Skip to content

Database Architecture

The thesaurus data that powers Code Thesaurus is stored in JSON files rather than in a database, so that language data stays easy to read and contribute to (see Project Architecture). The Django application does, however, use a database to track how the site is used.

The site runs on Django and can use PostgreSQL as its database (see the manual install directions).

Database Models

While the thesaurus data is loaded from the JSON files at runtime, the app records site activity through a small number of Django models:

  • SiteVisit - represents a single visit to the site, recording the time of the visit, the URL visited, the visitor's user agent, and the referer.
  • LookupData - represents a language comparison lookup, storing the two languages and their versions along with the structure that was compared. It links back to the SiteVisit the lookup happened on.
  • MissingLookup - records when a lookup couldn't be completed because something was missing (such as a language, structure, or concept that doesn't exist yet). It records the type of item that was missing, its value, the language context it happened in, and links back to the SiteVisit where it happened. This data feeds into the issues the team files for missing information.

Future Work

This page is still a work in progress. More detail on migrations, relationships, and how the recorded data is used may be added over time.