Codecov Report

Merging #1513 (4983a88) into main (36faa94) will increase coverage by 0.07%.
The diff coverage is 54.84%.

@@            Coverage Diff             @@
##             main    #1513      +/-   ##
==========================================
+ Coverage   70.42%   70.49%   +0.07%     
==========================================
  Files         375      376       +1     
  Lines       58440    58891     +451     
  Branches    20919    21193     +274     
==========================================
+ Hits        41155    41516     +361     
- Misses      14262    14316      +54     
- Partials     3023     3059      +36     

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

@speth ... at this point, this is ready for a review. While the line count is a bit larger than expected (a lot of this is due to deprecations, new unit tests and refactored code that otherwise doesn't introduce anything new), this PR should remove a lot of lint from the Func1 interface, which will be a lot simpler after the deprecations flushed out. As an example, I created a new Log1 functor, which only required a handful of lines.

The central enhancement are new factory functions (while eliminating the need for magic numbers), where I stuck to MATLAB nomenclature for the time being. Every single Func1 class is accessible from clib at this point, which I hope will be propagated into the new experimental MATLAB interface. Rather than using a single clib (or C++) function signature, I grouped as appropriate, which should make things simpler to use going forward.

@speth ... thanks for the review! I definitely appreciate your suggestions, and implemented almost everything. There is only one minor item left: I can't really think of a good name for the miscellaneous category (other than 'other', which is not an improvement). Everything else should be taken care of.

Rebased due to merge conflict with #1511

Resolved the final comment.

Edit: also renamed func_new_basic to func_new_essential in clib.

@speth ... thanks! I believe I took care of everything.

I like the new organization for the related Doxygen documentation. However, this still doesn't inform the user what name to pass to newFunc1 or func_new_xyz to get a function of a particular type, so I think we need to find a home for these lists in the documentation, even if they risk being out of date if new Func1 classes are added. I'd say that this is somewhat analogous to the list of thermo models that we maintain here.

I agree in general. At the same time, the generated doxygen documentation actually does a really good job summarizing everything, although it's probably not easy to find. As the code is largely absent from Python, this would have to be done manually on the website ...

Right, I think this documentation page is clear that newFunc1("???", c) is what should be called to create a new Const1 function, but how is the user supposed to know that ??? should be constant instead of const or Const1 or something else? This is almost obvious for a lot of the function types, but not all of them.

If you don't want a list in a single place, then an example in each derived class's docstring would be an alternative option.

@speth ... I added lists of available types to each of the four doxygen groupings, and mentioned the type of each class in (improved) class docstrings. I hope this addresses the remaining comment 😄