wehub-resource-sync
45 lines
1.1 KiB
YAML
45 lines
1.1 KiB
YAML
# Project and environment settings
|
|
project_id: "<your-project-id>"
|
|
location: "<your-region>"
|
|
|
|
# Data and storage settings
|
|
data_path: "{base_path}/data/data"
|
|
input_bucket_name: "ken-rag-datasets"
|
|
docstore_bucket_name: "ken-rag-datasets"
|
|
bucket_prefix: "raw_pdfs/"
|
|
vector_data_prefix: "vector_data"
|
|
rag_eval_dataset: "ken-rag-datasets"
|
|
firestore_db_name: "rag-docstore"
|
|
firestore_namespace: "hierarchical_docs"
|
|
|
|
# Index settings
|
|
index_id: "llamaindex_rag"
|
|
vector_index_name: "google_hierarchical"
|
|
index_endpoint_name: "hierarchical_endpoint"
|
|
indexing_method: "hierarchical"
|
|
qa_index_name: "google_qa"
|
|
qa_endpoint_name: "hierarchical_endpoint"
|
|
|
|
# Chunking and embedding settings
|
|
chunk_sizes: [4096, 2048, 1024, 512]
|
|
chunk_size: 512
|
|
chunk_overlap: 50
|
|
embeddings_model_name: "text-embedding-005"
|
|
approximate_neighbors_count: 100
|
|
|
|
# Document AI settings
|
|
docai_location: "us"
|
|
docai_processor_id: "f1713ecadbbf91ab"
|
|
document_ai_processor_display_name: "layout-parser"
|
|
create_docai_processor: false
|
|
|
|
# Authentication
|
|
service_account_key: "llamaindex-rag"
|
|
|
|
# API settings
|
|
fastapi_url: "http://localhost:8033"
|
|
|
|
# UI settings
|
|
streamlit_host: "0.0.0.0"
|
|
streamlit_port: 8080
|