Dockerfile

# ── Base image ────────────────────────────────────────────────────────────────
# Use Python 3.10 slim; HF free CPU spaces have ~16 GB RAM
FROM python:3.10-slim

# ── System deps ───────────────────────────────────────────────────────────────
RUN apt-get update && apt-get install -y --no-install-recommends \
        libglib2.0-0 \
        libsm6 \
        libxext6 \
        libxrender-dev \
        libgl1 \
    && rm -rf /var/lib/apt/lists/*

# ── Working dir ───────────────────────────────────────────────────────────────
WORKDIR /app

# ── Python deps (cached layer) ────────────────────────────────────────────────
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# ── Pre-download the model during build (avoids cold-start timeout) ───────────
# HF Spaces has no outbound rate limit during Docker build.
RUN python - <<'EOF'
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
TrOCRProcessor.from_pretrained("microsoft/trocr-large-handwritten")
VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-large-handwritten")
print("Model cached successfully.")
EOF

# ── Copy source ───────────────────────────────────────────────────────────────
COPY app.py .

# ── HF Spaces requires the app to listen on port 7860 ────────────────────────
EXPOSE 7860

# ── Run ───────────────────────────────────────────────────────────────────────
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "1"]