This is a small side-project to index the entirety of Wikipedia in ElasticSearch, and to search it in a web page.

Temporarily hosted on: https://wikipedia.momoperes.ca (using a free-trial of Elastic Cloud, so it will stop working by April 2020).

• api: A Spring Boot REST service to query ElasticSearch
• ingest: A multiprocessing Python app to download and index a dump of English Wikipedia
• search-tool: An Angular 9 Material web app to search the index using the REST API

• Dependencies

  • Maven 3.x, JDK 11+
  • Python 3.7+
  • Pipenv
  • Node.js 12+ and Yarn

• Ingest some documents. In the ingest directory:

  • Create a file named .env and add the following config:

    ES_HOST=http://my-elasticsearch-instance:9243
    ES_USER=elastic
    ES_PASS=*******
    

  • Run pipenv install
  • Make a directory named dumps
  • Look in the master.py file. Adjust the BASE_URL and POOL_SIZE for a better mirror depending on your location, and CPU count
  • Find a suitable dump date on https://dumps.wikimedia.org/enwiki/, e.g. 20200301
  • Start ingesting with pipenv run python master.py 20200301

• Run the API. In the api directory:

  • Set your terminal's environment variables like in the .env file above
  • Run mvn spring-boot:run

• Run the front-end app. In the search-tool directory:

  • Run yarn
  • Run yarn start
  • Open http://localhost:4200. You should be able to search for stuff.

• Run docker-compose pull
• Run docker-compose up -d
• Open http://localhost:8081. Note that the API may take a few minutes to start up, as the image first compiles the program.