main.py

import logging
import os
import shutil
import tempfile
import base64
from typing import List

os.environ.setdefault("HF_HUB_CACHE", "/tmp/hf_cache")
os.environ.setdefault("HF_HOME", "/tmp/hf_home")
os.environ.setdefault("HF_HUB_ENABLE_XET", "0")  # <-- disable xet backend
os.environ.setdefault("MPLCONFIGDIR", "/tmp/mplconfig")

os.makedirs("/tmp/hf_cache", exist_ok=True)
os.makedirs("/tmp/hf_home", exist_ok=True)
os.makedirs("/tmp/mplconfig", exist_ok=True)

from fastapi import FastAPI, UploadFile, File, Form, HTTPException
from fastapi.responses import HTMLResponse, JSONResponse
from fastapi.staticfiles import StaticFiles

from fastapi.middleware.cors import CORSMiddleware
import mmcv
from model.run_inference import infer_images  

logging.basicConfig(level=logging.INFO)

ALLOWED_MODELS = {
    "regnetx4.0gf+detr3d",
    "regnetx4.0gf+petr",
}

app = FastAPI()
app.add_middleware(
    CORSMiddleware,
    allow_origins=["https://yaghi27-imagetobev-lightweight.hf.space"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)
app.mount("/static", StaticFiles(directory="static"), name="static")


@app.get("/", response_class=HTMLResponse)
async def root():
    with open("static/index.html", "r", encoding="utf-8") as f:
        return HTMLResponse(f.read())


@app.post("/infer")
async def run_inference(
    model: str = Form(...),
    images: List[UploadFile] = File(...),
):
    model = model.strip().lower()
    if model not in ALLOWED_MODELS:
        raise HTTPException(status_code=400, detail=f"Invalid model '{model}'. Allowed: {sorted(ALLOWED_MODELS)}")

    if len(images) != 6:
        raise HTTPException(status_code=400, detail=f"Expected 6 images, received {len(images)}")

    tmpdir = tempfile.mkdtemp(prefix="bev_infer_")
    img_paths = []

    try:
        for idx, upload in enumerate(images):
            data = await upload.read()
            bgr = mmcv.imfrombytes(data, flag="color")
            if bgr is None:
                raise HTTPException(status_code=400, detail=f"File '{upload.filename}' is not a valid image.")

            out_path = os.path.join(tmpdir, f"cam_{idx}.png")
            mmcv.imwrite(bgr, out_path)
            img_paths.append(out_path)

        logging.info("Starting inference with model=%s on %d images", model, len(img_paths))

        bev_paths = infer_images(img_paths, model=model)

        output = []
        for p in bev_paths:
            with open(p, "rb") as f:
                b64 = base64.b64encode(f.read()).decode("utf-8")
            output.append({"bev_image": b64})

        return JSONResponse(content=output)

    except HTTPException:

        raise
    except Exception as e:
        logging.exception("inference failed")
        return JSONResponse(status_code=500, content={"error": str(e)})
    finally:
        try:
            shutil.rmtree(tmpdir)
        except Exception:
            logging.warning("Failed to clean tmpdir %s", tmpdir)