Dockerfile

FROM python:3.12-slim

ENV DEBIAN_FRONTEND=noninteractive PIP_NO_CACHE_DIR=1
WORKDIR /app
ENV PYTHONPATH=/app/backend:$PYTHONPATH

COPY . .

# Install GPU-ready deps
RUN pip install -r requirements.txt

# Preload spaCy + NLTK data so runtime doesn't download
RUN python -m spacy download en_core_web_sm
RUN python - <<'PY'
import nltk
nltk.download('punkt')
nltk.download('wordnet')
nltk.download('omw-1.4')
PY

# HF Spaces uses port 7860
EXPOSE 7860

# Single worker + a few threads = nicer on GPU VRAM
CMD ["gunicorn","-w","1","-k","gthread","--threads","4","-b","0.0.0.0:7860","backend.app:app"]