RAG-Chatbot mit Vektorspeicher

# Generated by: fastapi-fullstack new --preset ai-agent

# backend/app/agent/rag.py
from pydantic_ai import Agent
from app.core.vectorstore import get_vector_store
from app.core.config import settings

agent = Agent(
    model=settings.LLM_MODEL,
    system_prompt="Answer questions using the provided context. Cite sources.",
)

@agent.tool
async def search_documents(ctx, query: str, top_k: int = 5) -> str:
    """Search the knowledge base for relevant documents."""
    store = get_vector_store()  # Milvus, Qdrant, ChromaDB, or pgvector
    results = await store.similarity_search(query, k=top_k)
    return "\n\n".join([f"[{r.metadata['source']}]: {r.page_content}" for r in results])