optimize get_classes_by_slot() in schemaview.py

[sujaypatil96]

The motivation to optimize this PR is that we are having to remove the Applicable classes section in the generated slot markdown/web documentation pages: https://github.com/linkml/linkml/blob/main/linkml/generators/docgen/slot.md.jinja2#L36-L46 for large schemas like mixs.

The optimizations that have been applied here are:

• Use set() for accumulating class names
• reduce repetitive calls to self.all_classes()
• reduce number of iterations over all_classes
• direct/induced handled in separate loops

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (51d2fce) 62.80% compared to head (54d9515) 62.90%.
Report is 5 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #300      +/-   ##
==========================================
+ Coverage   62.80%   62.90%   +0.10%     
==========================================
  Files          61       62       +1     
  Lines        8506     8530      +24     
  Branches     2239     2239              
==========================================
+ Hits         5342     5366      +24     
- Misses       2553     2554       +1     
+ Partials      611      610       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[sujaypatil96]

After wonderful optimizations by Jonny in #296 I thought this would be an appropriate time to take advantage of his modifications and apply it to large schemas (like mixs) that need this performance improvement.

[sujaypatil96]

@cmungall thanks for approving this PR. I'm still benchmarking if it's easier/faster to return ClassDefinition instead of ClassDefinitionName (corresponding changes will be required in slot.md.jinja2).

I won't merge this in till I have results from the benchmarking. I'm experimenting with these changes on the mixs schema.

[sujaypatil96]

I think we are good to go with this PR as is @cmungall.