9O7ZFYA9Md+yZ98BxgwbSEJiMj36Pk/lihYCSmVfku3j48Py+S8BkHLmHC+9sohTcYn4eHszYsjjNL21vsvyyo+j+QLMeHU5/97/PSkp5wgJCSKglD8frJxD7OkERk95lczMTLKysqgcbmHYU49SMbyCK1PLU1HkC3Di5Cmmzl7KudTzBAaUYuyIQdSuUcVleeWnMPnml5PR65tf3NPnLqNtZBPatmwKQPQnO1j9/iYAbmtUjxefe8J+F4qCxozmRvNdF/05Gz7ejtlsxmaz0bFdcwb0vc++gaLROJJvWlo6Dw74Py5fzuD8hYuULRNC16g2PN2/N+B59S0oX0+s78o10bzx9ofUrFbZ/ryObVvQ7+EeQPbvstffeBebzUat6lUYP3KwfdNokcJ4btQMfvz51xz7CT09oA+RtzcCYPnbG/hk624AotpH2n/HSNH7+rvDTJm1hAsXL0EWBAaWYuQz/WjbsmmBcxspeu4y5/VE+a1FVJPiV9Da313W/a6k5pOIiIiIiIiIiDiNPnYnIiIiIiIiIiJOo+aTiIiIiIiIiIg4jZpPIiIiIiIiIiLiNGo+iYiIiIiIiIiI06j5JCIiIiIiIiIiTqPmk4iIiIiIiIiIOI2aTyIiIiIiIiIi4jRqPomIiIiIiIiIiNN4uzoAEZHi9PTIqRz+4RjeZjPe3t7UrlmF5wY9Qt2IGqSlW1m1Npptu/aRlHyW4OAAGtStzWN9ulM3ogbffn+EN9/ZwNFjv2Eymdj64VJXpyMiIiLikJ6PDWXok31p3+r2fI/JsNlYtuoDPt8Zw7lzfxIQ4E+d2jWYPHoIgQGl2Pz5bqbOXspD99/Fc4Metj/vxUlziahVlX8+ej8ALbv0xc/PFy+TyX5MsyYN+NfEYbn+z8uXM5g4YwE//HycuPgkZkwcWmCMIuKe1HwSkRLn6f596HNfV9KtVha9+T4vTJrDh6vmMnT0DLKAqWOeoVaNKlzOyGD3vw+wa+831I2ogb+/H3d3aU+XqNa8tnSNq9MQERERKVKr39vE/gOHWfDKGCqFW0g5e469X32f45jgoEA2fLyd3vd2IcwSmu+5ls6dyM21qjn0/za6pQ4P3tuFiTMW3kj4ImJgaj6JSInl5+vLPV3b896Gz9iy49/8dvIU61bMJjgoEABvb2+6dGxlP/6WurW4pW4tDhw84qqQRURERAptzNT5xCckM+HlBZi9vOjSsTUvPt8/13GHfzhG25ZNqBRuAaBcmdLc06V9jmPCLaHUqlGVZavXM27EoBuOzcfHmz73dQXAy6xdYUQ8lX66RaTESktL56NPdxIeVp6vvjlEy2aN7Y0nEREREU8xfdxzhFlCmTx6CF9EL8+z8QTQ6JabWb9pG+9++Bk//PQrGTZbnscNeux+tu/6iuMnYp0Ztoh4EF35JCIlzqIV77H87Q/x9fUhomZVZk4azvyla6gbUd3VoYmIiIi4zKMP3k3Z0sF8vjOGpavWYTab6dmtI4P7PYj5qquSKoZXoMddHVi84n3+NSn3Pk4AT42YgpfX/57T576uDOh7n9NzEBFjUvNJREqcp57obb+8+4oypYNITD7joohEREREitfKtdG89e5HADRuUIe5017Ay8uL7nd2oPudHciw2dj/7SEmzlhIpYoW7r2rY47n93uoBw/0G86hIz/nef5Fs8c7vOeTiHg+NZ9ERIAWTRvx+htrOX/hIkGBAa4OR0RERKRIXX3nOchuHvV7qEe+x3ubzbRqfiu333oLvxw/mWu8TOlgHunVjYXL3yUkOKjI4xURz6I9n0REgC4dW1HtpoqMnDCbo8d+I8NmIy3dyrZd+1iyah0AmZmZpFutXM7I3v8g3Wol3Wp1ZdgiIiIiDilXtjSxpxIKPGbth5+y/8BhLl5KIysri4P//YkD//mBhvUj8jy+z31dOXkqjoP/PXpDsVmtl7PnVFlZZGTYSLdasdkyb+icImIsuvJJRITsO9vNm/4iK9dGM3bqfJJSzhISHEiDurV5vE93AL4/9CNDXphuf84/7snerDNmy9suiVlERETEUY/16c7chatZsWYjnTu0ZOSzT+Q6ppS/H4tXvM/vf5wGoEL5svR/5F46d2iV69js4/3p/3BPZr6+MtfYoGEv5bjaqnrVyrz52uQ8z9P7nyOJi08CYNy017L/HTGIbp3bFSpHETEuk9WanuXqIERERERERERExDPpY3ciIiIiIiIiIuI0aj6JiIiIiIiIiIjTqPkkIiIiIiIiIiJOo+aTiIiIiIiIiIg4jZpPIiIiIiIiIiLiNGo+iYiIiIiIiIiI0/w/Qnallvy1KlcAAAAASUVORK5CYII=",
"text/plain": [
"<Figure size 1200x500 with 2 Axes>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"fig = plot_fidelity_comparison(\n",
" holdout_windows,\n",
" synthetic_holdout,\n",
" title=\"Sig-CWGAN: Real vs Synthetic Distribution\",\n",
" n_samples=min(500, len(holdout_windows)),\n",
" flatten_method=\"flatten\", # Flatten entire path for comparison\n",
")\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"id": "a72451ab",
"metadata": {
"papermill": {
"duration": 0.008753,
"end_time": "2026-06-14T16:58:54.273190+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.264437+00:00",
"status": "completed"
}
},
"source": [
"**Interpretation**: Overlapping point clouds confirm that synthetic paths occupy\n",
"the same region of feature space as real data. Sig-CWGAN's strength is capturing\n",
"path-wise properties via signatures -- the signature-based metrics below provide\n",
"a more rigorous assessment than visual inspection alone."
]
},
{
"cell_type": "markdown",
"id": "13111a91",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.008548,
"end_time": "2026-06-14T16:58:54.290394+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.281846+00:00",
"status": "completed"
}
},
"source": [
"### 10.2 Compare Signature Distributions\n",
"\n",
"**Paper Evaluation**: Compare expected signatures of real vs synthetic paths.\n",
"This is the same metric used for training (Sig-W1)."
]
},
{
"cell_type": "code",
"execution_count": 39,
"id": "88aa4c59",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:54.308712Z",
"iopub.status.busy": "2026-06-14T16:58:54.308458Z",
"iopub.status.idle": "2026-06-14T16:58:54.311965Z",
"shell.execute_reply": "2026-06-14T16:58:54.311643Z"
},
"papermill": {
"duration": 0.013469,
"end_time": "2026-06-14T16:58:54.312434+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.298965+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def evaluate_sig_w1(\n",
" real_paths: np.ndarray,\n",
" synthetic_paths: np.ndarray,\n",
" depth: int,\n",
" config: dict,\n",
" normalise: bool = True,\n",
") -> dict:\n",
" \"\"\"\n",
" Evaluate using paper's Sig-W1 metric.\n",
"\n",
" Computes:\n",
" - Sig-W1 distance (expected signature RMSE)\n",
" - Per-coordinate analysis\n",
" - Marginal statistics comparison\n",
" \"\"\"\n",
" # Compute expected signatures using paper pipeline\n",
" print(\"Computing real expected signature...\")\n",
" real_sigs = compute_signatures_batch_paper(\n",
" real_paths, depth, config, normalise=normalise, desc=\"Real\"\n",
" )\n",
" expected_sig_real = real_sigs.mean(axis=0)\n",
"\n",
" print(\"Computing synthetic expected signature...\")\n",
" syn_sigs = compute_signatures_batch_paper(\n",
" synthetic_paths, depth, config, normalise=normalise, desc=\"Synthetic\"\n",
" )\n",
" expected_sig_syn = syn_sigs.mean(axis=0)\n",
"\n",
" # Sig-W1 distance (paper metric)\n",
" sig_w1 = np.sqrt(np.sum((expected_sig_real - expected_sig_syn) ** 2))\n",
"\n",
" # Per-coordinate analysis\n",
" diff = expected_sig_real - expected_sig_syn\n",
" per_coord_mae = np.abs(diff).mean()\n",
" per_coord_max = np.abs(diff).max()\n",
"\n",
" # Marginal statistics\n",
" real_mean = real_paths.mean()\n",
" syn_mean = synthetic_paths.mean()\n",
" real_std = real_paths.std()\n",
" syn_std = synthetic_paths.std()\n",
"\n",
" # KS statistic on marginals\n",
" real_flat = real_paths.flatten()\n",
" syn_flat = synthetic_paths.flatten()\n",
" ks_stat, _ = stats.ks_2samp(real_flat, syn_flat)\n",
"\n",
" return {\n",
" \"sig_w1\": sig_w1,\n",
" \"per_coord_mae\": per_coord_mae,\n",
" \"per_coord_max\": per_coord_max,\n",
" \"sig_dim\": len(expected_sig_real),\n",
" \"ks_statistic\": ks_stat,\n",
" \"real_mean\": real_mean,\n",
" \"syn_mean\": syn_mean,\n",
" \"real_std\": real_std,\n",
" \"syn_std\": syn_std,\n",
" }"
]
},
{
"cell_type": "code",
"execution_count": 40,
"id": "4a342a53",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:54.330510Z",
"iopub.status.busy": "2026-06-14T16:58:54.330372Z",
"iopub.status.idle": "2026-06-14T16:58:54.461229Z",
"shell.execute_reply": "2026-06-14T16:58:54.460821Z"
},
"papermill": {
"duration": 0.140387,
"end_time": "2026-06-14T16:58:54.461580+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.321193+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"=== Sig-W1 Evaluation (Holdout Data) ===\n",
"Computing real expected signature...\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Real: 0%| | 0/592 [00:00<?, ?it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Real: 100%|██████████| 592/592 [00:00<00:00, 9854.18it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Computing synthetic expected signature...\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Synthetic: 0%| | 0/592 [00:00<?, ?it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Synthetic: 100%|██████████| 592/592 [00:00<00:00, 9991.78it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"--- Results ---\n",
"Sig-W1 Distance: 0.4200\n",
" (Paper reference: ~2.76 on S&P 500 log returns)\n",
"Signature dimension: 780\n",
"Per-coordinate MAE: 0.004960\n",
"Per-coordinate Max Error: 0.122825\n",
"\n",
"Marginal Statistics:\n",
" Real mean: 0.000109, std: 0.015508\n",
" Synth mean: 0.000530, std: 0.014002\n",
" KS statistic: 0.1172\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# Evaluate on holdout data\n",
"print(\"\\n=== Sig-W1 Evaluation (Holdout Data) ===\")\n",
"eval_results = evaluate_sig_w1(\n",
" holdout_windows,\n",
" synthetic_holdout,\n",
" CONFIG[\"sig_depth\"],\n",
" CONFIG,\n",
" normalise=CONFIG[\"normalise_sig\"],\n",
")\n",
"\n",
"print(\"\\n--- Results ---\")\n",
"print(f\"Sig-W1 Distance: {eval_results['sig_w1']:.4f}\")\n",
"print(\" (Paper reference: ~2.76 on S&P 500 log returns)\")\n",
"print(f\"Signature dimension: {eval_results['sig_dim']}\")\n",
"print(f\"Per-coordinate MAE: {eval_results['per_coord_mae']:.6f}\")\n",
"print(f\"Per-coordinate Max Error: {eval_results['per_coord_max']:.6f}\")\n",
"print(\"\\nMarginal Statistics:\")\n",
"print(f\" Real mean: {eval_results['real_mean']:.6f}, std: {eval_results['real_std']:.6f}\")\n",
"print(f\" Synth mean: {eval_results['syn_mean']:.6f}, std: {eval_results['syn_std']:.6f}\")\n",
"print(f\" KS statistic: {eval_results['ks_statistic']:.4f}\")"
]
},
{
"cell_type": "markdown",
"id": "8adf105f",
"metadata": {
"papermill": {
"duration": 0.009068,
"end_time": "2026-06-14T16:58:54.481575+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.472507+00:00",
"status": "completed"
}
},
"source": [
"### Signature Analysis: Why Factorial Normalization Matters\n",
"\n",
"Without factorial normalization, signature levels have vastly different scales:\n",
"- Level 1: O(1) magnitude\n",
"- Level 2: O(1/2!) magnitude\n",
"- Level 3: O(1/3!) magnitude\n",
"- Level 4: O(1/4!) magnitude\n",
"\n",
"The paper multiplies each level by k! to bring them to comparable scales."
]
},
{
"cell_type": "code",
"execution_count": 41,
"id": "2fdf2c98",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:54.501346Z",
"iopub.status.busy": "2026-06-14T16:58:54.501210Z",
"iopub.status.idle": "2026-06-14T16:58:54.528098Z",
"shell.execute_reply": "2026-06-14T16:58:54.527600Z"
},
"papermill": {
"duration": 0.037076,
"end_time": "2026-06-14T16:58:54.528395+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.491319+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"=== Signature Level Analysis ===\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Sample: 0%| | 0/100 [00:00<?, ?it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Sample: 100%|██████████| 100/100 [00:00<00:00, 9374.84it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Sample: 0%| | 0/100 [00:00<?, ?it/s]"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\r",
"Sample: 100%|██████████| 100/100 [00:00<00:00, 9621.51it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Path dimension after augmentations: 5\n",
"\n",
"Level 1 (5 dims):\n",
" Real magnitude: 0.6064\n",
" Synth magnitude: 0.6095\n",
" Mean error: 0.0003\n",
" Factorial factor: 1\n",
"\n",
"Level 2 (25 dims):\n",
" Real magnitude: 0.3743\n",
" Synth magnitude: 0.3754\n",
" Mean error: 0.0011\n",
" Factorial factor: 2\n",
"\n",
"Level 3 (125 dims):\n",
" Real magnitude: 0.2353\n",
" Synth magnitude: 0.2334\n",
" Mean error: 0.0017\n",
" Factorial factor: 6\n",
"\n",
"Level 4 (625 dims):\n",
" Real magnitude: 0.1494\n",
" Synth magnitude: 0.1463\n",
" Mean error: 0.0020\n",
" Factorial factor: 24\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# Diagnostic: Signature level analysis\n",
"print(\"\\n=== Signature Level Analysis ===\")\n",
"\n",
"# Get expected signatures (already computed)\n",
"real_sigs_sample = compute_signatures_batch_paper(\n",
" holdout_windows[:100],\n",
" CONFIG[\"sig_depth\"],\n",
" CONFIG,\n",
" normalise=CONFIG[\"normalise_sig\"],\n",
" desc=\"Sample\",\n",
")\n",
"syn_sigs_sample = compute_signatures_batch_paper(\n",
" synthetic_holdout[:100],\n",
" CONFIG[\"sig_depth\"],\n",
" CONFIG,\n",
" normalise=CONFIG[\"normalise_sig\"],\n",
" desc=\"Sample\",\n",
")\n",
"\n",
"# Analyze by signature level\n",
"path_dim = compute_path_dim_paper(n_assets)\n",
"print(f\"Path dimension after augmentations: {path_dim}\")\n",
"\n",
"count = 0\n",
"for level in range(1, CONFIG[\"sig_depth\"] + 1):\n",
" level_size = path_dim**level\n",
" level_real = real_sigs_sample[:, count : count + level_size]\n",
" level_syn = syn_sigs_sample[:, count : count + level_size]\n",
"\n",
" real_mag = np.abs(level_real).mean()\n",
" syn_mag = np.abs(level_syn).mean()\n",
" level_error = np.abs(level_real.mean(0) - level_syn.mean(0)).mean()\n",
"\n",
" print(f\"\\nLevel {level} ({level_size} dims):\")\n",
" print(f\" Real magnitude: {real_mag:.4f}\")\n",
" print(f\" Synth magnitude: {syn_mag:.4f}\")\n",
" print(f\" Mean error: {level_error:.4f}\")\n",
" print(f\" Factorial factor: {math.factorial(level)}\")\n",
"\n",
" count += level_size"
]
},
{
"cell_type": "markdown",
"id": "176e66c7",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.00913,
"end_time": "2026-06-14T16:58:54.547283+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.538153+00:00",
"status": "completed"
}
},
"source": [
"## 11. Visualization: Real vs Synthetic Paths\n",
"\n",
"Compare the distributional properties of real and synthetic paths."
]
},
{
"cell_type": "code",
"execution_count": 42,
"id": "36079ac7",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:54.566385Z",
"iopub.status.busy": "2026-06-14T16:58:54.566244Z",
"iopub.status.idle": "2026-06-14T16:58:54.570430Z",
"shell.execute_reply": "2026-06-14T16:58:54.570074Z"
},
"papermill": {
"duration": 0.014769,
"end_time": "2026-06-14T16:58:54.571055+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.556286+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def plot_path_comparison_unconditional(\n",
" real_paths: np.ndarray,\n",
" synthetic_paths: np.ndarray,\n",
" n_samples: int = 20,\n",
" asset_idx: int = 0,\n",
") -> go.Figure:\n",
" \"\"\"Compare distributions of real and synthetic paths.\"\"\"\n",
" fig = make_subplots(\n",
" rows=1,\n",
" cols=2,\n",
" subplot_titles=[\"Real Paths\", \"Synthetic Paths\"],\n",
" shared_yaxes=True,\n",
" )\n",
"\n",
" # Convert to cumsum for visualization\n",
" n_lags = real_paths.shape[1]\n",
"\n",
" real_cumsums = np.stack(\n",
" [real_paths[i, :, asset_idx].cumsum() for i in range(min(n_samples, len(real_paths)))]\n",
" )\n",
" synth_cumsums = np.stack(\n",
" [\n",
" synthetic_paths[i, :, asset_idx].cumsum()\n",
" for i in range(min(n_samples, len(synthetic_paths)))\n",
" ]\n",
" )\n",
"\n",
" # Plot real paths\n",
" for i, cumsum_path in enumerate(real_cumsums):\n",
" fig.add_trace(\n",
" go.Scatter(\n",
" x=list(range(n_lags)),\n",
" y=cumsum_path,\n",
" mode=\"lines\",\n",
" line=dict(color=COLORS[\"blue\"], width=0.5),\n",
" opacity=0.3,\n",
" showlegend=(i == 0),\n",
" name=\"Real\" if i == 0 else None,\n",
" ),\n",
" row=1,\n",
" col=1,\n",
" )\n",
"\n",
" # Plot synthetic paths\n",
" for i, cumsum_path in enumerate(synth_cumsums):\n",
" fig.add_trace(\n",
" go.Scatter(\n",
" x=list(range(n_lags)),\n",
" y=cumsum_path,\n",
" mode=\"lines\",\n",
" line=dict(color=COLORS[\"copper\"], width=0.5),\n",
" opacity=0.3,\n",
" showlegend=(i == 0),\n",
" name=\"Synthetic\" if i == 0 else None,\n",
" ),\n",
" row=1,\n",
" col=2,\n",
" )\n",
"\n",
" # Shared y-axis range so visual spread is comparable across panels\n",
" y_lo = float(min(real_cumsums.min(), synth_cumsums.min()))\n",
" y_hi = float(max(real_cumsums.max(), synth_cumsums.max()))\n",
" pad = 0.05 * (y_hi - y_lo if y_hi > y_lo else 1.0)\n",
" shared_range = [y_lo - pad, y_hi + pad]\n",
"\n",
" fig.update_layout(\n",
" title=f\"Sig-WGAN: Path Comparison (Asset {asset_idx})\",\n",
" template=\"ml4t\",\n",
" height=400,\n",
" yaxis=dict(range=shared_range),\n",
" yaxis2=dict(range=shared_range),\n",
" )\n",
" fig.update_xaxes(title_text=\"Day\", row=1, col=1)\n",
" fig.update_xaxes(title_text=\"Day\", row=1, col=2)\n",
" fig.update_yaxes(title_text=\"Cumulative Return\", row=1, col=1)\n",
"\n",
" return fig"
]
},
{
"cell_type": "code",
"execution_count": 43,
"id": "31a7d47d",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:54.591096Z",
"iopub.status.busy": "2026-06-14T16:58:54.590959Z",
"iopub.status.idle": "2026-06-14T16:58:54.740599Z",
"shell.execute_reply": "2026-06-14T16:58:54.740178Z"
},
"papermill": {
"duration": 0.160544,
"end_time": "2026-06-14T16:58:54.741367+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.580823+00:00",
"status": "completed"
}
},
"outputs": [
{
"data": {
"application/vnd.plotly.v1+json": {
"config": {
"plotlyServerURL": "https://plot.ly"
},
"data": [
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"name": "Real",
"opacity": 0.3,
"showlegend": true,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "G3sHPDL8bzzX7Zg8J1nSPGf03zwCn+o8P4HhPBd2DT3PAyk9nY0aPcbvQD2BTzo9sjtMPV4kbT3pCXY9g75zPQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "LQLRO5JgKjyZm4482TacPHThpjyxw508oC7XPAglBz2sXfE8/xAfPbpwGD3rXCo9l0VLPSIrVD2831E9dFdUPQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "+L6DOxy2NDyb7E880kFlPEwGUzwV7qI8hQnaPCEdvTy68AQ966D8PKc8ED1TJTE93go6PXi/Nz0wNzo9mm5qPQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "P63lOx8NDjxWYiM80CYRPFf+gTzHGbk8Yy2cPLbx6Dwtsds8kIn/PHStID3/kik9mUcnPVG/KT279lk96Es+PQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "+rPZOtkuQjuAgfI6DiYRPO5cfzwmhEU8ZoavPN1FojxAHsY8zPcDPVfdDD3xkQo9qQkNPRNBPT1AliE9ombpPA==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "uKmqOjBsRjkc3+s7biZkPKZNKjwm66E8naqUPACDuDxYVPo8tw8GPVHEAz0JPAY9c3M2PaDIGj1iy9s8hMzfPA==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "MtyRuq40wTs40U48cPgUPIpAlzwBAIo8ZNitPLyp7zxpugA9Bt78PLvmAD0lHjE9UnMVPcYg0TzoIdU8YtLPPA==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "u6vlO74MYTz2Myc8Tl6gPMUdkzwo9rY8gMf4PEtJBT3l/QI9nXUFPQetNT00Aho9ij7aPKw/3jwm8Ng8/FClOw==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "wW3cO2J4UTu+5k08rGUzPHEWezyQXL88pyfRPNuQzDxLgNE8kPcYPXuZ+jyd06A8v9SkPDmFnzxwtQC7cm4zvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "IGNnu7pfvzuWXYo7kN8MPCBBiDw3DJo8a3WVPNtkmjyv0/o8Cn7DPFhwUzybcls8kNNQPD5kDrwq/E69qikIvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "pYgZPCYP/jtYuEY8hC2lPJv4tjzPYbI8P1G3PArgCz1vauA8kaSGPLOlijwtVoU86Bapu/iFQL3xZvO8Lz0OvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "kgjUusu+NDti0jA8kGhUPPc6SzzWGVU8v/vKPBqmkzzwgOc7d4X3O2BH4jseFG68IuhmvaIVIL1YnzS9dKqovQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "imGPO3RTSzyi6W48CbxlPOiabzxIPNg8o+agPIpBDjzNQxY8wqQLPAyTU7zeR2C9XnUZvRT/Lb1SWqW9bP2GvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "r6IDPN04JzxECx48I+onPOZjtDyCHHo8jCGNOxMmnTv854c76KGNvA80cr2QYSu9Rus/vWpQrr2E84+9F01nvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "t1gOO6dE0zrQHRE7HCVlPKTz7DuoR3S7mj5Uu8e6frtAc8+8Xo6JvTxKTL3y02C9wMS+vdpnoL3iGoS9doV9vQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "jtkSulBGMTjujkE8SMelOzBQwbupS7G7wInGu1c+4bwjAY69xi9VvXy5ab2GN8O9oNqkvaiNiL2BNYO91eJPvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "8+0dOoe8Sjx6Irg7/vSuu3fwnruOLrS7iqfcvHDbjL1g5FK9Fm5nvdIRwr3stKO99GeHvc0Pgr1tl0291XMcvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "qN1APLxkpDu8ssK7Na6yu0zsx7v6luG8TBeOvRhcVb3O5Wm9rk3DvcjwpL3Qo4i9qUuDvSUPUL2N6x69BWQdvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "lFbdu4MbkbxhGo2852mSvOcCIb0BM6a9wcmCvZwOjb1kadu9fgy9vYa/oL1fZ5u9SCOAvfgiT71wm029jpplvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "vIszvHmJK7yEKDa8FFgFvZhdmL2w6Gm9ZnJ+vfqTzb0UN6+9HOqSvfWRjb29m2S9JXgzvZ3wMb2670m9J4NgvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "NyQAOvgxJ7lL6rC8IOyBvcAFPb12j1G9giK3vZzFmL1H8Xi9+UBuvcy4N700lQa9rA0FvcoMHb03oDO906MvvQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "tfApum3rtLxp7IK9UgY/vQiQU73MIri95sWZvdvxer2NQXC9YLk5vciVCL1ADge9Xg0fvcugNb1npDG9egXhvA==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "55uvvIeYgb2OXjy9ROhQverOtr0Ecpi9F0p4vcmZbb2cETe9BO4FvXxmBL2aZRy9B/kyvaP8Lr3ytdu8Get3vA==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "G2MrvTchybykNPK88OeKvRQWWb0jfCC91csVvVCHvrw/gDi8IGIyvEsvibwlVra8XV2uvCRosLtymc47fHLUuw==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "/6SNPCQjSTyM2dS86cs2vGhvLjsluqw74z6YPBSG+jwklf086ZbNPA9woDzXaKg8FlYVPUk2RT3L1BA97kHGPA==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "s02ku0U/Mb3yCum8Ia5vvOjsRLxgnqk6LcJZPEzgXzysx/87iFgWO8oeVjstB508ksf8PJYElDy6c+I7ShjVuw==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "j7UcvYb3v7xJhx28IYzlu0e1zjuC9JU8kgOZPK4KUjz0ee87i64HPJoaxjyA7RI9BBi9PLlgQzxIKsO6tPJpOw==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "L+dyPHmn6jwLBAA9OIw2PdCvZz1YN2k9OjhRPc2kOj0xoT4928J/PYfRlz2QQXs9vI1NPTycFj26VCs9U0VYPQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "w2diPH6UhjzYpPM8BPYqPYx9LD1ufhQ9Atb7PGXnAT0PCUM9QulyPcSHPj3w0xA938SzPNo13TyGixs921YmPQ==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#0a1628",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x",
"y": {
"bdata": "4gQrO/ZwgjwmuOQ8NsfnPPvItzwhooo86ZqSPB5vCj1RTzo90+0FPf5zsDz6IQU88ANYPCrjxTzUeds84ILXPA==",
"dtype": "f4"
},
"yaxis": "y"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"name": "Synthetic",
"opacity": 0.3,
"showlegend": true,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOVwJ+rk5BRq8UGcMOs5tvDuNXgC8nEQFvHDrYbuQZ4S5JhZQvGCDgbyx6dy8jxb7vIYGYL2fpX294pSMvQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOVcU1zpB0O46igbxux6ut7zK5L+8k88PvbUcN706QCm9LpJOvepnlr18is+9BTLPvRyIyL3UBQK+ykYcvg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOWIQzDtMz0e7osyUvEA1pbyNCPO8tIfyvCwO4bxkYuy7aKoKvJClbLxbq+y8JlmWvJAOBL3+efO8rx/QvA==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOevPSLyCwgu817O7vADMmDgmq6m8AN/vuuMRo7wCMpa8VOxcvDKvybtldGq8jlOgu3p2BTya7yA8NrSFPA==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOZ4+kjrkhhE8JrRrPMcBPDzYLU88IBnaO7t0Bz0y1A09kuvRPPBlgTw1hQA8Ak0ZPBtXGTxka107aokQPA==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOSv0lDuIBgs8AIWHPE9h7TzIM+Y8effaPI5/Cj1mzKk8cHbFPJ8X7jwSNgo8uBEOu6KjfryeWha92o7hvA==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOWVvlDyAFz09ckv7PPPHCz3gIVU9toI2PZ5Ecz3gaXI9p5GKPcAmtj1s+/M9uuLhPb/P0z0cZQU+od4TPg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOR+21LxQM6i8S6OcvDJQpLx2VsC8eRk4vcqbe72bF569MgDcvSHO7b3299+983oNvkq3Hr4LSRy+UUYzvg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOQ60XzwNrIg8DiMcPWwOCz1hlzQ9xUeiPJNiKjxQz1k8MKWpOrepMLyv9gy8qNmpvE3TtbxIHDe97t8NvQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOcsiuzrgrUa86PQ1PJzADLw/4i28TGykvJuyErwyu9O8JmIHvTULEr3e7Bu9ArfTvIr7FbyFrem8u6jJvA==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOaT29zvc7rE6JgAiPJX5hjy4izE9cpbVPAxSFT20CHU93gu8PS5F3T0WleY9ev0APqiBAD52NQ8+J+AbPg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOUkzrDxymiM95jkLPazAVzxY2JI843ejPFgbijs8dew7Cli9OyLcMTy8E1885ATKPF4kkTywWyQ7qiXHPA==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOQkDzbpXAxm89oaFvCjbm7vhbaW8OAaFvOsPwLwj36u87tyivIZNJ7z3BMy8RGG6vLGbW7zELy67WKphOg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOeoOTTsQun88GTCePMoOjTwcG8Q8SAy/O4TIaLvp9/G6bcYXPHX2GjwvwbU7YeehPBq0yzzCAiU9OUxCPQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOR8n5zsuO4084VYCPAxg2zxqbD09DnFrPdV7Qj3eRgQ9JAIyPS7McD1+gV49htGSPYx7lT233Kk9AF3DPQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOZms2Dxapv08Xp4rPcstdz1ag5g9l2aXPdkyzz00Cuc9N6HsPd6ABz7v6Rk+dFQvPvTyOD4nwDk+PnA/Pg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOcTKLzyM2T2741MhPPhOQDxyJfQ7xxj2O+mCGz3+NH09Hi1iPSyDsT20b8w9fT/LPUyBsT1Ces49mGzdPQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOVtqRrzrKgo88Wx9vD0Xk7zjIYe8JL7AvJ5s6Lz5LQy9DbRUvddpgL1gzJG9e5qdvdrUvb0HJr29NknCvQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOYItz7vmy+S8Lh0MvWA7lLyVuS28aMw/vYzSF72cuVS9WRF+vUVXlL24Eoa9JRO0vV7AqL2ok6y9JSe5vQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOUXcPjyQLQy6csG0PLtD+zxjFew8+mCzPAZF0DxY/EU9TV5HPYDppD255pk9OUbAPQIywz0OouU9BcoGPg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOTGBPLzshvu8yeclvWGYlr3AdqS9JPOUvZC/wb3iRvC959YJvqR9BL7m9Au+42sPvnXnFr4wICq+FWhDvg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVObmYMTwIl+i6Vq20u4i9H7xAH7m73EPiO/xFZbs1evS6yEfLuixDCr0XeLW85mGMvPiBKbvDX5I85qDFPA==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOX6YmDnOzIE80oI4PLpeyDyf44k8hmN3PASYPzzrF088gGM7O1wccbysV+68XXpDve2U9bxM0Bu9HCYUvQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOQQPAzxe+6+8oGyEOxRHgDyUjnY8CZjFPJLaIz0GzoM9UAxXPX5jKD0UfEk9kGYNPc5gVj3+tYM9DWaMPQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVORCy5Ds0WCs9fg83PS50Nz21s0k9AJBYPajFJT2XzDI9R5RKPVyILT1Cc/I8NYTiPDejQj160C89jPBXPQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOZQP3Lsot5a8p4hgvJJArLzigMW8MCMFvXyXAL1gWii9imHOvLQdbr3oKXi9svBpvSnidL2JaUW9ZDwkvQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOQgjlLveG0W8amqwvMW23rzjuBC9moRBvTJMbL3ZXz69PnCJvRdal73iPsC9Ml25vfuFt73dI9y9Z5HlvQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOU0HPrxIEHO8FmDUu39n9Lu+61m7z2PeO0jcibyMnrG7unORvOUOVLxtRTi8Q1MevRvGBL1YPuq8VRpfvQ==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVObA7yLv5y6C82G/ouyT3cjxgJms67qJUOrlC27yu6D+9iuBxvdbKkb27IaC9+GzEvaGI3b2Kqu+9SZcBvg==",
"dtype": "f4"
},
"yaxis": "y2"
},
{
"line": {
"color": "#C87533",
"width": 0.5
},
"mode": "lines",
"opacity": 0.3,
"showlegend": false,
"type": "scatter",
"x": [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15
],
"xaxis": "x2",
"y": {
"bdata": "XOvVOap8ULuvB8285sspvV+EJ72Jen297Dukvauusb1L4e+9kT0Lvqy8Hr5SgCm+isZHvv9GVb5dr2q+9nJzvg==",
"dtype": "f4"
},
"yaxis": "y2"
}
],
"layout": {
"annotations": [
{
"font": {
"size": 16
},
"showarrow": false,
"text": "Real Paths",
"x": 0.225,
"xanchor": "center",
"xref": "paper",
"y": 1.0,
"yanchor": "bottom",
"yref": "paper"
},
{
"font": {
"size": 16
},
"showarrow": false,
"text": "Synthetic Paths",
"x": 0.775,
"xanchor": "center",
"xref": "paper",
"y": 1.0,
"yanchor": "bottom",
"yref": "paper"
}
],
"height": 400,
"template": {
"layout": {
"colorway": [
"#0a1628",
"#D4A84B",
"#1a2d4a",
"#C87533",
"#10b981",
"#ef4444"
],
"font": {
"color": "#334155",
"family": "DM Sans, DejaVu Sans, sans-serif",
"size": 11
},
"hoverlabel": {
"bgcolor": "white",
"font": {
"family": "DM Sans, DejaVu Sans, sans-serif",
"size": 11
}
},
"legend": {
"bgcolor": "rgba(255,255,255,0.8)",
"bordercolor": "#e8e8e6",
"borderwidth": 1,
"font": {
"size": 10
}
},
"paper_bgcolor": "#FAFAF9",
"plot_bgcolor": "white",
"title": {
"font": {
"color": "#0a1628",
"size": 14
},
"x": 0.5,
"xanchor": "center"
},
"xaxis": {
"gridcolor": "#e8e8e6",
"gridwidth": 0.5,
"linecolor": "#e8e8e6",
"showgrid": true,
"tickfont": {
"size": 10
},
"title": {
"font": {
"size": 11
}
}
},
"yaxis": {
"gridcolor": "#e8e8e6",
"gridwidth": 0.5,
"linecolor": "#e8e8e6",
"showgrid": true,
"tickfont": {
"size": 10
},
"title": {
"font": {
"size": 11
}
}
}
}
},
"title": {
"text": "Sig-WGAN: Path Comparison (Asset 0)"
},
"xaxis": {
"anchor": "y",
"domain": [
0.0,
0.45
],
"title": {
"text": "Day"
}
},
"xaxis2": {
"anchor": "y2",
"domain": [
0.55,
1.0
],
"title": {
"text": "Day"
}
},
"yaxis": {
"anchor": "x",
"domain": [
0.0,
1.0
],
"range": [
-0.25897797048091886,
0.20818634927272797
],
"title": {
"text": "Cumulative Return"
}
},
"yaxis2": {
"anchor": "x2",
"domain": [
0.0,
1.0
],
"matches": "y",
"range": [
-0.25897797048091886,
0.20818634927272797
],
"showticklabels": false
}
}
},
"image/png": "iVBORw0KGgoAAAANSUhEUgAAArwAAAGQCAYAAABMPLOTAAAgAElEQVR4XuydB3wUVdeH/9mWXghNuhV7712xvjZsWECxoaCiUlQURaoIqAgiKqiA2MUK9obYey9gBwSpIT2bre/vfyazbMImmUndbM58X95IcnfmznNnN8+cOffcJJ+vPAzdlIASUAJKQAkoASWgBJRAghJIUuFN0JHV01ICSkAJKAEloASUgBIQAiq8eiEoASWgBJSAElACSkAJJDQBFd6EHl49OSWgBJSAElACSkAJKAEVXr0GlIASUAJKQAkoASWgBBKagApvQg+vnpwSUAJKQAkoASWgBJSACq9eA0pACSgBJaAElIASUAIJTUCFN6GHV09OCSgBJaAElIASUAJKQIVXrwEloASUgBJQAkpACSiBhCagwpvQw6snpwSUgBJQAkpACSgBJaDCq9eAElACSkAJKAEloASUQEITUOFN6OHVk1MCSkAJKAEloASUgBJQ4dVrQAkoASWgBJSAElACSiChCajwJvTw6skpASWgBJSAElACSkAJqPDqNaAElIASUAJKQAkoASWQ0ARUeBN6ePXklIASUAJKQAkoASWgBFR49RpQAkpACSgBJaAElIASSGgCKrwJPbx6ckpACSgBJaAElIASUAIqvHoNKAEloASUgBJQAkpACSQ0ARXehB5ePTkloASUgBJQAkpACSgBFV69BpSAElACSkAJKAEloAQSmoAKb0IPr56cElACSkAJKAEloASUgAqvXgNKQAkoASWgBJSAElACCU1AhTehh1dPTgkoASWgBJSAElACSkCFV68BJaAElIASUAJKQAkogYQmoMKb0MOrJ6cElIASUAJKQAkoASWgwqvXgBJQAkpACSgBJaAElEBCE1DhTejh1ZNTAkpACSgBJaAElIASUOHVa0AJKAEloASUgBJQAkogoQmo8Cb08OrJKQEloASUgBJQAkpACajw6jWgBJSAElACSkAJKAElkNAEVHgTenj15JSAElACSkAJKAEloARUePUaUAJKQAkoASWgBJSAEkhoAiq8CT28enJKQAkoASWgBJSAElACKrx6DSgBJaAElIASUAJKQAkkNAEV3oQeXj05JaAElIASUAJKQAkoARVevQaUgBJQAkpACSgBJaAEEpqACm9CD6+enBJQAkpACSgBJaAElIAKr14DSkAJKAEloASUgBJQAglNQIU3oYdXT64lEei+86G4984xOP2U41pSt1t0X3ufdwX23mMXjBk5pEWfR0N1/oefl+KIE87FH9+/j3Zt2zTUbuu0n9G334PikjLcPXFknV4fzy/6658VOOH0i/DF+y+hTU52PHdV+6YEEoaACm/CDKWeSDwT2JRfgElTH8Brby3B2nXr0SY7G7vt0hODB/bH0UccLF0fdvPtuPC807H3nrvW6VTmPPYsbhl3N5b//BE8Hrfsw+fzo/suh2KbHt3w6bvPR/b7x1/Lsd8Rp+Hlp2fjyMMOlJ9vzMvHtPsfwetvLcGKf1cjKzMD22zdHWf3PhH9zj0dGelplfp136z5GDVhKoYNvhSjRlxb6XfLV6zCnoechOysTHz3yauV/qhTMrfZuhumTRpl+TyvHDIKTz23UNo7nU5069oJZ5xyPEYMHYSUlORa9/Pt9z/j6JP74p+fP0ROdlakfV2FNxAIYNacp/DM86/g97/+QUpyMrbbtgfOPPUEDLjo3Aj/WjsWZw1W/rsaE+96AHfefvMW492UXf131X84qNeZ+HLJy+i0VYdKh/7+p19x1P/Ox/777oG3XprflN1C554HYfb0iTjlf71qPC6vj/GTZ+Cp5xahsKgYB+y7J+6cMBI77rBN5HWDrrsVHdrnYtytw5r0HPRgSqC1ElDhba0jr+fdpAT6XHgVfvvzH4y7ZRh6br81NmzchE+/+BZdu2yFvn1Oa5C+/P7ncux/5Gl4/YW5OPiAfWSfn3z+NS4ZdCPy8gvwy5dvoX27XPn5vCeew42jJmHFLx+LMK5avQYnnHExwuEwhl9zGXbZaQckJSWBkaiXFr2Fo484BIMu61upnwcefQZOOfFoPPHsQvz8xZsiouZmCm9ysgdXXHIext86vJJk1kV4l/7+Jx6cdjsoE19+/T1GjJ6My/qfg4mjb6iVX0MKL49/7kXX4JMvvsGQqy7FwQfsjezsTPzw01LMeuRJjB81LHITU2vH4qiB3++H223cKDX3Nm7Svfj1tz/w1Jx7t+gKbwx5rT393EK8vfBx7NRz2ybrrlXhHXvHdDwy/1ncd/dYdO/WGbdPuQ8/L/0DXy55Celpxo3jh598iQsGDMWvX72NtLTUJjsHPZASaK0EVHhb68jreTcZgfJyHzrtcCBm3DUG/c7pXe1xq6Y0UDavu3EcPv/qO3Tu1BGjb7oWI8fehTE3X4dzzzol5n522vdYXHJBH4wYOlB+P2nqgyKt/6z4F1dedgHOOPV4+fllV9+Etes24JUFD8u/+w0Ygi+//kEesUZHQM2DBIPBSkL7xVffy2so0QccdQbuGHMjTjzuiC2E99pBF+HBOU/i6w8WolvXzvL7ukZ4GXV+9blHIscYctN4LP7gU3z/yWuYfM8svLDwdSxfuRptc9vgpOOPxpiR14pcmPIdDeyMU0/A3AemSF/22G1HOJIcmP/Ui3C7Xbjg3NMxasQ1IvyxtgcfeRI3jZ6MN196FAfut1elJpThouISiWgzuj5+yr149vnX5IZj1513wNiRQyIR9U+/+Ab/O/MSLHr2Ydw8eopEivfda3c8MnMSGIHnMf5e/i8OPXAf3H/PhMjNypRps/Hmux/guKMPw5MLFiIvbxOOOuIgTJt0WyQNYeFr7+DeBx/Fst/+FIk9YL89MGnMCGzdo2ukv7nd95abhVfffA/ffPczrhnUHyedcHSllAbpx22T8cXX38MfCKBHty4Ye8tQnHDM4bIfXpuM8n/3wy/yROD0U07A+FuHIjU1RX5v9vXs3v/D/Q8/jsLCIjn/6VNuq/FR/h4Hn4QbrrscF553RiW+ZWVe7LjPMXjthbl48OEnkJOThQmjNt9MlZaW4abRU/DGO0uQX1CIjh3a4+J+Z8tNHDc+lXjo0aex+r91yMnOlBvD+bPvlt/xZm/GrPmY+9gCrF6zFlt37ypPYMw+7H7Q/8AIuLl1aN8Wv3373haXiNdbju32OAojhg0Cr39ueZvy0XPvYzD1jlvQ//wz5WehUAhb73q4pjHFfJfpD5VAwxNQ4W14prpHJbAFge33PEqifvdOGR2RgaqNooWXf3wPPe5stGubKzLpD/hx021T8O0PP8s+qhPey6+5GWvWrheJ4nZKnwE458yT8M/yf5FfUCR/cLnttM+xuPiCs3HTsEEiaPzDe9OwK0UyrGyDh49GenoqJo+7SVI1mPv55CPTIy81JfPtl+dj5Li7sd3W3fHg9Any+6rCe831Y/DlNz/gs/deqPbQTGmoKrwjx96JF15+E0u/eQf3zJyD/ffZA926dMLKVf9hxG2TcMiB++HOCTfJPmuK8FLWLrngbJx9xklYuuxPXDl0FB6+bxJ6n3xszP70OqWfyN1LT82qEdVtE6aKkE6fPBrbb7c1Hpn/NOY98Ty+WPySiKcpvPvuvTtuveFqudEgi4yMdARDIYy6cTDSUlNw1bBRci7TJhspIJRIMu/f90zcOf4meMt9uPDyYUhJ8eDpuTOkzbMvvCqyv9OO26K0zIu7pj+EP/9ajg/efAYOh0PaUHgpbRT/g/bfGyWlZXJzFJ3De+IZF0tKwcjrr0ayxw1G2ZnacsiB+2Lt+g3Y97BT0fvk4zB44EX4d/V/coN2fK/DK/X1zumzcckFZ2HMzUNQ7vPhjL6DsP8+fMRvjE3VjU8bdj3gBHz09gLstnPPSr9++vlFmPHgo/j47eckQnrJlTdIhNSMTN9+531Y9Pp7eHDaBDk3Xgvc35mnnYj3P/wMfS8bgjn3T8Geu+2EjZvy8clnX+OKS86XY9xx9/149oXXcMfYG7HTDtvi+5+W4tobxmD6lNGRvHorEd6vv/0Rx5x6gbDeY9edIv3nz3buuZ1Efc2N74We229bLQsr70VtowSUgDUCKrzWOGkrJVAvAm8v/giDh41GfmEh9t5jVxxy4D447eRjsdfuu0T2Gy28/ONMMfjp8zfQpfNW0mbpb3/hoF5nYNb026sV3vlPvYAbbr1DUhUozT12PUxyd//6ZyVGjpmCL95/Gct+/xsHHn16JPXhx1+W4fDjz8FjD03Fqf87JtKfnnv3ktQLbicce3jk8XJxSSl23PsYiQ4z3/jv5Sux/5Gn46cv3sBWHdpL+2jhZWTw5LMvw4dvPSsCU1V4757xCH5d9rtIZnVbtPAy2vzVtz+i76XXyU1ErNe98/7HuOKakfjrxyWyy5qE1+8P4LXn50QOff6l12Krjh1wzx23xuwObw7OPetkkf3qNkb5mDt954SbcVHfs6QZx+OQY8/GkYcdgEljR0SE95UFj+Cwg/eTNsxT5rm+u+hxUIS5Pfzo05g5+3F8+/Er8m8K7z0zH8Hv3y2O5Nn+svQPHHLsWfjqg4XYftseW3SLkc+uOx2Cj95agF122l5+T+HlTc6NQ66ItK86aY3iyRzty/qfu8U+Kd1MZ/n2o0VwuVzy+5deeRuXXnWj9K1tbo709YGHH8dv374bkdJH5j+DWXOelGsx1mbeCHAfZgqO2e6ksy7FiccdKZFT8mQkmBFe8+Zk4HW3SGSdEl91e+zpF+VJwFdLXt4i75uR4233OBKPP3wPjjnykMhL+YSETzNeePIB+ZkV4X3tzffR97LrsPTrd7BVR+P9wO28S65BKBjCs/NnRn7G/hYUFOLpecaNim5KQAk0HgEV3sZjq3tWApUISO7pNz/gy29+lEfxlFrm9PJRMrdo4X1o3lOYdv88yY2N3rrueDDunniLCC8n7lBQuPFRM4WI8rn3oadIhJdCMGjIrZJ2QEntscth+PnLN/HqG+/h1vFTseKXj0RCqhNeijHl8pZxd8HjduGZR++TYz365POYOXt+JWE5/vT++N9xR2Ho1ZdKm2jh3X/fPXHexdfIvhY8dn+dUxoog4xO8rz4xcfq9909TqSIE+3umvEQfvv9L4lYm9uq3z6VSGdNwstUg+g84OEjJ2Lt+vV4/KF7Yl7B5Hje2afUKLy/LvsTBx9zpjBizra5MQ2DbF588sGI8K5c+gkyM9KlyXsffIIz+16J9X9/FRHERa+/iyuH3Ip/l30qbSiRLy56s9IkRD4e77Dt/pg/626cdMJR+HXZHxg7aTq++ubHyE0LX8sxNNMRKLxsHz0Bq6rwzpz9mKQsMHXjqMMPkhsiU5gvGni9pLnMuX9y5PzWrFsvTw/MdA/29bW3FuP9156KtKEUDxkxTiYQxtpMBtFc2O7Pv5mjfrq8J8yJbOMn34vvf1yK5x6/X3bFnPVz+g9Gl84dccxRh0rahzkplDdvx/e+UN4Lxx51GHodeTBOPuFoeeLCiXBHnnhezP5wwqd5s1Ff4Q2HwpH3EQ82ZMR4ec9y8qhuSkAJNC4BFd7G5at7VwLVEhg14W7Mnvs0Vi37VCJk0cI7e+5TmP7AlsLbpefBkpZA4aU48XE2N7fbiW237i7/zVzDfuecJlLIHNDZ906Un/NR/NWXXyCPfAsLiyNRq9pSGvpfMRzl5eWRP9R8NMvHtuajce6bwrXdNt3x9YeL5FhVhZcCduhxfSQN4O4ZD9epSgMnMd0/dbyw6tq5Y2Siz29//CPRzcnjbsRpJx0nkcXPvvwWjAaaVRnsTFq7/pY7sHrNmkopGtGDSI4U1JokxRTeL5csxA7bbY64UnhXrFgt7M1I5po/vohEHHkjxMh+/r/fRw75yuvv4YrrRmL1b59FhHfBi6+A+zY33ky132Y/PPHwNPzv+CPlGuh1xMEYfs0AdO7UQcSUQjzvgTsjgkvhfebRGSKF5harLBmF7M13PsT7H36Ktxd/jDvG3CBpABRejgVzjs3NFF5WTzhgvz0jObyMWJsbhZepAit+/Tjme8OUzx8+fV0mfJnbmInTMO3+uZWuO17j3H787HV07dJJ/pv5sm+/9xHe//BzLHr9HRxz1CF4dJaRp8vo7+IPP5NzeeWNxXC7nFj82lP4+5+VOOqk8/HJO89HhD5W56wIr52UhosGDofT6ap006AfmUpACTQOARXexuGqe1UCtRLgo13K1b/LPpEoZEOkNPCgVw29TSapUQbO73NqZJIMc0oLCosl4nbV5RdGorF8DR/BMhoYa9JatPBSXA8+5iyJUHJCkLl5vV4cf/pFePnpWTj0oP22EF62u3rYbfh56e+S/7rtNt1tlyWrmsNrHpulwcbcMV1yOc3NnFhmCu9Pv/6Gw47rgz9/WCJCbG6xypLVJrwPPPwEbh4zpdpJa4wgskwZUxoYjTcnPZkpDUcdfqDkZddHeKdMmyXR+o7t28mpMLJJwWdKQ2pKsuTAfv3hQmy3jSHb5vkzam1GdK0Kb/SFTOl8b8mnkp/KlAaW3frmw0WRCY2xUho4wc6O8HKSZ7edD8FjD90TiUZT6Hfd/wQMvKwvTjz2yErvrYHXjsTJJ/aSfPSqG1OJ+lx4tcg1r7vozbzRe2z2VHBMmNJw/bVX4PprB1T73mU6C3N6q8vv5gutTlpj2wOO6o1+556B6668uNbPC22gBJRA/Qio8NaPn75aCdRKgH8AWWS+37m9sfuuOyIrMxM//PwrbptwD/bcbefI49iaJq0FggGMGDVZJq3NuHMMzjnz5GqPy0f/194wTn7/+eIXIpHfN975AAOuHiGPdKNzRNmOdU8prKxMwBntu+7cE26XS1Im+Fh63712kxxfit7Hn34twlN1O/uCq9C2bRvJMa4a4WVbTh7a5/BT5WXn9zktIryccLb0tz/lddVtsSatmW0ZvT2ud39wghxzin/+9Xece/E1ck6m8LIOMmfOc8IQI5osxcYobV2El+W7+vQfLBUKWJaM+dhZWRlblCXjwgkUQk5ao+Cbk9a+fP9l9OjepV7CO/3+uZJmMOG24SgtLcXg68egR7fOEoVn/3bYq5fUKL5yQD+JeDIa+9GnX4FyZ0d4WQKMlT226dEVmwoKMfSmCdi6exfJmzYnrbHixdVX9JdJa9feMHaLSWt2hZfjyrQOVs8wFwR59Y3F6D9wuFRFiL5hYVteP3MeW4DvP3kV98ycKykkTFPhtTxt5ly89d6Hktbz0itvIT+/EAexjFxWJt585wMMv2UiPn33BamPS4HnUxVWQel15KHwlnvxxVffwe8PRkryMe1hn713kyooHrcbuW023zxVvTFgn2oqS7Zu/UYwT54T8Nhf3ZSAEmhcAiq8jctX964E5HH/5HselMfBf/29AmVeLzpv1VFyLa+/9vJIeaaqZcmYs3jdjeNFrJiTyLJkFJBpk2+rMcJE0dvtwBOllBn/0JtbQWERttntCKSnpYoIRtfNZZv1G/Iw7f45eOPtJTK7nY+rd9h2a5x60rEYeOn5SPZ4sPN+x+LKARfGjIJxUhAnzC375l0RCy48QQllDq+5UQIpFawQYS48UdcqDdGXFnNNZz70mPyItY1Z/o0VA6IXmpjx4HzJPeZj9+iyZFVXWqstwstjUCoffOQpPPvC5oUneu6wjUxQO/fMk4UdH5+Pm3yvVEzYlF9YbVmyuqQ0UCLPOeMkkT3KPB/bU6zNSV5MjWB5Lo55dlYGbhwyENfcMLbSoglWIrycVMV60WvWrhNJZO7r7aOHR0SP1ybzwb//8Re5gTjj1BNjliWzE+ElX5ZVYwk+pipQXDnhq7zcL08Wqm5m3vrzTzyAX5f+gSeefVmecPCGjRP/xt06FLvvsiM4kZGR8WW//SXVInhtD7/28korG3KC4EPznpFJnlmZ6dhtlx1xzaCLcOxRh8phmV88YtQk/L18FXLbZMcsS8Z2jEizlnDVhSeiawZz4t5zL78h7xHdlIASaHwCKryNz1iPoAQahMCKlauxx8H/k2oH/AOuW+skYNa2jZbIRCPBm0RGU4ddMyBSOzqRzpETOA846nRMHn9TRKYT6fz0XJRAPBJQ4Y3HUdE+KQFAKg9kZqZj++16YNWqNRLxYnrEkjeeVj6tmEBrEF4O73c//oLffv+7xvSdlnoZMOWHi2MMvLTy6oUt9Xy030qgJRBQ4W0Jo6R9bJUEWGT/jrsfkFWh2rbJweGH7o/bb7teCurr1noJtBbhbb0jrGeuBJRAYxBQ4W0MqrpPJaAElIASUAJKQAkogbghoMIbN0OhHVECSkAJKAEloASUgBJoDAIqvI1BVfepBJSAElACSkAJKAElEDcEVHjjZii0I0pACSgBJaAElIASUAKNQUCFtzGo6j6VgBJQAkpACSgBJaAE4oaACm/cDIV2RAkoASWgBJSAElACSqAxCKjwNgZV3acSUAJKQAkoASWgBJRA3BBQ4Y2bodCOKAEloASUgBJQAkpACTQGARXexqCq+1QCSkAJKAEloASUgBKIGwIqvHEzFNoRJaAElIASUAJKQAkogcYgoMLbGFR1n0pACSgBJaAElIASUAJxQ0CFN26GQjuiBJSAElACSkAJKAEl0BgEVHgbg6ruUwkoASWgBJSAElACSiBuCKjwxs1QaEeUgBJQAkpACSgBJaAEGoOACm9jUNV9KgEloASUgBJQAkpACcQNARXeuBkK7YgSUAJKQAkoASWgBJRAYxBQ4W0MqrpPJaAElIASUAJKQAkogbghoMIbN0OhHVECSkAJKAEloASUgBJoDAIqvI1BVfepBJSAElACSkAJKAElEDcEVHjjZii0I0pACSgBJaAElIASUAKNQUCFtzGo6j6VgBJQAkpACSgBJaAE4oaACm/cDIV2RAkoASWgBJSAElACSqAxCKjwNgZV3acSUAJKQAkoASWgBJRA3BBQ4Y2bodCOKAEloASUgBJQAkpACTQGARXexqCq+1QCSkAJKAEloASUgBKIGwIqvHEzFNoRJaAElIASUAJKQAkogcYgoMLbGFR1n0pACSgBJaAElIASUAJxQ0CFN26GQjuiBJSAElACSkAJKAEl0BgEVHgbg6ruUwkoASWgBJSAElACSiBuCKjwxs1QJG5HFrz8FqbePz9ygi6nE1t1bIfjjjoIF/c9HR63u1FOfs4TL+GxZxdh8cuP1Lj/E/sMQkFhsbRJSkpCh3a52HXn7THwoj7o3nUry337d/VavPHuRzj95GPQLjcn8rrSMi+OOX0AbrjmYpx5yrGW96cNlUCiEPjki+/wxIJX8feKVSgt9aJtbg522mFr9D7paBywz+6NcpqL3nwfKcnJOO6ogyvtf8bsJ/DGe5/g1adnNshxv/ruZ/yy9C/0P+/USvv7/qdlGDR8PObMGIede25br2ONHD8diz/6MrKPnOwsbNujCy46/zRb/Kr7jOKOLx48Cl07dcSEWwbXq6/6YiUQrwRUeON1ZBKoX6bwUvg6tGuL0rIyLP7oC7z/0Vc494wTMGTQhY1ytnaEd+vuXXBBn1MQCofxz/JVeHzBK9KnJ2bdgfbtci3174tvfsR1N0/Go/ffjp7b9VDhtURNGyU6AYrnxKkPY+89dsYJRx+MjIw0rPpvPT7/6gf03L4Hrht4QaMgGDhsHHKyMjB5zLBGFd5Z8xbguYVv4+0XZlc6zrI//sHYKQ9gwi3XYNseXet1jhTeL7/9GaNvvFL2syFvE5558Q38s2I17h5/PQ45YC9L+6/uM0qF1xI+bdTCCajwtvABbAndN4X3qYemYOvunaXL4XAYF199K1auWov3Xn64UU7DjvDut9dulSIb737wBW69/V5cddl5uPCcUyz1T4XXEiZt1MoI9L1iBJKQhPkPToTT4ah09n5/AG63q1GINLfwNuRJUXi///m3SlHp9RvycOZFw7D/3rtg6oQbLR1OhdcSJm2UoARUeBN0YOPptGIJL/s3afocvPzae3hjwYPIzsqQLgdDITzx7Kt49e0lWL1mg0Rojjv6YAy8uA+SPR5pU1xSigfnLsDX3/2ENes2IiM9DbvsuC2uuux89OjWKXLq9RFe/jE5rd+1koLAyPT8Zxbh/Y++xMpVa2T/2/Togkv7nY6D9ttT/r3k469w07hpW2B/ds5d8viWKQ3XD74IhUWlePHVd1FaWobdd+mJG6+9GJ06to+87tsffsXs+c/jj79WIBAIoF3bNjjy0P0weMD58TSk2hclYJkA30c9t+2Ou8ZfX+1r8jYVoPcF1+HKS85B37NPqtSO7+N5T72EhU/OQE5WZuTR+7577Ywnn3sN6zdskvf9tVf0xb577SqvvfDKkfIeit6OP/oQjL3pKpgpDTMm34ypM+fj56V/yPusT+/jcc7pJ1R6TX5BIWY/+jw+/PRr5BcWo0un9jjvzP/h9JN6Sbt7HngMz770ZqXX8InQwifuRXUpDfysePL51/Dbn8vhSEpCj26dJbXriIP3qZZPLOFl43MuHQ6n0wkGE+rzGdWty1YRrsceeRBmz1+A1f+t34Irj6mfUZYvfW0YZwRUeONsQBKxO9UJ75CRU/Drb3/hjQUPSO4st9GT7pd0B0ZV99xtR6z49z/Mmvcc9t97V0wcdZ20Wbt+Ix557EXst/cu8oeqsLAEL732Ln5Z9heefngKcttkS7v6CO/3Py/DoGHjcUnf3rjioj6YPH2O5Bx22qo9QqEQPv78ezy38C1MmzgCB+67O7zeciz55CuMmfyA/GHfpiKS3aF9W/h8fhHeLp06YOtunXHS8UegrMyLB+c+i44d2uLh6WOlv/yjf9bFw2R/Z5zcSwR/9Zp18oexsdI+EvF603OKLwK3TLhXbgj5tOSYIw9Ex/ZtY3bwtjtmYtkff+OZR+6K/J7vNb4n9tilp8gqN+aarlq9Fnvu1hNXXnou0tNS5Qb4g0++xvOPTkWbnCzwhvXGMfcgKzMDI667RF6XlpoC5r5SeGxX0oMAACAASURBVF98dTHat8tB7//1QveunSQ/9rW3P6iUHlBSWoYB196G4pIyXNLvdHTt3BFffvsTnljwGoZeeaEIMnP/H37sBXntYw9OlONQQHmOsYT3hUXv4M775slNLAU8LTUZy35fjtTU5C1kOxpSLOEt9/lw0jlXYYdte+DBqaPq9RnFeRXkWlhYjJQUDy7oczLSUtPwxHOv4s+/V0a46mdUfL23tDf2CKjw2uOlretAwBTeefeNR7eunSS6+d6HX0h0JHoi14+//I4rho7F8Kv74+zTjo8ciW35R3P+A7fLh3usLRAM4pTzrsal/c6I/OGwI7z77LEzbr1+oMjs8pWrJfrMD/q5M8Zhxx22iXnMEWOmwuF04I5RQ+T3taU07LLjdnh4+piI3L/2zocYf+csLJh7t/wxNV9v/rsOqPUlSiDuCKzbkAfKLAWQGyeF7rfXrjjpuMMiEVn+3BTE+ybfHPn5x198h+tH3YUH7x4lN8Cm8K5dtxEvzJ+K1JQU+VlhUTFOOHsQRg4bgFNPOEp+VlNKw5PPvy43podW5L4yxeqcS6/HbjvvgNE3DpLXP/rUQsye/xzmzRxf6XPn7pnz8c77n+KVZ2ZKikZ1ObxVhZeTV0/rew323XPnLfKKaxs0Cu93Py3Dc/OmStMNG/Mx+9Hn8O4Hn8nNMOdCxNqsfkaZXFes/A8L5t2Ftm2MSbdr1m7AGf2HRLjqZ1RtI6W/j2cCKrzxPDoJ0reqVRrM0+LjQ0ZKzO2hx57HnMdflMkfTFMwN5/fj6NOvRTDrrowIsKMGPFR4srVa1FQUIQwwmA+4FmnHiepA9zsCK9ZpcE8Jh9LXjewL4454iD50e9/Lce8J18GJ6JszMuX1ItgMCSP/J6cPVna1Ca8jBQzYmxu3Gf/K2/BfVNGYt89d8GGjZvQ55Lh6Ln91hLh5SSf6qJhCXJp6Gm0IgK83j/76kf89Mtv+PLbX1Dm9WLQxedIpQFzu2DQzdi6W5dIPv2NY6aClQXM95gpZrk5mVvkrVJ4+/Q+DgMuPEt2V5PwLlj4Nt5fOAeOqJzim8beg8LiUtx/5y3y+iuvnyBPYubNnFBplD7/+kcMGTkZT8yeJJPRrAqv+flw17jrceiB1iaZmQeuWqWBP09J9kj6x2UXnCnnUZ/PKJNrVkY67p10U6XzPe7MK0SoyVU/o1rRGzYBT1WFNwEHNd5OyRRePpKkwG3KL8LTL7wOpg3wZ3y0x23KvXMlvzXWJBbKLD9wL7vgDCz55GvwjxP/uB112AHIycqCw5GE4aPuEnFklMeu8DJyPOBC4w9Hu7Y5YCqCOcHmv7XrRUwZoT3njBOwFX/ndMqjTD6CZUTWivBWLUvGGdbnX34jpk64AQfvb+QC//Tr73j06YX46tuf4S33SWTpqsvOieQKx9vYan+UQF0I5BcWYejIKfj9z+USKWVuLrcXXnkH9zzwOF5+fDr8gQDO7D8U1w3sV+lxf3Xls04+72qcfNzhkjpRm/DGKkt26+33Yc269ZEUo/MH3IDl/66By+WsdIrhMCS/nmLMm1KrwvvOks8wauJ9W1RxscKPwvv1979iypihQBKQnZmJzp3aR0o61vczyhTeWGXJqnLVzygrI6Zt4pGACm88jkqC9SlWDi9zXvsNvBk+n0+EMSUlWf5wPL7gVXmE6HRU/iNDJDnZGZKDx/QGTlZ75F4j99Xcju59GY478uA6CW/VKg3R+2WuLv8IM/LMPEBz46PW5f/+16DCa+6bf1B/Xvon5j75Er754Vc5hkZ7E+yN0cpPh09omNY0a+pt2GPXnkLDfOzf/7zT5LOBtXs5WS0zIz1Cq6mElxHiYDCIW4cPjDlSW3VoG/ncilWWrGpKA/N/r71pEu4cOxyHHbS3rdGvbtKauZP6fkbZEV79jLI1dNo4jgio8MbRYCRqV6qbtGaW/mIdzvPOPDGSwxcd9Y3FZOgtd0pZs2kTN5fiMXP9mL9XlwhvTcLL2c9MtVi88JFI/i1nb5/Rf6gsMGFGeL/7cak8Bn1o2mjJBTS36haeiBXhrXq+P/z8mzyajc5rTNTrRM8rMQn8vXyVVDWput0+9SG88uYSLJh7F7p23rzAy133PYrPvvpeUpT232c33Dr8ikovtSq8g0fcAU7Giv6c4I6qW3iiaoRXUqKeWSiT0aL7V/U8+Nkw7+mX8cEr8yr9qqrwMoXj1POvwd577CTSa2erTXjr+xlVF+E1+6+fUXZGUts2JwEV3uak30qOXZ3wmrV48/ILZRYwV1wbO+VBLP7wC5zd+zjsueuOcDodMnHio8+/wdAr+4Plc55+4Q3MfPgpTBo9VCo1MBI64a7ZyC8oAkvqNLTwLvv9b1xyzW0yIY6zl9euz8Pke+fI49ic7MyI8G7KL8Qp5w/GiccchtNPOloehW63TTcEAsGYK61VFd63Fn8ipc8Y/enYoZ3MDmfqB9s9O+fOSlGuVnLp6GkmAIHe/a6VaiRHHbq/VDmh+H3+1Y/g9c5SXFUXhvjrn1XoN3CEnDkrmOy603Z1Et6p9z+K19/5GKNuGIT2bXOkYgMrpVgVXlZpuHzIGHkfshTZdlt3gdfrw/KV/+H7n5fi7vE3SL9Y4YFCyslju+28PZI9bmy/bffYVRpeeQd3zpgn7/ETjzlUJt39/ucKiRRXN/GMx6hNeOv7GcXPXis3EvoZlQBvyFZ8Ciq8rXjwm+rUqxNeHt+MzN4w+GKceeqxUiWBjwcXvvG+lCTjBzGXIeZsaj7mZAkiVmSYPW8BXn/nIxQWlUj06PKLzhYJ3m2nHRpceNnPN9/7GHOeeBH/rd0gs8x7n9QLnCn++dc/RISX7Z5f9LakZfB3FProOry15fCy/NjcJ1/Er8v+xsZNBcjKTMfuO2+Py/ufLeKsmxJoiQRYSWDxh19h6e9/Sc1cliBkVZJeRxyIfn1OitTXjj638y8fAbfLJZVZqm5WxIyvYWkyRpG5YANTqKrW4a26tHDVCC/3UVRcIu/7JR9/LfvLyEiXiapcFp0TZLlxAisl9v2PvpAyZbXV4aUgP/7sK1InmDfFXOXxkr6n15jmUJvw1vczKroOb9WlhaNzePUzqiW+A7XPJgEVXr0WlIASUAJKIG4I/Lt6Dc659AYpWXjGycfETb+0I0pACbRsAiq8LXv8tPdKQAkogYQgwHq9LEHGnNi/V6zC8/OmyqN+3ZSAElACDUFAhbchKOo+lIASUAJKoF4EOEnsofnPoXvXrXDjNZdUWpSiXjvWFysBJaAEAKjw6mWgBJSAElACSkAJKAElkNAEVHgTenj15JSAElACSkAJKAEloARavfB+/d3PmDJjLvz+II7vdbAsdVl1Y8HwGbOfQkFhkZSGGnrVhbKil25KQAkoASWgBJSAElAC8U+gVQsvy0b1uWQ4Jo8eih7dO+OKIWMxZNAFkVV/zOHjUorZWZlSA5bFxEdOuBcsafPEE0/KUrT9+/eP/5GOgx6WlpYgLW3ziklx0KW47UJ5eblcW263O277GE8dCwYDUtKupW8DBlyOiRMnokuXLRdqaOnn1tD9DwT8MuYej05ss8K2pKRYPn9ZFk632gn4/b7aGwHIy9toqZ02angCubltbe20VQvvsj/+waRpj2DufeMFGpe6XLtuA665ol+1EFkD9rgzLsebz8/CgmcXqPDauNxUeK3DUuG1zootVXjt8UqE1iq89kZRhdceLzvC26HD5pUC7R1FW9eVwLp1a6DCa4PeR599i0VvLI6s9LPk46/wzpLPMH7k4Gr38to7H+Kt9z6V5So1wmsDNgAVXuu8VHits1LhtccqUVqr8NobSRVee7xUeO3xaurWKrw2iX/46Td45c33I8LLFXDe++DzaoWXEeFRE2fg3jtultW/KLzhcAh9+/a1eeTW2ZwSl5ysjx+tjH4gEJBHj06n00rzVt8mKcmBRHhSe9llAzB27Bh07ty51Y9pbQBCoSBCoTBcLldtTfX3AMrLvUhOTlEWFgkwpczKxpQGjfBaIdWwbVR4bfKkwN4x7RHMs5DSwNV/ho+6C+NHXoOe2/WQI1F4uV1wQfUpEDa7lNDNvd4ypKSkJvQ5NtTJ+Xw+SZfRP+bWiTInv6VvzOEdN26s5vBaGEgzjcXt9lhorU3Kykrl81dzeK1dC1Y/T1R4rfFs6FYqvDaJcsJDn0uux+TRQ7A1J60NHYfrBvbDnrvtiJWr1qC0zIsdt98aG/Lycc2Iibh+8MWVqjNoSoM94JrSYJ2XpjRYZ8WWmsNrj1citNaUBnujqCkN9nhpSoM9Xk3dWoW3DsRZcmzKvXPh8/lxQq9DcNVl58lenn7hDSxfuRojrrtUVv957NlXkJOdFTnCnBnj8Obrr+mkNRvMVXitw1Lhtc5Khdceq0RprcJrbyRVeO3xUuG1x6upW6vwNjFxjfDaA67Ca52XCq91Viq89lglSmsVXnsjqcJrj5cKrz1eTd1ahbeJiavw2gOuwmudlwqvdVYqvPZYJUprFV57I6nCa4+XCq89Xk3dWoW3iYmr8NoDrsJrnZcKr3VWKrz2WCVKaxVeeyOpwmuBVziEUMCHcNCPkNNaRSGdtGaBayM0UeFtBKg17VKF1x5wFV7rvFR4rbNS4bXHKlFaq/DaG0kV3s28pAIDpdaUW67aF/AhKRxGksuDJJcbSM6wBFiF1xKmBm+kwtvgSGveoQqvPeAqvNZ5qfBaZ6XCa49VorRW4bU3kq1VeMMBvxGxDfgqBDcAhAKA0w0H5Va+u+V7knNzTedESmngCrEHH98XXTt3lNrV3bt2wtibrkZum2x7FxGA9z/6Ap9++T1uHnq57dc25AtUeBuSpoV9qfBagBTVRIXXOi8VXuusVHjtsUqU1iq89kaypQlvOBhAyFeKcCgIhPn//J+KL/7LrLsd5m/CSCIO+Rl/J/8A90GZrSy3lNzaFytJNOE94uT++OSNJ4TbuCkPoE2bbFx7hf01BFR47b3vEqa1Cq+9oVThtc5Lhdc6KxVee6wSpbUKr72RbAnCG/SVIVTxxbNzeNKQ5HCCyyiK0HI5Rflv4zuQZMhuxX9X/h0siW11FBNVeHm+C19fjC+++RETbrkWhUXFmDTtEXBxLd4nDB5wPg7cdw/8s2IVJtw9C2Vl5XA6HRhx3WXYdaftNcJr722XOK1VeO2NpQqvdV4qvNZZqfDaY5UorVV47Y1kPAovo7em4PJ7kisZzuQ0ODyp9ZJVe2Rit05U4Q2GQrj19nvR6/ADcNxRh+D2qbNx2EH74MhD9kN+QSEuv240nplzN7zlPrhcTnjcblmIi1Hhh6aPVeFtiIurJe5DhdfeqKnwWuelwmudlQqvPVaJ0lqF195IxovwhgM+BMtLwWguQkGRW/MrnpY9rq/w/vHX8s1pF/aGql6tt9+2xxbLR5s5vD26dca69XnYduuumD1tjMjsqX2vRqeO7SOv2bBxU+R39zwwXxbg4hL3f/29Em+/+LAKb71GpwW/WIXX3uCp8FrnpcJrnZUKrz1WidJahdfeSDan8FJwI6kKDiecnlSJ5LIiQrxu9RXeeDovCq+Zw1tcUoobR9+NvffYGZf3Pxunnn815s68He1ycyp1eeYjTyHZ7caA/meLuPfqfSkWL5yrwhtPA9uUfVHhtUdbhdc6LxVe66xUeO2xSpTWKrz2RrI5hJeS6y/aAIc7ZXMUlzm5LWBLVOEl+r+X/4srho7Foifvw933P4rUlGQMvbK/RHl/WfYndtlxO0ya9jD23WsXSXv4+PNvccPou2TSm05aawEXb2N0UYXXHlUVXuu8VHits1LhtccqUVqr8NobyaYWXoouQiG4s9pXTDCz19/mbp3Iwku2t0yYjt136YmTjz8Cd86Yi1+X/YlQOCwT08bdPBh//fMvbp86C106dUSHdrl49a0leH3BLBXe5r4wm+v4Krz2yKvwWuelwmudlQqvPVaJ0lqF195INpXwBr0lEtV1Z7WDMzndXifjqHUiCW8cYW2wrmgd3gZDaW1HKrzWOJmtVHit81Lhtc5Khdceq0RprcJrbyQbXXjDYfgK10s018OobgvfVHjjewBVeJt4fFR47QFX4bXOS4XXOisVXnusEqW1Cq+9kWxM4Q16i+EvzhPRZcWFRNhUeON7FFV4m3h8VHjtAVfhtc5Lhdc6KxVee6wSpbUKr72RbAzhZR3dQNFGwOGEO7OtvQ7FeWsV3vgeIBXeJh4fFV57wFV4rfNS4bXOSoXXHqtEaa3Ca28kG1p4A2VFCJTkw5PVLmGiutFEVXjtXV9N3VqFt4mJq/DaA67Ca52XCq91Viq89lglSmsVXnsj2VDCGw4F4C/cAIcrGa6MNvY60YJaq/DG92Cp8Dbx+Kjw2gOuwmudlwqvdVYqvPZYJUprFV57I9kQwhsoLUCwrFgqMDjcyfY60MJaq/DG94Cp8Dbx+Kjw2gOuwmudlwqvdVYqvPZYJUprFV57I1kf4Q0FfBULSKTCncBRXZ5nyFuCgLcIzuytLAHOy9uIDh2stbW0Q21kiYAKryVMDddIhdceSxVe67xUeK2zUuG1xypRWqvw2hvJughvOBRCoGQTwkG/TEpLcrrtHbQltA6HRXAputwcKelwpWTCH/Bb6r0KryVMDd5IhbfBkda8QxVee8BVeK3zUuG1zkqF1x6rRGmtwmtvJO0Kr5G+UAhXehs4UzLsHawFtA75vWA5tVB5KRzJGXCmZsDh8kR6rikN8T2IKrxNPD4qvPaAq/Ba56XCa52VCq89VonSWoXX3khaFV5KYKB4E5ypmXCl59g7SJy3DgcDCJaXSB5yktMlkhtrNTiRYTgsnY1GeC1havBGKrwNjlQjvA2JVIXXOk0VXuusVHjtsUqU1iq89kayNuGl5FF0k1weI083yZrw2etF87QOlpca0dyAD64UI5qb5HBt0ZlQeQkCpYVy7knp1ipQtBThfeyZhXjx1XflnJM9Hoy47jLstftOtgfkr3/+xeq163DYgfvIa++45yEcvP+eOOqwAyzty+stx6I330ef3idI+6LiElwy+FY8N+8eS683G6nw2sJV/8Ya4bXHUIXXOi8VXuusVHjtsUqU1iq89kayOuGVMmNFeUA4LGXGoh/r2ztCfLUOBf0IlhWBEutwp0haRsxV4JjDW1Yoouv0pMKZliUMEimlYelvf+GW2+/F/AfuQHpaKvI2FcDn82Orju1sD9o7Sz7DN9//ghuvvbROwptfUIiBw8bimUfultcHQyH8/sc/2Knntrb6osJrC1f9G6vw2mOowmudlwqvdVYqvPZYJUprFV57IxlLeLkccKi8TETXmZxmb4dx2loi1YzShgJwpmTCkZyOJMeW0epQgEJcKCkOrtQsuNKyKkW1E0l4P/r8G8x78iU8NG0skpKSIiNXUlqG8wfcgBfmT4PL5UJxSSnOv/wGvDh/Os4bcAOOP/oQ/LNiNdas24DRN16JHt06S/v8wiJ079pJorRffvMjkpM9+P6nZcgvKMKQQRfi6MONaC/lmJHlYDCIbbfuhluvH4h7Zz2Ol197D7vstD322m0nXHDOKZEIbzgcxgNzn8GSj7+Ew+HACb0OxcXnnx7zSlPhbeI3oAqvPeAqvNZ5qfBaZ6XCa49VorRW4bU3ktHCy8gnZdedkSu5uomwSW5uRToC5TVmNBdAyFcGrhJHIXakZMJVzfknkvB6y30YNGyspA8csO/uOObwA7Hf3rvJsI+d8gCOPGRfSUl44ZV3sOLf/0Raz754KAZceBZOPOYwvPfh5/j8qx9w89DLRWKrRngpunfcNgSr/luHYbdOwYK5U7F6zTrZ94zJI+FxuzHr0QXIycrECb0OqRThjU5pePWtD/DW4o9x17gb4Ha7UFBYjOys2BMmVXib+F2rwmsPuAqvdV4qvNZZqfDaY5UorVV47Y0khTfZmYRgST4cyaynm2tvB3HamtFcRmodnhQ4U410hFgbJTfkLQIcLpHc6oTYfG19hde7YYWkiTT1ltKuOxAVxTWPz+jpz0v/wJff/IQFC9/CJX1Plwjtdz8txWPPLMLd42+QSOstw6/A9tt0F+FlRLhNThaYt3vPA/NFXmMJ7z577iLRWG4nnXslXnvmASx6433Mnr8AnbfqID8vL/dh/713Q78+J1crvGMmz8SRh+wfiRDXxE6Ft4mvLBVee8BVeK3zUuG1zkqF1x6rRGmtwmt9JMP+chRuWIXUtAy4M3NjTtiyvrfmb8lqC8y7ZaSa0VyKbpLDuWXHwiFJb2BbVmMwhLj2WsKy/1DI0om2lElr0Sfz/sdf4unnX8ODU0fLj5nGcP3gSyTd4NH7J8rPKLxz75uAzIx0/LNiFabcOwf33zUqpvBGT1ozhXfh64tFsBkVjt6q5vBGR3hHT5qJow/b39IEOBVeS5dnwzVS4bXHUoXXOi8VXuusVHjtsUqU1iq8tY8kqxJw4QhGGr1hJzJz2lXK4ax9D/HVIuQvF3kNs9pCKkW3+nQMf2k+gqWGELNtrKhnrLMjL1Z1cGS2t3TyLUF4Kaxl3nLs3HNbMNJ7/5ynsWHDJowecZWc4+PPLsITC17BZRechbN7H1+j8H765Xd4492PMfamq6Vd1SoNpvAyvWHAtaPw0PRx6Nq5o6RTFBQWoX27XJx90VAsemqmvD5aeF97+wO8+d7mlIZN+YUSYY61qfBaujwbrpEKrz2WKrzWeanwWmelwmuPVaK0VuGtfiS5Mpq/JB/hgN9YOCI5FbWVJYvn60LKipUVShcpr45aJtj5CzdInV07dYR5jEDRRjjTskWS65vSEE88//xnpURomVfrdDiwTY+uuGXYFWjX1ii9tiEvH737DsYbz82SiC636iK8pWVeXH3DBKnywAln3/24tFJZMlN4uQ9GkmfNexaBQBAulxPDr74Y++21K26/exZ++OU3HLjvHri8/9mVJq1Rxhd/+AWcTgdOOu4IXHRebxXeeLiYVHjtjYIKr3VeKrzWWanw2mOVKK1VeLccyXAoiEBJPhgJdafnVBLD5hBeY4JYMRAOAqioDhCVXxqpGMCfMd3V/F3Uv2UlNLdHRDfJnVzr5evb9J8IK6szWNmYvsAJfOwLl0826w8nkvDWxoF5uR988hXG3Ty4tqZx83uN8DbxUKjw2gOuwmudlwqvdVYqvPZYJUprFd6okQyH4S/ZJMvkMqoZayngphRe5tZykpixmlkmkmQRi4oJXDKRi/IblsfrsvG7/Gjzv8NsHwrDkZIGh9NC3m0ogPK81UjO2UoWz7Cy8eYg4C2GJ7OdTHqL3lqL8N45Yy4++eJbTJt4k5QdaymbCm8Tj5QKrz3gKrzWeanwWmelwmuPVaK0VuE1RlKkrawQbqYu1JDT2tjCy3xhii5XNJNliRmRdW65mlljXH+MJHPxjOTczpZydaV9cZ5MZKsu7aG1CG9jjEdT7FOFtykoRx1DhdcecBVe67xUeK2zUuG1xypRWrd24WVOq794kwibKy271mGtj/BSqhk15UpkVSd/Sf3bsiKEQ6EK0W3aur6swsDFJjzZRvmrmrYwqzYUbZTIMsuy1STkKry10Wze36vwNjF/FV57wFV4rfNS4bXOSoXXHqtEad1ahVeWwS3JF7lkVNfqVlfh5bGYbxAO+hD0lUmdW36FAgGEA+VS01dWNLOQX2u1r1bb+Ys2ioBbqSls1OstgCujraVV5VR4rY5C87RT4W1i7iq89oCr8FrnpcJrnZUKrz1WidI6kYRXcllDQSAcAieeGV8hmezF7+FQwPgeDsHlSZNKArGWy61pbOsivMZkLkfksT9TAXxFGxFixQSnWyaTOT1psogDc2CTHE2TwsDz9OWvkVzlWPnK0RwY/aUYc+lkVqywuqnwWiXVPO1UeJuYuwqvPeAqvNZ5qfBaZ6XCa49VorRuicJLYeTEslBEbg2plclaDqcxuYsLKCQ54HDyu9MQ2ySn/DsMBxyuukmlXeGV0l4OB5LcKRLNDZUXyeQxByWzoiwYJZxCKefl98qlxbSHJHeqlEJrjI03A7681XBnd6g5qsyJfMV5CAX9MinNbj6xCm9jjF7D7VOFt+FYWtqTCq8lTJFGKrzWeanwWmelwmuPVaK0binCK5O5vMUIektE0JwpaUhKMqTWlFm70dq6jGFM4Q2HRQhZmou1e43vAfgL1ho5uxUrkzFyK9UWapmEFgr4EfZ7EfSVIuTzwuFOEfHlvpJcyfVe9EKitYUb4MntXFH5ITYJVqwIlhXDndGm1ghwdSxVeOtylTXda1R468D66+9+xpQZc+H3B3F8r4Mx6OJzttjLho2bMOqOmVj62184+fgjcf3gi6SNCq894Cq81nmp8FpnpcJrj1WitI5r4WU9XG8JguXFSEKS1IR1pqTHXv62iQakKG89UpLdkjpByQXlNhySqK2IrMMly+5SFt1pOTVWfLDaZSP665W6wPxv5v6ylq6RB5yMJAvL/JrHYu5yqLwMnpyO1R4+UFog+c0iulxdrS4bbwIC5QiCpdRq31rCSmu1n0XLa6HCa3PMmDfV55LhmDx6KHp074wrhozFkEEXYI9de1baU3FJKbhSyW9/LMfylf+p8NrkbDZX4bUOToXXOisVXnusEqV1PAqvGclltJSC60zOsCV1jTE2RgmuTSj3+5GaniVSS8FNougyfSJqY14sKz4wJ7cxNka7w5Rffg+UIxQMVIgvc4FTjEhwlT6xHyw5xoUhXBmxc3BZ85dLArMUmp3V1bhv5keLkPvKpU8cO4lGW8z3VeFtjCul9n2q8NbOqFKLZX/8g0nTHsHc+8bLz5996U2sXbcB11zRL+aeFr6+GL/9uUKF1yZnFV77wFR47TELBgMIcZJPC98GDLgcEydORJcuXVr4mTR+9+NFeCmUUprLWyKPz7nsrZTvauaNFRWCrLDgcIoElpX7kJaWXm1aQXneKri5AENTZSR8ogAAIABJREFUVluoiKYaEWAKp0+qLkgk2JUsE+ECxZskDYSVIKpuvMGQihVmPd2oVdyqw7/5WMYxWYHC4WG02SMpGDy2SLbfZ2kEVXgtYWrwRiq8NpF+9Nm3WPTGYkweM0xeueTjr8Al9saPjL28ngqvTcBVmmuE1zo/FV7rrNhShdcer0Ro3ZzCy8fzwfLSqLxcRnOtLWXb2OxlOd/SAskPdqblRASupklr5Rv/hTu7o0R/m3uTihR+n6QVMOrqysgREY3eyJ4RXf6cMh8rKsz2nOBmiLQht/yi0Jty66Tcx4go2xXe5mbWWo+fm9vW1qkn+XzlFWv52XpdQjT+8NNv8Mqb70eEd/FHX+K9Dz63JbxMi+jXr29C8Gjskygv9yI5ufIHV2Mfs6Xu3+/3w8E/WJyprVutBPi4MxG2AQMGYMyYsejSpeUs8dlc3IPBoFGmq4klLVBahHCgDA53mtSgrU62mpqLRJpZLizJYUw2qxKp9Xq9SEnZ8vO3fONKJLfpXK34yQ0l9xtOksWB5b2WZC4PLD+QPOXI7yoWETbek8b7sqGixnKOlHkuWZyWU+0kOrkZYZ9DwYoJcylwuo38YatbonymWD3f1tCuVQsvUxrumPYI5tUjpYHrgffrFzsFojVcQHbOsboPXDv7aC1tKbySs1bHEkSthVOinSeFd+xYCq+mNNQ2tozqM+DQlMIbKM4TobSbJ1rbudTn94w2S0S3GtE19+31lknAISJy4RAY2U1u2y3mcrzMsw1JdYliOFihoUJepYQaPZbfuYXD4P+J2sp/R/2O/wonIRTySz6zTNyrw3LDjMxSdCnXrEFsph1U5SarrpUXy+8p/Yzk6qYETAKtWniZ89fnkusxefQQbM1Ja0PH4bqB/bDnbjti5ao1KC3zYsftt45cLZrSUL83jqY0WOenKQ3WWUkESnN47QFLgNZNndLgy19rTERLyYgLeiK6JQXSF1d69haP/at2MjqlgWkD5Xn/IaVdty3OxZh4VyzyylJkDXG+LHVm5DkXy2Q5WTCiopav4cchoCIHX/5bviQnQSpdsN4vbzKqpjbIa0MBGKuoFVUsbZxVJ6mOBiGl5Frts++4uLwbpROtWnhJ9Mtvf8KUe+fC5/PjhF6H4KrLzhPQT7/wBpavXI0R112KYCiEMy4cAq+3HP5AAJkZ6Zh2+wh8+vESeezcv3//RhmcRNupCq/1EVXhtc5Khdceq0Rp3VTCS9nybVoNdxYndDV/SharHPhluV9romuOtym8rELgL1iP5LabnyJQSFn2i0LKXGSZfGfj8X9N1xQXf4CsFBcWsQ2xRi9TDsrLpEYvS7axNBq4wIZkSHChDX4lSTRaFrqIMQlQ8qjLihDy++BKy6p/GbVwGAFvEULeEqMfGfbyQxPlfZXI59Hqhbc+g6t1eO3RU+G1zkuF1zorFV57rBKldVMIL6sG+ArWwtOmc1zk6vry1wGoPtpZ09hSeFPcTql6kJxr5IgbE++KZbEJSq4rNTNmekNdrhnul7LL5XylDJoIrEOSH4w8YCBQViIpCFzYwpXKChe1T/xjpJgRXQoxy5BFR4rr0k9DnItFxB1Mu2A/XB7LVRrqckx9TfMQUOGtB3cVXnvwVHit81Lhtc5Khdceq0Rp3djCa4pVcptOcYGMKRVGjdy6RZmL8tbBkxSCJ7u9REYlN1dWfsto0Lq7zLel6FIa3Rm5lgTalE4yZ/kxUzpN8MzVDnmLRHR5/hTd+uTnsnqDpG6UMcWCq8gxxaKybFstSxYXF4d2whIBFV5LmGI3UuG1B0+F1zovFV7rrFR47bFKlNaNKbys7crcUNaljYdNZNdCnm51ffUXbUBxQR7SM3Ok1q3k0HIiWjUluep0zuGwiC7zXym6dUmJoNgGK9IKmELLiWdM4eDPnGlZhujWo8+MaHMiHvsoop9SkU4R44RVeOt0FcT1i1R46zE8Krz24KnwWuelwmudlQqvPVaJ0rqxhJdyKI/Y03PiAlV9ZJeC5y9cL3JXVlKItJwOcKXUnjZg98QZeQ2Ucknf3AaZ5MbjU8z9pYVwehiF3nLRCat9lAlzjA6XGRFtpm1YWUlOhdcq4ZbTToW3HmOlwmsPngqvdV4qvNZZqfDaY5UorRtDeH2b/pOoIqsxxMNWH9n1Fa6XUmGUXQqp351e40prdTlfTjxjqTauhMZc3XjZmLIQ5pLKrDYRCoowM0dYJsNZ3FR4LYJqQc1ajPB6y3145/1Pseq/dWDBcXMzqyo0B3MVXnvUVXit81Lhtc5Khdceq0Rp3aDCGwrCm7cKnpytqq3x2tTc6iq7ZlRX0jGSHBLd9GR3QGHeWmRkZCOJCzDYEL9Y5y2lwIrzpA6yO7OtRMSbewsF/EYFCF+ZTMLjoiDMy61rZQ0V3uYe0YY/fosR3utvuwtZmRnYpkdXOKJWVerX5+SGp2Jxjyq8FkFVNFPhtc5Lhdc6KxVee6wSpXVDCS8nTPkLNyA5t4ulCVZNwa88fw3c1dSdren4TF/g5s5qL+LHUmOe7I6yBHJJ4UYku90AF+xguS+XRyZsGd/dxldFabCajsElfYPeUrgyc2OWC2sKPuYxJL/XxxJnpUYVCDclN61O+cNV+63C25Qj2TTHajHCO+zWKZg64camoWLxKCq8FkGp8NoDBUCF1x4yXXjCHq9EaN0QwhtgHdfyUnhyOsYNkrrILoWWuceerPZwJKdFyo0xssvKB5RfnyMlktLACCjr8fI7J3AZ/+2XiLAhv5tFWFY1S0qSqgZSZiwtR+reNtcmywtLHd9SOFxuODxpIrl1WcGtxpsHv6+5TlGP20gEWozwjp50P0Zdf0VcLbWqwmvvqtQIr3Ve+fkF8HjcSEtLs/6iVtxShbf1Db6PE6V8ZfBwgYTIUz+jxqux1C1rvcqCt8Z3TvuvWMyAq+JyUpRRiSF+Fhjw5a+pdkWx6kaYUV2eGmWXG+WXj/YZ5RXZLS+TRTOiV1qrbl+mCDM9wJRgCnGSi9FfNzxkFfWEtSmuOi78YaYqUHK5CAWlnhPP6lOxoba+a4S3NkIt7/ctRnhvu2Mm/l29Fvvvsyvcrs35QgMuPKvZqKvw2kOvwlszL7/fj8KiEhQWFXOOMoLBEFJSkpGdlSmr++lWPQEV3tZzdTDSyEf1IUqr0wMPH9OHuYpX2PBafochvfy5bPy9SLDxe76/nKxBmxYflRjYRcqum4s0WFzhzIzqUmzNxRdCXL7XVybl1Mgp5POK7HKzIrzVi7BfIr912UKMHAcDxjDI8sEVYyVjtvnf/J2MWsWywvwu/w4FI6kK9V1kwk7/VXjt0GoZbVuM8D782PMxiarwtowLjb1U4d1yrPjHl4LLLwouxTYrMx2hUEiWrQ4EgigoLEJxSanksFN+k5M9LWfQm6inKrxNBLq5DhMOSaUBii4nIrEeK6cu833i8SQ3V68a7Lh2ZddXsE7ybaPrBEcLriyq4PPCUyG79RXeupyo1NQtyZclhB0ulxFglxxhY8U1ufuQyXNJSHLIP7b4t+TlMqWiGTYV3maA3siHbBHCGwgGsfC1xTjz1GMbGYe93WuE1x4vFd7NvIqLS1BUUoqSCpGlzDKaa25Vc3hNMab8cqP48jWyRKduUOFNzIuAj9MpuiwxZS48YD5Sb4gc3nigZqQxtKl9olU4JBLLFIboqC7PgSunyYIPmW2NyK6/fItUjfpEeO1yCpQWIFhaAGd6G2O54jjfjFSOAEJB5jMHpJQZUrPjvNfaPbsEWoTw8qRuGjcNk24bYvf8GrW9Cq89vK1deL3e8kg0Nz09DZnpacioJlWhpklr5eU+ifoyKpyRnibym5pat+VG7Y1g/LZW4Y3fsalLz6TCQGkhn4HLimCsJVt1SwThrU12yYFVJJiaEAoG4PSkwZ1VOeeYUW9KGhd9qE52ya4phJfi7S/ZJBH4eFm4g+cudXkptSFDbCNfFf/mhDejSgW/G1+hOCi1Vpf3jr6megItRnjnPP4iunbpiCMO3rdSJKw5B1eF1x791ii8zMstKmZebgmcTodEZa1EZq1WaeC+Kb9MfWAqBOXX6XTaG5gEaK3CmwCDWBGppOhykhQrAdRUQ7UlCi+Fy1+4URa2oJxWjewyMhsRXL9XJmaRgcOTErM+sBHZ9cOdWbPsNrbwcjIZl2NmP1nFwUp5s8a8YskxWG5Euim3TKEQkY0S2uh/x+qLpjQ05gg1z75bjPAefMIFMQl9+ubjzUMOgAqvPfStSXgZzd2Ytwk+f6BCctPh5uQai5tV4TV3R7EuKDRygWWiW2YGGEVuLZsKb8sd6cjSr6WFxrKv/LIwQaqlCa9MtispEPnybloNV3IaXGltkCQ3qGEEKbguz2bBddf81EZSPVhlgpHdqJSG6q6Exojwsg4uI7qslsCIbl0ntjXE1SvXESfteYulP46UDDhdyUa5sjqkfqnwNsSoxNc+Wozwxhc2ozcqvPZGpTUIb1mZFxs35ctM8BSPB23aZNeplJ5d4Y0eCeYFFxQVg31hxDc7K8OWbNsb1ZpbkwMnFnF1xFCIlSeCCHEySzAok/Qiv+OMbU67j1SRMspLcXa9TFyp+INV3X9nZSaG3A8YcDkmTpyILl26NOQwxOW+eG1wtS5GNGXpV5u1XZtLePl4nBUSIuXPTJmqKAgRS64ohSG/D0l8Leu78tIO+KQ+LhxOqRjhymov8mqlEgFzZJnuwQgxZZclxFwZuTWOc1FRvrHSWh3kr+qOKZesyWv2wWGxskRjXIgiuWXFIv+c0Mj0l4aoyavC2xij1bz7VOGtB//WKLyBQEV5mYpyP0aJH8NSjP/e8rv5s7KyEuTktEnIR+6m6Pp8fricTnCiZVpqCsq85VJPl2kMdkqL1Ud4zUuaUsmILyO/Lpez0cqbUVpLS8vkXNnvIOtmVsgtrwmmWPDL4UgyvidVfHc4JM2D1SicDuM720dfR6ZDbPHzqGvN7w8gJztDXt/St9YivFJxgbmeGbl1ntRUH+Fl5YBwyCeCZHVZXKYOBFkOjRPoUtMhddFYOIvXonmjFvX5J597oSCkogLbcUIUwnBndYA7g5PUUkQ+5fG7rwz+4nyEygpkfyyX5k7P3rxqWNRSwEwd4MaIKpcNlpSGGmRX+l1agJKiTbLSGqPnXLABUY/4+TMrkVCWCgsUU969ckymXDTHxuPzpoPRXFk+OCW9wVd9U+FtjpFt3GO2GOE94pSLY5L44JV5jUuohr23JuE1xKkIIUblKqJuJhojYmBE4rhFRxCif+YtK438caEEJns8UmKL3/nvhog8NPXFQNFdt2GjlA2j6HLymFFaLCMi9pRBsispLauIuGbK+da0NYTwRu+f/WQfmPNrRH3rXt4sWnBLy8okfzgtLRWpKclISU4W8TTltjHH1CznRuHt2rlDUw99oxwv0YWXsugvyoMzJU2ik/XZ7AqvSFJ5KVirlqWuHK5kI1LL69VczCBGCSwKKSOqEkVNy445gS7WeUhea3Ge7J/pCs7kDHDls9q2MCsxlG6Cv3iTsVgG0wVSMuBwG0sAm7IbPVkt1j7NfoORz7RseANhWWmNkVBZXrhiEpcsMlExeYtlwkz5NXNcKcc8rlReKCuCK6ONZQa1naud37OPjOTKmDldkgcdazKjnX3W1FaFt6FIxs9+WozwMmJmbgF/AB9/8S1WrlqLi8/v3Ww0E114zWoAFCWzBmx06Sy74M2UBkYeuW+f3y/fy30+SGTU5aoQYLcUk6cM28l7tduf+rSnQC7/d7WIblpKCjq0byuMahJZRm7Ikl+UQpNprH40tPCax6ha3szsQ01iGktwKfaMYKelptYq71Y4k2eZ1yvCzGusumgtbx4ouhR3cwIg2zOH14wMWzlevLZJVOE1H4HzltiVmWs5qlrTOFkRXpbqksUYyo2bba7SBXcykioWqUhyp0g5Ly4vzDZMoXEkpxppBUy5YOoA+5yWbSuaSdFl1DYc8COEEJIz20u1CbubGclkPyQq63Ajuc1WUo84HOAqcVumMchrSgsRCgUNQee5WKzSULWSAReNMMt0sf/cX1NvUlPYWyL9EMlNzWiQ66em8+CxArJIhm6JRKDFCG8s6Fx9bdzNVzfbeCSq8JpSRrBmNLAhINeWw8uJV4YA+yMyzBQKiq8pwGZUuLkeX2/YkIcVq1ajpNSLbp23EtFldNPuxkltZmmxWDcTjSW80f2srrwZBZcCWlrmRUlpqYyJx+NBSsUNiNvtqiYn17wp3Rzxryn3NuDzobjUKzcNbrfTSAEpK5fUCKY+yFOAZDeS4JDroai4WG6AWM6NzItLSlBUXAZvmRcH7LtbnXKl7Y5bY7dPROHlI/Agl7ptyEfg4RDKNq1D0F+GZE5ycycbFR2cLnAiFQXRzHOFyyOTl4wVvkJwcBITS04xFSEUNB6HM+/T4ZRIp79wA3z5a6W9J7ujREet5NXKtREOy4ppMgnNVwYHV4LL2UqqTtR3k0f45SUIFOUhye1BStuulRZlMEq5UdCTjAoXVdINGmPSWn3PqcYbmrIiqfxA9hTdmip2NEQ/jNSSUqn5zEvDkWmsUKdb4hBoscK7fkMeRoydhjkzxjXbaCSS8EYLWH0feVc3ILUJb6zXMWpXNRrMf1OkzHSIaCFujIuB0ee16zaI6KakpKB7l63Qvl3lWpj1Oa6ZLsJJXJxgRv6Uf0p9Y0W4KbVMBeBx/IEA1q7biNVr1qKgqARulxNtsrMk0soILqO5VfNvq8vJZbvNq7ka+biVcnIrotyyslwojMyMNKklzOi+LC3qcMjYUvg35uXLV7mvHB63B21zs0V+/ZzwFgggt00OMtJTpZYxj8tzaulbIgkvo5CU3frk6cYaTwopJ4FBorC8IviouwiBok0S+aT8yjEz28KTngMkOSO1VSm10RvTFMxcUJlEFgoaE8dSs6R6ghn5pUxSIEW+eNyonFpzf8wL9hetl+MxyspoaPRKZ2Y7/i6ymW8WY9HjCmc2vstPmA8f9EkUmtFi9pHvD07yCwXK5f3CBZN5Hk53ish5dRPI6iK8zEGWGsD+MjlmU0xOE3Fn5QdXspGb3ACT7Kr7XOCxJM3FV2akSSSnyTLGjKZrSkNL/zTdsv8tRnhP7DOo0meEx+3C0Kv6o9fhBzTbqLR04a36iN1cvauxgNZFeKvrCyO/FFGmQ0SnRZi5wdE5wpSpSn/kKiSMgmQImTHRzvwyf8781Ly8AuQXFiIrKxM9unZu1FJfPp9P6vVSrvm4lrKZk5MdyYk1cmMdMumLf/TMXNnIz6tM2mLqCKWW0dFoueV/8wYnyBnjQUN809JSkJmejrTUZBFRpgwwimqlvBnHwcynLS/3ok1OjqQbREe+a1tZzhwfRnu5L0aYzZQFjjX7s2HjJrg9Lrg4q91lRJkp5UyDyM5K15SGerxxGfmU5V4bYJM83eI8WSSB+Z52Nz5GZ+Q1eoGAUCiAUGkhfMUb4U7PlUlfAU6M9JXCKWYYNkqaJRsLVIQDrGdr1LQ1or9GBJi5u9HSK5OwSpibWihRUzonj+tKzRDBi57tH5Ejpke4kuFK31wnmAIuE8hYiQGAp00nmUxVdfMXbZSJbOZdoVHpoWKjfHOVL/mq+O9w0CjR5nAjye2G0+mRiLSUAyveBH9JoSzL60pNgyM5C+6MnGrzWq0IryzQwEin34swv3gT6k6GMzlV8me58WbAcsTbxuBT6OVGRibksVybC/6C9XIjYKRzGAtCCA9OuKuDCBsCv1lyzZsYiYYnOSLXDRkgLcdG77VpSyDQYoSXf/DMjdElPv5s7q2lCq85gYlyYUZza5tE1RCsG1J4q+uPKcHRMmyWxjIjjWZpK1MajX8bk+34RT6UXH53O53o2qWzpC4wktwYGwWbE9pYTozfORbFlL5yn0gtI6D84n+zLScNmlLOagjys4ov85z4byNC7BJBDgWD8AU2507z5x6XS1IVeDz+0ZV+lJSKSCYnJ4MCXlxaJjcUnIhHqXRTCipmojPdoaioRH4vf7jDYaSnpyJvU4FEjQO+INweN9xuB7Kzs5CRZqQiVI34mkEuHp/R5KyMdNkf3/OmePP4VatcsD3F3VteLpHeuvwBbIzxrM8+myXCy6coeasNmUnLqrlqgtwYUkYpbSFD3uTGkf8dgL+oQlgycozUgRjbFrnW3GfUJKqqiwNQAP1lBXAlZ8iSunyzUlq8+WtFPJMzjYoH1W2G+FYs6OA3oqIiv5xMxln+XBUsOjeVEiyTo4pEjpnuUHVyFEWaohypn8uVucrLJf/Xk9OxklSzX4zQ+gvWSWUFijCjs4zKyoQxRpYZuXW4ROxkxS+ZVMdJalsylGg2c3orJtpxHEK+cgS9rHzByXVBkXEu6+tmbdyKqHZM4a1YrliiuAGvjKXcHHh4c8DFLiqnYohoywQ+3hRk1ikvueo4RSo/BMrh5hLLHso10xk2RcZbotu86TEn2AX8mxeSILMKdpTiqn2W11ByfaXyejNKL6upSWm4chFdjgevJ/MGKZjU+hbwqc9nV0t4bYsR3knT5+Cm6y6txHT0pPsx9qarmo1zSxJeyoGRm1ssf5wouhQIs6xY1ZJiptREfzdBU/zqsppXUwhvrIvBfNRdW94vZW/Fv6tR6i1HanIy2rfLRU52puSykh03K5O8rFyQlEQKIwWXcp6elipCTdksLimTR6vGo31jcp/PF0BGRirat22LtrmxIw/mYhdr1m2I7IOiS0EwJpmlynH4xai3mZZgiG6Z9IURUz49KS0rrxBfD9wuF7xeryyiwRQIyrFEfpAEJhGkp6VIWgGl3LzOGBkv93mNKHx5AE6XA21z26BDu1ykpqREJqaZFT6kJmsgKH2g5LIfpuRaEVldeMLKVbdlG4pEef5/SM7tIhE9H0tOeUuMnEk+umfET75CDH0akUlJEXBKRIwyxchwwFuIUHmZsXIYZUlqKJsP6WP3LTKuvNGsWM41ukRY1YgfRYbRX1/ev2D6QNDvRygcQFbXXWxNKKOsUVh5TrXV/pV8YJk0VRwRX/OxPvdRvmGlsdhBeSmS23ZDcptOFRUPDG5k5i9hubFiOJleIe+eJMnpZX6vCG6F3Nb06N6cvCapFbJgRfYWUm1+VlPeAsX5CJTmIVhWYtSmzciBN5iErDbtjfQIieKWiRw7OV6M4rpTLecayyIPZYVy7hLxZQpIHSKuzNHljQWvG95UcMx9RRtEOmsqtSbnWiHAUnHC/O+K73Id8WYhxFXWkozrQ55gsAyctyI1xGGw9yQbolulQoemNNTtMyWeX9VihPeqG27H/XfeUollvytuwhOzJzUb35YgvGZJLEoEhY9vfpbP2lwjfXPuWNWyYrHKizFayElF3KLTBsxJZTV96DWX8NZ2gZgT0Xhe3Tp3EtFyOB0S4ZXoJR8ZulxSkq20rBRlXp886s9tk43MzAxLH/QUOjOCS6ljXdz0tDSRTwotH+MztcB4jJ8u4sjx4nE5hnxNXl4+1m/ciPyCYhFx5rVS9Bip37ApHyVlZfC4PIaApqXIY3+XrDIEOJ0uJHvckRQAj4dSa0wY47Ey0llKLV2OZ14RlG9KdGFRETbmFciCGn5fAEwxYD3Rzlt1QPeunZCaujlqy/PKzDDOITr/mCkNzLunjPPvIiPTFF/+DXKw7JLTqBOckZFm8I9RU5cMzQUr5HvFghYUZUafE2FrygivSGPpJnjadIY3b7VIgIuPyxmR9HJZVi+c6blGHiyjhCK4FWkPHIsK0TPyVXNqlUfL4xMOS0oExc14tO0WoWQ92yCjfrmd5foIhh2Ay43Ahn/gzmyHlLbdGjXfM8DVzMpLjGh2xeN35vjy356MXJFxia6SRXqWyBVr6zqSU+DJbGtUFohmWAsQY3ncEskjJgMRXZuVHiR6WlogpeC8xYVwhsrhoVxm5hrRXwsr2tXUTd40iPiWFUoKCCs5WFn0gTcQHGOKuxldZ0SXE/3cGW3rnSssJcxKjQU5GDzgtS1LNLuNleyqprbEOkcVXsvv2BbTMO6Fd9Gb7+OVNz/An3+vxHbbdIuApcB16dQBd44d3myw41V4KQNmNJcyZUquKVN8jF3fjceIlUNrlhYTAZZZ9ptLi8WT8LLva9aux9/LV0q0qmPbXKSlp4rgsc9meSymzjDPlZLH0niUUv53fmER8vMLKx73pyEnh5O8UkTcKHSMhvrKy+H1+eWRu7esXKKglDk+fqek8RqmzJppJYxqMieZfWBFAiNVgJNSIGLr9fpQ6vVKusO69XkoLC6SfNsc5hd364JOHdujTQ5r7GZJrnF01J6RWW+5Dxs3bsKGvE3ILyiSCXJt2+Qgu0KezbQWvo7tGWmmEMsEOi4QkeSQHF9WyuBJcj/cn9PllMgtJ/IxykvR5WvI2IhOb/7Oa9Hsi9/nQ5pIf7LU8C8qLEJKagpSU4xSZ6bQkr+xUluoIjLNBSucsliFsagFS7xpDq+d9zSjapRIViEo++9PwJFkSAafLJgRR4fTmLDkY+SWEpdTsXRriSHDjBxykg/bmY+YZTGDuj8KNqsrSNWEJIfxGLoin9SV3lbSBZgawHJi8KTKNeHxJKNsw3IES/KR0q6H9LOxNj5q9xdsQICTnFiXNysXnpxOxjkzrScUkLQCX9FGiS6mdtzWVnUBcjTLqHF/Un+3gnF9z6moIA8ZWW0Q8pVuXpmMK9xRousQna3aH94QSC60ywM3I9CxahpHTUiTmxmmafnK4CvcUCG/vFGo32ZUqyiUm3IXF6VgGboYfantKCq8tRFqeb+Pe+Fds3YD1q7fiPsefhqDB5wXIcw/rNv06NKsqyvFm/CaS8pSxIxorlFWrOoEosa8TKsrLUaBCQUDyMnJiSw0UZe0iPr0naLKKOOKlauwdsMmpKZ40KXTVmiTkyWP/E3JtXMMni/l8b816+EP+OWlzH2l5FHUeHMXiJ3qAAAgAElEQVSRnc2IZ4Y8zaT8yiIVLpeU+WIUluPGaDLzXs1IuSPJiHwy3BoMBOV3zH91ikyEJGLLsWV7SvKm/AL5bx6PbUtLS6WKAQWbxzInlVGqJZKaniZ9ZH94bH43+mwci8dhtQb+0chgxJainhE7iso83vUb84Qt21PUc7KzJALOKC7H3uibu9L7lexWr1kHvscDwQA6tG0jeZ88H95YMIpuRn2N9IvqJ1VpSoP1q5aRNC7jzLSFsv9+k2oGyW06yw5k4qZUBGBeKSOWPsnZDBbnSYkoT24npOR2k2oAjORRUCUC6XJvzsPlBLgYq3iZubmx5IoRv/L8NSKOfNQv1To8qQiXe0XG3ZltJcrs3/SfTITjI+qqdXhZbcG7/v/svQmYZNld3fmPPSL3pTKz9upVagmxSNZgmWEzGJvNMx4sBiQYIRaBEGLYsRhZA0ISi8RmA0KDMWCwLGSMYQxmsbGHDzPA8H1gEEJSd6tbXV1bVu6x7xHz/c59N/JlVCwvMqs6K1v5pOpc4i333Xcz4txzz/+cq2JUU7MrE7Oho3rQ6Wb3nL4UiUOraZn5tYGsNoVpWkpPpgW85IVLfw0BlZJYEFJRK6lr4n4icUT2tf9++jW88pqFnVVaGdZfhFscvTaGe2HMMGFxBW65HiNOmyR7QRvc7QjossnN4ohFk0zgSJRTUAe2bEe8l1PAG/095aTsed8D3nBHwhZhn3Tpwtn7on/vB8ALiEOXu7O7K1YRXeXsrEv64t9hdFV3u3O9tVi+kLd43AEawBXbII/du3V9gBdM6TPPXLdipWIsqwMIV1fO2JWL58RsHhZ090sUQLMA170CH1oxLfdfPL+m7wGCm9u7AqEUnElmQIBSLCaQjUMCQBF5pDyIGw0rl8oq+JqbnZXEAHYZWQNA1U9gfLEd54LpQi6wvrFp3W5M7HqjUbNazTH86JFhgGFU7ywa6wqEYwEm4/2E873lOhwD6Ox3uhj2jLATA6X7gjMncXDMNn3Ntf1rMNlhna4H5bDTM1M59RX7Mlb8fsOCT04Bb7S/mmZhU3pFlnr5HsZ0UOKZKtkD31exuMmMA7jFbaexhNVt1rV8n5pbdsvSntkFNAfJXa4YranJrnSWzYZraFAkKk2lAGTLMvMrDnThCAB4Ke8J6AIAVfS1ty73A78MPyh4gut5D10xpFO4Fhxe7gJgZeldDHgibS3Y5ey0Jafn5RbB3wuyARWONRu6NmCLPnOFfG03KagUBOC93y+sI84TTn/adhHCsJGHYCKjPXmzUS4NkqdUS3pGkibIl/horh0q7AP4YnfcbfcK0mgv/Qo4Tc2dmUh/Pehe1XYB3aSTadylPjwFvFFH1snZ78QA3j/7i7+2H/zxfymD+n/78z9qf/2hJ+39v/579vY3v/HYevs4AK8r7nGWXCyrw+zxoQHAWVyY1zL1vfJuPWpH90safOKaB8DeYoz2O4/dwYlr3nWB45EG0B8KSqjVBToLhaLdur1hW1t7li8VbTqbFXiaIVKXicDcjPSzLM+joQV0egbRuw24JX3HvLK0DjiFdUT3ijSgVquKvezy+0ZDS/Kci3PyYY7eFhutYqVqmRQRytho4ZgZs26sY922++qLWKZygLwpFabNBOlluCqQPhZPoPfN6vzyyg/8ZvlaLFUUwNCsNy03RWHalFgiJkGNesOy2bTAKiDSMaSdQFYxLcZY2uDdvNhmiuGYEMDs4tLgJgzOKcEDV++ZG3UsOHlGWWAcEE0/nl09IxA9LLCDiUmhVFbfsx/3zdjgPDxvD37DrhmngHf8E6nv3tKEE5mAi2adOeBO4O2aABCwY0q1GmCtxbJ1Y/e2WSIu6zEDQDcqjpkEXKrIDaAbFG75n/l6wHdWnLLaAVBxtlANsX5c36eIOUBTENgNb6OS1mT/1W5JroTeFplDfxDDqB4D5BIgobB0FemlrFXds0R6Wn2nqv9WXXHBWJi16zVrN8pyFRBoleNLPNDsgu7jKiCjuI/iMLTTXaP/vHduVkzoUVnOUfcUyZaMYrRa0ZAn3C3W11nDOeaYSVKztO2eLx67R9hgppGPUHSWzCGhOHq4R7g5p4D3CA/nPj30xADer37jW+yHvvdb7a3vfE+veO1VX/dd9r6fe9exde3dALyAFl+IA3Dr1y32/8zNslRO1Txm++gmn0vJwlE6O6qGV7IIVfc3XBpZAYuumgCc6w80tU1rtzrWVCqYczVoN9tWU1obrgcZMbkLs7N6DZ1fp+tsrACIlYpLESvjR9tqCRA7o3dxrQJW+NTGOl1rNDvWiXXk3AArnE05lwr0tOkkiWAZ2X5ValWBbgq7KHSbnc1ZKpGWtRlLtcvLC5IZAOCwBdtnl7s9PbTTuzasVqna4tKiXTx/VvKAcEEhQBDwByDn2QMAAbR+xVSa7cB72INHWGZ+j4wDMMvEAB0wIJfCM+QKnmEFlHg3Ci8j8MCV6yJN8EzzKJmB997lq6QVGPkHrG0UFwb+LnybaLvGeS7bc3Kg8G2fPT4Nnhj6t9ntWnXjqsAfYIxJmjWx0JruMbEwsAJfClaYHqjpdKzwlsXSWbF1ei9C0wq7xt9frcibkyUBfVT+4+AQ/JOHqr53rCEgyAUa8K+h6wFgAMoUoElPjP0Yv4MdnF0JAGSsp5dl1QP8zN/fQOYP5rhSkDsC55ZXL3KIIQwgALm2+azVd65LWhPHdiuwJKMASrIK2Fy0oTNLAtTcf3XjGUtOzVpm+WKkZXQfYiG4H/jCqh/qVTkLEH/MP5wT7ubmNbxRV/16gRwdLMjm0FixLuUmzgLz+88iSjuZRAB407NnjgROpRUObNlg2sPOHlHaEXWfU8AbtadOzn4nBvC+9pv+qf3iT7/dwm4Nr3rdd9v7/sU7j623Dwt4KVSCneUrbz6+8MYX4jjLqDsLc2DjAAAACMADDgF605RVkLiIA8vVvmP6LccoPnKxrfuRvf1vgp5Fdf6u+16v+v0AH9ie122wvL7vfesYSX6GFc1mnV+qrGLi7qv7PkjYqtVd4RYRw3WKmmAtszadw83AORKgN4UpVDFXq63zAIS91RYAE03oxtaOtVodORbEEnExl+xD32G1BXMqK6xySS4EAE0YU/oYULiyvGiXzq9YOp2VzODW5pbdWnd6UxKeaFcmm9HyO0CWJLDF+Xk5DQwqDPSRzfSFL1Qb9uHDhAbXCPqB5w0LLXZfHrlpAT3pgiNuvpDxmWdvKNiCczx05aLufy8YU2dXV8S+4qWLLpl+8OETYTbWA27AL+ehHQBgL13hOF5TkR7yhAHaX8Z+2GeX43NZd4/hcauivSBtzyfS0Rb6nDGscxRL9tCDl45Vzx/xMYzd7aguDeq/kE0ToLJy8yMWT7MMP+sssbqA0tWeib+zBBvNjrEEDfhMsgQ9YsnYMbSbAs3h4jHpgQVyq64IDY/Tbsu6TDbbTWuVtt0yOkygJp9dOTKwXyI3w8KEVidkdSYJgVkbiUS304sWBiDGA+mFY1dJIGuJUUTagOSiUy0KVMIex2IJAWG0yfXdG0qFS82vWe7MJYvn5no+uI1ANyw2MZNz4I/3HrSqxS1Lz6/KvxYgBmDndfn2HqIQrOcXTD+1GgHwzbqvQ3yNw4PKO0g4GUnTjYXA75cC2ul5QjsmY1Y9sJf/chCa4Z4B0hVvVQcQdoV7PSeK4BkweaCPmWwglzjsxuSlWdmzJJKLqbkjFUdGacMp4I3SSydrnxMDeF//7W+zd7zlm+0tP/DTYnj/3z/7S/vlX/lNe8+PveXYenxSwAvg2d0riN2jsh5WbtwmR4B80RUEBWlgsHICGKGDHdx14Qnhr+HvvasCYAowSbAB4JHvvQVWMhEXIBQYTqfvSPPideWMBwDVg9bezyHw6n/nk7J2dndsdnbOuTco8CDtpAhVx7Yq6IFAhDQFDTgdNG13tyBZQn43rw+quVlsq2YCEN1WRT+AHH0sANEVhMUFrM+tnZHMw4Mx9tMyfblsOzu7SjXjQq1mW0VazU5LDgeAUe4LsElxytYuRusdW1ycszPLS5IleEcDACjsp2OIYVYTPdY1bNsWZkJpgwdvtI3rUTQX3tDC8ry5J8fmlsVwAfRhfAG8wyQBg8YU10OjC0hnhFAI+uz1W5JBLM3P2flzq3qmXI/2eeApYIpkotlSP3rvZkUSN10CGvKJfLEk+cT8zHQgrXAV9LDxPhTDfd8VG97WpAkW3Qzmv1qtwx3Z3MJMUJjnx3J3f0xTyWfmxm7NFdnRJiZFn/GKl2mSctK3owBeluABnEqhQrpQK1lj56blzj0qnS2gU9KAhbXI3QSAa+Q3xf4O0voOOxGSgmZ5xxKZWYFApQGmskFUblNMKbpgazdchCzsaWirbV93+luW+TUJcqa+rL74n1tN/j5bloyhn63LW5aves/zAQ6EOAC4StsC+wClTq1iraorqiK2l9Nnly+6wr0+kEogB6lu9fympWedRZp8iSnsa7dcMEJo84VgTA6cxINCsEOOS6zfAL7BJEHkAFITPGUBvzDcAbBVMAMAl6Bh3XMqlE7mvmdinzbHxFMkdtSiroM37vTKet+Qdtnpmh0oblliauFw9SRooIOgDWmjkS4cUVscdfCfAt6oPXVy9jsxgPevPvi4vfOf/4Lt7BXsyqVzBlP1I9//HfaSFz16bL0dBfB6wIrWFmAj/9QItmCuAKoofSXvwXywA3Q4R5QlKUCGq74HHFC8VBeYlBNBxlVBA558LK2LoW1aE01s2yVy0fYwaHMglXjL8bZD4eVoHhDHVgBO7a7uKV8syNZLgQjZbKA77li90dKSNQAyq+u5pf/ZqSkrVatieWvVmmUyKWu1zW5cv2XletXiFlMfnT+3JncAWEWANccC4Lkm6WmwxxSNoeWlOAsGOJdzIQf0C6AMoM290xcdHAtwVMhmJD2RhjQoCuQY/yzCjgee4QSfpaSJTer+BvWlA4xF50ow51wJaMOt9dt6djDHMKherzqs6KsfMPs/Cg90YbdpA5MLxoGXAvA6YxPwz70wmaJv6D9vJ8ZYYEKxs5dXP6JdXlpasMX5ud4qgXfA2M3npVVmY5VCYyywEGMywHjzKxk+Gtnrp+lbtNee/e7/w+4vtnMa5rIY8E940SORxuWxvVlEvPBhAC8gC31uan7VEqmMwEZt6xnrNquWO/eYmDDAEMA1s+TcGKJsKiyqFi09vxKpEAg22duIwXgKdNdL0vkig+BcrihtSq4E7cqeGcAMuUMQ/wvr2ihsyBu4PzHLt9kHJzRqFckzphZW+0BnUzIJpZnhNkG6Gqw0bggtfF5XBM6QYuCPi5OAA2ktSTNUYMd74PY1MdW4M6h4DoCOLAOJUiIjGcOoTW4OgZRC0oiIHrou6CNgU4OiNn5uA+jpt2pBYBvmW+fN8HXKAd0RlnBew8s9MiFhItA/0YgyLqLuo8JDfISbNZEBsrvzMc/0ZWhyAWB3PsbBv+BnxhAym3tpNefvRwWWciepW2Nv01KrD0S91dP9TkgPnBjAS38CAD/wN0/qDf2TX/ICgQSqz49rGwV4AZu7eefTCsiFaRy38YG+ly9IWwnzCp+1FhT4DKtO9+d0rJcDtmIm2+3Aast5ygKIwkDZm/jDuHmPU68X9r/j92L8GnUrFMq2DSuaLwrEzcxOq7gK71pvOUVbYFjz+aLt7uat3oSFa1m5WhGw5EMwR6IYyTbB8mSpUrc6b4gxs1yGArK0JAGAsk67K3AJQG3hQJGIqa2wwddu3rZquWrnL6zZA5cv2dmVJYUnAGKbzbYAWq3REOgHGKG9BcjRDthSNMJcC+CMqwGADUDIdQGogFMHVF38bnjrX44fFHvL/k4vXBEog3nxEwaAnS/U8yAYRrlYrNj65qbacRZXhXNrIwsQPeDj/Dx/H0EMOGbSQNEc+3APsFt41QKeBzkueL307Y1NeQdT+MfEAZ/e6enpA9ZiPpoaxhiwyrjiOYl1zmV7vxs33oe9zmoG4x+niyiTK87z8Vq0BgPYKu1YeuGswA6gsr591eK5RcsuX1AXCxDvrYvFjLJJBhCkXUVhdbk+LBxFYbCPTr+bdAll6C1rZQHK7OoVFyCxc0PgDK2uggrwYqWYqbxnrcKGpeZWA2CUDZwfuo7BDcArkgqAU9tikjQl2i7hrT/+tweQSUsr5xW92y4XrFHZtQy2ZfMr6huxotIXBx7CrPbk18VK8lmTWXR9e9hN8bVIJwT4Zy2uwj5f1OdAtgCuivzQ/wfaZ2n/g7AKb/MGuwmwDeKVef5IRXg/V7BC8AwGySn6i9Z8+IOeA6l6d2HzhY94CYuNDgoZpVWWa0PJOo2y/jklnks/E3iXbIPkuYChRmZzCFlIlNtwANtNhvR9i8jppHPLqRQVzBFDBnO6Pa964EQAXm/r9OhDV/ThvbWzZ7/8/v+gQIr/8hs/d2wPZBDg9YwZH9TIFoZ5l4YbDTCC5cOTlDcymD6YPZiu/s2FAriIV6fxdD6qPizB+cm6oqp9HW5gmZUv2vburuXzBfcmA+NmMHDor2B7PQPnvufNqFol/tYVyAGYiKdFM5nPl6xYKbto3HpTtlsAW7HIybTNa/l/QUVj6SResU3DlqzRbAu0sgwNQFUwRTKh3wP284Q5sFRJ8VjXVJxGkte5lWUxvLt7RUum4nZudVVFYABhAb5GG9WfNestS6V9itmUQh5wZNgtFFWotbQwZ4uLThISBrdRbbf6nwfn9FrU/WCJOz88+ou+PEgG/DJmkBgAxNGmUjjHvTrWlwjoqd7z1IphIG3x2mjapHjgSjWY9BD/29FExMsfwo4G/h6GrRQwthg/WLoVS1UcmXqMsItEdoEU9B9aPhhp2Fue5+rKsq551I0xfev2pkB3lPN9PAJeLJ9IIwMg+up3wCVhEuHIXLxpsytXIj0Ssbo1CrSCwrEBR2m5GsYRsL1320yFbi6AAt9efq/ErBz+s7OWSGWtUdi28rW/EQiaufTiO8IOdC+NqvxqAcWA5U7AZKIJBRgDFmFV/VK8d2lQ0ElxW0VwrogJbS6BFRXnEYzLQjJtMUX5ujHbLGwgIbbMmbMqQPOFT9ybZ3ZhcqMA/kgdG+wkB4RaRase/SBbzDEA9xCbj0sW4CR+WJG5uSBdzEkqBro08HlSdF64SYrJDgEwFXtcryiSWhHFocJHB3KLGqfSZoelJqo5cVIIsfHtltNgk4ZG1C+BEUeYaPQmPIGvtAe2TJycbt1NnFyb0tau5l1RHTIVJj3eQu8Qz+P0kPuzB+57wPsffuf/kZQBvStFQV//mi+1t/3Ie+wVL/8k+8av+V9VxX5cG4AXMPjqV7+qV/UPQIHN9RrPQW0DnMAoAgJJ+wLoAEAvnFsTA8dSr5MXkOrlggCUTiV9ZT1YMu5aWmDRsZGwsvuFYkHVv3c5KBY1UwZkLi8sKmhBqjivjWO5iQKPoGAIdrRIcVitLrDIHByZA64EOGmV67gElGxzc8/Q5ZIiNovV13TOta1jVmu6dDHsumamHMM8lc5Ydior/abXvqInhlFkuZE3oXgybt0295ayZKDVAnRVykWLxVO2vDhvZ8+t2ZnFBQFq2YsRoxvILLgP+kznDQrgAGn4yZ47uxppAnLY8dTvSsCzHMRQAuYAjMgDqvW6ZVJJu3ThXK8IkfEBw8qHgdMlVwUuXYiIk1H0/6PNMM/IDniOFIvhuTtsCxeHDd0p+PBzNm+bshajDRS2nVlePHBvLoIYnfCugPYLHnlwIo3xsDYw2YUN5/mNYns/3gCv/HQTKS31Uv0OOxqPJS01f5CtA7ylly4ETNqdvaxlXBhQktcIIMjNqaBLv6cgKVjilxwANhQmkulUoyYQo4KtwOpMBWHozAOgBYtJuwiL6FjMskE7uBb+qwAj6bsrexrrANresjcAFTACA4jeu4HutiyJhNhDooVVUEu6MCswhChUrYtjBEEGS2ctnp1TkVlcTKGLReZ83i6MdpDaBshJTM9ZMjPjiqKQCAQODYd9Lzju43im0v/C/raa6odas23T88tyoOjfNMkobgn8D2PKw8f0xx4jqUCz7BhbCvrK8trlZyZfUaJ89ZwDGUpb4wv3jris7nz88CT9CpBnAig9cwBsBagZW31yChVayqbPTdaZ+Kg4+XR7XvXAfQ94v/Srv8Pe+qY32Itf+LD9f3/+Afu2N7/Lvve7X2//4HP+x2N/EADeWxvb9tmf/TkBU8nyE64LccWe9vSKCbds4+zHnLCfZf9KtSpQIqurTFqMHq+JeUyltFyHNRYspkIlYqb4WvmkplMsxPUcDnARABxVa1Xb2ckHBW0dW1xYFEvGUjMspmcH++2k+hlItJf8wRMdmy9UrFgpWrlUtRZFVPWGkrdWVxbszOKSYnWdpKLhPgzb2IC5xC/kB4BxlvMTsa6trq6qHTDTADTuEyDjPHFhZKcEjsWcKtYWlnNaLB9FSfQfcomdvT05KxDOAMDyOtdh3r2eheWa4fCGezWIdL1C0VqBZ274OmhjAaaAtKzs0lzwA/ftGGn0s66Q0IM8L4+QxZfkAzO9SRUAGn0y28KcS41jDAISD8tcc30cK7gP2H3PSKMBZjVCMpEgxa9/LG1sbtkTT13V64RvuCLAw5vYR2F7P24AL44VuAYoPrdrFFXxxkCUK0vw6F5VQY+Od/uawhEAegp9aFNR77SqsgRDX0khW71krUrZEryndDsWT2RU3AVbqhCCQKIgl4BYXB68rKawBC1g0mlbkuAEQGIq65auKTTKb+h9L718UUBCBVZ6rWLN/C0xyd1YUseIUZZtXTsoxnJBFdZlKbype4ynAK4seSdVwEbbm0Qcd9sunQ2mDvDL70lFoxhuZnHfQsv74nonAcBvIulCLYq7Vt+9ZvFkztILKwJYAu4ssx8xAexevcdEPi+rMPWqlfLblkk4KzGKAXsTk9CJvH8xE5J+dpXnB4CENUc7TeyxpBAhFtbpvguSteBWcdQQCGmXw0VrSqwb/V5CGwG7kmqMCR1hZYGJHs8bkMsKhQof+TuYi17cGflZnO54rD1wIgDvr/7Cj/Y66Qu/7A322+9/97F2mr84gPf6zQ37iq94tUCBwIWqz7sCqiz1SnpQRcvKUnDFdvNFI4qVQik0pYBFWFpcAtpdZpXumHq9Jn/JhM6bEiBWoEQMsFyy/F5e7Ckf9JWyi99Udngi4SJn0aCy1Bcww8gNKBJDB41fLTZdLENjr9WhVF7VvWbVSl1V+XjZUryVSaVscWneziwt2pmlBVsmGjiXEYsKCHdMNIECcWmF8Z/l3q0bt0UidedmXbsBdaWi3VgngADAPGuXL50XU4gvLcCOc6niv1A8YIflvWR94EKYnSyXcTCoSqbhNLgpFZkBKlnydFrclNjn2WnicafUdlmfhZK+JnE8OOzg43rbu3sC7U4be5CBDcsj0oF2eFCIiPemhcVG+gHY53xh2QITEOzHpF0OpDHDVitVoKdJVdNppeOMDVfYBqM8qA30r7dZ4z4Yy/19eOPmLdvZKwa67OmB+0zSlxub22rjILb34wHw8kHcyN8Wq9os7AgwAkwBdXIHqOBj29AkGOCSnFnuhRlQGCZHAvlXo1eMW7eBbjZvydyUpRcvCDQ69paUNcIHStZpYp0IMExZlySx0pbAQZICLBWcpVx0LI4CiqktOmCNny6pVwRTiCXGPjBuXVhWLS+xjN0QAywnh27TkgBMbKtEFiR13W4c+0BYXv4PIHGODTFkCegvY0nLzSO/OKizF8tYLSoVDrcHwJlPRNt3Egi0syQSbj/rNKWBPpbiPwAfrLaW6YMQDgVtSLcegK5glUwyj8BJAqs0GHHdo1hxN8kQYGTZnknA/IoY0edy2y9aawkAe/cH5A9MFFjWl3NFoyrmOxZIVbz3MP0mTS7ylZCNHcVmnSoaZWKUcVK4N5ZhDkwTNAHjO3eHTRvPt1XYEhiXLGHAJsKn1RAD3di55aQlTNbE/OJmEUgpEslTScNzOTifo2vd94D3H33lt9iPvv07e93xrf/HO+0nfuC7ez8//MCl56ir7rwMgHd7t2CvetWX60MdppEPGz6US8WSFcrE2bokrN3doiXiXWt1uyoAQ6uKjRNMKouEANVsJhsEA6DBjTuGlw8S9Jn4w5YoHNvRz+kEAFkmPUoSo3hsKkNELZ8vvPnGBGpbpFuJKYzbmeVlW1le0DWu37ytwq+trR3L51laJy2sa0tL87aysCQgurKyrFQy+eTyDx/bWFxhD7gXwCQCVmUvFixp46WKlhMwDUPsrKhwfuhYrVq1y5cuim1G0gB7CAhUAUbgo7vvVNA+AH4X5vCc9eZrd9qvcQ5v3wUjSLsUdpBMimlmklFCqlEs62c+nHAb0IcwrHoyId00/4Y5Hhx2oHmgyz0uLcxHsqPDNo0+GgQ2uT8HnDsaNxQVwo5zv/1tp3CNIsa1lWVNxJhUMD49wJVel+x5AHYqqf5yOvJZOTZE2VSoWChqkuECK/YL42DUkUPwzBlfAHHnwjC4eG7c9TzbKzeJkMb9+Q54AZIwV+hX6zs3tESdml5WKEO7tCcQ2yrvWmrxrMAiH/gweL0iIqWrOUBLiALnA7ji2tCfQEYRFGBXHrXVogqu6nsbulZm4axcFlSclsHiLqaUsY7cGGasUctbHK1mOhtICWIuzKVZtVad4jV5J0qigLaWYiVnZ2UWCyr0BZhglynC0vI7rKRbJXORxMGKmWzvsCWD7m27pWq0qyp8QgfqEs/w16UfFDqBpVdog3Gub1y15OxSz2IMBrNR2iUdw+IpQGlXAMkFY6BvT1uXdomBduER2gR6nee4Y8MJ+qDYLRFYxXndblyOGYD2xOyyntXdKhrr//txxXDO+aBcKojAiKkQwE1CxPgHr8OkAwAtnbMkcc7thoB7cpaxlHHPs6/vAKFMotBqe3/icX/DR31dk96fVFsAACAASURBVLFKQeAUzTerGgqjIIo6FFW870kMa+tdOyhQa5ohx2HsZ2aGuoGcaniP+qTuv+Pve8D7+V/6+pG99ru/+p5j61UA71PP3LC/82mfpmIqirj4QId67bQ61ug0rF5tyhzde8BSgLS2tiJ3CQBQo9aw6ZlpxdMC2PZIAStVtLzS7HQtDtitVXtgD3kCXrEz03ha8h7btWqjLkAMkAHwpMV2OKAMk5wVMxzXNRqkk9VqVmk2bHZq2i5cWJMOenV5SWwgYNBF59alM3ZBDY7xc8vrXYErWD3vJUsbPFAdVfQQTlqD8YbNxRILGCubtMDbl37hWv6cky6Hh627AEiAsHC7uC8FTlTchIQiLPoeJr4Gs55J2crCgq2urUhKEbZi85KQg1/39dC93wc6MNhm7gcGdhIWGa029x0GvDDASCH4PecLO3f025vxvLytGCD/xq11pfIxkfGuGv7roGe2s7snqzKcEkbp0cN/fPQrz5RxwTEAUuQjbLDNAHjCPBxAdvv4pLhJ/4g920v76I/nM+BVQRaaTAUbNAVAZC2V39LSvbXrLgp47ow1tm44K7CpOUvPODAFOOb9BJ0ixyWm5y0zt+YiclsNFYg1SjsWgyVGH4tuMjNt8dy0xeNp6XszWJ7lZtUOMYCVvEIcGrs3rd1oaGUHgJFbfdhSgV+tLMoaNTG6nAt5hHMsKDgnhrRjhgF+Dng5cNYs7yqoQIwjsbHYmqnIKCm9sP9eE/4gac0zd3JyAMAFhUqyEQP8MkZwiwgmA7B5gN3a7adlgTbI9op7xZIMIOujiVUcFoRCaMk/YKP5qoKzQCccZTxTpFcvbGr5XR7ec2csJVb+Tu29wh9wvWhUFcpxUIMfcyy67hlZ3H7YhLT+9FkypXqB6el5i2slzutaYXZh0Z1cxctOYIDl/mBx6zYrlsAFA41uPG6tWlngUvI9AGOawBg+jLxfsgP9PT9eOiMUKBOlb/w+AqdafXTxzir0o60QQfWSVjIAu6mpRXkLh1l8abblSUxxmrtPxhV63kHFiGqzQjrwWG5aN/3cMvCT9Mvpvofrgfse8B7utu7+Ue/7td+xX/2//5OY0m/8mi+zz/mMTzUA73/8/T+xl7/sZdKSsnxVKJdUyOZiCEzBCBRtAZ4unF2z6Wk0uEmxMC6AoKA3EaVm6Q2DwAXHSNYadYFYWEEKzWC1nPsC76vOSQFmGeAD0wfQuHnrtvOx7XR6MbecI5nAp3baZmdnbW46YxcvXrD5MdX0LL17mzM0t3wPW4PODnbRO0mMs0zzTwPA22p3Aw3oPkiGbXShHBScEDSwv6QOaznIYSDqE5Z1V7F0IKRjWMGWD+ZAX721s2sbW9tyoKg3m/qQAUQKJCKdQDtMQEbwnN0HS1BMFjBS8RiMatsuEd07ImRkEOCs12rq34WFeemZYXR59v1At78fPOj0Hsoe1AKcJ7X7gv3FOYRxE5Xt3X/WVYVRwPD6mGLaxPlwcuB3vMZzZ7zCANPWMxPYDDKZ4XyAeLTPXvISdWzcj/v1+/Ci0UVnCDiLZ2esQ6EVALjVEOOWSE9bN9YVy9us7FoCO7KV89au1aShhdGNkQ42NSeQCysGGwxo67aC9ymW7eVikLTs6oO9oiXAAddFQgFg9gAWQEpFPowyYBEvW2ldE0mrYTkWaGOREcAGe1eFBsvk8cQB71cYyGGV+AJ4AvkdaUIViyxpgPvXwve305ZkadgmhhN2rwMIdnrlZnEvKOTK2tyjr7D03MHQi/5zAZ6RXnBNgO9INjYIWpAeVNIubzcWFACKze4ERaeugE6rco2GtSoAfSLAk5acXbbkNAVUsOEQHl1LZrA0m1I7BASly3besZoEaEkeUOhsvaRpDulr8UHXe3WvGBHmd9972AFLD36DCUanpfAdPfPCljTbydkFS86uWhJ9rApoXUGgI+9d7LArqo0HccSOmZ9k8x7G/E0rwc0XTxKwQfuTaY1NPgwpOOQr8prkzILFmSD16be8DpigEACvS59zEh8HcOkHCqdTTqrDWCP173R7XvXAKeCN8Div31xXsRzRxrBTX/+tb7X3/8t32a/92q/ZBz70pL30ZZ+q5XIK1WBw0YvmslO2trokhhBta7UKAChLQrCTL1glsMhKpNK2vU2KGtZOzneQwivsv2ZnXdwqIBF2VtIECg6ScckE0PiSNMTMHVAnJ4TclMAwAAeLK2QTqUDf5kErS/loWQGTUX1OvV4TEOQsy1w0MNIBpXfFYNkosmOZ0QFiAE74ZwAvsgqYP954RzHEPp0N5hIGGdcB9+Y6/oFNYq0zbl9kGPPz87qncBADkwgmLMg5kAL4mGZAcH+hGOwmHwa4G4S3UU4JMOC4IgAGcZfo1+iO74XBezBpASRGtfviLLC9FLChnY3K9vYYGhwnsLErlDReePYAVdhe+sMz07DJLCFWqs5/99KFsyM9iMN3B9u7MH+QxT9s/xz3cR7wnl9btdK1D7q4WjSHsIqAjnZHoIZqdibIWhnBmH9m0VILa9ZC3lDcls5UC+z4zzZKFoulxdBlzlywzNIFgU7+ZmGI0QDDvsqua++2ACpL+JwDXS/AAgmTc22geDZhMewHly87n1Ssrcq7YiBdmhr7dSQTUGEbEoG9dafpTdB2z8I2A3eYjjTBcmkY4CAg2yukFUTtyprMVdJ7WzLeB8dtHItfsCr3CapJJK1NMhn2aYRPTC86L98RfrSco83ko90UsynAGdyrgHggbYDMcIwvUbtBUIX31tXvkGSQQBYw294tJwDJ9eKGNbauKxUP/XB6/qwKCKW1DZjuXppaAGzH3T+vD7QlO/iG5MC0AjjcpMI7c/AzmmMnk8GGrOwY3jSWdNMC10fdfEqd3EJGpNRJhrK7LtZW4w0QzLNBo96o69m6CcO8EjIbe7fc5ysTxgDcOucG+tOB20Hx2qeShqM+0fvv+FPAG+GZwO7C+H3z616tvb/nbT9hX/R5n2VXn/qw/flffcS+6Au/wLK5rHSuLFtn0uhGO1Yu45iwo2VoHAz4EEhlXBEVQKtGpXSrK2BIodja6oqs1xKxuGWyGcvwZidtcEJhD365XLP+tmPL0EXy9nnxwjlbXnKODIDJUUDWOzIARABqvgJ/kHMDQJf9ACqck6V2wAqAFoaOsAvujUQ0Ve3PEFLgNMg+5UzJWom4NRp1SyRSvWInz/yNY4gB2UgPuNdxWyS7rZ4l27izOSYdsAYIG9ROgCmMPOCXQji+Ao7pVyYuYnBUsFe29Y1NW1k5owlCvyTCe+r633NefH2Z+MB8c/0oUdTj78jtEdXuy5+P9qDFPQzb68/h/aYB8nzPfWFfhh2fZ/EZUzdurtuNWxsKbLlw/mwvXEYSoIbTIIe/Mj4uX0S7engniKj9dq/3A/C+7Z9+p2Vu/rnV9jatUymaxQFGSUtkXbwv4BD7qCQFXumsJSi4cbhTy7etyo6LA56akyQBfWWMItpa0WlTWa+AGZ1dtszyRUupAM0ViQFcq7efFsubI2mKCSxFqXG3fA3QAfR4CQDgo1lwhWwc4xlYWEkY3WZpW6CIFDgH0mAOHRPpNLbOdkwazFrRWa3lZu/Q2tLvblnf6TcBvmhNu7G4Jv6DNucBW7BWcUfsNBpTisXQffoNUImbBfePdpeNCUAa14gsdQN3bmKKayV3L4GsoRcQcYgxKDtJNMK4HzTrQSQxhVRpaanpQ7GZyB6mFzRxOEws8FjAO+Hgdow5IRKu32BWKe4LF7RFOSX9roldp6VnJFu0IcwGkxbYbaKy+z2LHfPfVJ+1SlvWaRDP3pIvtbe860UuR2jYKeCN0EknbJdTwBvhgf3kz75X+scv/5LP194//jO/bA9cPm+Vvdv28+/9dbt46bIlMKjtxq2hKmhnC9SKEambUpoX4QuKyaVAiA8PWYslrVJtaimXoiHA5ZwY4qnAlcHbiDkNlCtRc7pdCpeRJFBs5Wy+nC0ZQEAyB6WI3Vm53H+77O8CLBpB9LAD1vwOUO7BGU4IsHvTObfM4yQVbumKDVmCPw8Fa1wfdtq/jp0VwJm2ck7+nZQNoIW7BlIUXAvGbfSf/JMDOyX29765JNXB/lCE5/tGLklBP3rQxjUdi+4A816hpFVBQODd2nj2nBc5QNTnwThgpYD2T2J5Fh5n/E2wwf5TwIlm+tzqcg+w+gjn3b2Cbe7k5f8Mw881GdPIdxhj/isTqle84m/bhQsuWewkb9/y2q+2rzr7IWvF09ZJ5qyVmrJuYsbagNpExtoqgopZrBu3TmDPRdV/J2bW7lIUZdbOzFk7OaPl9HZmXsVXAN5Et2mxVt0StYL+fjuWtHYQk9uOJS3Osn8iZc30nMW7bUu2StZIL1o7NaX3n2SrbK3UrJwT9B7QqlmiWbZOKiv9b4zl+1hCIJSz83obizOq51NT1knc+Tcv/S7vbFjEcE4KxVosVbetk8wOPEb7tWtmFL4F+n//zGmHO0dF7YFhpf3tOKltB5eHurGUtdFpcm/NonViTiMcb5Qt3ipb3DrWTuS0TyeRO3rqF/KNoKjN0JW2Gxbv8K9pbZ53PGXdxOD37BgMK/vTrm5L98Ozpa+jbgA4v9oX9ZjI+3Xawb003GcVYzWeUn8O3Ni/XbdEu24dWP9ExrrsP2SLtRsafzwPxkWUzY/HofsGOmP2Q7ImKzIx74zlrv3Pr/7qKJc53ecE9cAp4I3wsP75z77XVkOA98fe/a/swSsXBXh/4X3/3h668qA0SsBRPHRzYveSArlKgpTu1jkRaFke0EohW7ttq4rEdUUKgCRsw/gZABoGr1GXkgFHBFpwPCDUg18XLzt8czrakhVKVE+j9805BwkK37J45DpHgyibB7+AIzGX1pUlVSoZV2raSd0A7Sy5z8uq63CJSNy7YnyLZRXlefDX3yew/jyHMFPPM2Xpn+tHmcxE7WfkCrD2nDfKM2ac0v5sJiVJy7DNRVM3NBZh+P1YPFhA6Nw4Nnf2bGVpQeONNgBokQg567yuxmW+WNE1Ab79E4/P/uzPsitXoqWJRe2X49jvq77qtfbVF6+LtcVyC8AIYOtYytqptLViWZIWZEeH1rPDCgJgUSsJcesGMdIAIYG8WNfqqQUVhiVbjokDWAGyUu2SJbAoA9jFHMBWagzL2YCvjlm825SVGK81krNOK8mxzbz8b5upWQHFDrGsAWhNtqoCi42EkyGx6XfdljWTU/vgtl2zJCAbw7F4SmCmt+Ht22lYolO3djyjfugHd8gJYPUyyEaDfZ1BL0A1Yy3uv+cscBDsAuhjFIAlE9ZMuklsQm3knhyzC8hMtquWalUtRjsSOWsxCUnyfn2ntor7oy9j7bbFmGTgD8zPArroeWPSW8c6MUt2atbkfOkp68TCIFdRiq6WQddwP2tSzHOylHXRQbfLliSMgxS7xJxzjRizsUqD1G2cjGvceca9zlgTkG87izz3bHEIiWvs8EzZRyCf5zpK4BuMW1nthcbTuDboddrhwayeAVeiSNJ9r9XA4O8neEU/A3v54H7Na08Bb6R+PkE7nRjAyzLmr/z737WNzR37zjd+lT1z7Zat396wV7z8k+95d0vSsL1j3/z1X6Frven7f8K++O87ScP1m5v2dz/3c+3c2hm7eP7cxBpHX0jkQwLQNFKkdHZtRX+QPjAh7Bkb1TZLBVsEIJTKAs9euuCX5nndaXNLAiRIEmDNWGre3qFIyskSADX8m5nhg8CxzJ6xDH/v30i93lcuA7BQZmoDnsLZCIA30htyn7xh3DEUSkUBc+MGE0CUpX102ThtHGXjPGw8+/5tkEuD3sO7XV2fyQhj5G5tw+y+Rp2fMYLlHnIL7yZB+7xPMLIXgCnSlShs8Ac/9Lg9+vADChgZtm1ubtuN9dsG+4sUgmALwlueTy4N73j7220l3ZIkoHTrceuyHJ+eU/FZp0nKWEbuBSx5p2bmLZbMOdsuwGgOrWdCWtNmkehhZzGWW71iUxdeoGXn8IY2uFncsQbFZnG0j0vOV5cioWbd6ts3XXEbBVu4NAjnxSy3fMXSi2cDuzB3RsUbFzZ7gRX9z1A62uJ2oPNtqrgKeYFS06oFLWsjJUC3Gt4kS0CKwapSiuI4IoyDorVWw5K40LQ70pGyJJ7MsSQ+eHKu1RctizcktVCsbswst/ZwcA81a+w5r+OwhRnXaxa2rVG4rXvAdg29qKy70PK2WoHTQcLiKaQOKRVWeRmHJBywnUwKkHilMtZCl12rWILX8PlVSIKWewIrNvfV/8xXFVs1fCFhW21t7N20RHbOMkvnnLZ4iBzgbksaorz3SAdcL1l9Z93aLaKflwIbttHEh/TS1YLsxJCY9NvJDbu2ZC/EHDdwOqIILR1EODt3Bx/nLPeOMQTOqaQhyhM+WfucGMD79h/9WRWE/dUHn7Cf/8nvF0j7lu/5IRWS3evt2g2K1t5p/+rd7xBw+4Zv+/5e0RrL+N/4+m8QQOz/8I/arv5CIs7FexZV+n7rB6+SGvgXQ+BvGPCDbaN9gJF6rWl7+YLV0InKYiwX6IgrwrMkeWE95s8F+MJ5QgVcs9M2NeuW1Xua4sBflb4hPY6CuGnM7NGw6vPRuRcUCnt2+eKloezoMP2tB9c9gBmOhYyg66UN6Ea9O0DU5zJqPyYFe/miwOo4DfKo81AEeWt9Q+dBs+u3YYDXv85zpFiL4+6mtvfa9VtingkMibIxkUJLzgoAj4Ixhs4XkDtpvzDGP3b1uj384OWxl2bice3GLQW/UJj5KS954dhjTsIOYZcGAC8V5WgcKzcft9rWVUvNnRWwAXB1iO0tblu7WXPWXRjy56Zk9pJZPCdNbn1v3WLNmlW2nhWoSs+uWmb5/L5WMtQpACkcIQJzWWuhHyaAQcu8XWvu3tTeFI258Aoq+gHeWHyRqNhWgZWS2byFmIq0gop93LMaNatuPOPS2QKwq0p7Cp9SaQUewOQq2CJgcGU1Bdhr1AVm5H/L/dI37bZlZxbk5ztO10rhGzHMKlIL9LkUoDXzm9bYvWVTZx8RSDWY6N116wRpYqrgF929b43F/Qp8A1YTLq7Wv88NG2cqvsPdolqyFG3OzgiAqSAx0PCiy+Y5Sd88ZBOgq5d7vsHs1tjbUCgJ/Y4Xrc6Tdkv/fvJSLu7a1MyCJTPIM6Kt1h32b8bb1+l5mXu+gH36Des0LO70jPvdFOpVAV36Gz/mcUlptM9fSyBXY2lKEzv+do6ynQLeo/Te/XnsiQG8X/n677F//Z4ftDd81zvs3e96s3rzVa/7J/a+f/HDz0nP/pt/99v2b3/j98QSftPXfrl97mf9bdmSPXv9tv3jV36J2Ca0uBtbO2K1wmA1agPDtlGcB91iGAT582ipuFpz6UZohVvtXriDd0jwjgl8ILD8CUBgUwFZPCZAA+MLGwtwY2l9AfeEPm1tPwglLlcFditnxNo5Nq/cY4h79lt8QMkn0/2TTVqlrGVp9llc2C8cGdY/4WuzvEfR11GKt7wXLOcYxwhHeWae7cX1Amuso2xie7tdOTKwjQO8+iDrdgU2YeXX+hwgjtKWdYIipnLOKm/MRh8wQbtx67YmOC950QvGHTLydVY3sDPjGUXdiL+emspEYpGjnvO49uu3JZM1WLvpkqO6XSvfetxqN56QvVZm7SGbOv+YnBCwCKsXd625fdWFOTQbAqap+bMCPIBBQJ00rdU9t5yLJ21g4wTbieUVzFr19sfkwQowTC+sqVCO4i6Bl4DhpcIdD1mszwBwgFc8ZFkKJp4VT10XzOCWjWFlYZJVlEXgSzIne0OkG3KFaFZ1fsXRoi+uViwxtywmdP8cgVWWQE7NGrgYpCmSBdgcDKXxZAC/VShHaVcsIQVfzj7LyQUAXxQAd1oEUDwjf+DM0kUBWYVlNGqWWTx/RziBbOFGgFI/fuRIodCKskurC9wcYOC5X0AZ7gcEPbiUOweK6Udsv1ScFzCRzvKLQKG084lt1Q+AXlwfGAfNAnZ00sg5/9rUlCVz01ZptCyXSugZAN4BxIydGPrwMNup+GlS6AIXiSBQQ3ZfIZuzA38j8sWFWXUJbpoIBNHM/f1EH8gmrFp0bgy5WeeyANClxoUwicASzFm43QnOXaBKVcWILio5ALl3wS3C39cp4D2ud8F7d90TA3hf+8a32C/+1Nt6gBc95dd+81vsfT/3rnvXO2PODOBFr/qKT/sMLd0DMvFn5XcUhcC+TaqZgu2FjcRtARYb66awfhf9J+wsKWa8yXsHhH5HBAdsnSTBSxP8kj7tZEkYlhIZAx6/gxK9ht0+EcVPPn1V78OX8fOdm5Ezw7jNB094NtzZXI0/zp/3bizn4wIQ9oId12Y/gfDgu/8rx3u2l2eFtja8D/3K76IAbBhywKYbN3g1HwyeGNZWz/YCEicJtxh17zCtyAWGjQsPdPk7xKMXcMzYJeDi4QePpqUlHY4xG2VS5O/h+SRp+IEf+IEDBXgAICJV8bzt3W+lYJXrH7bqxtNKOMuef8RyZx8WK1p85r9bp163LkVB2TnLrT6oZXQq6utb1yyWJHCiJWYMUJOELQZEKgmta1MXX2TpxXMWjyUUFwuzjMwhOb0soAl4bDfrllk8Kzs0WZoFjgjyiwWwBIlmikMubltj6xm1M6FQirTzgqUwjYhbWazB4sYd20nVPyCMws101jILa0Ggg2OLBaC7HWsQjhBP2dTs4sGCtGDVxwHAHXeP01w36VSxAuEB6BVL7UIX2KobDuwrjGJqXn0Ec5qaxVIwFjCx5SCCNu3cKvoAmQunqCjQA+QJk0uCmWKUszM6b+851irWrrF0X7V4ylt8JTVhaRLTXC1aLJWVpZZ8dclGaOHug7ShqqAE7k0Fzaq7YpIRE1gXqIU5p3ixWbdatW5TC2csThgOQFzBILDLVWc1ifwMfXfgqez62imJVYgWpNl5f2W5NAQAl2fIcQBPQHRUBhnLsPr2LUtkc5ZeOHcgAEQyl0pe4yU1u+TuuV4RyCWFj7HBRG2Qld249/Uor58C3ii9dLL2OTGA98d/5peUMPZHf/oX9oav/TL7pV/5TXvho1fsm772VcfW4wBeimle+cpX2rPXbtr5c6tiW1mm3tvLK1p47cyinTu7Nhbw9DOpgEJ0y/j1PvzQFWfSXyyJKaUAjGIebMxgaaNugGXAGd6u3oc1DMT62WFfLe9+H3jstmBlSEGblS54EhnHwaQ1txR+GDYcYAhbe5TlfI6HHeccg8CoZxobdZcite+oMDjemP65vbmtuGIf0oDO1vv1KrQCx460sysDBA8Dk7C9rqBxOfJE5G6zvQDaZ6/fsof6ortxntjZzVsY6IbHH/1G4SUrHkfZkCusLC9FlkU8nwGv8EarafXdmw6I9jGL1c2rVrv5hFKkusmUZZfOCxwARuKdljVLW5aYWrTM6kOWXTpr7WpF3r2wfwBabLtUGJVMWmpqQVZeMLpda1mrTCrailg0QBJgJ0Ub4g7AwQCHpQTS8aKRreJGQsGhYwszZ64INPdvkiv4wILQi7SLNqgwrVGXjlORxHgC110QA4G+zXLecosuUEOAS4xlSkCJ+5B8YYS37qAxChPdquYV7CHgHotZfetZgav08gW3PA9zjESivCfNcHJqwSWhYS0mz9opS2YcSEV2AgiW9/GQ4jJAORMOUu180h3Mp67DeWtld16YT/TV2MwRHsKz4twzS0atVYzCMFhxCo6ZhBS2Be7juTmr1aqWxqmDyOPsjAOmoGRW42BYmwQx1ATyfTob8hQX5MAxU9JKo9OWfzPnmF06wMhG+Zvv+e3C8LIKwcSC8YrXcrMqQl8WZdjqJTNirRu765pcMf55DpEM2aM0ZsQ+p4D3iB14Hx5+YgAvHqe/9P7ftD/6k7+QH+tnvOKl9tpX/yPJCI5rA/DCwr3mNa8REO3XHwJSr167aZvbO/bA5QtirCYBqMgW/uYjT9rWzp590ie80Bbn5xQu4TfYQJauBhU99YOQ3TwfHh0VW+lNNAiIAKAD1nycr2eD+xliRQsHVfNhoOY1nFGAaxjw+vYdVvvs2V7adNjlfMAZ4JLjeS4+pQyZB/pTJhaT6mOZUMDM80zCTgoARSYcAsAN9z2/c369zkKOscxX+v7GDZafY/KhnWSD7WVcUPwVhVUedW4mbgSaILOgrTwrNOue0R12LAwtKxyHkfWEz/nkU8/oPsZtkvbgwRlBzz3uXMf9er+kob89yBIEBALQFX69Vc6L1QUEpmA0k2nFCcPRtWt5p49FyQDQQc4we0bJVAKzKoaqmjsHBWVEyBYtkZy1WCouXa5LXq1at1a1xPSsJXLzKgBLomUlAQzLLaQMrDxlpq1R2LAEQTqA3QC8Re1fAVuB521rl3ZV4JWaXrLsymUtxQPICJ7AI9jKO9Ktyvu3tKPUOXyKYaCjFjuF2wX4q+3e0kQgkUGPnLL04gXFOndqJUvNr/WW9hWqsbuue+V66YWzvQkAjDoexWHNcJT713MAsFOol0xbAskG7gGNmhlFcuYK3wDkvEcgKaF4LpWZthaAG0/jVkMAGAkLEwbcN7C2m1l7SBMmwDUBEwB1r6UNp6QBQJGmkKxGe+hHF0qSsdQiOu2cY3cB4ti64Z2MB+8IbbBirFWIVnca8r4xAbCv722YNWH9ce8wS82vWHpmSddnAkSk9WEmMVH6vX+fU8B7mF67v485MYCXgAPYs/tpCwNe2sWSLuD20oVzB5oJuHnqY1e1TAszgi4XfS6gqlf4RWlISO/qJQcAsUK+YIVKxV7y2KOWYDlSU2CsamJifpFAULHPvmEm0ieZCUTF4/JPBbii1VVqWyB34HdHdTCIAlwHAV46ahLQ3P/878Zy/lNPX7V8qWQLc3PyufUhG4cda4BDgDTn4d+wjefo44x7YLjOB5ojMAgumZubkV56EhsyJi9PP3PNHnnoaNIC2kFUNSw+IquvsQAAIABJREFUOkmAbtQJG+mEpMNFdRQZ1EdM+IhUvjgE9DNhKQSuJlcuHfybO+yzO+7jxgFe2geIgp1lKXvY5uOHAQdoNt1y+rS1y3nJA6T1LO1aYmbRsmhWWf7GVSBJ4di6Vbeuu5UNqt7R/3dalgBEzq/qeGlQKXBLYleYtFajZLjvOnu0toAqBXaEWrjkrrYLyyBtTZGzrkoe5hCW0WkyK1p6F8CiWI6gijhtYom/Zo3CugB5ZuGCgHoLRtK6lsLxoJLvFaKRsOUKmRxY82EDKiwLAiycRrUrMImGlfc/AJzTKTvAzu8AaMmZZfWXYnczM9YqbmnCQcyzdKroUDNTYthZhgfMwYbTFwDxoX//LNGT9iaXBwAmMbdN7a6IYDS/Ap14oscD94sZsbEwoDgYIJvoVEtya6ArM/PnVGioSYFzlXXWYM26lbduyIc5NT1nU2sPWywzrfZKOz0130uwG9Re+pJzwsLST7hY+M2FPRQFomGgJeFgUsXnGffUhr1F3jG4EI1xSJEgEyYkOImc608FktSKIqG1yqB0lY5jzLtdsdr3Ss6gv7Omey8+3Z4/PXBiAO/nf+nr7dNf8TL74r//mfYpn/jYffEE+gEvjQKA8UE9iHWE+QK4AxpgyihMkol8UEAhf9NCQV6vHnh5IMoS/Ob2nj14+XyvmEhujQFI3th08bWADIrQ8vlCT88JU0diFzpLtKGAXwDypFX04zod0AbQG8b2DgO8/rxRQPOgNhxmOR9QChMLm4vuNRmPayneTxzG3WuU1ynkwk1jEi0q54VxLxSK8sYFOOJ6wOb1wEgiPBs8TBbBRAhW/7DSAjw7kS4o0KRatYceuDwR6Ka9URnaUX1JrDFBK74okEklf2PIWgDTJPsxgXy+Sxr6+4jleljQMPDo3weQAwOLmwHFWqSvid0Tg+uKlmrrTwk0waQZCXZ7Ny2WnrLMmUuq9AdQCBy2GrLwahZ3LTk1K00rDB3JZ829bYux0haPWRdNqnSlS9btNqzbpLiqqeV1wBvXVoEbulOYZRhlCsBSLJ3nLB4A0F6BloqpAi0pGl/A++41mSYk0DTjkFDZlhsF9hTEJvtYX8gFSSIoFANctpqWANRnpyXNoC+aZeKGd8xIk+P3xAvLJQEgnOpJSTiv+rO86yJ20baij5WfK8Vd7ivscm37umJrk3Nn3PJ74A3siArnS057PKgl5Q22tpcCNoAldewoYLOsYjRs0DQJmSKS2emCvZ7W6Y3v3LAlyyRi1ti5KV02sdHZtQctu3yhJwFh9QB3hKHFaXDFjZpkK5klpyl3EcQOsANs25U9Mc2yYiOa2szSs8saM744Tgl9pR2x40w6kvNrkjYMSmkDkCOjSM+d6a1sMJHheCYmPvUvynvyJPucAt5Jeutk7HtiAC9g5D/91z+23/7P/00a1C/4e59hX/j3Pt3Org2fQd/rRzAI8HLNUUU3fGADCgG0Hgh5cALgpYBsWHU8S9VoJ9GC9rNeMIXXb9yyjz17w65cOi+gBVjx1+nX6qKfvRtesoP6mPvHnswX7fl9xgFe9hsHmkc90yhsL0AJkEtfU2wHAxtmt+ljJDOTuASMapN0ygEom2Q8hl0aPHD0sghJIog0bjQlwwjLIbw8gnviOfBhyzjbXz3Yd87o/S7kpsHvGKOMHc/o8runPvbsxIzxsBWPSfqBfa9eu6FJWqvd0SoJE0b+RsLP7eMN8NIvWsIvbkuzO6pIKGzHJVCGu0Eqa51qXuCG1wEe7XbD0vPnVNHvQdy+pZgr6mJuDsvaKG0LJCYBMgHTDBiW5hN5g9wIAiAouYnz9e02atbIbxnFSmBZXBHiWHRlZtQWdK/ysU06azPnm+o8VMMbwLO0/rTOlwmkGWKJG3i9zjmdK6lrsLeA3FRahU6QAt5JgPbFKcajIIy0L1jWVt2xwyqkazsHAyJ+S9vOrQFHDOQGaG0DRlvx4TgPVPY0WWBywbXpJ86F1zCFgfureV13T0GohJPi7AdOwDw7v3MKxZzGed+bF3kJjgRugtC/Keq3Xh3ILPf78GJ7V739lCZFmZXLllt9SJIUWF/FRwN8Q64HmjQEwJZr1Hauy5kDcC/NryQWmrFLi9uuEWLUtXhmVv3ZLu8qfKKjUI6uJhaSgMj9wTlQOAKIFYDga8+H2K1sIM1Iza70NLw8C8YvKxmHka+Mei86BbyTvlPf//ufGMAb7sqnnrlm/+ZXf9t+57/8kf3x7/7ysfWyL1r739/4hjvaALuH5nZY1fyNm7c104fNQ1Iwat/wyZ+9flPeuYVSyR64fFGSBpg8ABzsLQAIWcXc7Kw9OmZJ+255yQ56AB64Air3gX3ZpqaiFdnB9pYrlTtA87iHzXI+E4qwVRdspWdzsUQD5I5aai+VyjrH3WJ7AYtouMOpaePuIwx4Ry3t82FJX3sA7L8yrpjk3N7cssX5ed2v8zOGAHOR0L1/iore/x3Jev12eIdljBljBFBMatvGfTE5YRKD6wlM70te9OjQUIqPR8DLGIJdq+/esvT8ymAfWiwBWQJvw97edoVsBFhQ/S45QsJiKmCDBZ4NAI73zb0zTSw8bmUNVti0TqUkW6847CBAFWYvCF1wMglXZ6HgifKekzcAClsNaxS2FKzRqpat3SiibxIYxIs3kU65FC4S5PgumbZkILsAwMIQW2rappf35SywfrClXh8q9hEQ26g7MOvZZEBaMi0GMgzqwvcH2PQAWLre7es6rys+SzmfYGKTizvq18T0vMUFftsBgHMpgZ16QfIJiq7C1l6exQba6rghgROTavGHgd5hwROA8trGU1bdeFZMcW7tQT0/sfKEN5CiR3R0iImWY0Q8IWbd2Yq1rdMoO+Y6lXEuCjDknba16HscQHBKq+bVb5IuMNGR/VgA7HvAv+MmB0G8OsDYy3cUlFLYlJzBeynrGsgcWAXj2QyzThv3ptv3+ingnbDDTsDuJwrwssT7x3/2V/Y7v/+H9pcfeNw+/e+8zN787a87tm4G8CJNeOUrv3Rg6hZFbJcunD3gD8of8vbOrj39zHXJDs6fW5P8AdaWZWq+jtPTfvTpq/KyvXZzXbZlMA0AHdLQAJcwfsMCDfo76256yQ56EJ7thTFtNOqRAS/nGgSaoz5sgBL9D3gjucuzuZN8eMD2MolYXjqaxy7M7I1bGwK9Ubd+H14AH58H3gEi6nnYj/FCmMMk9z7o/IMCUaK0A5Y7agGgdLnFklZxPJMLWIfNvnV7w65cGtyHH6+A1/c/YFZL7XjYBgC3g5yAeNte4hdBMF0FSsDCAgzRqMJiovOVxdiEjgZcH5mB/HIVcuDlAw2nY2WZO7C+gr1Di8myudjbgLVVsENgq+WsyirWgKnEuzaeshiMbywpGzPsumAIJS+oVeTFm8nOWmblgthGiqcARF5XO2h8uoI4JgEAYCe58PICJysIEtIGeLrSz0gzWErnWJjmxPSSdLHSIzNxHAC4emCfAi/5zDonhUk2QDcMq3eJGHWsnB1IZAtpiMcmrbG6s33NareeUhFcTL7FS04SMzW3L91QIAiODg1r12vWruyK5U/OrwaWZEyYgsARFUQyJgDP2MMtaGKmEA50vyTeZWdH6odZkWjVSxo7KtRDXwvA7bTFZPtCOckcitsu1CJk/TaqnziHXw3QWGBsBBOk1MrRayAmeb6n+977HjgxgPcn3vPL9p//4E/twSsX7Ys+7zPs7376/3DXNaiTdreXNHzeP/iCgSER4aVgQA+FaDC0bAAQQOuTH33GlpdJVHMWVnywA0zCANgDYR/PynmvPnvT8vm89FGXL56XFGIQUO4PNBh2j/ea7QX0LMxN2+Li5OCRQA4YyyhRuoBkH21M/zIhgV2MEqIwqG9IM6O/D+sE4c85Sts96LqDgifGrRoMe7b0CUwv4+So22GL0WC5+dsdNplj/PH3ga6cZ9ULMAk1mOfKvWDXFt6YtHl7q6Pe33EfH6VobVgbvSODbMvEssLCHZQC+GO9zAFbKDSZKugCKKsYKHqU60gwASCs7AkYYtNFewTaggItvHfFsoptxVM1ewcIFDj1xWSUYQVhCZIWsMTOkn952+q7t7V8zjgApOEznF4+32MBxz1X3yYVhHk9aqPuPHo9iI0B0uMC5Lg1IFVwjg3RU8sE9LASa9ZcUh3SBIrMApA9qJ2y8cJyrtVwEosEBWzj30cBvUwMYD3ZRgFegB/WZxSeSWPcCOzlAKSk2CE5kSa6KcmCwh5yOHWQ6JaTnlbFZbNB3DrMN3reegn+2hXcDZBgKAq4VpSFnAe+gwrRmJQgwWC8ylUikKYgcwBEhx0fVDxYL6tIEoAsDbrS8pwfsaQ1gf1al0hqThbMPaQlJzkQP+mF6ATFuPF1+vr90QMnBvD+/Ht/49g1u/2PLKzhpSoe7Wz/srVCGj52VdZTgDb26Y8M7rczYz+ArwfA/qvzxW1Jx7u0MGfLy0vy1R1XxR8ONBiU3Obvi3MDTClwm3QJOspw/tBHnlBQxag2DDsPDCd9Mgz0epDLBzaShYX5ud6pcBqA5T0MO8pJKBhUgVSE5LFR/TDJeYYlrR2WrSWsBGlAL545ygMbss9hitGQldy6vaXxH954rjDHKqhMJiybzYpVH1aMB+uOVZy3keNY/h4uXVj7uLAlO8JjG3iod3Nwy9awsYQR1OQCkZo5Y6npeQeGJ9gAJwK6KlBDn0vh1p0b4FQAGNYWKypS2CRZyDgtcB9AUsV/4LwgIBxPWRd7tZQD0twLIQWdbtNaxT1rFm4rYjk5e8a5PnBuJgABuCHEQW1QCIYLVnDL6+5nSStksxbYrcEw86/dsQ6S03ZHrDjMcq/gbIJ+EhMfsIliw0kV0wQg6xjzeML1I6l5AMbAik7xzzDFhF6M2cKgdxDgBRjK4xepB6ERQTGfPy163trOTfkqsyqAFEFtjMfF3ErWgYQDDTTJdKTu0U7009PzlsrNiyFmPI3bPPDlXLI5S+fEBMMWe9kJVm2MK+/awIoCxZhMigR68TNu1RW8US9sq52cS/WC1rF4PO0cLBLYvTnZhQoz2zxXVg5qjvFn7OX2P0PGtf309ZPRA/c94GWpnjc0QN+gDSB5XFsY8PYvWwMyAaPoL2H3AJEvfuyRkSEHowqlOB9SCP5wkS7w8+WL5ySXiFpQBNvLe/04ptSzvbRnEjuscc+BorX1jZ2BE4Nxx3rgKX/cwOqLoiiALv0LYOL3w5wnDsuO+nY98+wNO392smQ4joWBCievXX32hp076xL43O/dp+9+QYu7Yq3mihOnpg6CBRhOIpaHLe3392P42jx/7PC8t3D4NX8cTfJezMOeyWGL0XhWHAtb7oEukz/6iGfIKgWMvC/G21/lSB9Y8QD0U5TZaLZ6nsAf75KGKH8/w/YB7FD8xgY4lS8vlfgsG7fqSrSSY0Ogx3UOBnemByo8oLIn4JAYAXRHtVUFY4EHryzFAj2oY0L3bSll6QdT2qi7tMZAHgBAVKzyzJKAWPnG45bMZJ0NmgVxuhlYvKxZkuQ1lY45PTsAtwPQVYKDvvfFZL0VhMCNAYAkSQhpaPgVi3WMBQAqJIuQRCKaVzzt1f3j31vcFHhkyZ5Jg2MekVo4xl5a6Fgs0tK9B72NeEaSMpbtld5XL7toX1wpBrCv4eckyzhS3wCQ2IGFZBsEVnSqe9YkrESTkrrlVi47KYMs6ZBx3Om7O2wcSJpAYSATmBg67qaKE/GO1ntlmzCVbQfQ0/g/80brGOrUzIJs5LDKY4zKN7rTVmGnm+z491uCPlwMMpIb+hUGWk4hgTznVMN7lHeV+/PY+x7wfuYXv9Z+5kfeYt/4nW8b2IN/+Fu/eGw92+/S4Ez/NyzJrDEId2BpFrYUq7BRcanDnB1g5rCH4gMecOBjhsNyiUkKijhfsVi2SxdH+5beC7YXwJtKpSfWs4YfMMBzKpcRMGL2DsiNyrwClEgOG6eRHjSgBgWLDBt43vKM8cBzCnsjcx70vExW2LyML7yPT8TDKYEtLcbTsZ5M/rh3CsFgrPtBK9dEv83zU7JbKuWOS7uvJJjh8OGL1Pw9HGwjQSTdgaEYfr/DFqN95Imn1H7kFehNGff8ffRbt3FfMNJudcM5UuD8cXtjW3p17p8wFu6L/pqfc76pJ307iqThqPfuNJkV50VLFT3gV4lddcfWpnMqIpO9GGlcge0UoBSQqphbwNkhdMBDwU8P/KK7rQX2ZTCgWWsTxtDpyCM6vCmworilJXIK9Jp7G/LtdUv2FUkUAE/ON5gkXu8I4P4gfTGZ+/t07gG9QIXge7ldIG2gCDCZsna9KlDH/XsNtfPVRSLROljIl0irXf2bNMEAdhLtmDBkp4Pld/obbSm+w67ADQAPI8p94AoR9jbGkFcex2FQWi9baXtdK40AuiQgd0B4ybgxxD3j6gATDEgETAMo6Wva6yYNLatv31A8swf7SB7oh1E2ejCrjK16fkP2brK6YxIRPC9n/ca/ltMI06/pnOWWL8mFgw2LOfpd2uXg/t3kbUvWaARy8MzE4mZcMp9L07vzveMU8I4bDSfv9fse8N7PXeoB76tf/WrpDwGTlUrNUqmEzc/P2ZmlxQMM6bjiHR+nyoc454MFY2meGOBBFf5hbeYkBUWwwxQHjWN66fthyWGHeS7elmwSPet+3HFLTgsuTKNkn/ySx3rgP2pbxhU+jTuPjxsexsT3W57x7AYViuECUSw7Bwq/8UbuZRkAt2wmLds4GHwvb5EtmVLa6pIBLCBrWXTLmqwkAAS9nIPoaefecNDCjEnX9k7eHnnoci/hbRBQ9LIazhl2gaA93gaN8XlmacEWgzaM6r8wowvTDIhnAhdFOuNT3rxVGoWInO/mrQ3rsAwci9mLX/jwKeAdN4AneF02W/WqdWHApDdNOIutwPcU0OQBCsVRLnAgmgPLBM24ExRK0+vAb5OitlTWsrMuiat/E8hSUdSMwC4xyAe0qqxmwPaK1Tt8qJH3v23mN1VclV0iGGPpQAS0Wy4P9MFKEUOSsBjompuK6mXCQHrdqH6UjADf38AhQjKOWNxNMiQvwAoOSzZkGR0BuThygljCqpWi5SiuU6BHCBTHHDh2zPboSSP9DvNMOAnMLVHLijUesDV2b7p7DJ6NKyrbku6ZzctZ/AQKQM+EBFbbOWEgQRncHs+Gc//Nwoau42QQWZHyCiLB15dJkRjtkhjj1OJ5p1H3RZbyUHb2eVqhwEe6XrIu7Tj7yFGG6umx92EPnBjA+0P/7OftTd/yNQe68Ht/6N321jfdaQn2XPWztyX7h//wf7LZmSlZgwEQAL0AikE6RBjKC+dWB77GB/uf/+XfSGfJ8m6UaNYweJ2koIhiLIDLqCQw348+Oeyo2l4AL2wM7Nyt9Q1LZ9JirvlZwDaIO+ZnGMr9uOOEAeB81f444Dnq+Q8rfIo6ZvqZeMBouEhunOWZvw6FeLC2MC4cDyvrk9kca+kY7GFaVvqHscK+sE3SviaTWt4FTIYji8PFjoBX0svy+aJS3PgZwOj3xx2BZ4FsJlw46dsR9gKGfaXoEi9sNM6+uBKgzvloP8AccMpYBsQTk41Wl+v0xwZzzz4d2DG8Ls7YA13axIYsYmtn13b2Cra2siyv6Znp3JFdKKKOgXu533EyvEPvi+hm2MymM/tnyVoeqoqYHazPvZd95M9N0VqjUrJkl+r6uoIj+jWonu2VK0QsccCfVprlYFkb3SeMdjyVc5rjAQ4KACLn4VuRlhc9qy+g6k1cm3Wrbl+X1RqAS/7CmYBFDHWKWMf8hmOtM1MCjYdlxvut2MJ97xLlHAguF4tyrQnrlD1wdm4FyK8cSO4xxAEYpniPSQOsKc+efpYlW8Dc9gN83wYKzbgvhCPSauPDvEc8NtIDgGda+lnCT+h3eTpP6l7RaVtjb13gVo4jgSMEkoUuNowURQKeYzFFQGNFxz04j+OuJDDegk0ex9wf7Hdm5rkYxqfXeA574MQA3jd81zvs3e9684Gu+Yqvf5O992d/6DnsroOXAvACeP6XL/nH0tQCVAFlAIar125q+XzQ1l/0wwc4XrqAD96Qmq3W0DjVQecLs7uTFDVh6H9ubTUyU7q1vSNQH4UZDreTPoIp3tretlxuStIOQBgMHeAf3a1ijkNxxy72eHihw6hwj3EDIlz4NG7fQa/DxANUmdwAyj1QjWr7xTGA3A8/8ZQKDnEd6I/sHQZ4YXI9QHZa7oq94JEHes30UcXhyGLGYxgAA0xZjUDLixxE3s1bO4oR5nf0O+PZJwCyrJiIx+SckMsBhPc1tXw2IdGAacY2jOMYI0h40CGfO7uqlY5Gs6FjvQ7as9wry4s9j2T6FMYWoEv7Go2WWGzfN/QbOnYCKLge7X7y6au2MDtrf+ulLz5leA8zmA9xjCJ4y7tW31m3OGNh4fx+dX5wPkl5HJwQe6aNr+6XgdTHfZWrxCE279LAJBrwhY2ZltgVajBjydxMj7EEFMIGAiy9a8FBYNhxhXOBI4QCK9I5AT8xjxTyAZ7SLINPSaJAsAQaXgIaZIMVAmoAaZhH51ebdEVVAGrOmcxI6+wS7yi4qzmHgghFaMO6iXuT/212OEgba0vmn10AkmMU9AVFfYBDgPugJLReoh8+yeh7+wArbePBOzeOtEu4K+1osiEXjFjsrsQEK4QCJp/Y41hMkyAB+QC0e2kK+l8S4QR6iTtWPDSTHcfyy8UBGc8RWP9DDOfTQ56DHrjvAe9v/t4f2G/93h/aUx+7Zg8/uA8g+WAFLL3rrd/xHHTT4EsAeIn7fd3Xfc0BVwD2HrRs7c/il9ZXlpdsZy+vN2iCJ7x7wWE8V31BEkxcVAuqSXSpvu0OiDTvsIYa1EP0AQl5AC40mslE7IAP77iJwbgH6yUgh4lIHuSRPO56YcszAPdLP+nFE1njobUGrOIq4LTH05oY9TOdtCMMeD1ARsoB082xHgQOkqcwfmBFe3Z2wcSBiVRYnnBzfdOWl+bFmPJ8+osUw6CZSOxSqaJobDaKfQCnnXbHtrd3rNpo2mMPP2Arq8sCz7SPic21G7etUCxaKpmyRDKhsc5KQTzWFduby2YF2P1qxub2rsYXMh5/jxzDigQsOADdaaSLNj01JRCOVOixFzx46tIwbgDfg9cBGXUYTXSnoaVtQI8vBhPOBQTxL6jT7L0O99fpiHn1y8xRmxkGvAfAa7vpYnhrZcXWwrICaAAxlfUnVdSEe8PAjaVtlt7Lu85qizS2VFaAXhZc/SCoi69xXn7GFPZJjsC9q2CrreIqisNS2XlrkzaGjrhVFygTq0tAQyKpIjSdI1iaj9oH4f2w6JIudYg2NyrgPcy1OUayAfS9sLdTwx0OFBRR2pUGGaY8s3z50JOe/rYyOcGHmUlI2Kps396uGtQJxwRssyuXXTRyA3eGmryLAeSS7KTvvUTnsH19etzheuC+B7zrt7fs9ua2/dTP/Yq98eu+vHeXfNA9eOXCsbI6w6KFfSP9snXYIsu/xof29RvrYksHpX4dxlXAOwkAqKJaUB1GHjCOIfUuD7CKAHl/f4OihUdNDKIM6cNYZHFe2MFrN9blDTtu85IF9vNs7qj0s/D5APXu+JIkDGGwyn7DzgPgBYhW+FepBtedGShx8PIUgCFAHLDonU0AjwDdfn9n2rW5tW2bmzv2iS95oYrU0EezWtBWLGtXgNYXxdFWpUFRM91sW6vdtO3tPSuWS2K7AdhYv6VTcXvs0UesUqWwrm0wuPPzswI81WrVdnYL+ntGvhKPmb5PJymoi9uFs2dtLfh78K4VgHf6zk9q+L135KDojfPCKK8sL9xVR5FxY+JevX5fShoi3CwgES9VdLKTbuiFBdYChjDq8cMA74G/PyQICmGoiv0EDNY3r5olM5ZdvuAsqUJ6ZdlgweIG8gYYXnnFBgVqXvqgpfN2U/pPWZcB4gC+cj7IWXLK6URVONbBkzZv8fS0peaWxTj6eGMV+8VijmWkEE/SEeKRl3SeSbdRoRv3GvD6thLGgRbWxf0e1Ea71/DIJeTEyQqa+dvu2QT+upPe86D9XZGcS7yjT2HaYebpU723UNhYLVhz94birWHo0YMr+S1gqE+L1u7Gk7i/znHfA977q7sOtmYc4GVvgibG+eQOu8dhrgIAynK5KgDVr7P0ANAzvv2etw6AAVL2K/ejygPkBrGX15I253/w8gVFJ7ME7oM1AHewhegqvXbU398gwMtrTAxgLvsr9aM8+6jAc9C5RhXvhS3PPMjtt2gbxcT7iQTtc8cPBqu0K3weD5DpE0AeS/fjfIu5j7/64EfsoSsXJSEYJq0A5MPSrm9uSi4Ag1qpVIzJ1ezstIArqXKrZ5Z6Lg6+oC18Tlh+xozTms+ra2n3Ex99xhKJmH348aftsRc+bNl0Wi4L7F+t1jVOVs4sqlANltf3Jx7WOEfcvLWuAjjaxTi6tb6pccFx42znTm3Jovy13Nt9BHqrxZGV+KNagO0VoNGxveOLyKIA3t71pEMuOc9ZmNfytmGmG4vHAvnDnAuBGGPPhSwDAAVgMnTBFHslKPgiZQ2dcEyMLnIHwDRMp78X+gb9aiI9Y5nlfT9qiv984ZsAd+Bpi55YfTGh5APdrPexDff3cwV4uSZyDVhcKP3U7BkVowFCxWAPYH+juDhMOnqZfMTkp5xUcpxkCrI3C+zpAnmJJglpJikHWelTwDtpj9//+58YwAv79N5/9x/tqaevqfLcbz/5w99zbL0cBfCG7cMmbWi/qwB9ANDgwx9QUKs7z1JABYwuoJM33N18UeCHpd6HH7gkhpXfc2y/3ynHAGApAjq7ckbApz/imGVwXmcDgHFdwO/V6zdtbnra8qWSzo9WEzDjj+8vuBoGeDnvUeQJAEYYyigV//3PIFy8t++UUJJlVhTLszATD6D0ThJMRIjTHQdWfXuYqHirMPlSKD4UAAAgAElEQVQJZ5iQpIcWrXEcz4DnQn/D6sJYk+A3aKNtPH/GCs+J5+W1wDCk6IHRywLQsf964PKdzDdjByBO+4b1NTpvwOzG1o7OyTVYjeF77g/dc1hnTD8hi6jW6/bApQv21DNXrdFoSzO8uMjEaS5SYeUp4J303eXe7H9U0CswWtgSWByktQ23eiLAGzpQyWWA6/yGipr4T2Jq1tLzK0OT6Q7TW571piDKJ345MFvRPWbPXAmidvdt1Rz7yNI6BYK71sZtIDdnGQoER+hz+9s3CPQ+l4DXtwedruKX8XIe4ubg9w27OEwK8vvvX24S6LkFcOOWSOWcp7PS/AZYkAXyFSYYfjsFvIcZ9ff3MScG8L7prT9uL3rBw/b7f/in9rrXvNJ+/w/+RGzRN7/u1cfWw1EArwcmaHVhsSbdAAy7u3ktzVDlDtAYVczlC8S4HiEDz964JaYtXyzJ8B/dMA4QYfbV+5x+5Mmn7OL5c70lcLSjpXJZujukF96BALAEa0exEm8Kjzx4xdlnBdHI/qsH056FTsS7trAwPBLzsPIE+hTguTA32wtVmKSfH3/yaYE0b+vFfXq/Y4Ai9zuqzz/wwY/Y0tKCgJxnc/vZ7WEgFIDMM4bNRQv7khe/UDrbGzdv4fojBh1AKzY/8NPFxcJPQMLWd4PkKV7mgP4VUOz0wHdqgcMTjn7WnPOiraWgElZ3mN8tgJg0OW8dx/6XL51X//X7BfufuVe5gKxvGpIc8AeTjU99+afIZzhcbDeMuYaNh1l+PmwnVdIQ7nsty6Nbndu33Zv02QiQlnd1jkGRtJzvsIA33BYFPTQAoBQyFS0xs2DpBdwnJpcThM+L3AHA1ak5RwdkE7gDsHTO0j2OD0gr+L1Y4MzBIrseCMSbt0R08rpYZNoGMI+yNfZuK6HNM8zHAXjVTifijtJkFZE19w4ncRBTXi9Zq1bWBAPJBJrmsB/xqEaIma/kLb1wTu9Zp4A30iM7UTudGMD7ld/wPfav/68ftNd/+9vsPT/2FnXy9/3wz9j3/ZNvPLYOjwp4aSDaSLZJ4m29/RKs2aUL5yc6FtazUCwLTAGUHn34So8N5rVLF87dwR4CdABSAHPYQI6FTQZ0eOP/ra1d2VnBvCFbwC94lL2ZGGhYvWbT1tdv20MPPjC00Oso8gT6lsQ5NLlhQIZGmOIoQJQHjHwFkFKABfjjNe4Db9/wBpsJi4reFIAWlo/w9o1uFXstwDD61Be9cN+3UX25vatjw8dJW9toyIfXAWSXEMc5aOuHn3zazq6eETsMyOWzwk8gaC/BCzDs2HF5YO7uyzknhOUpPGdCS2TnFYuN1QKHJxw+oAQEKinC8uJAthnASj9hiUcf0a7Z2RmBY9hzADurEeFgC/rYg2DAMeOUAjdY7enpaSsWS/bRjz2rsSdgHKReUeDGCgTFaow5XCDQq3Ouc2tnTovWju2d8M4LAx6whQozZhM3D2u64iZToIHnuRuA9wD4DQqeAIrxdMbSc2sCjIOcCQbdi7TAQYIZryeJ6VVynU8IawpUMyFQclrgnQvYjmcy1mm1ek4Lg+KYAWOyMmtULL18aaj/bbht2HXJCzeVsWMDvBM/eJP8YVxQhbB0kLDGBEvexkGBok+km/TSgOb67k1NLNo+g3rSk5zuf9/2wIkBvK9941vsF3/qbfb673ibvfP7vk0fht/25nfZj7/ju46tcycBvDQyaiEajBcgDXYRBg9JwiSuAp6RAzjAsAGcOUc4kWwYm/r0M88qmACA7IvtAGaAJ8DF7HROhVThaNuo9mZRooUP41DhB0BYAgLg3NzG6zYlTSrL9gBH/GflQwurmsmI9QSUtVttLeWTAAarja0WbKbvA5hRseeBfhWg68Ih0vLABYTChMPAA+LQnp5dhYnp9iYL9CvXTiacpzDFXA58J6xQKEkHy3OizTg4eB/eMEvL+ZEIeCnL/lcnFeDYnUDmkM1lBVYlr5jDq3a0Z6qfcABaYXQ3NrakCWZloH+j/wC5jDEALpOkQZ7Og6QqYWu1sPdw+BqerQZo+wAN9MYAcQrdYKync1MCwEwiH7i8r4k8tjeEu3Dh5wPD67vBR9qOkyaM6zbAMxKA9NzKAa/auw14w+1Af9rI3w78dnFfmFU6GbZa4U2AK2C0KZKSR22G/YbHCUu20KiIiVRSWyojH9hEelpWaq7f8Ofl59k7ADeWWrWtq2JDMwvnxkYDN3ZvWXJ22ar1hlxyotonjnsu9/p1+obIa8A/xWSkqfnJA30kLbWCRVxq3FFlEP0Thdj04ECNe33fp+e/dz1wYgDv//mDP23f/ob/zf7rf/sze/+v/66Ka/gwPO7gCUDJa17zmshPaJRPLqAKhg5gSYJVWP8ZxVUAYAvICzNygFGAF0Vm4cCLfn0wAAamjiIkLe3PzghwIY3g2thGDbPCot0UHj3y0JWh/cA+165dt7n5efnvPvaCh4bKBKJODAZdTH6y6xsCQX65HyDpl9n7tbVMLjxo/OjTz8oreHXljIAu+8KOuiIqJz0gKAOQBkj2G+cAxH3gbx7HXt3WVlZ0b1wXsOi9aT2bS7/6JDOeC+w/ABeWlt8D6rgOdmUcC0sL0B3k9tHfB3v4He/sye7sUz7xRWKQo8gr/HlgtH38MJ684TATxhesLc+HD01ALpOC/mK+/jbBvF+5dF73MkhOMWzQhOOLAcBY3KFVp69ge5nA0O/Xb962xx69IpeIk749nwAvzwKrrW6rdYdP72GeE8VFbMnZM4E7R3NgtPBhzj3sGICV4mq7bYvHU85RITulIAOx2M16j8md1FatNzEQYK5Yo7ilrxmS2maXpD/F5xcLQMIQsD074PVbLVhDBVdTSmnjq8DfADkGqWeNWNam551+/15u3pLNhVng+OJkYSruIxKY2GP/NRR/PLBNOFwwAcDSrVpQ/yAHgbVOzWB/thBZLjHpPZ9KGibtsft//xMDeMNd+eEnntYy6Mtf+gkKKziubVKGl3aG44DD7Qb0AHQANsPSz4a5CgC4YOScDdTSHZIB2FyKmWCJw0VNnA8AAwAD2AFeeGOCPfvvH/iQQAqgMQzufJv7o4w9ILlw7k5bIi+PmM5lZDQurefmloA4oNIlhKX2JQfptA2bGLioYf4dTGcDtHIvFD+1mi1FO8O+6gMymXSMah/48+fJ5wu2sb2j6zPZwEoNz1gAXrFQsnqrZS94+IrOATsDWxxOhtvY3BEYpugL+QmuHH7b3Ny2T/mkF99hPeeCOHb1vMPuFJyfe/nw40/Yxuau/a2XfuJYKYsvlvNaYCYr9Cca6wcuXzjSnwft+cu//rA+Z+kPWOkL584eWC0YdQEANJMmnudLXvSCkW4V/edhkvTERz+moAmYeDTafiwiC0GGwhh2riAzAj8nfXu+AV6B3kpBWtVhEbSTPDNYPYAvrHE3mbnngLcHSvGYrTjnBbS4gLlh4HKS+7ljzDeqVt+5oWCE1MKawDRAsV2vSZ8qNwGkEj6uFxlGfkOpbryPwxpT9BYLwG9Yi5y/+bSlEjFLJJFYxB3wxGWCIq4D3+/HDvNal+AGeSh39Rxdels7sGML/azftdRHLqo4cK8gXAJZWJD4Fv7qgyF8/LGihENtiseTYr+RiihYg6KzeNK6rbqLam7V9bMcNpIEW2QU13w3tlPAezd68f46x30PeGEbR21YKR3XdhjAS1sBtt4nl+9hJWEko0QJh10F+IDn2GqtLrYbRm7QBiAG/CjVqlhSwIBf8sem6szyglhNABNsIgVUM1NTtrO3d0C60H/ufuszABx6WH8f/aAu7NKgAIZKVWClV+QW+MUCZpBu5otFe/DyRUkN8KVVWk/cpbSh+fTpbPQhQBU2FbDH76/duNlrO/IEvGjDQJm+4xkQikDE8crSoqQJnB9W/AWPPGi1ak2AyjOLPCOXAOfaUCw6f10mCqw4MPkCIJI49vCDV4LCB9f3HnhGcTqgn9EEMzkqV2s2aBLBPmh+eZ79WmD/nI7iccz4oAANz2XQLi4YL3/pSyKl8oW9hz2jTr8h54hSuAmw5pkiG2HyxzMCZMMkyxov+PuB6UWmweQnlXKhFid9ez4CXgd68wJMR0kTCz/bRmHL2q2GxacWLJN97uKNFWhRyUtaIe/WcQzlIQcky/n13RuSUMTMAVDpgbumAAuW8vclD0mlyHkLMGedhlyioiAFJBb8qzZbLvgH4Np1UcLh7wViiRYOktb8PgKlgFBew3otZMUmiUEI3B5GO7sfa7wfgwwolr8xzDr+uAr0GExuETYhyzG+0jeddgB8086ZIZUZeuyox3MKeA85eO/jw+57wPuZX/zakd33h7/1i8fWvYcFvDQYoEpMqqQZZybTCgHIWNL1ccZhbe6wzvCaSFhPACBABDaYrx958mlrNppyYoBJ80vggFJAF9G3wzaWzikYYjmejSVwPH6JvIURDd9bvy3ZqGAOAB8Sg/WNLYEkQBBJez76lmuh94VpDmttfTsLhaJ96PGndE+LgSewB6sAJMAXAAkGuz+pbWNzy5565poAMzZsaGwB814HPQ609jPx9D2TA/qVYq5xThvcg09aA/SFQ0Rcylggr5jKCeAPYuB9P0TVRLsJABHQBVmKASwpSrxw/qyzwAvkMhQ7svnJA+wyhXn+ZyYrTMAGuVUMGk9cl2NhzdHo7uwV9Fx84Ztn8pGLzEwjI4nZ7MyUxpuPn2YCsnrm3i/VPhdvNM9XwCvQW97T5EkxvHdha1QKVs9vWnZ2yVIzw91f7sKl7jiF9wy+18BXcop2w+KEMrSaArGKbU5myKjT8j7sLKwvfxut8o6CFHryigD8woxXCnuWm51X+ILikWFeI24AULGvJ2ATWIYBDgBwu1EXSBcDnEo7PXAEb+NTwHsCHvaETbzvAe+E9/Oc7n4UwOtBwyi7q1E348HyJDfsNZEAGEAcbC5MGowsIA4m0dtx+fOOS1VjP18AB0hCQ3zz5m371Jd/8h0a3UE+vP16XYAVDJ6Lu52z7e1dqzebYmAff+oZORggueAfYHVtdfmA769s3PYKcjigOIwPB0Cv72+AKwVmHBtmxAFeHEcfuQlE1+LxxAG/2Y888ZSW/wCCo+y5uJZn4lmBoE9oL0vyFIFF2cLRwpwLhtrLKcJa4GHnCrPZaJoB9RTp9SQhIZAKEwxDDlMMa4q8BekC9x+WbgC0Ae9Mivzkga/0c4/5TiTElIeP60lA2m0noUkgoYGZRRcZ77lWJGJxXZeJBWw5rxVLJU2eprIZq9WbihD2QNffu+uXfW/uKP17v+7zfAa8DvTumsUSI6Nnoz4bX7QWb9VcYMXM0kRetVGvM2q/HvBNTzlHh3vA+KITpmgvkZvRZMElv5VVtEXSm2QH/K/VFAsMKE7m5u9g00ulomWTcceGNqu6LblFIAcABA5oO/ZqsMQcQ+Hbvdb/3o1nMugcig+G/W02JJEQQ04SXjIAwKGCOH/8KeC9V0/j+M57CniP0PdHBbxHuPShD2WZ2gdVSBc5v58uM8y5YZxDxP/f3pmAyVVWef90bb13Ohtkg4RFUGRAx2900G/EwQEXRhRnogghgkRA1iQgSJAJJGxBgQACA4iAu4CMggqon4CDIw4ioMgSFhMSkhA6W3enl1q/539u3051par6fetWdd17+3958oSk33vrvb/3dudXp857Tk/PDnn+pVdk1sxpGjXWVAstpTayX32pxhO4PqKwKNWFAzm0iIaiRBqkB2kXe8+ZpR/h/+WFV2TfvfbQCGI6k9lZAWHzNh2PtAOkbrS0IAoYl23buqWtzfnYG1KP6Gp+RNytWwwhhWAjxcL9oQ5RdFMk3M2AiBzv/7a9jX7wb9i4STcAIqcXEoc3D4jGmkTkIbyInqJFLyQe4jd71kyt91vuwD26b2yG0z6iUdm4cZNM2303aW5udCL4uZxKuFabaGqU5sZGlWr8HdIEwN1N3YB8ukKLN0lILSjcQIe1wRsVSKxbt9hNOclPA8G1ELXHGiEyjDcZWKf8FtRupNlpkjJB1wXzcdMZ3BQP8MH54DN7j5HPWsXfIHU+MezCC7woOaV1Uj22kh1RpQElzFDKKp3UtrUmndqqudSZAdQNRvvgFt1MVerjdy+viQg5RBVl0iKI8CJiGYlqpQetFJFG+h/ybJ0qEOgE1zR19nD1gsKyZCqBqQHtPobfIc2O9GKbXIN2J9Oc5ZiTO4u0CWwYLFYyzct91etcrcGsqRBOigh+JuYLcAZ5zjxCRSAwwlsqtSGoKQ31fIogDsU+BsfH0es3vrXLRqdSFSLw90hLgChBmNLZrEwfiqbm77B377WY8CLCDImCiL/7oANUsgpLskGi0JQAc0akr7evX6PRyDOFZCHtAlFXzQfOq/vrVl94fd0G3XSGPOWdc3FyQXEPOK+UhL6+dr3+QwBphRiaVMtwXwPiX9ikYfXr6zS67FZxyE8HyI+KIqUjFourwEPuIZHF6gzjtRDhBG/IH2QV91MYqcc4bBxDhLy7d4ezAa2jXeKxqL7RGI1D/vOaH5XH+uF1wQbSWpgeUvicu5H0P//1RXn72/bWNyBuRz53PbFukNximzc1t3cHah9HNHUC4yDf7LRWz58o9q9dDektVpYM0oeSYvj4HZ29ahFxLXe3qEoB8UWKASK+VRdf1PpN7pQ0CJvzUb2zIRh5rxBVbaGsdYW7JDF5lpZ0G8jkRpQlw7kYA+HLJgcF0VzN4dVc+KxEUdM20TwcAdY3K6iU0RAZtQue/RPhgzNQYk6ZOBviGlrHNk3GBwRCP4XACC+iee6BXfi/+9+nZe0bb8oJn/tk3RYpiBHe0WCV2uiUn5cKwYLoIjqIj/bd+q6F+aJuhNRNHXCFF2Lnpg9AViAtkJyXXnlNJnainNjIkmyYM4TMzS1G2SxUQ5g6ZbKmCZi073Wj104EtFtTOiBo5fJf8bqYa2HaSWGksRhTJ/q4Xdtg56cXQIDXvbFR9pozS0UvP3qaHxHFP+aFrYVxHeRM7z1nD33JUpHQwvm4VRyQBoO20we+fd/h6Koph/xrYh7PPveilk5DDnph575iPAoj6ZBZvIlB1RA8W2AFNliTUpsvXe6IlCM1BJsv3YPCO9p3tv++joih7ry3aJmbfxfl6vCifJVKdWNr0fxerRSAWGaN8lJ3im+rRFs7nUoINToKN21pGTBUktCNaRlJbl6nG8uSkYS0NDl58NpQZqjaAyK4aLTh1rh1p4k0BrQ41hQAlD2LxjW67Ao1yqd57UhXIyRVuSxTGqqC0VcXCYzwFqOG2rzLLji9bkDDKLyAWWqjE/IvnY/ZB0rWhS3M+c2PSm7dulUGk875bvoAXs8tyYaPypEP6ubc5i8sIoNuzjFykDEPiG6pEm7550LQEUFe9epq2X/fvUpGQG0epPxKG/nnaQ7ytu3ahQ0pIyh1lp/v6sr7Wkjv7FklXzI/hzd/EKKgiLBC1MtFQnEO0hPwcb9bxQFRbLTwxcZCSDPWwN1saHLviKTj3rB+iKhjPcrVXsY1XUktFkHGs/Lq6tdlnzl76pqUqueL+8AzCSGe1Ok0CilsaEHhNVlB/41xxVDly91Rb5gHa9J4Arm9+BVr7nBkTT/GHtDd/whkJjqmFK1bWy1SmuO7Y5tTR7fG4js85xw2sw1Fb4eqFyDq3bN5g0ya83cqv5r6kE6qiGtDBzeXtcxmLufj/yTKRChDNLRABDjeMRnJwBppdkqDVacsWLXWoNLrUHgrJeff8wIrvIgwnn/JSvnWDcvqRjeswgugpTarIYe1o6O9LPP8zmuQLUgvOmIhh2zG9OnDDSyKlWTLb42LF3G7vKGtrtPOuGM4fza/KUKxCRVGQLGRDVFpN+3C64MDicZH+JBuN4cV14S8jdbVbLQocTHhdSOhEMDGeExmzZzhVDgoqE2czeVUNCGGyEFGFDe/Zq3+YzfUFc6EATaqIf0Dbzpwb24DklI1pXHNUpF0N+8Wa49rITUBa1yqm5uWyWtM6MbKQiHOfzNF4TVZSX+OQRRRP1of+ogdlRyGa6rGIVEoK7Vrs4SyEV5UNEBkEpKGBg792/USsY7dJA7xjDpSlty+SdMevHaDG42ss7lt21DE2a4qz2jXNvk68nV7t3ZJZstqaZ32NkdSob2ZlFPNoGAzl7OhCzVtE2WrOSBKn+rfLnFU3lDRRmmwlJM3rLnGjgAHMRJM4TV5soI1JjDC+9G5p+a/gZVEPCaLTpsvh/3Te+tGPMzCC6ivr1uvuaujddIqtgBIIUA5M4jWYHJQmhubZMpkROda9SNsvGEpLFvmXgepEIlYTJLptH5sXy7HFtKDGrf5aQflckERTcaB3NhqHH998WWJoHIDNrwZ5LDmv2apKDHGbN68RZKptA6HbEL8EC2GSCM/F5FbrItKY5HaxNgk51a8QFmzwpxYyHphu+lCHm7Fi8JNZfnjILaQa0Tld80l3hlBLqy+kb/pLf9NjnsN3C8i0vll8grnl5/iQeGtxtPsj2s4O+qdmqrIpUS00t08NZyvGktoZQ68CUskGocit070FuKsdWtRvxWNChB5RHS3RH4vuqkhChtH62LUbK3h4UZ8GyfPqlmHsFLTdzet9a9/UXN+mybNKlqeS1MZ3FxW3QgnOg5dzYo1dYA0J7eh+UW7RtJxOOuXHBLqlP6OCDfqAQfloPAGZaXM5xkY4YUkuQfkBvmD9T7CLrzgC3HFx9am5WjcTUlvdW3WaOpB79xfc1URMc5kUpLO5ASlupD7W6wkmxspfUPbxe4lnZ3la3bmS49pLiiqJyAdwo1U2j5HhTnIW7dtk333nrPLBjWT6+ZHiTHe3QSGPDwIL1IjsMkNgg7xBDN3LUq1YHYrNWBsuZzYwrQA/YeqoDwbRNndVFbqfsAT6wbJyK+qgPEjIt8TOkrmW2u7YqRAaLrLBBVdk2cOb25QUxm1oNlpzeSJC+YYRA11c1V+kwFUKMjmJCrZ4RJTw5JbJmfWze9F1FErHSAaGYlKuqdLJblazTHKkR7oel0aJ86wqoXrdeXyqzQMbl6v5csQ2UZVi3KHtgrGBrgdW4f4TCya+6w1g7NpSUzYtfQiroE3FbhOtKWj4rxtrwxszqfw2tAKxtjACK+L0y107/45iJ3WgvFoOLN0u7SN1qK2WHkvRP7czm641uo1r8v0aU7HrMIDkuaWvHIjpaXKpOWfi4ggZPq1NWtHzQXNPy8/7cJ0PfJzWPNzkAs3k5lezx2Hyg3xGNoa92utXJgf2EH6sDGv3PHa6rWy56zpKsJupQbU0S2XE5t/PVRuQD4vauXi/GLl2Uq9fn4kHSkYaBDibjizrd6A1+jp6dWGE7YH5hyNNGjqRtCP8VCWrFprlOrv1eocjah/WyTlYbTXSff3alMHN51Cc3yR+4prTppR82jkwOZ1kpiwe9Va4Y52v4VlybCpLz3QoyXI4oblxtBpDrWUEckt9sYAJdKQJoJoebHyZYj0otU0IvGoKVx0w2I247yxQb1h5Ftrh82oExHHGxmkvIz4s1Oz2/3aaBxMv07hNSUVnHGBEd5HHn9Srr/1e9LVhZIzO/O5WJas9g+b26Utf1e8+6rlNiVhTH55smJlyfB1iC4+qofg5UcTIdFo04uOZ4WHm5+Lj90hd2hpDGkr1xWu8BomQo1zSuWw5l/PjTSatM51z8sXaKRxzJqxu4orhB/5xvghPlp0FaL57HMvKTukCeD3wlJo5Z4QtyPd9GlTy6aO5F+jVCRdS8c1N2onOLf6hltZo9ZPKVMaak3Yf9c32bRmM2vNZ8XH+QN9ktzyhrOJTmv6Ormsmtc6lPtrc91yYwe3vKGyWetUCufnWO+IsmT4O6RYoIavVqvI5STeMcWolBqitdhwWKrZx2jlyyCziBhn+rudrm+xxuF8Yrx50cj7UA1gZ89BTpts4Jd2UnPLp+mf0YIYY4ZKqjVAgKOSgxhHY3otJ5fYqVtselB4TUkFZ1xghPfo+Qvlyv9YKPvvO8c3dMdDSoMLu7CurvuxOYRmtPJebsOFWLRBf+DmpwS4H1+XkjTIFT4WdzuU5dfedWuwunMs16q42EMzWvTaJIc1/7omrZgLBRo/3JFji5x0dBR7x/776CVLVWlwXy8/ugrJRPmzwkYfZUV3qEEF1q8xHtdudm5Dh1LnuRFkpBOViiCjmgI2mI31QeEda+L1f71qC2/hHaU1ArpD4q0TtIUvor+oUKAVDRpbPTfOcF8vuW2jRjtR8aCWRzHhxeshKgv5jLdOlGR3l84l1rKzIVGpOUE8Ee1FtDbeNlnr9eYfbm50onOaymhhTq8KbENkeLMiOrkl2icbCXdZTsMS7GzK03xk1Bgezkfe2bgDOcml3sRQeGv5NNbn2oER3hNO/6rceeOl9aFU4lXHk/ACAfJN8U4bubnlGhwUw4UNcC1NCclknXJZiERCmkwObDTDDx/ktEKMy9XeLZXXWup1CqPXleSw5l8bpduQa1vYiQxjXIFGZQLJibY4BgOU+IJ45keJSwlvqegq3oDg/GLVDtz5lWtQUWoDXWFVBbzJqLQdtslaVzqGwlspueCeV2vhBRlUkEC0EkIXbWp1Io3Y/DawQ1MRvHaLG5be7Zv04/1adjErJbyYA4RwcOsGQW1dRF1x34n2KUU3tRU+MXgzoC2jESFunzRCIBF9TW5Hfj4K/7qlyxzJRARXI7P4D+kLO7ZrwwzINtoou1Hdqr4R0OYSSJcY2lSXRrONjNYtRjWJ/Eg+hTe4PxtKzTwwwvuzhx8TfOz74UP/UXeru8ceM6fVbVXGm/ACND4Cx4Yvm4/N3QVa9cqrMm333Y1a6xYuKl4XH++jssBox2itkAvPR+Syv39QS2TZ5LCWmkd+ubR8gdZ881xW0tnccIewwmu4UeIJHW0jUhpMoquFJd3ca5s2qICsI78aZdbKVVUYjX89vk7hrfG2B0oAACAASURBVAf1+r7mWAjv8PdQT5f+L9IPhv+ud4t+RI6IaDUOiDXkrtJGHKPNoZzw6rm5nDh5xbtpfd6Uu4lvlE1temo2I+mBXkl1bxaJRCTR0qHCir/HrwZXbLXhB9IT0M0N6bjIy0WK4tDv+Pk40CPZgX4nx7e5XTfCxVon1uzNgKZIqAC70WDn/6Od4WhXPtpzMZ6+HhjhvfsnD8t/3nGPRsM0QjZ03HPH1Z7W66ln/ipX3XCHpFIZOeKwQ+TUEz6zy/UgEhddcaO8uOo1OfKIQ+XcMz6vY8aj8HqBXSyH18v1Sp0LyYT0oovXaIebg4xKAzOm7z6ie9do55b7+ksv/03a21q0/S3SFfDjHT/cTWr0IjUDHwGiikTvjn6t3OBWPxgtuprf3c6kVW/hPeB8yC5yiJGqYtLFzgunap1L4a0WyeBcZyyFF1QQfUz1blYhRDUHHMhlhcBByKpxaPc5pEw0l691XslrjSq8QxdFikW0qV0j2k5jkG3amljFUAU2PSyyknH+jNQEyH9DNOrkQff3SLRlgqYoaCTX0VuVW+06N9omw1xO6xZri2TJSqavW6PAuGYU+b4QZnjykDjjz/pH/Uk79CX9vUErcGiOsjtH/X/n7zSnV/8+4ozLq+7BCG8lT5m/zwmM8B513Fnyzesult2mVK9oN75J5p54jqxYukhm7zlDTl54iSw8dZ4c9M79Rqwaon6vrl4rq15ZI2vWbqDwVvhMj5XwYnqQtrc2bymZ11osB9lL3WEXiSvQyWRSN28hWormFLY1etdv2CjbtveohBdLjyi3BLgP5EkjIl6uLFmpa2AzXSW1lyt8LKpyGoW3KhgDdZGxFl6Fk8s5zSpicd2wpdLb143CsxIb+rNXiKieUM3IsTsfU+HF+GT3WxLFpj3U1c3lBCIOmVSpjcSG5daRyVjRKhnocJdBlztEZysUeG0BjRbJOVGRTqFrXSympdQaIK0NDcPBBFVdeLAr0/ga8oSHRN3Z8AZBd66pv+PPQxvh3L9TCUYd57ad0Xyva8rz/UEgMMJ76jnL5aavXVjRR+mlUL/0ymq5cuXtcsc3lusQRJGRp3rmyccVPeX+Bx+RVa++TuGt8NkdS+HFFAvzWsvlsLq3hDJdiAzbljkq1lkM+a84KqlUMDAwoLmyo1VpCJO0VvhY6WkUXi/0gnluXYR3CJVWN+jv0fJbkF/8Pza15ac8eKGqNWtz2WGp9nKtSoQX52AjG1IRvIg8fuZmdmwVlDPTHOWm1uHoeLl70ihyJiPZTFJiTSOj3Yj6ImcY16pWZH3EXIYkGKlnPMJFIDDCe+PtP5DNW7bL4R86RDcFucd7Dj6g4hV5/Imn5YGHHpEVFy/Wazz2uz/Krx97QpYvOcNYeDFw3rx5Fc9hPJ04MNAvTU213YlcyBPpKHjDn8lktcObWy6rVA4yJHXd+je1e9tohyvQiMS6tW/zn83Rzi/3dUSIMcdKZNnL6wb1XI3khOBYsGCBLFu2XGbOHP35C8HteroF901OPL5rXW9PFzY8GRu9NM810aL5pvrxe6pPYu1TDa9Qfhgix3iNarU97u/v05+/Nm/mVeRRhaHD+z1p1Ya+7U5+bGOzdnvTdF4tK4boqyO5+H+NGqPJTjSmfxcv0swCaQ5Ie4i2dg53eKsK+KGLhOVnSjWZBP1agRHe0758WVHWiPqWOta+sVEuvfrWol++9rLz5KlnnpefPfzosPCi1u9vfvsHK+GF9Bx77LFBfw7GZP6oPNCIxgpjfGzevFUiUacLmMmBtIRt3T0l83mxCQySi6iuW2Whkk185eaC18A1R8vZNbmf8TDG4R/8iAwaT1x88cUyY8aM8bBsnu4xiwYF2azEYjsDIJ4uWOHJTgexQYm1T9GGCVlEfosIWiWXhySqcOZtlqvkOjhncHBAGhtH3/RbeH2ULcv0bxcRZ++MpjYUDNKc2qHcWfdLw3m2yItFtzZsXtPKDCKZVJ9IKimRxiaJNCJfuM352lCebf7lUTEi07dNEkU2kTkR5G1OI4vWzqqWdkNtdx7hIhAY4a0Eeyab1Y5VxQ6U1Vr16hq5YuXtcidTGirBa33OWKc0WE8w74RiZboq2QRW6RxGq8Nb6XXDeh5TGsK6sqXvq54pDbtIYWrQ6TCG3NJoTNK9WyUxsTq7/NEKGRUQirXstVl1mxzeXe4PNWxHfIqyqwy6+bTuuW4kWTeUaa7vrk0fILPZIamPNLZKrLm9aF1cRIUHt22Upil7FL1ljbb3OqXRkN+LjX9eD25a80rQf+cHRnife+HlovQOfMfbKqaK6MDcE8+VFUsXyhxsWlu0TM4+5Tg5+MD9BdHhvv6BEY0umMNbMWo9MUjCi/kinxtl0PCDG2XBEEWsZBNYJdQovHbUKLx2vMIw2k/C6/LUzV3ZjEYbU91d2qK4GgeirEgHQBOHSg8vwlvpa5qehw1k2cEdWpIMtRVizahJ3DaymkMuKwNda7VW8M7KDyNfAZHeVM8W7VznbGzbWdHJdC7Da5lK2p7C8T4nEBjhPWXxsmGU6Ez12up1sues6fLtm4unOphyf/Lp5+Sq6+8QtLH9yGHvl9NOOkZP/eF9D8matevl/LO/IIgUH338Qt35n0qntVzUysvOl9//7jGVoPnz55u+3LgeFzThxWJt2tQlyVRKpk6ZPKaVCyi8dt8qFF47XmEY7UfhBVfIqRvtRR5u4+RZVcGNKgXpHVu0DS/KokXRmW208l55r+xFeDVKa/FaXm4YHdHS/b2SHeyVhnizxLDRLe9e0Y4ZZdLQ8a7U4XR52+o0lNAWxXGRiPM78oNNuDHC62UV/XluYIS3EN+GN9+S79/7oJxzev1kk3V47R7qIAqv3R1WbzSF144lhdeOVxhG+1V4hyOE2zepWEGAG6eMXhPcaE2yackM9mvOKtIBnM5gTdrW160NXOo6NsKr5buQj4y2vOmkZNJJaezcvWQbXqO5VzAIUd/MYN/QvTY695lolnTPZt2sptJf5gAn1O/NZlKaR4yucG7d4MhQtze365v7Z1eGKbwVLJjPTwms8ILrkuXXyeUXnV03xBReO/QUXnNeFF5zVhhJ4bXjFYbRfhdefS77e7RxQyY1IC3T9tFNWdU8IKSu/EJMXfmNJlp2+di/pPDmclpSLYuWu/hd83VzGkFFaoAr1Wg9jBzlqrb6tYCh9wnZT/ZrPjFkONo6SRIdU4witvkvBQl2qkLgV2qEFA/XFq5SMxGLW+TQGhMIjPCi8cPwkRN5bc06ueenv5Rbr11aY0SlL0/htUNP4TXnReE1Z0XhtWMVltFBEF6wRrQUKQ7p3i3SMnN/5yP1WhwQV1cKU85mbTf1oSHRJH19fdLS0uoIHkQZEdz0oMpeRNMkEk66RMxJAyh24D6ijS01a39sjCWb0ahvcst6yWZTEm+botHeCMqdRb1V7XBlODNUlcJ4ThzoewKBEd7jv7RkGCZKNc2avrucNO9o2Wt2/epVUnjtnm8KrzkvCq85KwqvHauwjA6K8Lq8semsf8Or0jLr7SqNtT4QwcwmndSHXHJABpIpScQi2kENNXA1chtr1MYZNgc25kGIUXvYDwci6LjPSKJRI8Coe+ZEuput85zz74cpDX5Y3erOITDCW93brs7VKLx2HCm85rwovOasKLx2rMIyOmjCC+6Irva+8bzEYi3SNG3voqW6arU+PVu7pK1zclU2n2nt4WxW4u1Oe+V6H+n+bo1ao0GG5uymBjT1Ab80aj0kvzZyT+Gt96pW//V9L7zP/vUleWP9m/Lxwz844u4fePhR2Xv2LHnn2/etPhXDK1J4DUENDaPwmvOi8JqzovDasQrL6CAKr8t+cPubMrhptcQ7p0vz1CptaBtlYW02rZk8I5mBHo2oeq0PbPJaJmO0XnF/jyQ6dx8x3JVfRLqRm6ypDwl0enNKTpY6KLwm1IM1xvfCe8b5V8hZJx8r++0zewRZ1OW97dv3yXVXnF834hReO/QUXnNeFF5zVhReO1ZhGR1k4XXXYKBrjaS2bZLGKbMl0blbTZem2sKr33eooNDf7ak+cDVv2qnDu1lr9RY7NPqLzm1D0d/81IfC6C+Ft5or449r+V54/+3zi+XHd12zCy3UBfz0/EXyX99ZWTeSFF479BRec14UXnNWFF47VmEZHQbhxVogzaH/zb9JLjUgjVNn1yw3thbCi/mjUkSqB0026refJv+ZhtQObl0vTZP3GLV6Q37qQy6b001vbgSYwhuWnxQ778P3wnvkMafLz37wjV0+ekCXtCM/e7o8eM/NdVsVCq8degqvOS8KrzkrCq8dq7CMDovwuuuBj+MHNq3W5ghNU+dozdlqHrUSXkfa04KGEGiyUayFcDXvw+RayC8e3LxOEhOnGbcZ1k1+gzujv9HO6rSGNpkvx4wNAd8LL1Ia5h51uBz6gf8zgsivH3tC7n/wUbn+yq+MDakir0LhtUNP4TXnReE1Z0XhtWMVltFhE153XVLb39IIZbRlgjROnK61cCs+UKpM68ympLd7u7S2T5BooqlmUgrJjE/YzVgyK74vwxOTWzdIpKlNYs3thmfsHMYIrzUy35/ge+FFru7ir35dPvGRD8oBb3+bJp3/5YVV8otfPS4rLz9PDth/n7pBpvDaoafwmvOi8JqzovDasQrL6LAKr0ZMUbt32wZJ9W6TRPskiXVMFXQCK3XkclnJoYvYkNxqRzH8ymX0PNSm7UsmpSkWFUHXsYbIUEMJp3uZ19q1+fMa3LJe4m2Tqh6hrvS5RU4vuqdhTjYHhdeGVjDG+l54gfHl19bId+/+uax65W/64O63zxyZ/9lPyD577VFXyhReO/wUXnNeFF5zVhReO1ZhGR1m4XXXCBustLlCelDindMk3jpxqFkEhBatcpMawW2QnEqr+0ubR8RiI5pc5Kc0qBijZTAaUCQHtOuYbuByu6vFm0bNfy33HPmmQcXQJLXj3UCvRsxNDwqvKangjAuE8PoVJ4XXbmUovOa8KLzmrCi8dqzCMno8CK+7Vum+bklueUNy0ajEmydoswhHbmNDEdzRu7eVzeEd6tKmrYoh0akBlWVEf7VRBRpUlOi+Vup58luDCgg+2iMnkCYSS4z6bUDhHRVR4AZQeD0sGYXXDh6F15wXhdecFYXXjlVYRo8n4dU1y+W05BaisfH2ydZpCLab1hAFRvRXJTiT1BQJbEjTTWn4PYoIsvNn5+/cP0eGHzFtUJHLWqcT1PIZNc3rpfDWchXqc20KrwfuFF47eBRec14UXnNWFF47VmEZPe6Ed2jhnDJgmzX6apOXaiu8xZ4T5Bbjl+jvacllnD87f5d2/j+XG5LgmIoxosWZgR3aoCLeNtEXj59JXi+F1xdLVdVJUHg94KTw2sGj8JrzovCas6Lw2rEKy+jxKrzu+qHTWapni0pv1KAKQTWE1+jZ0coQkN+0irHksk6t3u5Nkh0ckFj7ZIl3TNac4WpuljOaW96g0fJ6Kby2RP0/nsLrYY0ovHbwKLzmvCi85qwovHaswjJ6vAuvu46p3i2aeoA0h3IlzMZMeMs8YNh8l+x+SzJ92zUv2NkkV73NcrbPdrm8XgqvLU3/j6fwelgjCq8dPAqvOS8KrzkrCq8dq7CMpvDuXElsNEv3btGNZhBfVDMqPPwgvO6c0KhC83vTg9KASC9SHwo2y0HeI1FUmyhdjq1az3KxvF4Kb7Xo+uc6FF4Pa0HhtYNH4TXnReE1Z0XhtWMVltEU3l1XMju4Q5I9myXW0imxlo4RA/wkvO7EUC84vWOr5v7GWzslkmjWWsJOy99BLZ0msbgkVOJ3boarxTNcmNdL4a0F5fpek8LrgT+F1w4ehdecF4XXnBWF145VWEZTeEuvJCQyM9ivm8QgkTj8KLzD4psa1DQHbHiLQXzzustlB/uGJL5DYi0Tavr45uf1UnhrirouF6fwesBO4bWDR+E150XhNWdF4bVjFZbRFN7yK4lIaap3qzQ0NEisbZL09fdLS0ur/tmvByK7SHVANBdNNvLTGfD3mYFevZdoY4tzC7mcdpPLZbNDlSOyIpLVPyNtAlFu241xbl5vtNO8SYVfeXJeIwlQeD08ERReO3gUXnNeFF5zVhReO1ZhGU3hNVvJzGCf5vcOZCPSPnGqRMYgJ9ZsZqVHocNcasdWaYgmJBKNOkKLag+aArFFu8xFmzucbnINqAMc0RJoyF12/4w6wen+bok2tlYUGWaE1+sq+u98Cq+HNaHw2sGj8JrzovCas6Lw2rEKy2gKr91Kdr+1XhKRrMphNNGiIjgWG8LsZjlydHZgh1PjF0LbENG5Q2hR7SG1Y5tGekerRYzIMCLHqANskwdM4fWycv48l8LrYV0ovHbwKLzmvCi85qwovHaswjKawmu3km4Or2TSkhncIZlkn6YEQHwjjS1G7XbtXrH2o5Fzi7JsSHOIlalFjDrAye2bNKc52tRmNDEKrxGmQA2i8HpYLgqvHTwKrzkvCq85KwqvHauwjKbw2q1ksU1rSBHIJvu0ExqOSGOzxDTym7C7eJ1Ha8vldFLFt1wtYq3EgFbHHVNHnTGFd1REgRtA4fWwZBReO3gUXnNeFF5zVhReO1ZhGU3htVvJ0ao0YJOXK7/ImY02tUgk0VJWIO1mUNvR+bWIIb6a11vkQE5zqvstTXFwK1gUG0fhre161ePqFF4P1Cm8dvAovOa8KLzmrCi8dqzCMprCa7eSowlv/tWyqIU72K8CjBxadEPThhYBOJCugU16kaZ2re1b7ED5M0hvJBqXWNvEomMovAFYbMspUngtgeUPp/DawaPwmvOi8JqzovDasQrLaAqv3UraCG/+lYcjv8l+SUzY3e5F6zh6eLNa57SSs0j390imr1ujvYUb+Ci8dVy8Gr00hdcDWAqvHTwKrzkvCq85KwqvHauwjKbw2q1kpcLrvgoaWWT6t0uijEDazaj2o3Wz2raN0tg5TRrymlnkvzKi2Yj2xhrbJJrXnY7CW/v1GetXoPB6IE7htYNH4TXnReE1Z0XhtWMVltEUXruV9Cq8eDUIZLpnsyQmzbB78TqPhvQiH7mw3XL+tNCdDvfnli+j8NZ50Wrw8hReD1ApvHbwKLzmvCi85qwovHaswjKawmu3ktUQXrwiOrhp1HTyHnYTqPNoFdp0yhHaEkfOLV/WPkkykXidZ8yXrzYBCq8HohReO3gUXnNeFF5zVhReO1ZhGU3htVvJagmvSm82K4Ob10rj5FnSgA5nATnQwS3Z3SWNE6cLOrGVOlLdXSLNHQG5K07TlACF15RUkXEUXjt4FF5zXhRec1YUXjtWYRlN4bVbyWoKr/vKg5vX6UY2v3dsyyeFqhOIUMdaJpRtQsGUBrvnKwijKbweVonCawePwmvOi8JrzorCa8cqLKMpvHYrWQvhxQwGt67X9r4oXRakQ5tQiJQst0bhDdJqms2VwmvGqegoCq8dPAqvOS8KrzkrCq8dq7CMpvDarWSthBezSG57U6LN7RJtbLGbVJ1HZwZ6JdPfLXFUcWgY2aiCwlvnxanBy1N4PUCl8NrBo/Ca86LwmrOi8NqxCstoCq/dStZSeDETbeQQb1LxDdKRSzub8GIdUySaaB6eOoU3SKtoNlcKrxknRng9cHJPpfCaQ6TwmrOi8NqxCstoCq/dStZaeFV6e7foJjbkxwbtSHVvkoZoQmJD3dkovEFbwdHnS+EdnVHJEYzw2sGj8JrzovCas6Lw2rEKy2gKr91KjoXwYkbocJbLZTWvN2hHum+7ZJMDkujcXSi8QVu90edL4R2dEYXXA6P8Uym85iApvOasKLx2rMIymsJrt5JjJbz6/djfo00c4h1T7Cbpg9HZ1ICTk9w53Qez4RSqSWDcC+9Tz/xVrrrhDkmlMnLEYYfIqSd8Zhe+Tz79nNxw6w9ke3ePtLe1yqLTjpf3HHyAMMJr9yhSeM15UXjNWVF47ViFZTSF124lx1J49XtycIdkBnaUbfRgdwdjO5oR3rHlPRavNq6FN5fLydwTz5EVSxfJ7D1nyMkLL5GFp86Tg9653wj2z73wskzoaJc9Zk6TZ597SZZcer38/Ic3Ungtn1AKrzkwCq85KwqvHauwjKbw2q3kWAsvZodGD+m+bm3pG8nbEGY38/qMpvDWh3stX3VcC+9Lr6yWK1feLnd8Y7kyvvsnD8ubm7rkzJOPK8k8ncnI4Ud/UR7+8S1yz933SCQSkfnz59dyjUJzbQqv+VJSeM1ZUXjtWIVlNIXXbiXrIbwqvalBQW5sLpPWRg8xVHFoaLCbfB1GU3jrAL3GLzmuhffxJ56WBx56RFZcvFgxP/a7P8qvH3tCli85oyT2X/z6v+WXv/m9rLz8PI3wNjQ0yLx582q8TOG4/MBAvzQ17Sz7Eo67qs1dJJNJfTMVi5Vuf1mbVw7mVbFJBp/YBP1YsOCLsnz5cpk5c2bQb6Xm889k0pLNZiUeT9T8tcLwAv39ffrzF/9m1ePIZVKidW8HeiWSaFH5jcQb6zEVo9fMZjNG4zgoOARCLbxr39gol159a9HVuPay8+SpZ56Xnz386LDwPvL4k/Kb3/6hpPAiInzR5TfI9VdcINN2n6LCix+4xx57bHBWvI4zTSYHJZHw7w+4OqLZ5aXT6bT+wxSNBqdPfT35RaMji8bXcy5eXnvBggWydOnFMmPGDC+XGRfnQkiy2RzfFBquNj41amz0x89fSG92cIdILieRprayLX4Nb6/qw8LyM6XqYAJ8wVALbyablb6+/qLL09baIqteXSNXrLxd7jRIaVi3fqOcc9HXZfmSM2W/fWbrNblpze7JZ0qDOS+mNJizwkg32md3lv9GI8J7+eWXM8JrsDRMaTCAlDekXikN5WaZTSe1ogMEONbcofIbicXtbqxGo5nSUCOwdbxsqIV3NK6Izs498VxZsXShzMGmtUXL5OxTjpODD9xfEB3u6x+Q/fedI11btsmZ518u555xglZncA8K72iER36dwmvOi8JrzorCa8cqLKMpvHYr6Ufhzb+DNEqZDfSINEQ14httarW7wSqPpvBWGagPLjeuhRf8UXLsquvvkGQyJR857P1y2knH6LL88L6HZM3a9XL+2V+Q2759r3zn7p9J54SO4SX71g3L5OEHf8FNaxYPMYXXHBaF15wVhdeOVVhGU3jtVtLvwuveDergQn5zyQGJNLXWrYEFhdfu+QrC6HEvvF4WiRFeO3oUXnNeFF5zVhReO1ZhGU3htVvJoAjv8F3lsprqgAoP8fapEkk02d2wx9EUXo8AfXg6hdfDolB47eBReM15UXjNWVF47ViFZTSF124lAye8Q7eXy2Yl1dMlDQ2RMe3cRuG1e76CMJrC62GVKLx28Ci85rwovOasKLx2rMIymsJrt5JBFV73LlHVIdndJfGOqRJtbLG7+QpGU3grgObzUyi8HhaIwmsHj8JrzovCa86KwmvHKiyjKbx2Kxl04XXvNtX9lv4vxLeWB4W3lnTrc20KrwfuFF47eBRec14UXnNWFF47VmEZTeG1W8mwCK9+vw/2CcQ33j65ZjV8Kbx2z1cQRlN4PawShdcOHoXXnBeF15wVhdeOVVhGU3jtVjJMwuveOVIcJJfV3F7k+FbzoPBWk6Y/rkXh9bAOFF47eBRec14UXnNWFF47VmEZTeG1W8kwCi8IZJP9kuruklhrp0Sb2+2glBlN4a0aSt9ciMLrYSkovHbwKLzmvCi85qwovHaswjKawmu3kmEVXpdCqneL5NIpTXNoiMbs4BQZTeH1jNB3F6DwelgSCq8dPAqvOS8KrzkrCq8dq7CMpvDarWTYhVejvakBreSANsWxlp2NouxIOaMpvJVQ8/c5FF4P60PhtYNH4TXnReE1Z0XhtWMVltEUXruVHA/C6xJJ79gqmdSgxJraJBJvqijiS+G1e76CMJrC62GVKLx28Ci85rwovOasKLx2rMIymsJrt5LjSXidaO+gdmpD1LehoUEi8Wbt1gYBloaGUeFReEdFFLgBFF4PS0bhtYNH4TXnReE1Z0XhtWMVltEUXruVHG/Cm08nm06p+GKDG36PxBIqvsMCXAQlhdfu+QrCaAqvh1Wi8NrBo/Ca86LwmrOi8NqxCstoCq/dSo5n4S0khegv5DeTGhBJJ4fkt9lJf4jFdTiF1+75CsJoCq+HVaLw2sGj8JrzovCas6Lw2rEKy2gKr91KUniL88rlcpJLDUgm2a+/57JZJ+2h2dumN7vV4eixIEDh9UCZwmsHj8JrzovCa86KwmvHKiyjKbx2K0nhNeSVzWjaQybiRHp5hIcAhdfDWlJ47eBReM15UXjNWVF47ViFZTSF124lKbx2vJjSYMcrCKMpvB5WicJrB4/Ca86LwmvOisJrxyosoym8ditJ4bXjReG14xWE0RReD6tE4bWDR+E150XhNWdF4bVjFZbRFF67laTw2vGi8NrxCsJoCq+HVaLw2sGj8JrzovCas6Lw2rEKy2gKr91KUnjteFF47XgFYTSF18MqUXjt4FF4zXlReM1ZUXjtWIVlNIXXbiUpvHa8KLx2vIIwmsLrYZUovHbwKLzmvCi85qwovHaswjKawmu3khReO14UXjteQRhN4fWwShReO3gUXnNeFF5zVhReO1ZhGU3htVtJCq8dLwqvHa8gjKbwelglCq8dPAqvOS8KrzkrCq8dq7CMpvDarSSF144XhdeOVxBGU3g9rBKF1w4ehdecF4XXnBWF145VWEZTeO1WksJrx4vCa8crCKMpvEFYJc6RBEiABEiABEiABEigYgIU3orR8UQSIAESIAESIAESIIEgEKDwBmGVOEcSIAESIAESIAESIIGKCVB4K0T3gx8/KPf89JcSjUbkS1/4rBz2T++t8ErhP23J8uvk6b+8JPF4TG/2muXnyr577xn+Gze8w//901/kG7f9UF5dvVaWXXCGfPiD9++08QAADqpJREFUO5+lp575q1x1wx2SSmXkiMMOkVNP+IzhVcM7bOmVN8n/PvUXmTixQ75/64rhGw3yc5ZOp+Wya74pzzz3orS3tsolXzlN9po9M7yL6PHOPjb3SxJPxPUqzU0J+dHtX/d4xXCd/v0fPyj3PfAreWPDJnnw7pukc0KH3iCfs13XuWvzVrnoihvlxVWvyZFHHCrnnvH54UF8zsL1fUHhrWA9163fKIsu/JrceeOl0rujT05eeIn86PavSVNTYwVXC/8pEJHPHv1ROfjA/cN/sxXc4br1b8rgYFLu+uH9cugH/mFYeHO5nMw98RxZsXSRzN5zhj5nC0+dJwe9c78KXiU8p/zpzy9IU2NCLr361l2EN6jP2QMPPypPPPlnueyrZ8lvf/8n+fH9v5Lrrjg/PItW5Ts58pjT5ec/vLHKVw3P5V56+W/S0dEmpyxeLt++6dJh4eVztusa499wBBtWvbJG1qzdMEJ4+ZyF53sCd0LhrWA9Ed3t2rJVzvzisXr2BctXypGHHyr/9x/fXcHVwn8KhddsjRG5/OD7/8+w8L70ymq5cuXtcsc3lusF7v7Jw/Lmpi458+TjzC4Y4lFr39go519ybWiE97yLr5FPfuww+cD73iV4o/Pxz54u9955tbS2NId4FSu/NYqIGbujjjtrhPDyOSvN7f4HH5FVr75O4TV7tAI5isJbwbLdcOv3ZOqUyXLMpz+qZ19783dkzp4z5OgjP1zB1cJ/CoT3hZf/JtFIRP9BP33B5yQRdz6O5LGTQKHwPv7E0/LAQ4/IiosX66DHfvdH+fVjT8jyJWeMe2ylhDeoz9kJZ1wkSxYtkP32ma1re/yXlsjyJWfKnD2mj/u1LgYAHzW3tbVILBaTz3zqCP7sLfGUFAovnzM74eVzFq4fPxTeCtbz+lu/J7vlCe81N90le82exR+6JVhufLNLdps6SfoHBmX512+R/faZI1847lMVkA/3KYXC+9+//5P87OFHh4X3kceflN/89g8UXhEpJrxBfs5OOP2rsmTxF4eFd94pF8ilXz2LwlviW379xk0yY9pumqO66MIVcuHik5kyVYTVLsLL56zkPyLFIrx8zsL1by6Ft4L11JSGzVuGP1r+yrKV8q9HMKXBBGVh1NLknPEyplhKwxUrb5c7mdKwyyNQTHjzBwXtOcNHzUd99J81LUpTGj5zmtx71zVMaTD45i/8xM3glHEzpFhKA5+z4stfTHjzR/I5C/63DYW3gjXEP7aLLrxK7rrpMunp3SGnLFrGTWtlOK5+fb2mfCRTKbns6ttk5vSpcvLn51ZAPtynFApvNpuVuSeeKyuWLlR+Jy9aJmefchwjWSUivEF+zu5/6FH5wx+dTWuP/c9Tcu9Pfyk3rLgg3A98hXfX3dMr6XRGJk2cIFu2bpfTz7tcFp92vPzDuw+s8IrhPa1QePmcmac08DkL3/cFhbfCNf3+vb/QTUSRSEROP+kY+fCh76vwSuE/7aSzlsqmtzYrq/e86wD58pknSHNTU/hv3PAOn33uJS2L09PTK4lEQv8h/8FtTrmtJ59+Tq66/g5JJlPykcPeL6eddIzhVcM7DJ+oPPf8K7Jte7dMmtQpC47/tBz10Q9JkJ8zlIta/vVb5em/vCgd7a2y7CtnyN5zWJas2FP82uo35ILl10pf/6DEYzH51McPk/nHfCK8D3wFd/bdu38ud//0Ydm8ZZtM6uyQQ/7hXbJk8QItS8bnbCTQTDYrRx+/UAYGBiWVTkt7W6usvMypkMLnrIKHz8enUHh9vDicGgmQAAmQAAmQAAmQgHcCFF7vDHkFEiABEiABEiABEiABHxOg8Pp4cTg1EiABEiABEiABEiAB7wQovN4Z8gokQAIkQAIkQAIkQAI+JkDh9fHicGokQAIkQAIkQAIkQALeCVB4vTPkFUiABEiABEiABEiABHxMgMLr48Xh1EiABEiABEiABEiABLwToPB6Z8grkAAJkAAJkAAJkAAJ+JgAhdfHi8OpkQAJkAAJkAAJkAAJeCdA4fXOkFcgARIgARIgARIgARLwMQEKr48Xh1MjARIgARIgARIgARLwToDC650hr0ACJEACJEACJEACJOBjAhReHy8Op0YCJEACJEACJEACJOCdAIXXO0NegQRIgARIgARIgARIwMcEKLw+XhxOjQRIgARIgARIgARIwDsBCq93hrwCCZAACZAACZAACZCAjwlQeH28OJwaCZAACZAACZAACZCAdwIUXu8MeQUSIAESIAESIAESIAEfE6Dw+nhxODUSIAESIAESIAESIAHvBCi83hnyCiRAAiRAAiRAAiRAAj4mQOH18eJwaiRAAiRAAiRAAiRAAt4JUHi9M+QVSIAESIAESIAESIAEfEyAwuvjxeHUSIAESIAESIAESIAEvBOg8HpnyCuQAAmQAAmQAAmQAAn4mACF18eLw6l5I7Bk+XXy9J9flLa2FhkYTMqsGbvLvx/1L/LhD/6jtwvzbBIggXFJ4KNzT5WmxkaJxaKSSqXlgLfvLfM/e5S8Y7+9xyUP3jQJBIkAhTdIq8W5WhGA8L77oHfI3E8eoec99ezzcslV/ylfOO5T8qmPH2Z1LQ4mARIgAQjvjV+7UPaZs4ekMxn5+cOPyXW3fF9uvfY/ZN+99yQgEiABHxOg8Pp4cTg1bwQKhRdXe+j/PS633vVjue/b10rXlm2y9IobZUdfvyRTKfnA+94lp5/0OY0G//vnF8udNy6XKZMn6iSuuekumTSxU0743Ce9TYpnkwAJBJZAvvC6N3H5Nd+UTDYrF517sjz59HNyy533SiqVkkwmK5//3FFy+IcOkedfelUuvfpW+f6tK4bvfd6pF8h5Z54oB71zv8Dy4MRJIEgEKLxBWi3O1YpAMeF9fd1G+exJ58ojP71dGiIN+o9SS3OT/oN13tKr5TOf+oi87z0HyU23/1ASiYQsOP7T0j8wIJ+ev1i+d8sVMmniBKs5cDAJkEB4CBQT3p/+4jfy0wcflW/dsEy6e3qltbVFopGIbO/ulZPO+g+588ZLpa21RRacvVROO+kY+fuD3iHP/OVFueam78i3b74sPHB4JyTgcwIUXp8vEKdXOYFiwrtm7QY5ZsGXVXgj0Yjc/K0fyfMvvSYNDQ3y5lub5d8/cbgcN/dI2bipS750zqVyz51Xy/0PPiLPPrdKLvnKaZVPhmeSAAkEnkAx4f3JL34j9w8J7/qNm+Q/77hHNrzZJfF4TF5+dY1cf+VXNMf3l4/8jzz2P0/JZReeKV+97Bvy3r9/pxz1sX8OPBPeAAkEhQCFNygrxXlaEygmvA/++nG5/bv3yb13XiPfu+fnsurVNfpRZCwWk+tu+a60t7Vpji8OnP8vHzpEvvXd/5Lzz/6C/N0Bb7OeA08gARIID4FiwnvZNbfpDV64+Ity7kVfl0Pee7D82ycO17/7/GkXyuLT5svBB+4v6XRa5p54rly5dKEsuvAque+ua6WpqTE8cHgnJOBzAhReny8Qp1c5gXzhzeVyumlt2ddukQXzjtbIys13/EjisZgsOP7fNG1h/pculI/9yz8NC++f/vyCXLLiZumc0C533cSPHitfCZ5JAuEgMGLTWjotP/vlb+X6W74vt61cKvvstYd8ceElcuqJc+U9Bx8gr61ZJ/NPXaKb3CC8OL75nfvkgYcelcM++F45+5R54YDCuyCBgBCg8AZkoThNewIQ3qeefUFaW5plMJmUPWdN15SFDx/6Pr0YUhiWXnmTTJ7UKbFoVCDFc/acOSy8GDP3xHPk+M/8Kz96tMfPM0ggdAQgvPF4XH9eoErDge/YR0445pOy/9v20nv94zN/lZu/dbfsMXOa7g14ftVrsujUecPCu6lri3zyuLPkR7d/XfacNS10fHhDJOBnAhReP68O51ZXAm9s2CSnLl4m99xxNT96rOtK8MVJIBwEUCXmV4/+Xq5e/uVw3BDvggQCRIDCG6DF4lTHjsBt375XHnj4t3LGgmPkiH9+/9i9MF+JBEgglAQWLlkheBO9Yuli2XvOzFDeI2+KBPxMgMLr59Xh3EiABEiABEiABEiABDwToPB6RsgLkAAJkAAJkAAJkAAJ+JkAhdfPq8O5kQAJkAAJkAAJkAAJeCZA4fWMkBcgARIgARIgARIgARLwMwEKr59Xh3MjARIgARIgARIgARLwTIDC6xkhL0ACJEACJEACJEACJOBnAhReP68O50YCJEACJEACJEACJOCZAIXXM0JegARIgARIgARIgARIwM8EKLx+Xh3OjQRIgARIgARIgARIwDMBCq9nhLwACZAACZAACZAACZCAnwlQeP28OpwbCZAACZAACZAACZCAZwIUXs8IeQESIAESIAESIAESIAE/E6Dw+nl1ODcSIAESIAESIAESIAHPBCi8nhHyAiRAAiRAAiRAAiRAAn4mQOH18+pwbiRAAiRAAiRAAiRAAp4JUHg9I+QFSIAESIAESIAESIAE/EyAwuvn1eHcSIAESIAESIAESIAEPBOg8HpGyAuQAAmQAAmQAAmQAAn4mQCF18+rw7mRAAmQAAmQAAmQAAl4JkDh9YyQFyABEiABEiABEiABEvAzAQqvn1eHcyMBEiABEiABEiABEvBMgMLrGSEvQAIkQAIkQAIkQAIk4GcCFF4/rw7nRgIkQAIkQAIkQAIk4JkAhdczQl6ABEiABEiABEiABEjAzwQovH5eHc6NBEiABEiABEiABEjAMwEKr2eEvAAJkAAJkAAJkAAJkICfCVB4/bw6nBsJkAAJkAAJkAAJkIBnAhRezwh5ARIgARIgARIgARIgAT8ToPD6eXU4NxIgARIgARIgARIgAc8EKLyeEfICJEACJEACJEACJEACfiZA4fXz6nBuJEACJEACJEACJEACnglQeD0j5AVIgARIgARIgARIgAT8TIDC6+fV4dxIgARIgARIgARIgAQ8E6DwekbIC5AACZAACZAACZAACfiZAIXXz6vDuZEACZAACZAACZAACXgm8P8B+2s/7VlJeNAAAAAASUVORK5CYII="
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Plot comparison\n",
"fig = plot_path_comparison_unconditional(\n",
" holdout_windows, synthetic_holdout, n_samples=30, asset_idx=0\n",
")\n",
"fig.show()"
]
},
{
"cell_type": "markdown",
"id": "3482a0bc",
"metadata": {
"papermill": {
"duration": 0.012669,
"end_time": "2026-06-14T16:58:54.767116+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.754447+00:00",
"status": "completed"
}
},
"source": [
"**Interpretation**: Both panels should show paths with similar volatility\n",
"and drift characteristics. Synthetic paths that are too smooth or too volatile\n",
"indicate the generator has not fully captured the return distribution.\n",
"The cumulative sum visualization amplifies differences: look for matching\n",
"spread (volatility) and symmetry around zero (mean accuracy)."
]
},
{
"cell_type": "markdown",
"id": "87a0b899",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.012349,
"end_time": "2026-06-14T16:58:54.791808+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.779459+00:00",
"status": "completed"
}
},
"source": [
"## 12. TSTR Evaluation: Train Synthetic, Test Real\n",
"\n",
"TSTR (Train Synthetic, Test Real) measures whether a model trained on\n",
"synthetic data can predict real outcomes as well as a model trained on\n",
"real data.\n",
"\n",
"For unconditional generation, we use a next-step prediction task."
]
},
{
"cell_type": "code",
"execution_count": 44,
"id": "59eee4e5",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:54.817649Z",
"iopub.status.busy": "2026-06-14T16:58:54.817501Z",
"iopub.status.idle": "2026-06-14T16:58:54.821948Z",
"shell.execute_reply": "2026-06-14T16:58:54.821588Z"
},
"papermill": {
"duration": 0.018179,
"end_time": "2026-06-14T16:58:54.822319+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.804140+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"def tstr_evaluation_unconditional(\n",
" real_train: np.ndarray,\n",
" real_holdout: np.ndarray,\n",
" synthetic_train: np.ndarray,\n",
") -> dict:\n",
" \"\"\"\n",
" TSTR evaluation for unconditional generation.\n",
"\n",
" Task: Predict sign of next return from path features.\n",
" \"\"\"\n",
" from sklearn.linear_model import LogisticRegression\n",
" from sklearn.preprocessing import StandardScaler\n",
"\n",
" # Feature engineering: use path statistics as features\n",
" def extract_features(paths):\n",
" \"\"\"Extract statistical features from paths.\"\"\"\n",
" features = []\n",
" for path in paths:\n",
" # Rolling statistics\n",
" feat = [\n",
" path.mean(),\n",
" path.std(),\n",
" path[-1, 0] - path[0, 0], # Total return asset 0\n",
" path[-1, 1] - path[0, 1] if path.shape[1] > 1 else 0, # Asset 1\n",
" np.corrcoef(path[:, 0], path[:, 1])[0, 1] if path.shape[1] > 1 else 0,\n",
" ]\n",
" features.append(feat)\n",
" return np.array(features)\n",
"\n",
" # Extract labels: sign of last return\n",
" def extract_labels(paths):\n",
" \"\"\"Label: was last return positive?\"\"\"\n",
" return (paths[:, -1, 0] > 0).astype(int)\n",
"\n",
" # Prepare data\n",
" X_real_train = extract_features(real_train)\n",
" y_real_train = extract_labels(real_train)\n",
"\n",
" X_syn_train = extract_features(synthetic_train)\n",
" y_syn_train = extract_labels(synthetic_train)\n",
"\n",
" X_test = extract_features(real_holdout)\n",
" y_test = extract_labels(real_holdout)\n",
"\n",
" # Handle NaN in correlation features\n",
" X_real_train = np.nan_to_num(X_real_train, nan=0.0)\n",
" X_syn_train = np.nan_to_num(X_syn_train, nan=0.0)\n",
" X_test = np.nan_to_num(X_test, nan=0.0)\n",
"\n",
" # Scale features\n",
" scaler = StandardScaler()\n",
" X_real_train_scaled = scaler.fit_transform(X_real_train)\n",
" X_syn_train_scaled = scaler.transform(X_syn_train)\n",
" X_test_scaled = scaler.transform(X_test)\n",
"\n",
" # Train on real, test on holdout (baseline)\n",
" model_real = LogisticRegression(max_iter=1000)\n",
" model_real.fit(X_real_train_scaled, y_real_train)\n",
" acc_real = model_real.score(X_test_scaled, y_test)\n",
"\n",
" # Train on synthetic, test on holdout (TSTR)\n",
" n_classes_syn = len(np.unique(y_syn_train))\n",
" if n_classes_syn < 2:\n",
" majority_class = int(y_syn_train.mean() > 0.5)\n",
" acc_syn = (y_test == majority_class).mean()\n",
" print(f\"Warning: Synthetic data has only {n_classes_syn} class, using majority vote\")\n",
" else:\n",
" model_syn = LogisticRegression(max_iter=1000)\n",
" model_syn.fit(X_syn_train_scaled, y_syn_train)\n",
" acc_syn = model_syn.score(X_test_scaled, y_test)\n",
"\n",
" return {\n",
" \"accuracy_train_real\": acc_real,\n",
" \"accuracy_train_synthetic\": acc_syn,\n",
" \"tstr_ratio\": acc_syn / acc_real if acc_real > 0 else 0,\n",
" \"n_train_samples\": len(y_real_train),\n",
" \"n_test_samples\": len(y_test),\n",
" }"
]
},
{
"cell_type": "code",
"execution_count": 45,
"id": "92754db3",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:54.848713Z",
"iopub.status.busy": "2026-06-14T16:58:54.848547Z",
"iopub.status.idle": "2026-06-14T16:58:54.933447Z",
"shell.execute_reply": "2026-06-14T16:58:54.932749Z"
},
"papermill": {
"duration": 0.099311,
"end_time": "2026-06-14T16:58:54.933968+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.834657+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"=== TSTR Evaluation (Holdout Test Set) ===\n",
"Task: Predict sign of last return from path features\n",
"Training samples: 3249\n",
"Test samples: 592\n",
"Accuracy (trained on REAL): 0.7686\n",
"Accuracy (trained on SYNTHETIC): 0.7331\n",
"TSTR Ratio: 0.9538\n",
"\n",
"(TSTR ratio near 1.0 indicates synthetic data preserves predictive utility)\n"
]
}
],
"source": [
"tstr_results = tstr_evaluation_unconditional(train_windows, holdout_windows, synthetic_paths)\n",
"\n",
"print(\"\\n=== TSTR Evaluation (Holdout Test Set) ===\")\n",
"print(\"Task: Predict sign of last return from path features\")\n",
"print(f\"Training samples: {tstr_results['n_train_samples']}\")\n",
"print(f\"Test samples: {tstr_results['n_test_samples']}\")\n",
"print(f\"Accuracy (trained on REAL): {tstr_results['accuracy_train_real']:.4f}\")\n",
"print(f\"Accuracy (trained on SYNTHETIC): {tstr_results['accuracy_train_synthetic']:.4f}\")\n",
"print(f\"TSTR Ratio: {tstr_results['tstr_ratio']:.4f}\")\n",
"print(\"\\n(TSTR ratio near 1.0 indicates synthetic data preserves predictive utility)\")"
]
},
{
"cell_type": "markdown",
"id": "7114848c",
"metadata": {
"papermill": {
"duration": 0.012339,
"end_time": "2026-06-14T16:58:54.958868+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.946529+00:00",
"status": "completed"
}
},
"source": [
"**Interpretation**: A TSTR ratio near 1.0 means a classifier trained on\n",
"synthetic data performs as well on real holdout data as one trained on actual\n",
"data. For return-sign prediction (a near-random task), ratios close to 1.0\n",
"are expected even with moderate generation quality. The key diagnostic is\n",
"whether the synthetic-trained model significantly underperforms the\n",
"real-trained baseline."
]
},
{
"cell_type": "markdown",
"id": "a56942d8",
"metadata": {
"papermill": {
"duration": 0.01258,
"end_time": "2026-06-14T16:58:54.983978+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.971398+00:00",
"status": "completed"
}
},
"source": [
"## 13. Summary Visualization"
]
},
{
"cell_type": "code",
"execution_count": 46,
"id": "fe1ac2b9",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:55.010219Z",
"iopub.status.busy": "2026-06-14T16:58:55.009965Z",
"iopub.status.idle": "2026-06-14T16:58:55.066830Z",
"shell.execute_reply": "2026-06-14T16:58:55.066251Z"
},
"papermill": {
"duration": 0.070686,
"end_time": "2026-06-14T16:58:55.067193+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:54.996507+00:00",
"status": "completed"
}
},
"outputs": [
{
"data": {
"application/vnd.plotly.v1+json": {
"config": {
"plotlyServerURL": "https://plot.ly"
},
"data": [
{
"marker": {
"color": [
"#0a1628",
"#D4A84B",
"#1a2d4a"
]
},
"text": [
"0.0420",
"0.1172",
"0.0004"
],
"textposition": "outside",
"type": "bar",
"x": [
"Sig-W1 (÷10)",
"KS Statistic",
"Mean Error"
],
"xaxis": "x",
"y": [
0.042001232504844666,
0.1171875,
0.00042026996379718184
],
"yaxis": "y"
},
{
"marker": {
"color": [
"#0a1628",
"#C87533"
]
},
"text": [
"0.769",
"0.733"
],
"textposition": "outside",
"type": "bar",
"x": [
"Train Real",
"Train Synthetic"
],
"xaxis": "x2",
"y": [
0.768581081081081,
0.7331081081081081
],
"yaxis": "y2"
}
],
"layout": {
"annotations": [
{
"font": {
"size": 16
},
"showarrow": false,
"text": "Sig-W1 and Marginal Metrics",
"x": 0.225,
"xanchor": "center",
"xref": "paper",
"y": 1.0,
"yanchor": "bottom",
"yref": "paper"
},
{
"font": {
"size": 16
},
"showarrow": false,
"text": "TSTR Performance",
"x": 0.775,
"xanchor": "center",
"xref": "paper",
"y": 1.0,
"yanchor": "bottom",
"yref": "paper"
}
],
"showlegend": false,
"template": {
"layout": {
"colorway": [
"#0a1628",
"#D4A84B",
"#1a2d4a",
"#C87533",
"#10b981",
"#ef4444"
],
"font": {
"color": "#334155",
"family": "DM Sans, DejaVu Sans, sans-serif",
"size": 11
},
"hoverlabel": {
"bgcolor": "white",
"font": {
"family": "DM Sans, DejaVu Sans, sans-serif",
"size": 11
}
},
"legend": {
"bgcolor": "rgba(255,255,255,0.8)",
"bordercolor": "#e8e8e6",
"borderwidth": 1,
"font": {
"size": 10
}
},
"paper_bgcolor": "#FAFAF9",
"plot_bgcolor": "white",
"title": {
"font": {
"color": "#0a1628",
"size": 14
},
"x": 0.5,
"xanchor": "center"
},
"xaxis": {
"gridcolor": "#e8e8e6",
"gridwidth": 0.5,
"linecolor": "#e8e8e6",
"showgrid": true,
"tickfont": {
"size": 10
},
"title": {
"font": {
"size": 11
}
}
},
"yaxis": {
"gridcolor": "#e8e8e6",
"gridwidth": 0.5,
"linecolor": "#e8e8e6",
"showgrid": true,
"tickfont": {
"size": 10
},
"title": {
"font": {
"size": 11
}
}
}
}
},
"title": {
"text": "Sig-WGAN Evaluation Summary (Paper Implementation)"
},
"xaxis": {
"anchor": "y",
"domain": [
0.0,
0.45
]
},
"xaxis2": {
"anchor": "y2",
"domain": [
0.55,
1.0
]
},
"yaxis": {
"anchor": "x",
"domain": [
0.0,
1.0
],
"title": {
"text": "Metric value (lower is better)"
}
},
"yaxis2": {
"anchor": "x2",
"domain": [
0.0,
1.0
],
"range": [
0,
1.05
],
"title": {
"text": "Classifier accuracy"
}
}
}
},
"image/png": "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"
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"fig = make_subplots(\n",
" rows=1,\n",
" cols=2,\n",
" subplot_titles=[\"Sig-W1 and Marginal Metrics\", \"TSTR Performance\"],\n",
")\n",
"\n",
"# Left: Signature and marginal metrics\n",
"metrics = [\"Sig-W1 (÷10)\", \"KS Statistic\", \"Mean Error\"]\n",
"values = [\n",
" eval_results[\"sig_w1\"] / 10, # Scale down for visualization\n",
" eval_results[\"ks_statistic\"],\n",
" abs(eval_results[\"real_mean\"] - eval_results[\"syn_mean\"]),\n",
"]\n",
"\n",
"fig.add_trace(\n",
" go.Bar(\n",
" x=metrics,\n",
" y=values,\n",
" marker_color=[COLORS[\"blue\"], COLORS[\"amber\"], COLORS[\"slate\"]],\n",
" text=[f\"{v:.4f}\" for v in values],\n",
" textposition=\"outside\",\n",
" ),\n",
" row=1,\n",
" col=1,\n",
")\n",
"\n",
"# Right: TSTR\n",
"methods = [\"Train Real\", \"Train Synthetic\"]\n",
"accuracies = [\n",
" tstr_results[\"accuracy_train_real\"],\n",
" tstr_results[\"accuracy_train_synthetic\"],\n",
"]\n",
"\n",
"fig.add_trace(\n",
" go.Bar(\n",
" x=methods,\n",
" y=accuracies,\n",
" marker_color=[COLORS[\"blue\"], COLORS[\"copper\"]],\n",
" text=[f\"{a:.3f}\" for a in accuracies],\n",
" textposition=\"outside\",\n",
" ),\n",
" row=1,\n",
" col=2,\n",
")\n",
"\n",
"fig.update_yaxes(title_text=\"Metric value (lower is better)\", row=1, col=1)\n",
"fig.update_yaxes(title_text=\"Classifier accuracy\", row=1, col=2, range=[0, 1.05])\n",
"fig.update_layout(\n",
" title_text=\"Sig-WGAN Evaluation Summary (Paper Implementation)\",\n",
" showlegend=False,\n",
" template=\"ml4t\",\n",
")\n",
"\n",
"fig.show()"
]
},
{
"cell_type": "markdown",
"id": "df8e3156",
"metadata": {
"papermill": {
"duration": 0.012621,
"end_time": "2026-06-14T16:58:55.092705+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.080084+00:00",
"status": "completed"
}
},
"source": [
"**Interpretation**: The left panel reports three distributional checks - the\n",
"signature-level Sig-W1 distance (scaled down by 10 for visual co-display), the\n",
"marginal Kolmogorov-Smirnov statistic, and the absolute mean-error between real\n",
"and synthetic samples. A near-zero Mean Error bar is good news (centred\n",
"generation), not a missing value - the bar is small because the generator has\n",
"learned the marginal location. The right panel reports TSTR: classifier\n",
"accuracy when trained on real vs synthetic. The two bars should sit at similar\n",
"heights for utility preservation; a gap exceeding ~0.05 indicates the\n",
"synthetic data is missing discriminative structure."
]
},
{
"cell_type": "markdown",
"id": "fe76e9cc",
"metadata": {
"lines_to_next_cell": 2,
"papermill": {
"duration": 0.012452,
"end_time": "2026-06-14T16:58:55.117607+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.105155+00:00",
"status": "completed"
}
},
"source": [
"## 13b. Additional Evaluation Metrics\n",
"\n",
"Beyond Sig-W1, we evaluate:\n",
"- Autocorrelation structure (ACF)\n",
"- Cross-asset correlation\n",
"- Stylized facts (kurtosis, skewness)"
]
},
{
"cell_type": "code",
"execution_count": 47,
"id": "8f846ab7",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:55.143610Z",
"iopub.status.busy": "2026-06-14T16:58:55.143380Z",
"iopub.status.idle": "2026-06-14T16:58:55.150900Z",
"shell.execute_reply": "2026-06-14T16:58:55.150551Z"
},
"papermill": {
"duration": 0.021606,
"end_time": "2026-06-14T16:58:55.151531+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.129925+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"======================================================================\n",
"STYLIZED FACTS COMPARISON\n",
"======================================================================\n",
"\n",
"Metric Real Synthetic\n",
"--------------------------------------------------\n",
"mean 0.0001 0.0005\n",
"std 0.0155 0.0140\n",
"skewness -0.9278 -0.0033\n",
"kurtosis 16.2203 0.2279\n",
"acf_returns_lag1 -0.2570 0.1929\n",
"acf_squared_lag1 0.4949 0.0512\n"
]
}
],
"source": [
"def compute_stylized_facts(paths: np.ndarray) -> dict:\n",
" \"\"\"Compute stylized facts of financial returns.\"\"\"\n",
" # Flatten to get marginal distribution\n",
" returns = paths.flatten()\n",
"\n",
" # Basic statistics\n",
" mean = np.mean(returns)\n",
" std = np.std(returns)\n",
" skew = stats.skew(returns)\n",
" kurt = stats.kurtosis(returns)\n",
"\n",
" # ACF of returns and squared returns\n",
" def acf(x, nlags=10):\n",
" \"\"\"Simple ACF computation.\"\"\"\n",
" result = []\n",
" x_centered = x - x.mean()\n",
" var = np.var(x)\n",
" for lag in range(nlags + 1):\n",
" if lag == 0:\n",
" result.append(1.0)\n",
" else:\n",
" result.append(np.corrcoef(x_centered[:-lag], x_centered[lag:])[0, 1])\n",
" return np.array(result)\n",
"\n",
" acf_returns = acf(returns, nlags=5)\n",
" acf_squared = acf(returns**2, nlags=5)\n",
"\n",
" return {\n",
" \"mean\": mean,\n",
" \"std\": std,\n",
" \"skewness\": skew,\n",
" \"kurtosis\": kurt,\n",
" \"acf_returns_lag1\": acf_returns[1] if len(acf_returns) > 1 else np.nan,\n",
" \"acf_squared_lag1\": acf_squared[1] if len(acf_squared) > 1 else np.nan,\n",
" }\n",
"\n",
"\n",
"print(\"=\" * 70)\n",
"print(\"STYLIZED FACTS COMPARISON\")\n",
"print(\"=\" * 70)\n",
"\n",
"real_facts = compute_stylized_facts(holdout_windows)\n",
"syn_facts = compute_stylized_facts(synthetic_holdout)\n",
"\n",
"print(\"\\n{:<25} {:>12} {:>12}\".format(\"Metric\", \"Real\", \"Synthetic\"))\n",
"print(\"-\" * 50)\n",
"for key in real_facts:\n",
" print(f\"{key:<25} {real_facts[key]:>12.4f} {syn_facts[key]:>12.4f}\")\n",
"\n",
"# Cross-correlation comparison\n",
"if n_assets >= 2:\n",
" real_corr = np.corrcoef(holdout_windows[:, :, 0].flatten(), holdout_windows[:, :, 1].flatten())[\n",
" 0, 1\n",
" ]\n",
" syn_corr = np.corrcoef(\n",
" synthetic_holdout[:, :, 0].flatten(), synthetic_holdout[:, :, 1].flatten()\n",
" )[0, 1]\n",
" print(f\"\\nCross-asset correlation: Real={real_corr:.4f}, Synth={syn_corr:.4f}\")"
]
},
{
"cell_type": "markdown",
"id": "8019870a",
"metadata": {
"papermill": {
"duration": 0.012938,
"end_time": "2026-06-14T16:58:55.179393+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.166455+00:00",
"status": "completed"
}
},
"source": [
"**Interpretation**: Key stylized facts to check: (1) kurtosis should be >3\n",
"(fat tails), matching the well-known leptokurtic property of financial returns;\n",
"(2) negative skewness reflects asymmetric crash risk; (3) near-zero lag-1 ACF\n",
"of returns but positive lag-1 ACF of squared returns captures volatility\n",
"clustering. Large deviations in any metric indicate the generator misses\n",
"important market microstructure."
]
},
{
"cell_type": "markdown",
"id": "d67edc09",
"metadata": {
"papermill": {
"duration": 0.012987,
"end_time": "2026-06-14T16:58:55.205212+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.192225+00:00",
"status": "completed"
}
},
"source": [
"## 14. Save Outputs"
]
},
{
"cell_type": "code",
"execution_count": 48,
"id": "364b1baa",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:55.231814Z",
"iopub.status.busy": "2026-06-14T16:58:55.231671Z",
"iopub.status.idle": "2026-06-14T16:58:55.235683Z",
"shell.execute_reply": "2026-06-14T16:58:55.235283Z"
},
"papermill": {
"duration": 0.017954,
"end_time": "2026-06-14T16:58:55.236049+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.218095+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"# Output directory\n",
"output_dir = get_output_dir(5, \"sigcwgan\")\n",
"checkpoint_subdir = \"checkpoints\"\n",
"checkpoint_dir = output_dir / checkpoint_subdir / \"sigwgan_paper\" / f\"depth{CONFIG['sig_depth']}\"\n",
"checkpoint_dir.mkdir(parents=True, exist_ok=True)\n",
"\n",
"# Save model checkpoint\n",
"checkpoint = {\n",
" \"generator\": generator.state_dict(),\n",
" \"expected_sig_real\": expected_sig_real,\n",
" \"train_mean\": train_mean.tolist(), # For reference only (no normalization used)\n",
" \"train_std\": train_std.tolist(), # For reference only (no normalization used)\n",
"}\n",
"torch.save(checkpoint, checkpoint_dir / \"checkpoint.pt\")"
]
},
{
"cell_type": "code",
"execution_count": 49,
"id": "6031f7ae",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:55.262683Z",
"iopub.status.busy": "2026-06-14T16:58:55.262538Z",
"iopub.status.idle": "2026-06-14T16:58:55.265424Z",
"shell.execute_reply": "2026-06-14T16:58:55.265095Z"
},
"papermill": {
"duration": 0.016829,
"end_time": "2026-06-14T16:58:55.265819+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.248990+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"# Save metadata — core fields\n",
"metadata = {\n",
" \"version\": \"2.0\", # Paper-exact implementation\n",
" \"generator\": {\n",
" \"name\": \"sigwgan_paper\",\n",
" \"architecture\": \"LSTMGenerator\",\n",
" \"paper\": \"Ni et al., Sig-Wasserstein GANs for Time Series Generation, Mathematical Finance 2024\",\n",
" },\n",
" \"training\": {\n",
" \"created_at\": datetime.now(UTC).isoformat(),\n",
" \"device\": str(device),\n",
" \"random_seed\": 42,\n",
" \"final_loss\": float(training_losses[-1]),\n",
" \"best_loss\": float(min(training_losses)),\n",
" },\n",
" \"data\": {\n",
" \"source\": \"sp500_index\",\n",
" \"description\": \"S&P 500 daily log returns (paper-exact)\",\n",
" \"start_date\": CONFIG[\"start_date\"],\n",
" \"end_date\": CONFIG.get(\"end_date\", \"present\"),\n",
" \"holdout_start\": CONFIG[\"holdout_start\"],\n",
" \"n_lags\": int(CONFIG[\"n_lags\"]),\n",
" \"n_features\": int(n_assets),\n",
" \"n_total_days\": int(n_total_days),\n",
" \"n_train_windows\": int(len(train_windows)),\n",
" \"n_holdout_windows\": int(len(holdout_windows)),\n",
" \"split_method\": \"temporal_at_date\",\n",
" },\n",
" \"normalization\": {\n",
" \"method\": \"none\",\n",
" \"note\": \"Paper uses raw log returns without normalization\",\n",
" \"train_mean\": train_mean.tolist(),\n",
" \"train_std\": train_std.tolist(),\n",
" },\n",
"}"
]
},
{
"cell_type": "code",
"execution_count": 50,
"id": "056db997",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:55.292528Z",
"iopub.status.busy": "2026-06-14T16:58:55.292379Z",
"iopub.status.idle": "2026-06-14T16:58:55.295968Z",
"shell.execute_reply": "2026-06-14T16:58:55.295585Z"
},
"papermill": {
"duration": 0.017885,
"end_time": "2026-06-14T16:58:55.296232+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.278347+00:00",
"status": "completed"
}
},
"outputs": [],
"source": [
"# Save metadata — hyperparameters, augmentation spec, and evaluation results\n",
"metadata[\"hyperparameters\"] = {\n",
" \"sig_depth\": int(CONFIG[\"sig_depth\"]),\n",
" \"lstm_hidden_dim\": int(CONFIG[\"lstm_hidden_dim\"]),\n",
" \"lstm_n_layers\": int(CONFIG[\"lstm_n_layers\"]),\n",
" \"noise_dim\": int(CONFIG[\"noise_dim\"]),\n",
" \"total_steps\": int(CONFIG[\"total_steps\"]),\n",
" \"batch_size\": int(CONFIG[\"batch_size\"]),\n",
" \"learning_rate\": float(CONFIG[\"learning_rate\"]),\n",
" \"lr_scheduler_step\": int(CONFIG[\"lr_scheduler_step\"]),\n",
" \"lr_scheduler_gamma\": float(CONFIG[\"lr_scheduler_gamma\"]),\n",
" \"normalise_sig\": bool(CONFIG[\"normalise_sig\"]),\n",
"}\n",
"metadata[\"augmentation_pipeline\"] = [\n",
" {\"name\": \"Scale\", \"scale\": CONFIG[\"scale_factor\"], \"dim\": CONFIG[\"scale_dim\"]},\n",
" {\"name\": \"AddTime\"},\n",
" {\"name\": \"LeadLag\"},\n",
" {\"name\": \"VisiTrans\", \"type\": \"I\"},\n",
"]\n",
"metadata[\"signature_spec\"] = {\n",
" \"depth\": int(CONFIG[\"sig_depth\"]),\n",
" \"path_dim\": compute_path_dim_paper(n_assets),\n",
" \"sig_dim\": compute_sig_dim_paper(n_assets, CONFIG[\"sig_depth\"]),\n",
" \"factorial_normalization\": bool(CONFIG[\"normalise_sig\"]),\n",
"}\n",
"metadata[\"evaluation\"] = {\n",
" \"sig_w1\": float(eval_results[\"sig_w1\"]),\n",
" \"ks_statistic\": float(eval_results[\"ks_statistic\"]),\n",
" \"tstr_ratio\": float(tstr_results[\"tstr_ratio\"]),\n",
" \"paper_reference_sig_w1\": 2.76, # For comparison\n",
"}\n",
"\n",
"with open(checkpoint_dir / \"metadata.json\", \"w\") as f:\n",
" json.dump(metadata, f, indent=2)"
]
},
{
"cell_type": "code",
"execution_count": 51,
"id": "b9047ce7",
"metadata": {
"execution": {
"iopub.execute_input": "2026-06-14T16:58:55.323419Z",
"iopub.status.busy": "2026-06-14T16:58:55.323280Z",
"iopub.status.idle": "2026-06-14T16:58:55.327026Z",
"shell.execute_reply": "2026-06-14T16:58:55.326688Z"
},
"papermill": {
"duration": 0.018037,
"end_time": "2026-06-14T16:58:55.327399+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.309362+00:00",
"status": "completed"
}
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n",
"Sample statistics (log returns scale):\n",
" Mean: 0.000365, Std: 0.013708\n",
" Range: [-0.065317, 0.055378]\n",
"\n",
"Saved outputs to:\n",
" 05_synthetic_data/output/sigcwgan/checkpoints/sigwgan_paper/depth4/\n"
]
}
],
"source": [
"# Save pre-generated samples (already in log return scale, no denormalization needed)\n",
"np.save(checkpoint_dir / \"samples.npy\", synthetic_paths.astype(np.float32))\n",
"\n",
"print(\"\\nSample statistics (log returns scale):\")\n",
"print(f\" Mean: {synthetic_paths.mean():.6f}, Std: {synthetic_paths.std():.6f}\")\n",
"print(f\" Range: [{synthetic_paths.min():.6f}, {synthetic_paths.max():.6f}]\")\n",
"\n",
"# Save holdout data for external evaluation\n",
"np.save(checkpoint_dir / \"holdout_returns.npy\", holdout_returns_original)\n",
"np.save(checkpoint_dir / \"train_returns.npy\", returns_original)\n",
"\n",
"print(\"\\nSaved outputs to:\")\n",
"print(f\" {checkpoint_dir}/\")"
]
},
{
"cell_type": "markdown",
"id": "6fff8317",
"metadata": {
"papermill": {
"duration": 0.012615,
"end_time": "2026-06-14T16:58:55.353052+00:00",
"exception": false,
"start_time": "2026-06-14T16:58:55.340437+00:00",
"status": "completed"
}
},
"source": [
"## Key Takeaways\n",
"\n",
"**What Sig-WGAN Offers**:\n",
"\n",
"1. **No Adversarial Training**: The \"discriminator\" is the expected signature,\n",
" eliminating mode collapse and training instability\n",
"\n",
"2. **Principled Distance**: Path signatures provide a mathematically rigorous\n",
" way to compare time series distributions\n",
"\n",
"3. **Factorial Normalization**: Rescales signature levels to comparable magnitudes,\n",
" critical for stable training and meaningful loss values\n",
"\n",
"**PAPER-EXACT Implementation** (Ni et al., Mathematical Finance 2024):\n",
"\n",
"1. **Generator**: LSTMGenerator with Brownian motion noise input\n",
" - NOT AR-FNN with independent Gaussian noise\n",
" - Noise: `z = 0.1 * randn(...).cumsum(1)` (Brownian path)\n",
"\n",
"2. **Augmentations**: Scale(2,0) → AddTime → LeadLag → VisiTrans(\"I\")\n",
" - NOT cumsum_concat → lag_added → lead_lag\n",
" - Critical: VisiTrans adds 2 rows + 1 column\n",
"\n",
"3. **Factorial Normalization**: Level k terms multiplied by k!\n",
" - Without this, higher levels have tiny magnitudes\n",
" - Paper: `sig[count:count+dim**(i+1)] *= factorial(i+1)`\n",
"\n",
"4. **Hyperparameters**: batch=2000, depth=4, lr=1e-3, steps=2500\n",
" - Scheduler: StepLR(128, 0.95)\n",
"\n",
"5. **Loss**: RMSE (L2 norm), NOT squared MSE\n",
" - `loss = sqrt(sum((sig_fake - sig_real)**2))`\n",
"\n",
"**Paper-Exact Data Source**:\n",
"\n",
"This implementation now uses the SAME data as the paper:\n",
"- S&P 500 index daily close prices (paper used Oxford MAN, now defunct)\n",
"- Log returns: `log(close_t) - log(close_{t-1})`\n",
"- Single asset (paper's `.SPX`)\n",
"- Period: 2005-01-01 to 2020-06-01\n",
"\n",
"**Paper reference values** (Ni et al., 2024, on S&P 500 log returns):\n",
"\n",
"| Metric | Paper Reference |\n",
"|--------|-----------------|\n",
"| Sig-W1 | ~2.76 |\n",
"| TSTR | ~1.0 |\n",
"\n",
"The Sig-W1 scale depends on signature depth, augmentation choices, and the\n",
"magnitude of the input series, so direct numeric comparison to the paper is\n",
"only meaningful when those settings match. The values printed above in the\n",
"evaluation cell are the achieved metrics on the run in this notebook.\n",
"\n",
"**Limitations**:\n",
"\n",
"1. **Signature Cost**: O(d^depth) where d = path dimension after augmentations\n",
"2. **Small Universe**: 1-5 assets max at depth 4 (exponential scaling)\n",
"3. **Single Asset**: Paper only demonstrates on single-asset time series"
]
}
],
"metadata": {
"jupytext": {
"cell_metadata_filter": "tags,-all",
"formats": "ipynb,py:percent"
},
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.12.13"
},
"ml4t_provenance": {
"source_py_blob": "29529be57cf90f95cd77509413ab0dac34866c2e",
"executed_at": "2026-06-16T02:46:15.232956+00:00",
"executor": "py312-docker",
"production": true,
"parameters": {},
"notes": "markdown-only header reframe (unconditional Sig-WGAN) over prior py312-docker run; outputs unchanged"
},
"papermill": {
"default_parameters": {},
"duration": 16.30759,
"end_time": "2026-06-14T16:58:58.083056+00:00",
"environment_variables": {},
"exception": null,
"input_path": "05_synthetic_data/03_sigcwgan_signatures.ipynb",
"output_path": "05_synthetic_data/03_sigcwgan_signatures.ipynb",
"parameters": {},
"start_time": "2026-06-14T16:58:41.775466+00:00",
"version": "2.7.0"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Reference in New Issue View Git Blame Copy Permalink