Checking behaviour when deleting searches in browser functional tests

Why are we doing this?

#2421 added functional browser tests. One aspect not covered is that users can delete searches. This can have quirky behaviour (#2532, #2534). It would be good to test it more, and to help specify its intended defined behaviour.

How will we know when it's done?

When we have tests that cover, for example (but not limited to):

Case 1

1. Add a search.
2. Select some concepts.
3. Delete the search.

Case 2

1. Add a search.
2. Select some concepts.
3. Add a non-overlapping search.
4. Select some concepts.
5. Add an overlapping search.
6. Select some concepts.
7. Delete the original search.

Case 3

1. Add a search.
2. Select some concepts.
3. Add a non-overlapping search.
4. Select some concepts.
5. Add an overlapping search.
6. Select some concepts.
7. Delete all searches.

What are we doing?

• Writing tests to cover at least some cases like these.
• Extending the existing test code to be able to handle deleting searches. This will need:
  • some means for a test author to be able to specify a search is to be deleted
  • adding a small amount of code (a couple of Playwright instructions) to actually delete the search
  • ensuring we can still compute the final state correctly

Defining delivery tasks guidance

Start by locating the functional browser tests added in #2421 and the existing search and final-state handling. Extend the test setup to specify searches for deletion, add the Playwright deletion steps, and cover cases such as deleting one or all searches while verifying the computed final state.