Refactor test suite

[gboeing]

This PR re-thinks the OSMnx test suite by doing the following:

• separate live/online API tests from local/offline coverage and unit tests
• add pytest fixtures and cache files to support offline tests
• expand offline tests for better unit testing
• minimize live/online API tests to keep them for compatibility rather than coverage
• CI now just uses local/offline tests
• Live/online tests run once a week on a schedule to maintain compatibility

At this point, my main question is about brittleness and maintainability. There's a lot of monkey-patched things in here to allow all the tests to run offline with high coverage.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.88%. Comparing base (6f405d8) to head (795cf55).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1380      +/-   ##
==========================================
+ Coverage   98.61%   99.88%   +1.27%     
==========================================
  Files          25       25              
  Lines        2594     2584      -10     
==========================================
+ Hits         2558     2581      +23     
+ Misses         36        3      -33     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[gboeing]

I'm going to close this. Too much monkey-patching adds too much technical debt from spaghetti code. My fundamental goal is to keep this package legible and maintainable. I'll keep brainstorming a better way to split offline correctness tests from online live API tests.