docker-compose.yaml

services:
  database:
    image: "openml/test-database:v0.1.20260204"
    container_name: "openml-test-database"
    environment:
      MYSQL_ROOT_PASSWORD: ok
    ports:
      - "3306:3306"
    healthcheck:
      test: ["CMD", "mysqladmin" ,"ping", "-h", "localhost"]
      start_period: 30s
      start_interval: 1s
      timeout: 3s
      interval: 5s
      retries: 10

  database-setup:
    image: mysql
    container_name: "openml-test-database-setup"
    volumes:
      - ./config/database/update.sh:/database-update.sh
    command: /bin/sh -c "/database-update.sh"
    depends_on:
      database:
        condition: service_healthy

  elasticsearch:
    profiles: ["all", "minio", "rest-api", "frontend", "elasticsearch", "evaluation-engine"]
    image: docker.elastic.co/elasticsearch/elasticsearch:6.8.23  # This version is not available on arm. Newer version do not play well with the PHP code. See https://github.com/openml/OpenML/tree/feature/elasticsearch8
    container_name: "openml-elasticsearch"
    platform: "linux/amd64"
    ports:
      - "9200:9200"  # also known as /es (nginx)
      - "9300:9300"
    env_file: config/elasticsearch/.env
    healthcheck:
      test: curl 127.0.0.1:9200/_cluster/health | grep -e "green"
      start_period: 30s
      start_interval: 5s
      timeout: 3s
      interval: 10s
    deploy:
      resources:
        limits:
          cpus: '1'
          memory: 1G
        reservations:
          cpus: '0.2'
          memory: 250M

  nginx:
    profiles: ["all", "minio", "rest-api", "frontend", "elasticsearch", "evaluation-engine"]
    build:
      context: config/nginx
    container_name: openml-nginx
    ports:
      - "8000:8000"
    networks:
      default:
        ipv4_address: 172.28.0.2


  php-api:
    profiles: ["all", "minio", "rest-api", "frontend", "evaluation-engine"]
    image: openml/php-rest-api:v1.2.4
    container_name: "openml-php-rest-api"
    ports:
      - "8080:80"  # also known as /api (nginx)
    env_file: config/php/.env
    depends_on:
      elasticsearch:
        condition: service_healthy
      database-setup:
        condition: service_completed_successfully
    volumes:
      - ${PHP_CODE_DIR:-./config/emptydir}:${PHP_CODE_VAR_WWW_OPENML:-/tmp/unused}
      - ./logs/php:/var/www/openml/logs
      - ./data/php:/var/www/data
    healthcheck:
      test: curl 127.0.0.1:80/api/v1/json/data/1 | grep -e "data_set_description"
      start_period: 2m
      start_interval: 5s
      timeout: 3s
      interval: 1m
    extra_hosts:
      - "localhost=172.28.0.2"

  # This is the new REST API, which will later replace the PHP REST API
  python-api:
    profiles: ["all", "rest-api"]
    image: openml/rest-api:0.0.1-alpha
    container_name: "openml-python-rest-api"
    ports:
      - "8082:8000"  # or localhost:8000/py with nginx
    depends_on:
      database-setup:
        condition: service_completed_successfully
      php-api:
        condition: service_healthy  # used for serving arff files
      minio:
        condition: service_started  # used for serving parquet files
      nginx:
        condition: service_started  # used for routing to MinIO and PHP REST API
    volumes:
      - ./config/python-server/config.toml:/python-api/src/config.toml
    healthcheck:
      test: curl 127.0.0.1:8000/datasets/1 | grep -e "anneal"
      start_period: 1m
      start_interval: 5s
      timeout: 3s
      interval: 1m

  email-server:
    profiles: ["all", "frontend"]
    image: foxcpp/maddy:latest
    container_name: "email-server"
    platform: "linux/amd64"
    ports:
      - "25:25"
      - "143:143"
      - "587:587"
      - "993:993"
    env_file: config/email-server/.env
    volumes:
      - ./config/email-server/:/data

  frontend:
    profiles: ["all", "frontend"]
    image: openml/frontend:v2.1.1
    container_name: "openml-frontend"
    ports:
      - "8081:5000"  # also known as / (nginx)
    env_file: config/frontend/.env
    volumes:
      - ${FRONTEND_CODE_DIR:-./config/emptydir}:${FRONTEND_APP:-/tmp/unused}
    depends_on:
      php-api:
        condition: service_healthy

  minio:
    profiles: ["all", "minio", "evaluation-engine"]
    image: openml/test-minio:v0.1.20260204
    container_name: "openml-minio"
    ports:
      - "9000:9000"  # also known as /data (nginx)
      - "9001:9001"
    env_file: config/minio/.env
    command: >
      server /export --console-address ":9001"

  arff-to-pq:
    profiles: ["all", "minio"]
    build:
      context: config/arff-to-pq-converter
    image: openml-arff-to-pq-converter-cron
    container_name: "openml-arff-to-pq-converter"
    env_file: config/arff-to-pq-converter/.env
    entrypoint: >
      /bin/bash -c "/run-cron.sh"
    volumes:
      - ${ARFF_TO_PQ_CODE_DIR:-./config/emptydir}:${ARFF_TO_PQ_APP:-/tmp/unused}
      - ./logs/arff-to-pq-converter:/home/unprivileged-user/logs
      - ./config/arff-to-pq-converter/config:/home/unprivileged-user/.config/openml/config:ro
    depends_on:
      php-api:
        condition: service_healthy
    extra_hosts:
      - "localhost=172.28.0.2"

  croissants:
    profiles: ["all"]
    image: "openml/croissant-converter:v1.4.2"
    container_name: "openml-croissant-converter"
    env_file: config/croissant-converter/.env
    ports:
      - "8001:8001"
    command: >
      web_api.py --host "0.0.0.0" --port 8001

  evaluation-engine:
    profiles: ["all", "evaluation-engine"]
    build:
      context: config/evaluation-engine
    container_name: openml-evaluation-engine
    entrypoint: >
      sh /run-cron.sh
    env_file: config/evaluation-engine/.env
    volumes:
      - ./logs/evaluation-engine:/logs
    depends_on:
      php-api:
        condition: service_healthy
    extra_hosts:
      - "localhost=172.28.0.2"

networks:
  default:
    name: openml-services
    ipam:
      driver: default
      config:
        - subnet: 172.28.0.0/16
          ip_range: 172.28.1.0/24