Enhance Python Func1 Interface

[ischoegl]

Changes proposed in this pull request

• Provide access to all C++ Func1 objects from Python with dynamically created classes.
• Remove unnecessary argument from advanced functor constructors (oversight in Refactor Func1 #1513)
• Replace Python callbacks returning constants with C++ functors
• Rename TabulatedFunction to Tabulated1 to ensure names are consistent

While the Func1 API in Python should mainly serve as an 'advanced feature', it can serve as a template for other API's (example: new MATLAB interface Cantera/enhancements#177).

If applicable, provide an example illustrating new features this pull request is introducing

In [1]: import cantera as ct

In [2]: f1 = ct.Func1.cxx_functor("sin")
   ...: f2 = ct.Func1.cxx_functor("Gaussian", [0, 1, 2])
   ...: func = ct.Func1.cxx_functor("sum", f1, f2)

In [3]: func
Out[3]: <cantera.func1.Sum1 at 0x10457cd00>

In [4]: print(func.write())
\sin(t) + \mathrm{Gaussian}(t)

Checklist

• The pull request includes a clear description of this code change
• Commit messages have short titles and reference relevant issues
• Build passes (scons build & scons test) and unit tests address code coverage
• Style & formatting of contributed code follows contributing guidelines
• The pull request is ready for review

[codecov]

Codecov Report

Merging #1521 (95ca64a) into main (03f5b82) will increase coverage by 0.06%.
The diff coverage is 85.47%.

@@            Coverage Diff             @@
##             main    #1521      +/-   ##
==========================================
+ Coverage   70.52%   70.59%   +0.06%     
==========================================
  Files         376      376              
  Lines       58918    58956      +38     
  Branches    21198    21197       -1     
==========================================
+ Hits        41551    41619      +68     
+ Misses      14308    14264      -44     
- Partials     3059     3073      +14     

Impacted Files	Coverage Δ
include/cantera/numerics/Func1.h	49.55% <ø> (+1.47%)	⬆️
src/clib/ctfunc.cpp	32.71% <33.33%> (ø)
src/numerics/Func1.cpp	33.45% <84.37%> (+5.49%)	⬆️
interfaces/cython/cantera/func1.pyx	90.47% <85.96%> (-9.53%)	⬇️
src/numerics/Func1Factory.cpp	96.70% <100.00%> (ø)

... and 4 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[speth]

I'm not really clear on the use case for constructing Func1 objects in Python, except for Func1Py which allows calling arbitrary Python functions. The latter makes it easy to define both simple functions as lambdas as well as arbitrarily complex functions including capturing outside variables, which the piecewise Func1 construction can never replicate.

[ischoegl]

@speth ... this is ready for a review. For the Python Func1 API, I kept the access outside of the main constructor. On of my initial reasons is that it is possible to recreate the C++ type names dynamically (see SO example), but I'm not sure that it's worth the effort. Another reason is that I want to keep this an "advanced feature" with the current Python callback behavior remaining the go-to solution.

[ischoegl]

I'm not really clear on the use case for constructing Func1 objects in Python, except for Func1Py which allows calling arbitrary Python functions. The latter makes it easy to define both simple functions as lambdas as well as arbitrarily complex functions including capturing outside variables, which the piecewise Func1 construction can never replicate.

I am aware that the Func1Py approach is a lot more flexible. This PR (where most lines are actually a clean-up for #1513) mainly illustrates that it is extremely simple to add access. Python can serve as a test-bed for other API's that lack the generic callbacks (e.g. MATLAB, or prototyping for C++ implementations). In theory, the C++ functors should also be faster than Python callbacks, although I'm not sure how much this matters. Regardless, I replaced the constant lambda functions with the C++ equivalent.

[ischoegl]

@speth ... I renamed the factory class method to cxx_functor; the factory recreates the underlying C++ classes dynamically, where type names are preserved.

Ad use case: the Python Func1 API will allow for assigning dedicated C++ functors from the Python interface. Beyond this PR, there is no additional Cython code necessary (all classes are created dynamically).

[speth]

Thanks, @ischoegl. I'm glad to see that this incorporates the simplification I had been trying to get at in #1513 to eliminate the unnecessary integer argument to the Func1 constructors. Besides that, just a couple of very minor comments.

[ischoegl]

Thanks, @speth for the prompt review. I can retain the old name for TabulatedFunction, but explained why I opted to update. Let me know ...

PS: there is one newly introduced ct.TabulatedFunction in a test in #1515, which I'll address based on what is merged first.

[speth]

Thanks, @ischoegl, I think this is good to go.