docker-compose-mutagen.yml

# supplementary docker-compose file
version: "3.7"
services:
  #blazorboilerplate:
  #  build:
  #    context: frontend
  #    # the below path is relative to the build context
  #    dockerfile: ./src/Utils/Docker/Dockerfile
  #  ports:
  #    - 54415:80
  #    - 54443:443
  #  depends_on:
  #    - sqlserver
  #    - controller
  #    - redis
  #  environment:
  #    - ASPNETCORE_ENVIRONMENT=Development #Consider changing this in Production
  #    - Serilog__MinimumLevel__Default=Debug #Consider changing this in Production
  #    - ConnectionStrings__DefaultConnection=Server=sqlserver;Database=blazor_boilerplate;Trusted_Connection=True;MultipleActiveResultSets=true;User=sa;Password=yourVeryStrong(!)Password;Integrated Security=false;Encrypt=False;
  #    - BlazorBoilerplate__UseSqlServer=true
  #    - BlazorBoilerplate__ApplicationUrl=blazorboilerplate
  #    - BlazorBoilerplate__IS4ApplicationUrl=blazorboilerplate
  #    - BlazorBoilerplate__CertificatePassword=Admin123
  #    - ASPNETCORE_URLS=https://+:443;http://+80
  #    - ASPNETCORE_Kestrel__Certificates__Default__Password=Admin123
  #    - ASPNETCORE_Kestrel__Certificates__Default__Path=aspnetapp.pfx
  #    - CONTROLLER_SERVICE_HOST=controller
  #    - CONTROLLER_SERVICE_PORT=5001
  #    - REDIS_SERVICE_HOST=redis
  #    - REDIS_SERVICE_PORT=6379
  #    - CONTROLLER_DATASET_FOLDER_PATH=/app/app-data/datasets
  #    - REGISTRATION_ALLOWED=false
  #  restart: on-failure
  #  volumes:
      # shared volumes between frontend and controller
  #    - datasets:/app/app-data/datasets
  #    - training:/app/app-data/training
  #    - training-autokeras:/app/app-data/training/autokeras
  #    - training-mljar:/app/app-data/training/mljar
  #    - training-sklearn:/app/app-data/training/sklearn
  #    - training-flaml:/app/app-data/training/flaml
  #    - training-gluon:/app/app-data/training/gluon
      #- training-autocve:/app/app-data/training/autocve
  #    - training-pytorch:/app/app-data/training/pytorch
      #- training-alphad3m:/app/app-data/training/alphad3m
      #- training-mcfly:/app/app-data/training/mcfly
  #    - training-evalml:/app/app-data/training/evalml
  #    - training-pycaret:/app/app-data/training/pycaret
  #    - training-tpot:/app/app-data/training/tpot
  #  extra_hosts:
  #    - "docker.host.internal:host-gateway"

  sqlserver:
    image: mcr.microsoft.com/azure-sql-edge:latest
    volumes:
      - dbdata:/var/opt/mssql
    environment:
      - MSSQL_SA_PASSWORD=yourVeryStrong(!)Password
      - ACCEPT_EULA=Y
    ports:
      - 1533:1433 #expose port, so can connect to it using host: 'localhost,1533' | user: sa, password: yourVeryStrong(!)Password

  redis:
    image: "redis:alpine"
    volumes:
      - redisdata:/data
      - redisdata:/var/lib/redis
      - redisdata:/usr/local/etc/redis/redis.conf
    ports:
      - "6379:6379"

# sometimes when building the controller it does not copy the contents of the datasets folder.
  # This is probably a caching docker caching issue, which can be fixed with:
  # 'make clean-hard'
  controller:
    build:
      context: ./
      dockerfile: ./controller/Dockerfile
    container_name: controller
    depends_on:
      - mongo
    # we map the port of the controller to the host.
    # This way we can either connect locally with localhost:5001
    # or from a docker container that accesses host.docker.internal:5001
    ports:
      - "5001:5001"
    environment:
      # The environment variables are set manually here in docker-compose.
      # But in kubernetes the environment variables are set as <SERVICE_NAME>_SERVICE_HOST and <SERVICE_NAME>_SERVICE_PORT
      # Therefore the services must be named accordingly in the kubernetes files.
      - AUTOKERAS_SERVICE_HOST=autokeras
      - AUTOKERAS_SERVICE_PORT=50052
      - MLJAR_SERVICE_HOST=mljar
      - MLJAR_SERVICE_PORT=50053
      #- MCFLY_SERVICE_HOST=mcfly
      #- MCFLY_SERVICE_PORT=50054
      - SKLEARN_SERVICE_HOST=sklearn
      - SKLEARN_SERVICE_PORT=50055
      - FLAML_SERVICE_HOST=flaml
      - FLAML_SERVICE_PORT=50056
      - AUTOGLUON_SERVICE_HOST=gluon
      - AUTOGLUON_SERVICE_PORT=50057
      #- AUTOCVE_SERVICE_HOST=autocve
      #- AUTOCVE_SERVICE_PORT=50058
      - PYTORCH_SERVICE_HOST=pytorch
      - PYTORCH_SERVICE_PORT=50059
      #- ALPHAD3M_SERVICE_HOST=alphad3m
      #- ALPHAD3M_SERVICE_PORT=50060
      - EVALML_SERVICE_HOST=evalml
      - EVALML_SERVICE_PORT=50062
      - PYCARET_SERVICE_HOST=pycaret
      - PYCARET_SERVICE_PORT=50063
      - TPOT_SERVICE_HOST=tpot
      - TPOT_SERVICE_PORT=50064
      - PERSISTENCE_LOGGING_LEVEL=DEBUG
      - SERVER_LOGGING_LEVEL=DEBUG
      - ONTOLOGY_LOGGING_LEVEL=DEBUG
      - BLACKBOARD_LOGGING_LEVEL=DEBUG
      - WIN_DEV_MACHINE=NO
    volumes:
      - datasets:/app/app-data/datasets
      - training:/app/app-data/training
      - training-autokeras:/app/app-data/training/autokeras
      - training-mljar:/app/app-data/training/mljar
      - training-sklearn:/app/app-data/training/sklearn
      - training-flaml:/app/app-data/training/flaml
      - training-gluon:/app/app-data/training/gluon
      #- training-autocve:/app/app-data/training/autocve
      - training-pytorch:/app/app-data/training/pytorch
      #- training-alphad3m:/app/app-data/training/alphad3m
      #- training-mcfly:/app/app-data/training/mcfly
      - training-evalml:/app/app-data/training/evalml
      - training-pycaret:/app/app-data/training/pycaret
      - training-tpot:/app/app-data/training/tpot

  mongo:
    image: mongo
    container_name: mongo
    ports:
      - "27017:27017"

    # limit log verbosity
    command: --quiet --logpath /dev/null

    environment:
      MONGO_INITDB_ROOT_USERNAME: root
      MONGO_INITDB_ROOT_PASSWORD: example
    volumes:
      - mongodata:/data/configdb
      - mongodata:/data/db

#     adapters do not need to be available from the host machine -> no port mappings

  autokeras:
    build:
      context: ./adapters
      dockerfile: ./AutoKeras/Dockerfile
    container_name: autokeras
    environment:
      - GRPC_SERVER_PORT=50052
    volumes:
      - datasets:/app/app-data/datasets
      - training-autokeras:/app/app-data/training

  mljar:
    build:
      context: ./adapters
      dockerfile: ./MLJAR/Dockerfile
    container_name: mljar
    environment:
      - GRPC_SERVER_PORT=50053
    volumes:
      - datasets:/app/app-data/datasets
      - training-mljar:/app/app-data/training

  #mcfly:
  #  build:
  #    context: ./adapters
  #    dockerfile: ./Mcfly/Dockerfile
  #  container_name: mcfly
  #  environment:
  #    - GRPC_SERVER_PORT=50054
  #  volumes:
  #    - datasets:/app/app-data/datasets
  #    - training-mcfly:/app/app-data/training

  sklearn:
    build:
      context: ./adapters
      dockerfile: ./AutoSklearn/Dockerfile
    container_name: sklearn
    environment:
      - GRPC_SERVER_PORT=50055
    volumes:
      - datasets:/app/app-data/datasets
      - training-sklearn:/app/app-data/training

  flaml:
    build:
      context: ./adapters
      dockerfile: ./FLAML/Dockerfile
    container_name: flaml
    environment:
      - GRPC_SERVER_PORT=50056
    volumes:
      - datasets:/app/app-data/datasets
      - training-flaml:/app/app-data/training

  gluon:
    build:
      context: ./adapters
      dockerfile: ./AutoGluon/Dockerfile
    container_name: gluon
    environment:
      - GRPC_SERVER_PORT=50057
    volumes:
      - datasets:/app/app-data/datasets
      - training-gluon:/app/app-data/training

  #autocve:
  #  build:
  #    context: ./adapters
  #    dockerfile: ./AutoCVE/Dockerfile
  #  container_name: autocve
  #  environment:
  #    - GRPC_SERVER_PORT=50058
  #  volumes:
  #    - datasets:/app/app-data/datasets
  #    - training-autocve:/app/app-data/training

  pytorch:
    build:
      context: ./adapters
      dockerfile: ./AutoPytorch/Dockerfile
    container_name: pytorch
    environment:
      - GRPC_SERVER_PORT=50059
    volumes:
      - datasets:/app/app-data/datasets
      - training-pytorch:/app/app-data/training

  #alphad3m:
  #  build:
  #    context: ./adapters
  #    dockerfile: ./AlphaD3M/Dockerfile
  #  container_name: alphad3m
  #  environment:
  #    - GRPC_SERVER_PORT=50060
  #  volumes:
  #    - datasets:/app/app-data/datasets
  #    - training-alphad3m:/app/app-data/training


  evalml:
    build:
      context: ./adapters
      dockerfile: ./EvalML/Dockerfile
    container_name: evalml
    environment:
      - GRPC_SERVER_PORT=50062
    volumes:
      - datasets:/app/app-data/datasets
      - training-evalml:/app/app-data/training

  pycaret:
    build:
      context: ./adapters
      dockerfile: ./PyCaret/Dockerfile
    container_name: pycaret
    environment:
      - GRPC_SERVER_PORT=50063
    volumes:
      - datasets:/app/app-data/datasets
      - training-pycaret:/app/app-data/training

  tpot:
    build:
      context: ./adapters
      dockerfile: ./TPOT/Dockerfile
    container_name: tpot
    environment:
      - GRPC_SERVER_PORT=50064
    volumes:
      - datasets:/app/app-data/datasets
      - training-tpot:/app/app-data/training

volumes:
  datasets: # shared volume between controller, adapters and frontend/dummy to transfer datasets
   # shared volumes between controller and adapters to transfer training files e.g. *.zip file

  training-autokeras:
  training-sklearn:
  training-flaml:
  training-gluon:
  training-pytorch:
  #training-autocve:
  training-mljar:
  #training-alphad3m:
  training-mcfly:
  training-evalml:
  training-pycaret:
  training-tpot:

  # shared volume between controller and frontend / dummy to transfer training
  training:
  mongodata:
  dbdata:
  redisdata:

x-mutagen:
  sync:
    defaults:
      ignore:
        vcs: true
    datasets:
      alpha: "./mutagen_datasets"
      beta: "volume://datasets"
      mode: "two-way-resolved"