improve error handling consistency

[wlav]

Originally noticed by @greenc-FNAL, see here: #495 (comment), there are cases in the current code where a C++ exception is thrown where to code can still be reasonably be expected to return to Python for error handling. This PR removes two cases of throwing C++ exceptions in favor of setting Python exceptions and returning nullptr.

[codecov]

Codecov Report

❌ Patch coverage is 83.33333% with 4 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
plugins/python/src/modulewrap.cpp	83.33%	2 Missing and 2 partials ⚠️

@@            Coverage Diff             @@
##             main     #525      +/-   ##
==========================================
+ Coverage   85.10%   85.49%   +0.38%     
==========================================
  Files         142      142              
  Lines        3645     3653       +8     
  Branches      626      629       +3     
==========================================
+ Hits         3102     3123      +21     
+ Misses        326      313      -13     
  Partials      217      217              

Flag	Coverage Δ
unittests	85.49% <83.33%> (+0.38%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
plugins/python/src/modulewrap.cpp	80.35% <83.33%> (+2.90%)	⬆️

... and 3 files with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0539dee...844d126. Read the comment docs.

🚀 New features to boost your workflow:

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

[greenc-FNAL]

LGTM—I assume return "" vs return nullptr was a conscious choice?

I asked Copilot to try to improve the test coverage for the flagged lines missing coverage here, and it came up with 8c28888: feel free to cherry-pick, take inspiration from, or ignore as you prefer.

[wlav]

LGTM—I assume return "" vs return nullptr was a conscious choice?

Yes; nullptr was never on the table, but I was thinking of going for std::optional there, then was afraid of triggering more coverage issues, so figured that since "" (no annotation) is always an error anyway and since this is not a public function, might as well deal with it this way.