Dockerfile.nvidia

# Pale Fire - Dockerfile.nvidia
# Custom Dockerfile for NVIDIA Quadro K2200 (sm_50) support
# Uses PyTorch 2.1.2 with CUDA 11.8 support which retains sm_50 compat

# Stage 1: Builder
FROM python:3.11-slim as builder

# Set working directory
WORKDIR /app

# Install build dependencies
RUN apt-get update && apt-get install -y \
    gcc \
    g++ \
    && rm -rf /var/lib/apt/lists/*

# Install specific PyTorch version FIRST to ensure we get the right one
# We use --index-url to point to PyTorch's wheel repo for CUDA 11.8
RUN pip install torch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 --index-url https://download.pytorch.org/whl/cu118 --no-cache-dir --user

# Copy requirements
COPY requirements.txt .

# Install other Python dependencies
# We use --no-deps for packages that might try to pull a newer torch, or just let pip handle it but prefer the cached wheel
RUN pip install --no-cache-dir --user -r requirements.txt

# Download spaCy model
RUN python -m spacy download en_core_web_sm

# Stage 2: Runtime
FROM python:3.11-slim

# Set working directory
WORKDIR /app

# Install runtime dependencies
RUN apt-get update && apt-get install -y \
    curl \
    && rm -rf /var/lib/apt/lists/*

# Copy Python packages from builder
COPY --from=builder /root/.local /root/.local

# Copy spaCy model from builder
COPY --from=builder /usr/local/lib/python3.11/site-packages/en_core_web_sm* /usr/local/lib/python3.11/site-packages/

# Make sure scripts in .local are usable
ENV PATH=/root/.local/bin:$PATH

# Copy application files
COPY . .

# Create necessary directories
RUN mkdir -p logs data

# Set env vars
ENV PYTHONUNBUFFERED=1
ENV PYTHONDONTWRITEBYTECODE=1

# Expose ports
EXPOSE 8000

# Health check
HEALTHCHECK --interval=30s --timeout=10s --start-period=40s --retries=3 \
    CMD curl -f http://localhost:8000/health || exit 1

# Default command
CMD ["python", "api.py"]