Generate repro test errors when model does not run successfully

[jo-basevi]

This PR:

• Adds more stdout/error logs to exception messages so they will be displayed as part of annotations on the PR workflow logs (to make it easier to find out what went wrong)
• Moves checking whether the experiments ran without errors as part of setup in a fixture rather than within the test. This means any exceptions will be reported as a test setup error rather than a test failure. This will hopefully make it more clear when test failures are repro failures rather than the model did not run.

Closes #170

[codecov]

Codecov Report

❌ Patch coverage is 84.21053% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 85.69%. Comparing base (b881d91) to head (9432980).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
...fig_tests/config_tests/test_bit_reproducibility.py	78.94%	4 Missing ⚠️
src/model_config_tests/exp_test_helper.py	89.47%	2 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##             main     #173       +/-   ##
===========================================
+ Coverage   68.39%   85.69%   +17.29%     
===========================================
  Files          14       22        +8     
  Lines         829     1244      +415     
===========================================
+ Hits          567     1066      +499     
+ Misses        262      178       -84     

☔ 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.

[jo-basevi]

Rebased changes onto main and resolved conflicts.

[jo-basevi]

The latest version of pytest-cov dropped support for subprocess measurement by default (see https://pytest-cov.readthedocs.io/en/latest/subprocess-support.html). As a number of tests for tests currently use subprocess to invoke the tests, this python code coverage wasn't being measured. Added measuring code coverage in subprocess call back in 9432980

[CodeGat]

We ran through this on a call, looks good!