Qdrant RAG System

A Retrieval-Augmented Generation (RAG) system built with FastAPI, Qdrant, LangChain, and OpenAI. Upload documents, store them as vector embeddings, and query them with natural language using an LLM.

Architecture

Client -> FastAPI -> VectorStoreManager -> Qdrant (vector DB)
                  -> QuestionAnswerService -> OpenAI (LLM)

• Document ingestion: PDF upload with automatic chunking (RecursiveCharacterTextSplitter)
• Embeddings: OpenAI text-embedding-3-large (3072 dimensions)
• Vector store: Qdrant with cosine similarity
• LLM: OpenAI ChatGPT via LangChain LCEL chains
• Sessions: Per-session chat history support

Setup

Prerequisites

• Python 3.10+
• Qdrant instance (local or Docker)
• OpenAI API key

Local Setup

pip install -r requirements.txt
cp .env.example .env
# Edit .env with your values
uvicorn server:app --reload

Docker Setup

cp .env.example .env
# Edit .env with your OPENAI_API_KEY
docker-compose up

This starts both the FastAPI app (port 8000) and Qdrant (port 6333).

Environment Variables

Variable	Description	Default
OPENAI_API_KEY	OpenAI API key	(required)
QDRANT_HOST	Qdrant host	localhost
QDRANT_PORT	Qdrant port	6333
LLM_TEMPERATURE	LLM temperature	0.7
EMBEDDING_MODEL	Embedding model	text-embedding-3-large
API_KEY	API key for authentication	(optional)

API Endpoints

Collections

Method	Endpoint	Description
POST	/vector_db/create_collection?collection_name=X	Create a collection
DELETE	/vector_db/delete_collection?collection_name=X	Delete a collection
GET	/vector_db/list_all_collections	List all collections
GET	/vector_db/get_collection_info?collection_name=X	Get all docs in a collection

Documents

Method	Endpoint	Description
POST	/vector_db/add_doc_to_existing_collection	Upload a PDF to a collection
POST	/vector_db/add_text_to_existing_collection	Add structured text to a collection
POST	/vector_db/query_collection	Semantic search on a collection
DELETE	/vector_db/delete_doc_from_collection_by_name	Delete a document by name

Chat (RAG)

Method	Endpoint	Description
POST	/chat/chat_agent	Ask a question with RAG
GET	/chat/history/{session_id}	Get chat history for a session
DELETE	/chat/history/{session_id}	Clear a session's history

Health

Method	Endpoint	Description
GET	/health	Health check (Qdrant connectivity)

Examples

# Create a collection
curl -X POST "http://localhost:8000/vector_db/create_collection?collection_name=my_docs"

# Upload a PDF
curl -X POST "http://localhost:8000/vector_db/add_doc_to_existing_collection?collection_name=my_docs" \
  -F "file=@document.pdf"

# Query with RAG
curl -X POST "http://localhost:8000/chat/chat_agent" \
  -H "Content-Type: application/json" \
  -d '{"query": "What is this document about?", "collection_name": "my_docs", "k": 3, "session_id": "user-123"}'

# Health check
curl http://localhost:8000/health

References

• Qdrant
• LangChain
• OpenAI

License

MIT