app.py

from fastapi import FastAPI
from pydantic import BaseModel
from typing import List
from sentence_transformers import SentenceTransformer
import uvicorn

app = FastAPI(title="Medical Embedding Service")

# Load model ONCE at startup
print("Loading Medical RAG Model... this may take a moment.")
model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
print("Model loaded successfully!")

class QueryRequest(BaseModel):
    text: str

class DocumentRequest(BaseModel):
    texts: List[str]

@app.post("/embed_query")
async def embed_query(request: QueryRequest):
    # Uses specialized encode_query for IR tasks
    embedding = model.encode_query(request.text).tolist()
    return {"embedding": embedding}

@app.post("/embed_docs")
async def embed_docs(request: DocumentRequest):
    # Uses specialized encode_document for IR tasks
    embeddings = model.encode_document(request.texts).tolist()
    return {"embeddings": embeddings}