Reasons for creating this PR

While debugging performance issues in global search, I noticed that Skosmos sometimes performs several identical SPARQL queries in a row, especially queries for concept labels. This is a waste of time and resources, even if the query can often be served from a HTTP cache (e.g. varnish or nginx).

This PR adds a simple SPARQL query cache that stores results for the duration of a single request. It improves the performance of some global search queries that return many hits from different vocabularies.

Note: This should probably be backported to Skosmos 2 as well. It should be very easy because the GenericSparql code is the same.

Link to relevant issue(s), if any

• n/a

Description of the changes in this PR

• Add query cache within GenericSparql class

Known problems or uncertainties in this PR

None.

Checklist

• phpUnit tests pass locally with my changes
• I have added tests that show that the new code works, or tests are not relevant for this PR (e.g. only HTML/CSS changes)
• The PR doesn't reduce accessibility of the front-end code (e.g. tab focus, scaling to different resolutions, use of .sr-only class, color contrast)
• The PR doesn't introduce unintended code changes (e.g. empty lines or useless reindentation)