docker-compose.yml

version: '3.7'

services:

  triplestore:
    image: redpencil/virtuoso
    environment:
      SPARQL_UPDATE: "true"
      DEFAULT_GRAPH: "http://mu.semte.ch/application"
    ports:
      - "8890:8890"
    volumes:
      - ./data/virtuoso:/data/virtuoso
    labels:
      - logging=true
    
  database:
    image: semtech/mu-authorization:0.6.0-beta.5
    environment:
      MU_SPARQL_ENDPOINT: "http://triplestore:8890/sparql"
      ERROR_ON_UNWRITTEN_DATA: "on"
    volumes:
      - ./config/authorization:/config
    labels:
      - logging=true

  identifier:
    image: semtech/mu-identifier:1.10.0
    ports:
      - "8000:80"
    environment:
      DEFAULT_MU_AUTH_ALLOWED_GROUPS_HEADER: "[{\"variables\":[],\"name\":\"public\"}]"
      DEFAULT_ACCESS_CONTROL_ALLOW_ORIGIN_HEADER: "*"
    labels:
      - logging=true

  dispatcher:
    image: semtech/mu-dispatcher:2.1.0-beta.2 
    volumes:
      - ./config/dispatcher:/config
    labels:
      - logging=true

  resources:
    image: semtech/mu-cl-resources:1.21.1 
    volumes:
      - ./config/resources:/config
    labels:
      - logging=true

  migrations:
    image: semtech/mu-migrations-service:0.8.0 
    links:
      - triplestore:database
    volumes:
      - ./config/migrations:/data/migrations


  aggregations:
    build: ./backend/aggregations

  file:
    image: semtech/mu-file-service:3.3.0
    links:
      - database:database
    volumes:
      - ./data/files:/share
    labels:
      - logging=true

  # module for sending delta messages
  # currently used by:
  #   - tika-text-extractor, for saving text when a file is 
  deltanotifier:
    image: semtech/mu-delta-notifier
    volumes:
      - ./config/delta:/config 
    labels:
      - logging=true

  search:
    image: semtech/mu-search:feature-dense-vector
    links:
      - database:database
    volumes:
      - ./config/search:/config

  elasticsearch:
    image: semtech/mu-search-elastic-backend:1.1.0
    volumes:
      - ./data/elasticsearch/:/usr/share/elasticsearch/data
    environment:
      - discovery.type=single-node

  models-init:
    image: google/cloud-sdk:alpine
    command:
      - "gsutil"
      - "-m"
      - "cp"
      - "-r"
      - "gs://abb-textgen-models/topic.model"
      - "gs://abb-textgen-models/RobertaModel_PDF_V1"
      - "gs://abb-textgen-models/NER-model"
      - "gs://abb-textgen-models/zero-shot-model"
      - "gs://abb-textgen-models/translate-nl-en"
      - "gs://abb-textgen-models/dutch-gpt-medium"
      - "gs://abb-textgen-models/dutch-gpt-neo"
      - "gs://abb-textgen-models/gpt2-retrain"
      - "/models/"
    volumes:
      - ./config/poc-ai-embed/models/:/models

  poc-ai-embed:
    image: lblod/poc-ai-embed:latest
    pull_policy: always
    volumes: 
      - ./config/poc-ai-embed/models:/models/
    container_name: lblod-embed-api
    depends_on:
      - models-init
    restart: unless-stopped
    ports:
      - 8081:8080

  # embed-ai-controller:
  #   build: ./backend/embed-ai-controller

  tika-text-extractor:
    build: ./backend/tika-text-extractor
    environment:
      LOG_LEVEL: "debug"
      FILE_RESOURCE_BASE: http://mu.semte.ch/services/file-service/files/
      MU_APPLICATION_FILE_STORAGE_PATH: /share
      MU_SPARQL_ENDPOINT: http://database:8890/sparql
      MU_SPARQL_TIMEOUT: 60
      MODE: "development"
      TARGET_GRAPH: "<https://rdfs.org/soic/ns#content>"
      FILE_DIR: "/share"
    volumes:
      - ./data/files:/share
      - ./backend/tika-text-extractor:/app
    labels:
      - logging=true

  swagger-ui:
    image: swaggerapi/swagger-ui
    container_name: "swagger-ui"
    ports:
      - "8080:8080"
    volumes:
      - ./docs/swagger.json:/swagger.json
    environment:
      SWAGGER_JSON: /docs/swagger.json
      BASE_URL: "/docs"

volumes:
  model_store: