RAG Knowledge Assistant

from fastapi import FastAPI
from sentence_transformers import SentenceTransformer
import numpy as np

app = FastAPI()
embedder = SentenceTransformer("all-MiniLM-L6-v2")

@app.post("/query")
async def query_documents(question: str):
    query_embedding = embedder.encode(question)
    context = vector_store.similar_search(query_embedding, k=5)
    response = await llm.generate(
        prompt=f"Answer based on context: {context}",
        question=question
    )
    return {"answer": response, "sources": context}