Dockerfile

FROM python:3.11-slim

WORKDIR /app

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

# Shared package
COPY shared/ ./shared/
RUN pip install --no-cache-dir -e ./shared/

# Torch first (heaviest layer — cached unless version changes)
RUN pip install --no-cache-dir \
    --extra-index-url https://download.pytorch.org/whl/cpu \
    "torch>=2.4.0"

# Remaining deps
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Pre-download the model inside the image — HF builds have plenty of time & space
RUN python -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('all-MiniLM-L6-v2')"

# App source
COPY app/ ./app/

ENV PYTHONUNBUFFERED=1 \
    PYTHONPATH=/app \
    TRANSFORMERS_CACHE=/app/.cache/hf \
    HF_HOME=/app/.cache/hf

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