ci: Refactor code coverage workflow

[stinodego]

Following up on #14532

This brings the workflow more in line with the existing testing workflows.

Changes

• Only run store the cache on main, avoiding cache storage space issues.
• Only run workflow when Rust or Python files have been modified.
• Make setup more modular (more easily see where things go wrong and how long each step takes)

(I did a random test refactor to trigger the workflow)

I'm also thinking this could be much improved by running coverage in the existing test workflows and combining the results. We should look into that later.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (815732b) 80.76% compared to head (f1ed81e) 77.90%.
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #14563      +/-   ##
==========================================
- Coverage   80.76%   77.90%   -2.87%     
==========================================
  Files        1326     1326              
  Lines      173035   173035              
  Branches     2439     2439              
==========================================
- Hits       139758   134795    -4963     
- Misses      32805    37797    +4992     
+ Partials      472      443      -29     

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

[c-peters]

Regarding refactoring and running in separate workflows, this would be ideal. The only blocker for now is that on Linux the coverage tool segfaults (see original PR).