Use tags to populate subject areas

[MatMoore]

Now that the ingestion code is setting tags in DataHub for the subject areas, we can start to use these for search/browse instead of the DataHub domain.

This work will allow us to assign entities to more than one subject area, so we don't need to worry about something not precisely fitting into a single category when redefining the taxonomy.

Known issues

I've used the test environment for testing this since we've had some OpenSearch issues on the dev environment. When I compared before and after I noticed a few discrepancies, but the number of entities in each subject area is almost the same.

It seems like the discrepancies are due to two things, neither of which seem like blockers:

1. Entries in the catalogue that are no longer being updated, but haven't been deleted automatically as part of a stateful ingestion, no longer show up in a subject area (example). These entries should be deleted altogether but for whatever reason they haven't been. This affects CaDeT databases, and some ESDAs (to be expected as these do not belong to an ingestion).
2. When ingesting CaDeT databases, there are cases where the database contains models that are in different domains. We used to pick a single domain to use for the database, but now we assign tags for all of them (example). Such databases now show up in multiple subject areas. I hadn't intended to change this right away but I think it is the behaviour we ultimately want.

Next steps

• Currently where there are multiple subjects tags assigned to an entity we just pick the first one. I'm planning to change this to show all of them. At the same time I can look into the css bug that's making it look wonky when the subject area tag flows onto a new line.
• There is likely some further refactoring that can be done in the home app in Find MoJ data as well, e.g. some remaining references to "domains" that can be cleaned up.

[hjribeiro-moj]

LGTM

[sentry-io]

Suspect Issues

This pull request was deployed and Sentry observed the following issues:

• ‼️ CatalogueError: Unable to execute list domains query / View Issue

_{Did you find this useful? React with a 👍 or 👎}