Correct `linkml_files.URL_FOR` - add URL resolutiontests

[sneakers-the-rat]

Continues: https://github.com/linkml/linkml-runtime/pull/310 Fix: https://github.com/linkml/linkml/issues/2047

Previously I had fixed some problems with linkml_files but didn't test the URL_FOR function. This is used by the ContextGenerator and RDFGenerator, and causes linkml tests and fail and breaks them.

This PR

• fixes URL_FOR
• switches tests to pytest
• Add requests_cache
• Unskip tests for URL resolution and correctness - these are ~slow-ish, but since we're using them upstream i figure we should test them.
• Skip fail-prone and unused github URL tests - I can't see where these functions are used within the library and they give us 403's constantly, we might want to remove those? i mostly care about the local path functions here, but also i don't think we should have broken stuff lying around if we can help it

edit: confirmed this does fix test_generators/test_rdfgen.py

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 64.35%. Comparing base (0f99716) to head (df1a56e). Report is 9 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #320      +/-   ##
==========================================
+ Coverage   62.70%   64.35%   +1.64%     
==========================================
  Files          63       63              
  Lines        8580     8612      +32     
  Branches     2444     2447       +3     
==========================================
+ Hits         5380     5542     +162     
+ Misses       2583     2456     -127     
+ Partials      617      614       -3     

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[cmungall]

Merged as the action appears broken, see:

https://github.com/linkml/linkml-runtime/pull/324

[cmungall]

Retrospective note: it turns out the SoT for this file is actually in linkml_model, and any changes such as this PR get wiped out by the metamodel release process. Clearly this is not desirable, and the SoT should probably switch to this repo