docker-compose.yml

version: '3.8'
services:
  solr:
    image: "solr:latest"
    container_name: solr
    hostname: solr
    ports:
     - "8983:8983"
    entrypoint:
      - sh
      - -c
      - |
        mkdir -p server/solr/configsets/nutch/
        docker-entrypoint.sh solr-precreate nutch
    volumes:
      - ./solr/data:/opt/solr/server/solr/configsets/nutch
      - ./nutch/schema.xml:/opt/solr/server/solr/configsets/nutch/conf/schema.xml
      - ./solr/conf/solrconfig.xml:/opt/solr/server/solr/configsets/nutch/conf/solrconfig.xml
      - ./solr/conf/solrconfig.xml:/opt/solr/server/solr/configsets/_default/conf/solrconfig.xml
    networks:
      - solr-net
  
  nutch:
    image: "apache/nutch:latest"
    container_name: nutch
    links:
      - solr
    environment: 
      - JAVA_HOME=/usr/lib/jvm/java-11-openjdk
    working_dir: /root/nutch_source
    ports:
      - "8081:8081"
      - "8080:8080"
    entrypoint: 
      - sh
      - -c
      - |
        crawl -i -D solr.server.url=http://localhost:8983/solr/nutch -s urls crawl 30
    volumes:
        - ./nutch/exchanges.xml:/root/nutch_source/runtime/local/conf/exchanges.xml
        - ./nutch/index-writers.xml:/root/nutch_source/runtime/local/conf/index-writers.xml
        - ./nutch/nutch-site.xml:/root/nutch_source/runtime/local/conf/nutch-site.xml
        - ./nutch/regex-urlfilter.txt:/root/nutch_source/runtime/local/conf/regex-urlfilter.txt
        - ./nutch/seed.txt:/root/nutch_source/urls/seed.txt
    networks:
      - solr-net
  
  python:
    container_name: python
    build:
      context: .
      dockerfile: Dockerfile
    image: python-mongodb-connect
    links:
      - solr
    depends_on:
      - solr
    networks:
      - solr-net
  
networks:
  solr-net:
    name: solr-network