Dockerfile

FROM nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu22.04

# Install Python and system dependencies
RUN apt-get update && apt-get install -y \
    python3.11 \
    python3-pip \
    python3.11-venv \
    nodejs \
    npm \
    git \
    build-essential \
    curl \
    && rm -rf /var/lib/apt/lists/*

# Install pnpm globally
RUN npm install -g pnpm

# Create user for HF Spaces
RUN useradd -m -u 1000 user
USER user
ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH

WORKDIR $HOME/app

# Copy files
COPY --chown=user embedding_api.py .
COPY --chown=user requirements_embedding.txt .
COPY --chown=user hono-proxy ./hono-proxy
COPY --chown=user vespa-certs ./vespa-certs

# Install Python dependencies with GPU support
RUN pip install --user --no-cache-dir -r requirements_embedding.txt

# Install Node dependencies
WORKDIR $HOME/app/hono-proxy
RUN pnpm install

# Setup Vespa certificates
RUN mkdir -p $HOME/.vespa/il-infra.colpali-server.default && \
    cp ../vespa-certs/* $HOME/.vespa/il-infra.colpali-server.default/

# Create startup script
WORKDIR $HOME/app
RUN echo '#!/bin/bash' > start.sh && \
    echo 'echo "Starting services with GPU support..."' >> start.sh && \
    echo 'echo "Python version: $(python3 --version)"' >> start.sh && \
    echo 'echo "Node version: $(node --version)"' >> start.sh && \
    echo 'echo "CUDA available: $(python3 -c "import torch; print(torch.cuda.is_available())")"' >> start.sh && \
    echo 'echo "GPU device: $(python3 -c "import torch; print(torch.cuda.get_device_name(0) if torch.cuda.is_available() else \"No GPU\")")"' >> start.sh && \
    echo '' >> start.sh && \
    echo '# Start embedding API with GPU' >> start.sh && \
    echo 'CUDA_VISIBLE_DEVICES=0 python3 embedding_api.py &' >> start.sh && \
    echo 'EMBED_PID=$!' >> start.sh && \
    echo 'echo "Started embedding API with PID: $EMBED_PID"' >> start.sh && \
    echo '' >> start.sh && \
    echo '# Wait for embedding API' >> start.sh && \
    echo 'sleep 15' >> start.sh && \
    echo '' >> start.sh && \
    echo '# Check if embedding API is running' >> start.sh && \
    echo 'if curl -f http://localhost:8001/health; then' >> start.sh && \
    echo '    echo "Embedding API is healthy"' >> start.sh && \
    echo 'else' >> start.sh && \
    echo '    echo "Embedding API health check failed"' >> start.sh && \
    echo 'fi' >> start.sh && \
    echo '' >> start.sh && \
    echo '# Start Hono proxy' >> start.sh && \
    echo 'cd hono-proxy' >> start.sh && \
    echo 'PORT=7860 CORS_ORIGIN="*" EMBEDDING_API_URL="http://localhost:8001" VESPA_ENDPOINT="https://f5acf536.ed2ceb09.z.vespa-app.cloud" npx tsx src/index.ts' >> start.sh && \
    chmod +x start.sh

EXPOSE 7860

CMD ["./start.sh"]