Link Rating - A link collector and rating app

This is a test application using Spring Boot & Elasticsearch in order to collect Links and display them based on rating and age.

The intention of this app is to be able to run it yourself locally and play around with Elasticsearch and its scoring.

You can check out the application at https://linkrating.spinscale.de/

Requirements

You need to create a github oauth app. You can do so at https://github.com/settings/developers

Next step is to spin up an Elasticsearch instance. You can use Elastic Cloud for this or alternatively just spin up an instance locally and use that one for testing.

Next up is to set the proper environment variables before starting up. I highly encourage you to use something like direnv and then place the variables in the .env file. The following environment variables are required

• SPRING_SECURITY_OAUTH2_CLIENT_REGISTRATION_GITHUB_CLIENT_ID=...
• SPRING_SECURITY_OAUTH2_CLIENT_REGISTRATION_GITHUB_CLIENT_SECRET=...
• ELASTICSEARCH_URL=https://user:password@your-elasticsearch-node:9200
• ADMINS="spinscale,xeraa"
• ELASTIC_APM_SERVICE_NAME=linkrating
• ELASTIC_APM_SECRET_TOKEN=THE_SECRET_TOKEN (optionally)
• ELASTIC_APM_SERVER_URLS=http://localhost:8200
• ELASTIC_APM_ENVIRONMENT=dev (only set it to prod when deploying)

The ADMINS variable contains a list of github logins that are allowed to approve links, which brings us the app functionality.

Note, that the elasticsearch user above needs to write to the users index, the links alias and will also create an index mapping, and an index lifecycle policy, which is used for the links being added. So it needs the manage_index_template and manage_ilm cluster privileges - you could also add the template and the ILM policy manually and would not need that privilege. You also need to have index privileges for the links* and users index, namely read, write, create_index and view_index_metadata.

Application Features

When starting up the app and elasticsearch and github oauth are properly configured, you can open up a browser in the / URL and will see a list of links, that are read from Elasticsearch. The interesting part is the sorting of those links, as this is based on the number of votes and the age of the added link. Newer links or higher votes links will appear higher and as they age out will score lower.

If you are not logged in, there is basically only two options. You can either watch the main page, execute a search in the top or watch a link detail page (which features a stable URL you can link to).

Things are slightly different, once you choose to login via github. Now you can vote for each link. However each link can only be voted once per user.

In case you are wondering why there is another icon. Links can also be deleted, if you are an administrator.

Once you are logged in, you can also submit new links by clicking Submit Link at the top. New links will not be added to the list, unless they have been approved by an administrator user, who can visit the /unapproved URL.

Used technologies

• Spring Boot
• Thymeleaf for server side rendering
• Spectre.css for anything CSS
• Elasticsearch for storing and querying the data