Dubletten in Paperless NGX ausfindig machen ...

# ... im Code bei "Update document fields" ...
                document.title = doc_data.get("title", "")[:500]
                
                # FIX: Leere Checksummen abfangen
                raw_checksum = doc_data.get("checksum", "")
                if not raw_checksum:
                    # Eindeutigen Platzhalter generieren
                    document.fingerprint = f"MISSING_CHECKSUM_{paperless_id}"
                else:
                    document.fingerprint = raw_checksum[:64]
# ...

services:
  # 1. Datenbank
  postgres:
    image: postgres:17
    container_name: dedupe-standalone-postgres
    restart: unless-stopped
    volumes:
      - ./postgres_data:/var/lib/postgresql/data
    environment:
      - POSTGRES_DB=paperless_dedupe
      - POSTGRES_USER=paperless_dedupe
      - POSTGRES_PASSWORD=dedupe_secret_password
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U paperless_dedupe -d paperless_dedupe"]
      interval: 10s
      timeout: 5s
      retries: 5

  # 2. Redis
  redis:
    image: redis:7-alpine
    container_name: dedupe-standalone-redis
    restart: unless-stopped
    volumes:
      - ./redis_data:/data
    command: redis-server --appendonly yes
    healthcheck:
      test: ["CMD", "redis-cli", "ping"]
      interval: 5s
      timeout: 3s
      retries: 5

  # 3. Das Backend (API)
  paperless-dedupe:
    image: python:3.13-slim-bookworm
    container_name: dedupe-standalone-backend
    depends_on:
      redis:
        condition: service_healthy
      postgres:
        condition: service_healthy
    ports:
      - "30001:8000"
    user: root
    environment:
      - PAPERLESS_DEDUPE_DATABASE_URL=postgresql+psycopg://paperless_dedupe:dedupe_secret_password@postgres:5432/paperless_dedupe
      - PAPERLESS_DEDUPE_REDIS_URL=redis://redis:6379/0
      # Eure Paperless-IP anpassen!
      - PAPERLESS_DEDUPE_PAPERLESS_URL=http://192.168.178.xxx:8008
      - PAPERLESS_DEDUPE_SECRET_KEY=change_me_to_something_secure
      - PYTHONPATH=/app/src
    working_dir: /app
    restart: unless-stopped
    volumes:
      - ./src:/app/src
      - ./alembic:/app/alembic
      - ./README.md:/app/README.md
      - ./pyproject.toml:/app/pyproject.toml
    # Der "Zauber-Befehl": Installiert alles frisch beim Start
    command: >
      /bin/sh -c "
      apt-get update && apt-get install -y build-essential libffi-dev &&
      pip install --upgrade pip &&
      rm -f /app/uv.lock &&
      pip install --force-reinstall --no-cache-dir alembic asyncpg celery cryptography datasketch fastapi greenlet httpx 'psycopg[binary]' pydantic-settings pypaperless python-multipart rapidfuzz redis sqlalchemy uvicorn &&
      uvicorn paperless_dedupe.main:app --host 0.0.0.0 --port 8000
      "
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8000/health"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 300s

  # 4. Der Worker
  worker:
    image: python:3.13-slim-bookworm
    container_name: dedupe-standalone-worker
    depends_on:
      redis:
        condition: service_healthy
      postgres:
        condition: service_healthy
      paperless-dedupe:
        condition: service_started
    user: root
    environment:
      - PAPERLESS_DEDUPE_DATABASE_URL=postgresql+psycopg://paperless_dedupe:dedupe_secret_password@postgres:5432/paperless_dedupe
      - PAPERLESS_DEDUPE_REDIS_URL=redis://redis:6379/0
      - DOCKER_CONTAINER=1
      - PYTHONPATH=/app/src
    working_dir: /app
    volumes:
      - ./src:/app/src
      - ./alembic:/app/alembic
      - ./README.md:/app/README.md
      - ./pyproject.toml:/app/pyproject.toml
    restart: unless-stopped
    # Gleicher Installationsbefehl für den Worker
    command: >
      /bin/sh -c "
      apt-get update && apt-get install -y build-essential libffi-dev &&
      pip install --upgrade pip &&
      rm -f /app/uv.lock &&
      pip install --force-reinstall --no-cache-dir alembic asyncpg celery cryptography datasketch greenlet httpx 'psycopg[binary]' pydantic-settings pypaperless rapidfuzz redis sqlalchemy &&
      sleep 10 &&
      celery -A paperless_dedupe.worker.celery_app worker --loglevel=info --concurrency=4 --queues=high_priority,default,low_priority,deduplication,sync
      "

  # 5. Das Frontend
  frontend:
    image: ghcr.io/rknightion/paperless-ngx-dedupe-frontend:latest
    container_name: dedupe-standalone-frontend
    ports:
      - "30002:3000"
    depends_on:
      - paperless-dedupe
    restart: unless-stopped

volumes:
  postgres_data:
  redis_data: