Dockerfile.hf

FROM python:3.11-slim

WORKDIR /app

# Install system deps
RUN apt-get update && apt-get install -y \
    libpq-dev \
    gcc \
    libmagic1 \
    curl \
    && rm -rf /var/lib/apt/lists/*

# Install PyTorch CPU first (cached layer)
RUN pip install --no-cache-dir torch --index-url https://download.pytorch.org/whl/cpu

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

# Pre-download the reranker model during build (no cold start!)
RUN python -c "from sentence_transformers import CrossEncoder; CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2')"

# Environment
ENV TOKENIZERS_PARALLELISM=false
ENV TRANSFORMERS_CACHE=/app/.cache/huggingface
ENV SENTENCE_TRANSFORMERS_HOME=/app/.cache/sentence-transformers

COPY . .

EXPOSE 8000

CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000"]