Modify sig fig feedback priority order

[axlewin]

The feedback priority order should now be:

1. Too many significant figures, but otherwise correct
2. Default feedback, if present
3. Too few significant figures, regardless of correctness

This makes numeric questions consistent with coordinate questions, where this was already the feedback order.

Note that we now always do sig fig checking (unless diregardSignificantFigurers is set) and overwrite the sig_figs_too_few feedback with the default feedback if necessary, instead of short-circuiting with the default feedback before the sig figs check. IMO this makes the code easier to follow since we're now only calling useDefaultFeedbackIfNecessary in one place, but we could add the short-circuit option back in between the two types of sig fig check.

Also moves the hardcoded sig_figs feedback tags to constants.

[codecov]

Codecov Report

❌ Patch coverage is 87.50000% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 41.01%. Comparing base (b4f8f70) to head (9a267dd).
⚠️ Report is 17 commits behind head on main.

Files with missing lines	Patch %	Lines
...c/cam/cl/dtg/isaac/quiz/IsaacNumericValidator.java	83.33%	0 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #818      +/-   ##
==========================================
+ Coverage   40.97%   41.01%   +0.03%     
==========================================
  Files         550      550              
  Lines       23555    23583      +28     
  Branches     2837     2845       +8     
==========================================
+ Hits         9652     9672      +20     
- Misses      13032    13039       +7     
- Partials      871      872       +1     

☔ View full report in Codecov by Harness.
📢 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.

[axlewin]

FWIW, even if this gets approved before next week's release, I'd advocate not merging this until the release after; I don't think that modifying this behaviour mid-STEM SMART residential when the content teams won't be available to test is sensible.

[barna-isaac]

Thanks for the changes! I'm pretty happy with this PR now, and feel reasonably confident this will only create the desired change, which is to raise the priority of "too many s.f." messages above the priority of default feedback.

I find I can now quite easily explain to myself how this is achieved: you sank the early return (which returns the default feedback) so it only happens after the too "too many s.f." check. (A byproduct is that the "too many s.f." now runs before the "too few s.f." check, but this won't matter as we can't have both at the same time).

I still think the validator overall is a little too complex. I think validateQuestionResponse is long enough it should only return once, and thevalidateWith family of private methods should consistently mark answers with "tooMany" as well as "tooFew" s.f. problems. Mostly though, I'm just happy we found a small, unobtrusive code change that achieves the new behaviour, and I'm very keen not to touch the numeric validator.

Thanks for the test changes too, and for extracting the constants.