RAG API¶

async rag_api.process_query(request: ChatRequest, db: Session = fastapi.Depends)¶

Main endpoint for the RAG service, processes a RAG query.

Parameters:

request (ChatRequest) – The request object containing: query, language, tags, sources, llm_model, retrieval_method, k_memory, response_style, autocomplete, rag, user_uuid, conversation_uuid parameters.

Returns:

The response from the RAG service

Return type:

StreamingResponse

async rag_api.docs(query: str, language: str | None = 'de', tags: List[str] | None = None, source: List[str] | None = None, organizations: List[str] | None = ['ZAS', 'EAK'], user_uuid: str | None = 'test_uuid', conversation_uuid: str | None = 'test_conversation_uuid', llm_model: str | None = 'gpt-4o', retrieval_method: List[str] | None = ['top_k_retriever'], k: int = 10, db: Session = fastapi.Depends)¶

Retrieve context documents for a given query.

Parameters:

  • request (ChatRequest) – The request object containing the query and context.

  • language (str) – The language of the query.

  • tags (List[str]) – The tags to filter the documents.

  • k (int) – The number of documents to retrieve.

  • db (Session) – The database session.

Returns:

The retrieved documents.

Return type:

dict

async rag_api.rerank()¶

Dummy endpoint for retrieved docs reranking.