Dockerfile

FROM python:3.12-slim

# System deps for audio processing (librosa, soundfile)
RUN apt-get update && apt-get install -y --no-install-recommends \
    libsndfile1 \
    ffmpeg \
    && rm -rf /var/lib/apt/lists/*

WORKDIR /app

# Install torch + torchaudio CPU-only (saves ~2.3GB vs full CUDA build)
RUN pip install --no-cache-dir \
    torch torchaudio --index-url https://download.pytorch.org/whl/cpu

# Force cache invalidation
ARG CACHEBUST=6
# Install remaining dependencies
COPY requirements-deploy.txt .
RUN pip install --no-cache-dir -r requirements-deploy.txt

# Copy source code
COPY src/ src/
COPY config.yaml .
COPY scripts/cache_models.py scripts/cache_models.py

# Pre-download ML models at build time (avoids cold-start downloads)
ARG HF_TOKEN
ENV HF_TOKEN=${HF_TOKEN}
RUN python scripts/cache_models.py

# Create data directory for SQLite + uploads
RUN mkdir -p data/samples

# HF Spaces port
ENV PORT=7860

EXPOSE 7860

CMD ["uvicorn", "src.stage4.main:app", "--host", "0.0.0.0", "--port", "7860"]