Micro optimization: avoid recompiling regex in incremental_coverage.py

[mhucka]

The function determine_ignored_lines() in dev_tools/incremental_coverage.py compiled a regex. Compiling regex inside a function that is potentially called multiple times adds unnecessary overhead. This PR pre-compiles some regex patterns at the module level in dev_tools/incremental_coverage.py to avoid redundant compilation.

In simple wall-clock timings on a linux VM, running dev_tools/pytest-and-incremental-coverage, I get an average of about 10% time improvement over the unmodified version of the file.

(Full disclosure: this was largely the work of Jules, which I've experimenting with.)

[codecov]

Codecov Report

❌ Patch coverage is 77.77778% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 99.62%. Comparing base (6689231) to head (cbe4c83).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
dev_tools/incremental_coverage.py	77.77%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #7953      +/-   ##
==========================================
- Coverage   99.63%   99.62%   -0.01%     
==========================================
  Files        1108     1108              
  Lines       99571    99574       +3     
==========================================
- Hits        99205    99204       -1     
- Misses        366      370       +4     

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[pavoljuhas]

After adding a blank line to every Python source in this repo, the wall time before and with this PR for

python dev_tools/check_incremental_coverage_annotations.py origin/main

changed from 9.1 to 8.9 s. This should be considered in context of a preceding pytest coverage run needed to create .cover files which takes about 5 minutes. The improvement in time is at this scale negligible; indeed re.search caches its compiled patterns so any difference is likely from cache lookup overhead.

The other possible benefit of this PR could be in making the code more readable, but it goes the opposite way by adding more module constants. The level of improvement here does not justify more work.