Fix CUSTOM_OBJFUNC support in Python layer

[riddhi1228]

This PR fixes the handling of CUSTOM_OBJFUNC in the Python interface.

Problem

When evaluating custom objective functions, the Python layer could raise
a KeyError because CUSTOM_OBJFUNC was missing from historyMap.py

Solution

• Added if (false) AddHistoryOutput("CUSTOM_OBJFUNC", ...) in CFlowOutput.cpp
  so updateHistoryMap.py registers it correctly
• Fixed updateHistoryMap.py parser to handle if (false) AddHistoryOutput pattern
• Fixed multi-line BGS_ENTHALPY call in CFlowIncOutput.cpp so parser works correctly
• Regenerated historyMap.py with CUSTOM_OBJFUNC correctly registered

Fixes #2586

[pcarruscag]

==================== Start Test: sphere_ffd_def_bspline ====================
/github/workspace/tests/TestCases/deformation/spherical_ffd
sphere_ffd_def_bspline: PASSED
execution command: SU2_DEF def_spherical_bspline.cfg > def_spherical_bspline.log 2>&1
test_iter=10
test_vals (stored): 0.002068
sim_vals (computed): 0.002068
delta_vals: 0.000000
test duration: 0.03 min
Traceback (most recent call last):
  File "/github/workspace/install/bin/continuous_adjoint.py", line 32, in <module>
    import SU2
  File "/github/workspace/install/bin/SU2/__init__.py", line 19, in <module>
    from SU2 import eval
  File "/github/workspace/install/bin/SU2/eval/__init__.py", line 1, in <module>
    from SU2.eval.functions import function as func
  File "/github/workspace/install/bin/SU2/eval/functions.py", line 92
    header_to_key = {v["HEADER"]: k for k, v in su2io.historyOutFields.items()}
    ^
IndentationError: unexpected indent

[riddhi1228]

Fixed — the indentation error on line 92 has been corrected and black formatting applied. Thanks for catching it.

[riddhi1228]

The black formatting check is failing on historyMap.py even though black 22.6.0 reports no changes needed locally and the file is unchanged on the develop branch. This appears to be a pre-existing formatting issue in the repo rather than something introduced by this PR. Happy to investigate further if you can point me to what the CI is reformatting exactly.

[pcarruscag]

It's ok. The bigger issue is that this historyMap.py file should be generated, not edited manually.
Can you try to run the script that generates the historyMap (it's somewhere in SU2_PY) and see if the tests still pass?

[riddhi1228]

Hi @pcarruscag, updated the PR based on your feedback:

• Moved the fix to CFlowOutput.cpp using if (false) AddHistoryOutput("CUSTOM_OBJFUNC", ...)
  as done in Fixing Custom Objective Function for Unsteady Adjoint  #2587
• Fixed updateHistoryMap.py parser to handle the if (false) pattern and a pre-existing
  multi-line BGS_ENTHALPY call in CFlowIncOutput.cpp that was causing an IndexError
• Regenerated historyMap.py — CUSTOM_OBJFUNC and its derived fields are now correctly registered
• Fixed the 2 failing regression tests from Fixing Custom Objective Function for Unsteady Adjoint  #2587: AVG_TOTALTEMP was renamed to
  SURFACE_TOTAL_TEMPERATURE in the C++ layer but 3 test configs were never updated

[pcarruscag]

There are errors in the tests
parallel_regression.py, serial_regression.py, and serial_regression_AD.py.
Please iterate on them locally, I cannot copy the logs for you each time.

Traceback (most recent call last):
  File "/github/workspace/install/bin/shape_optimization.py", line 313, in <module>
    main()
  File "/github/workspace/install/bin/shape_optimization.py", line 184, in main
    shape_optimization(
  File "/github/workspace/install/bin/shape_optimization.py", line 289, in shape_optimization
    SU2.opt.SLSQP(project, x0, xb, its, accu)
  File "/github/workspace/install/bin/SU2/opt/scipy_tools.py", line 128, in scipy_slsqp
    outputs = fmin_slsqp(
  File "/usr/lib/python3/dist-packages/scipy/optimize/slsqp.py", line 207, in fmin_slsqp
    res = _minimize_slsqp(func, x0, args, jac=fprime, bounds=bounds,
  File "/usr/lib/python3/dist-packages/scipy/optimize/slsqp.py", line 423, in _minimize_slsqp
    g = append(fprime(x), 0.0)
  File "/usr/lib/python3/dist-packages/scipy/optimize/optimize.py", line 327, in function_wrapper
    return function(*(wrapper_args + args))
  File "/github/workspace/install/bin/SU2/opt/scipy_tools.py", line 434, in obj_df
    dobj_list = project.obj_df(x)
  File "/github/workspace/install/bin/SU2/opt/project.py", line 248, in obj_df
    return self._eval(konfig, func, dvs)
  File "/github/workspace/install/bin/SU2/opt/project.py", line 211, in _eval
    vals = design._eval(func, *args)
  File "/github/workspace/install/bin/SU2/eval/design.py", line 151, in _eval
    vals = eval_func(*inputs)
  File "/github/workspace/install/bin/SU2/eval/design.py", line 365, in obj_df
    grad = su2grad(obj_list, grad_method, config, state)
  File "/github/workspace/install/bin/SU2/eval/gradients.py", line 97, in gradient
    grads = adjoint(func_name, config, state)
  File "/github/workspace/install/bin/SU2/eval/gradients.py", line 214, in adjoint
    function(func_name, config, state)
  File "/github/workspace/install/bin/SU2/eval/functions.py", line 133, in function
    func_out = state["FUNCTIONS"]["COMBO"]
KeyError: 'COMBO'

[riddhi1228]

I ran updateHistoryMap.py as requested, but the script appears to produce a different historyMap.py than what's currently in develop — it adds 28 new keys and removes 14 existing ones (including RMS_ENTHALPY, RMS_TEMPERATURE) which breaks the regression tests. This seems to be a pre-existing issue where the generator is out of sync with the current codebase.

The CUSTOM_OBJFUNC fix in CFlowOutput.cpp and updateHistoryMap.py parser changes are correct. Should I:

1. Submit just those two files and leave historyMap.py regeneration for a separate cleanup PR?
2. Or is there a specific version of the generator I should be using?

[pcarruscag]

ok let's go back to what you had before then, I don't want to waste more time with python scripts