Dockerfile

# ---------- Base Stage ----------
FROM pytorch/pytorch:2.7.1-cuda12.8-cudnn9-devel AS base

WORKDIR /app

# System dependencies:
# Added 'poppler-utils' for PDF processing 
# Added 'libpcsclite1' which some OCR tools occasionally peek at
RUN apt-get update && apt-get install -y \
    libgl1-mesa-glx \
    libglib2.0-0 \
    poppler-utils \
    && rm -rf /var/lib/apt/lists/*

# Create local folders to ensure they exist for the volume mounts
RUN mkdir -p /app/data /app/output

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

# ---------- Dev / Test Stage ----------
FROM base AS dev
ENV PYTHONPATH=/app
ENV PYTHONUNBUFFERED=1

# Install dev dependencies (pytest, etc.)
COPY requirements-dev.txt .
RUN pip install --no-cache-dir -r requirements-dev.txt

COPY . .

# Pre-download the Blackwell-friendly model
RUN python3 -c "from doctr.models import ocr_predictor; ocr_predictor(reco_arch='vitstr_base', pretrained=True)"

# Pre-download the embedding model
RUN python3 -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('sentence-transformers/all-mpnet-base-v2')"

CMD ["python", "src/main.py"]