Resolve original condition ids in reinit

[BSnelling]

Updating the function _condition_reinit_target_value to resolve conditions ids in the same way as _eval_nn does, while preserving gradient tracing.

The context in which I noticed this bug was a JAXProblem built from a PetabImporter instance. The underlying petab problem contained conditions which modified a model state variable. The PetabImporter changes the names of conditions and condition targets as part of import which is why _eval_nn needs to resolve current condition ids to originals, and why _condition_reinit_target_value should do the same.

[codecov]

Codecov Report

❌ Patch coverage is 96.87500% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 78.52%. Comparing base (ac032df) to head (beed9f1).

Files with missing lines	Patch %	Lines
python/sdist/amici/sim/jax/petab.py	96.87%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3230      +/-   ##
==========================================
- Coverage   78.59%   78.52%   -0.07%     
==========================================
  Files         318      318              
  Lines       21106    21122      +16     
  Branches     1487     1488       +1     
==========================================
- Hits        16589    16587       -2     
- Misses       4509     4527      +18     
  Partials        8        8              

Flag	Coverage Δ
cpp	72.05% <75.00%> (+0.05%)	⬆️
cpp_python	36.53% <0.00%> (-0.03%)	⬇️
petab	48.36% <78.12%> (+0.06%)	⬆️
petab_sciml	16.32% <68.75%> (+0.04%)	⬆️
petab_sciml_benchmarks	14.89% <43.75%> (+0.03%)	⬆️
python	68.90% <75.00%> (-1.47%)	⬇️
sbmlsuite-jax	31.59% <12.50%> (-0.03%)	⬇️

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

Files with missing lines	Coverage Δ
python/sdist/amici/sim/jax/petab.py	91.94% <96.87%> (+0.43%)	⬆️

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

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

[FFroehlich]

@dweindl does this look good to you?

[dweindl]

Thanks, looks good to me.

I was just wondering whether it might not be easier to just store the mapping during import instead of reverse engineering it later, but fine as is.

[BSnelling]

I was just wondering whether it might not be easier to just store the mapping during import instead of reverse engineering it later

I tried this out but found it didn't simplify the code much (or speed it up). I think it's slightly easier to follow if all the resolution is done together.