104 lines
3 KiB
Docker
104 lines
3 KiB
Docker
# Dockerfile for GLAM Heritage RAG API
|
|
#
|
|
# Multi-source retrieval-augmented generation system for heritage data
|
|
# Connects to: Qdrant, Oxigraph SPARQL, TypeDB, PostGIS, Valkey cache
|
|
#
|
|
# Build:
|
|
# podman build -t glam-rag-api .
|
|
#
|
|
# Run:
|
|
# podman run -d --name glam-rag-api -p 8010:8010 \
|
|
# --network host \
|
|
# -e OPENAI_API_KEY=$OPENAI_API_KEY \
|
|
# glam-rag-api
|
|
|
|
FROM python:3.11-slim
|
|
|
|
# Build args for cache busting
|
|
ARG BUILD_DATE
|
|
ARG VCS_REF
|
|
|
|
# Labels for OCI compliance
|
|
LABEL org.opencontainers.image.created="${BUILD_DATE}" \
|
|
org.opencontainers.image.revision="${VCS_REF}" \
|
|
org.opencontainers.image.title="GLAM Heritage RAG API" \
|
|
org.opencontainers.image.description="Heritage Knowledge Assistant with DSPy RAG" \
|
|
org.opencontainers.image.source="https://git.bronhouder.nl/kempersc/glam"
|
|
|
|
WORKDIR /app
|
|
|
|
# Install system dependencies
|
|
# - curl: healthcheck
|
|
# - git: some pip packages need git
|
|
# - build-essential: compile native extensions
|
|
RUN apt-get update && apt-get install -y --no-install-recommends \
|
|
curl \
|
|
git \
|
|
build-essential \
|
|
&& rm -rf /var/lib/apt/lists/*
|
|
|
|
# Create non-root user for security
|
|
RUN useradd -m -u 1000 -s /bin/bash glam
|
|
|
|
# Install Python dependencies first (better layer caching)
|
|
COPY requirements.txt .
|
|
# Install CPU-only PyTorch first to avoid massive CUDA download and runtime issues
|
|
RUN pip install --no-cache-dir torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
|
|
RUN pip install --no-cache-dir -r requirements.txt
|
|
|
|
# Copy application code
|
|
# Structure:
|
|
# /app/
|
|
# main.py
|
|
# provenance.py
|
|
# dspy_heritage_rag.py
|
|
# ... (all .py files)
|
|
# specificity/
|
|
# evaluation/
|
|
# optimized_models/
|
|
# data/
|
|
COPY --chown=glam:glam . .
|
|
|
|
# Create directories for runtime data
|
|
RUN mkdir -p /app/data /app/optimized_models /app/benchmark_results \
|
|
&& chown -R glam:glam /app
|
|
|
|
# Switch to non-root user
|
|
USER glam
|
|
|
|
# Environment variables (can be overridden at runtime)
|
|
ENV PYTHONUNBUFFERED=1 \
|
|
PYTHONDONTWRITEBYTECODE=1 \
|
|
# Service endpoints (default to localhost, override in compose)
|
|
QDRANT_HOST=localhost \
|
|
QDRANT_PORT=6333 \
|
|
QDRANT_COLLECTION=heritage_custodians_minilm \
|
|
EMBEDDING_MODEL=all-MiniLM-L6-v2 \
|
|
EMBEDDING_DIM=384 \
|
|
TYPEDB_HOST=localhost \
|
|
TYPEDB_PORT=1729 \
|
|
TYPEDB_DATABASE=glam \
|
|
SPARQL_ENDPOINT=http://localhost:7878/query \
|
|
VALKEY_CACHE_URL=http://localhost:8090 \
|
|
POSTGIS_HOST=localhost \
|
|
POSTGIS_PORT=5432 \
|
|
POSTGIS_DATABASE=glam
|
|
|
|
# Expose RAG API port
|
|
EXPOSE 8010
|
|
|
|
# Health check
|
|
HEALTHCHECK --interval=30s --timeout=10s --start-period=60s --retries=3 \
|
|
CMD curl -f http://localhost:8010/health || exit 1
|
|
|
|
# Run with gunicorn for production (multiple workers)
|
|
# Using uvicorn workers for async support
|
|
CMD ["gunicorn", "main:app", \
|
|
"--bind", "0.0.0.0:8010", \
|
|
"--workers", "2", \
|
|
"--worker-class", "uvicorn.workers.UvicornWorker", \
|
|
"--timeout", "120", \
|
|
"--graceful-timeout", "30", \
|
|
"--keep-alive", "5", \
|
|
"--access-logfile", "-", \
|
|
"--error-logfile", "-"]
|