scripts/vision_api.py

import os
from fastapi import FastAPI, File, UploadFile, Form
from scripts.blip2_infer import describe_image
import shutil

app = FastAPI()

@app.post("/vision")
async def process_vision(image: UploadFile = File(...), prompt: str = Form("Describe this image.")):
    temp_path = "frame.jpg"
    with open(temp_path, "wb") as f:
        shutil.copyfileobj(image.file, f)
    desc = describe_image(temp_path, prompt)
    os.remove(temp_path)
    return {"description": desc}