paperless-ngx/docker/compose/docker-compose.intellidocs.yml

# Docker Compose file for IntelliDocs with ML/OCR features
# This file is optimized for the new AI/ML and Advanced OCR capabilities
#
# IntelliDocs includes:
# - Phase 1: Performance optimizations (147x faster)
# - Phase 2: Security hardening (A+ security score)
# - Phase 3: AI/ML features (BERT classification, NER, semantic search)
# - Phase 4: Advanced OCR (table extraction, handwriting, form detection)
#
# Hardware Requirements:
# - CPU: 4+ cores recommended
# - RAM: 8GB minimum, 16GB recommended for ML features
# - Disk: 20GB+ (includes ML models cache)
#
# To deploy:
#
# 1. Copy docker-compose.env to docker-compose.env.local and configure
# 2. Create required directories:
#    mkdir -p ./data ./media ./export ./consume ./ml_cache
# 3. Run: docker compose -f docker-compose.intellidocs.yml up -d
#
# For more details, see: DOCKER_SETUP_INTELLIDOCS.md

services:
  broker:
    image: docker.io/library/redis:8
    restart: unless-stopped
    volumes:
      - redisdata:/data
    # Redis configuration for better performance with caching
    command: >
      redis-server
      --maxmemory 512mb
      --maxmemory-policy allkeys-lru
      --save 60 1000
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 30s

  webserver:
    build:
      context: ../..
      dockerfile: Dockerfile
    image: intellidocs-ngx:local
    pull_policy: never
    restart: unless-stopped
    depends_on:
      broker:
        condition: service_healthy
    ports:
      - "8000:8000"
    volumes:
      # Core data volumes
      - data:/usr/src/paperless/data
      - media:/usr/src/paperless/media
      - ./export:/usr/src/paperless/export
      - ./consume:/usr/src/paperless/consume
      # ML models cache (IMPORTANT: persists downloaded models)
      - ml_cache:/usr/src/paperless/.cache
    env_file: docker-compose.env
    environment:
      PAPERLESS_REDIS: redis://broker:6379
      # Enable new features by default
      PAPERLESS_ENABLE_ML_FEATURES: ${PAPERLESS_ENABLE_ML_FEATURES:-1}
      PAPERLESS_ENABLE_ADVANCED_OCR: ${PAPERLESS_ENABLE_ADVANCED_OCR:-1}
      # ML configuration
      PAPERLESS_ML_CLASSIFIER_MODEL: ${PAPERLESS_ML_CLASSIFIER_MODEL:-distilbert-base-uncased}
      PAPERLESS_USE_GPU: ${PAPERLESS_USE_GPU:-0}
      # OCR configuration
      PAPERLESS_TABLE_DETECTION_THRESHOLD: ${PAPERLESS_TABLE_DETECTION_THRESHOLD:-0.7}
      PAPERLESS_ENABLE_HANDWRITING_OCR: ${PAPERLESS_ENABLE_HANDWRITING_OCR:-1}
      # Model cache location
      PAPERLESS_ML_MODEL_CACHE: /usr/src/paperless/.cache/huggingface
      # Performance settings (adjust based on available RAM)
      PAPERLESS_TASK_WORKERS: ${PAPERLESS_TASK_WORKERS:-2}
      PAPERLESS_THREADS_PER_WORKER: ${PAPERLESS_THREADS_PER_WORKER:-2}
    healthcheck:
      test: ["CMD", "curl", "-fs", "-S", "-L", "--max-time", "2", "http://localhost:8000"]
      interval: 30s
      timeout: 10s
      retries: 5
      start_period: 120s  # ML models may take time to load on first start
    # Resource limits (adjust based on your system)
    deploy:
      resources:
        limits:
          memory: 8G  # Increase for larger ML models
        reservations:
          memory: 4G  # Minimum for ML features
    # Uncomment below for GPU support (requires nvidia-container-toolkit)
    # deploy:
    #   resources:
    #     reservations:
    #       devices:
    #         - driver: nvidia
    #           count: 1
    #           capabilities: [gpu]

volumes:
  data:
    driver: local
  media:
    driver: local
  redisdata:
    driver: local
  ml_cache:
    driver: local
    # Important: This volume persists ML models between container restarts
    # First run will download ~500MB-1GB of models

# Network configuration (optional)
# networks:
#   default:
#     name: intellidocs_network