name: diapredict

services:
  airflow-db:
    image: postgres:16.11-bookworm
    environment:
      POSTGRES_USER: airflow
      POSTGRES_PASSWORD: ${AIRFLOW_DB_PASSWORD:?Run python scripts/init_env.py first}
      POSTGRES_DB: airflow
    volumes:
      - airflow_db:/var/lib/postgresql/data
    healthcheck:
      test: [CMD-SHELL, 'pg_isready -U airflow -d airflow']
      interval: 5s
      timeout: 5s
      retries: 20
    restart: unless-stopped

  mlflow:
    build:
      context: .
      dockerfile: docker/mlflow.Dockerfile
    ports:
      - '127.0.0.1:5000:5000'
    volumes:
      - mlflow_data:/mlflow
    healthcheck:
      test: [CMD, python, -c, "import urllib.request; urllib.request.urlopen('http://localhost:5000/health', timeout=5)"]
      interval: 10s
      timeout: 10s
      retries: 12
      start_period: 30s
    restart: unless-stopped

  jupyter:
    build:
      context: .
      dockerfile: docker/jupyter.Dockerfile
    working_dir: /workspace
    environment:
      MLFLOW_TRACKING_URI: http://mlflow:5000
      PYTHONPATH: /workspace
    init: true
    ports:
      - '127.0.0.1:8888:8888'
    volumes:
      - .:/workspace
    command:
      - python
      - -m
      - jupyter
      - lab
      - --ip=0.0.0.0
      - --port=8888
      - --no-browser
      - --allow-root
      - --IdentityProvider.token=diapredict-local
    restart: unless-stopped

  fastapi:
    build:
      context: .
      dockerfile: docker/fastapi.Dockerfile
    user: '${LOCAL_UID:-1000}:${LOCAL_GID:-1000}'
    environment:
      HOME: /tmp
      MLFLOW_TRACKING_URI: http://mlflow:5000
      MLFLOW_MODEL_URI: ${MLFLOW_MODEL_URI:-models:/DiaPredictClassifier@champion}
    ports:
      - '127.0.0.1:8000:8000'
    volumes:
      - ./app:/workspace/app:ro
      - ./src:/workspace/src:ro
    depends_on:
      mlflow:
        condition: service_healthy
    healthcheck:
      test: [CMD, python, -c, "import urllib.request; urllib.request.urlopen('http://localhost:8000/health', timeout=5)"]
      interval: 15s
      timeout: 10s
      retries: 8
    restart: unless-stopped

  streamlit:
    build:
      context: .
      dockerfile: docker/streamlit.Dockerfile
    user: '${LOCAL_UID:-1000}:${LOCAL_GID:-1000}'
    environment:
      HOME: /tmp
      API_URL: http://fastapi:8000
    ports:
      - '127.0.0.1:8501:8501'
    volumes:
      - ./app:/workspace/app:ro
    depends_on:
      fastapi:
        condition: service_healthy
    healthcheck:
      test: [CMD, python, -c, "import urllib.request; urllib.request.urlopen('http://localhost:8501/_stcore/health', timeout=5)"]
      interval: 15s
      timeout: 10s
      retries: 8
    restart: unless-stopped

  airflow:
    build:
      context: .
      dockerfile: docker/airflow.Dockerfile
    user: '${AIRFLOW_UID:-1000}:0'
    command: standalone
    environment:
      AIRFLOW__DATABASE__SQL_ALCHEMY_CONN: postgresql+psycopg2://airflow:${AIRFLOW_DB_PASSWORD}@airflow-db:5432/airflow
      AIRFLOW__CORE__EXECUTOR: LocalExecutor
      AIRFLOW__CORE__LOAD_EXAMPLES: 'false'
      AIRFLOW__CORE__DAGS_FOLDER: /workspace/dags
      AIRFLOW__CORE__PARALLELISM: '2'
      AIRFLOW__CORE__SIMPLE_AUTH_MANAGER_USERS: admin:admin
      AIRFLOW__CORE__SIMPLE_AUTH_MANAGER_PASSWORDS_FILE: /opt/airflow/state/passwords.json
      AIRFLOW__API_AUTH__JWT_SECRET: ${AIRFLOW_DB_PASSWORD}
      AIRFLOW__CORE__EXECUTION_API_SERVER_URL: http://localhost:8080/execution/
      AIRFLOW__LOGGING__BASE_LOG_FOLDER: /opt/airflow/state/logs
      MLFLOW_TRACKING_URI: http://mlflow:5000
      PYTHONPATH: /workspace
      TRAINING_PYTHON: /opt/airflow/training/bin/python
    ports:
      - '127.0.0.1:8080:8080'
    volumes:
      - .:/workspace
      - airflow_state:/opt/airflow/state
    depends_on:
      airflow-db:
        condition: service_healthy
      mlflow:
        condition: service_healthy
    healthcheck:
      test: [CMD, python, -c, "import json, urllib.request; d=json.load(urllib.request.urlopen('http://localhost:8080/api/v2/monitor/health', timeout=5)); assert all(d[k]['status']=='healthy' for k in ('metadatabase','scheduler','dag_processor'))"]
      interval: 20s
      timeout: 10s
      retries: 15
      start_period: 90s
    restart: unless-stopped

volumes:
  airflow_db:
  airflow_state:
  mlflow_data: