Handle reserved symbol names during code generation

[FFroehlich]

Summary

• avoid reserved name renaming during SBML import
• drop reserved name check in ModelQuantity
• sanitize symbols when printing C++ and JAX code
• adjust sbml tests for new behavior
• fix reserved symbol handling during code generation
• allow BNGL conversion for empty-compartment tests
• fix JAX SBML test module import
• revert SBMLSuiteJax import path
• use semantic-results-jax directory for JAX SBML suite results
• fix parameter ID names
• fix parameter ID handling
• fix JAX sanitization of variable names

Testing

• pre-commit run --files python/sdist/amici/jax/ode_export.py
• pytest python/tests/test_sbml_import.py::test_import_same_model_name -vv -s
• pytest python/tests/test_events.py::test_handling_of_fixed_time_point_event_triggers -q
• pytest tests/benchmark_models/test_petab_benchmark.py -k Boehm_JProteomeRes2014 -q
• pytest tests/benchmark_models/test_petab_benchmark_jax.py -k Boehm_JProteomeRes2014 -q
• pytest tests/sbml/testSBMLSuiteJax.py::test_sbml_testsuite_case_jax[00525] -q

fixes #2461

---

https://chatgpt.com/codex/tasks/task_b_685d2802e714832b968ced9c529d9411

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 80.27%. Comparing base (08bf630) to head (4e43dc7).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2853      +/-   ##
==========================================
- Coverage   80.38%   80.27%   -0.11%     
==========================================
  Files         329      329              
  Lines       22410    22413       +3     
  Branches     1514     1514              
==========================================
- Hits        18014    17993      -21     
- Misses       4385     4409      +24     
  Partials       11       11              

Flag	Coverage Δ
cpp	75.41% <92.85%> (-0.11%)	⬇️
cpp_python	33.79% <42.85%> (-0.05%)	⬇️
petab	39.21% <85.71%> (-0.04%)	⬇️
python	74.44% <92.85%> (-0.11%)	⬇️
sbmlsuite-jax	33.83% <57.14%> (-0.19%)	⬇️

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

Files with missing lines	Coverage Δ
python/sdist/amici/cxxcodeprinter.py	95.68% <100.00%> (+0.19%)	⬆️
python/sdist/amici/de_export.py	97.43% <100.00%> (+<0.01%)	⬆️
python/sdist/amici/de_model_components.py	96.40% <ø> (-0.03%)	⬇️
python/sdist/amici/jax/jaxcodeprinter.py	90.62% <100.00%> (+1.73%)	⬆️
python/sdist/amici/jax/ode_export.py	98.03% <100.00%> (+0.08%)	⬆️
python/sdist/amici/sbml_import.py	93.20% <ø> (-1.00%)	⬇️

... and 1 file with indirect coverage changes

🚀 New features to boost your workflow:

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

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[dweindl]

What happens if a model has a parameter t that is not time?

[dweindl]

I'd be happy to handle reserved symbols only during code generation, but I don't think this will work well in general. The problem is that we don't know anymore whether the given symbol is a user-defined symbol or an amici-internal symbol. There are quite a number of (currently unhandled) names that will cause problems, that we can't handle here (d{}dt, xdot{}, stau{}, any C++ keywords and typenames, any parameter names from amici's model functions, ...).
I think the most and only sustainable solution would be some string-based indexing that is evaluated at compile-time that avoids any collisions (see also #2226).