Dockerfile

FROM python:3.10-slim

RUN apt-get update && apt-get install -y --no-install-recommends \
    libgomp1 \
    && rm -rf /var/lib/apt/lists/*

RUN useradd -m -u 1000 user
RUN mkdir -p /home/user/app/model_cache && chown -R user:user /home/user
USER user

ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH \
    PYTHONDONTWRITEBYTECODE=1 \
    PYTHONUNBUFFERED=1 \
    TRANSFORMERS_CACHE=/home/user/app/model_cache \
    HF_HOME=/home/user/app/model_cache \
    ACCELERATE_USE_CPU=true \
    TORCH_CPU_ONLY=1

WORKDIR $HOME/app
COPY --chown=user . $HOME/app

RUN pip install --no-cache-dir torch==2.4.1+cpu \
    --index-url https://download.pytorch.org/whl/cpu

# openai SDK used as a thin client pointed at Groq's own API —
# completely separate from HuggingFace's Inference Providers billing.
RUN pip install --no-cache-dir \
    "numpy==1.26.4" \
    "sentence-transformers==3.0.1" \
    "sentencepiece" \
    "openai>=1.0.0" \
    "fastapi" \
    "uvicorn" \
    "faiss-cpu" \
    "pydantic>=2.0"

CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]