mirror of
https://github.com/hanzoai/chat-rag-api.git
synced 2026-10-30 18:45:41 +00:00
zeekay
74110d03f8
Extracted from the deployed chat-rag-api:0.7.8-hanzo image; adds a Qdrant vector store targeting hanzoai/vector as a drop-in for pgvector. - app/services/vector_store/qdrant_vector.py: QdrantVector(QdrantVectorStore) with the id helpers (get_all_ids/get_filtered_ids/get_documents_by_ids), delete-by-file_id, and the route's Mongo-style file_id filter translated to native Qdrant filters. file_id stored in payload metadata (unique point ids). - factory.get_vector_store: mode='qdrant' (lazy import; pg/atlas deploys unaffected). - config: VectorDBType.QDRANT + QDRANT_URL/QDRANT_API_KEY/QDRANT_DISTANCE/ EMBEDDINGS_DIM(1536) + construction branch. - health: qdrant_health_check (client.get_collections()). - requirements: langchain-qdrant==0.2.0, qdrant-client==1.12.1. Verified: pytest tests/test_qdrant_vector.py (4 tests) against in-memory Qdrant; factory builds QdrantVector with all route methods present.
38 lines
730 B
Plaintext
38 lines
730 B
Plaintext
langchain==0.3.12
|
|
langchain-community==0.3.12
|
|
langchain-openai==0.2.11
|
|
langchain-core==0.3.27
|
|
langchain-google-vertexai==2.0.11
|
|
sqlalchemy==2.0.28
|
|
python-dotenv==1.0.1
|
|
fastapi==0.115.12
|
|
psycopg2-binary==2.9.9
|
|
pgvector==0.2.5
|
|
uvicorn==0.28.0
|
|
pypdf==4.1.0
|
|
unstructured==0.16.11
|
|
markdown==3.6
|
|
networkx==3.2.1
|
|
pandas==2.2.1
|
|
openpyxl==3.1.2
|
|
docx2txt==0.8
|
|
pypandoc==1.13
|
|
PyJWT==2.8.0
|
|
asyncpg==0.29.0
|
|
python-multipart==0.0.19
|
|
aiofiles==23.2.1
|
|
rapidocr-onnxruntime==1.3.24
|
|
opencv-python-headless==4.9.0.80
|
|
pymongo==4.6.3
|
|
langchain-mongodb==0.2.0
|
|
cryptography==44.0.1
|
|
python-magic==0.4.27
|
|
python-pptx==0.6.23
|
|
xlrd==2.0.1
|
|
langchain-aws==0.2.1
|
|
boto3==1.34.144
|
|
|
|
# Qdrant (hanzoai/vector) backend
|
|
langchain-qdrant==0.2.0
|
|
qdrant-client==1.12.1
|