docker-compose.yml

# docker-compose.yml — P2.6: local development with PostgreSQL
# Usage: docker compose up -d
# The backend will connect to Postgres instead of SQLite when DATABASE_URL is set.

services:
  db:
    image: postgres:16-alpine
    restart: unless-stopped
    environment:
      POSTGRES_DB: ${POSTGRES_DB:-raglab}
      POSTGRES_USER: ${POSTGRES_USER:-raglab}
      POSTGRES_PASSWORD: "${POSTGRES_PASSWORD:?set POSTGRES_PASSWORD in your environment or .env file}"
    volumes:
      - pgdata:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U raglab -d raglab"]
      interval: 10s
      timeout: 5s
      retries: 5

  chroma:
    image: chromadb/chroma:latest
    restart: unless-stopped
    volumes:
      - chromadata:/chroma/chroma
    environment:
      IS_PERSISTENT: "TRUE"
      ANONYMIZED_TELEMETRY: "FALSE"
    ports:
      - "8001:8000"

  backend:
    build:
      context: .
      dockerfile: Dockerfile
    restart: unless-stopped
    depends_on:
      db:
        condition: service_healthy
    ports:
      - "7860:7860"
    environment:
      # Switch to Postgres for production
      DATABASE_URL: "postgresql+asyncpg://${POSTGRES_USER:-raglab}:${POSTGRES_PASSWORD:?set POSTGRES_PASSWORD in your environment or .env file}@db:5432/${POSTGRES_DB:-raglab}"
      CHROMA_PERSIST_DIR: /data/chroma_db
      UPLOAD_DIR: /data/uploads
      PDF_IMAGE_INDEXING_ENABLED: ${PDF_IMAGE_INDEXING_ENABLED:-false}
      HF_IMAGE_EMBEDDING_MODEL: ${HF_IMAGE_EMBEDDING_MODEL:-openai/clip-vit-base-patch32}
      # Embedding / LLM keys — override in .env or CI secrets
      NVIDIA_API_KEY: ${NVIDIA_API_KEY:-}
      HUGGINGFACE_API_KEY: ${HUGGINGFACE_API_KEY:-}
      GEMINI_API_KEY: ${GEMINI_API_KEY:-}
      FRONTEND_URL: ${FRONTEND_URL:-http://localhost:5173}
      # Tuning knobs
      MAX_UPLOAD_SIZE_MB: ${MAX_UPLOAD_SIZE_MB:-50}
      MAX_EMBED_BATCH_SIZE: ${MAX_EMBED_BATCH_SIZE:-96}
      ODL_TIMEOUT_SECONDS: ${ODL_TIMEOUT_SECONDS:-120}
      PIPELINE_CACHE_TTL_SECONDS: ${PIPELINE_CACHE_TTL_SECONDS:-3600}
      WEB_CONCURRENCY: ${WEB_CONCURRENCY:-4}
      LOG_LEVEL: ${LOG_LEVEL:-info}
      LOG_FORMAT: json
    volumes:
      - chromadata:/data/chroma_db
      - uploads:/data/uploads

volumes:
  pgdata:
  chromadata:
  uploads: