# Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import numpy as np

import paddle
import paddle.distributed as dist
from paddle import nn
from paddle.distributed import Replicate, Shard
from paddle.io import DataLoader

BATCH_SIZE = 4
BATCH_NUM = 40
IMAGE_SIZE = 16
CLASS_NUM = 8
np.random.seed(2024)
paddle.seed(2024)


class RandomDataset(paddle.io.Dataset):
    def __init__(self, images, labels, num_samples):
        self.images = images
        self.labels = labels
        self.num_samples = num_samples

    def __getitem__(self, idx):
        return self.images[idx], self.labels[idx]

    def __len__(self):
        return self.num_samples


class DemoNet(nn.Layer):
    def __init__(self, mesh, shard=True):
        super().__init__()
        self._mesh = mesh
        self.linear_0 = nn.Linear(IMAGE_SIZE, IMAGE_SIZE, bias_attr=False)
        self.linear_1 = nn.Linear(IMAGE_SIZE, CLASS_NUM, bias_attr=False)
        self.relu_0 = nn.ReLU()
        self.relu_1 = nn.ReLU()
        self.relu_2 = nn.ReLU()
        self.shard = shard
        # shard the weights of this layer
        if self.shard:
            self.linear_0.weight = dist.shard_tensor(
                self.linear_0.weight,
                self._mesh,
                [Shard(1)],
                stop_gradient=False,
            )
            self.linear_1.weight = dist.shard_tensor(
                self.linear_1.weight,
                self._mesh,
                [Shard(0)],
                stop_gradient=False,
            )
        else:
            self.linear_0.weight = dist.shard_tensor(
                self.linear_0.weight,
                self._mesh,
                [Replicate()],
                stop_gradient=False,
            )
            self.linear_1.weight = dist.shard_tensor(
                self.linear_1.weight,
                self._mesh,
                [Replicate()],
                stop_gradient=False,
            )

    def forward(self, x):
        x.stop_gradient = False
        out = self.relu_0(x)  # triggle backward partial allreduce
        out = self.linear_0(out)
        out = self.relu_1(out)
        out = self.linear_1(out)
        out = self.relu_2(out)  # triggle forward partial allreduce
        return out


def create_data_loader():
    images = np.random.rand(BATCH_NUM, IMAGE_SIZE).astype('float32')
    labels = np.random.rand(BATCH_NUM, CLASS_NUM).astype('float32')
    dataset = RandomDataset(images, labels, BATCH_NUM)
    loader = DataLoader(dataset, batch_size=BATCH_SIZE)
    return loader