Add ab09hd and tg01gd, and show that n,m,p can be removed

[mccullerlp]

Hi all,
I wanted to test using ab09hd as it can ensure relative error during reduction. Additionally, tg01gd is useful for maintaining descriptor systems and reducing them back to a standard form. I've added wrappers and python interfaces. It's less work if you are willing to pull directly, but I could split this into separate PR's. The docstrings are possibly too much of a copy of the slicot documentation?

Additionally, I added an auxiliary wrapper for ab09nd that adjusts the dependency resolution between n,m,p and a,b,c. This is a general technique that can allow the n,m,p arguments to be dropped from any/all of the wrappers. I know this changes the interface, since n,m,p are exposed to the external API, even though they could all be determined from the python-side wrappers anyway. Including it since I already tested it, and it may be of some interest to simplify the interfaces.

[bnavigator]

Thank you very much for your enthusiasm and work on Slycot.

• Please rebase your contribution onto current master, so that merge conflicts are resolved and we can run the CI.
• Please adjust the docstring to the numpydoc standard, see @KybernetikJo's open Pull requests Change analysis.py to numpydoc style #206, Change math.py to numpydoc style #211, Change synthesis.py to numpydoc style #212, and Change transform.py to numpydoc style #213 as examples.
• New wrappers can leverage calculating parameters implicitly, some do already, some really old ones don't. Whenever I made a new wrapper it was a hit an miss for me though about what f2py could deduce in the right order and allocate correctly. For this reason, it's mandatory that you create meaningful unit tests.

[bnavigator]

.*~ This is already covered by *~
What kind of tool creates files with # in it?

[bnavigator]

Remove the leading ... parts. The wrapper only returns two-dimensional arrays.

[bnavigator]

Hi @mccullerlp, glad to see you working on this branch. My comments from December 2023 are still valid. Please make a rebase. There are lots of changes listed in this PR diff that are unrelated.