ENH: special: Add inverse of digamma function#25188
Conversation
There was a problem hiding this comment.
Thanks again @fbourgey !
I get similar refguide check issues in #25161 . Could you weigh in there @j-bowhay as they come from numpydoc?
Tests and python wrappers here look fine to me. What is missing cython_special testing. There are two places that need to be updated accordingly:
- scipy/special/tests/test_cython_abi.py
- scipy/special/tests/test_cython_special.py
|
You will need to add a skip here: Line 33 in 656dbc5 |
Thanks for the review @dschmitz89! Is there a policy that new scalar special ufuncs should also be exposed through cython_special, or should this one remain Python-only? I think psi is already in cython_specialm so this might be natural. If so, I will add it to SciPy's Cython API. I have decided to change the name to digammainv for consistency with other inverse functions. |
Indeed. New private functions don't have to go into |
OK, thanks. Let me know if this the correct approach. I ran |
Reference issue
Closes #17507
What does this implement/fix?
digamma_invfrom xsf. It was added by @dschmitz89 in ENH: Add inverse digamma function xsf#130Additional information
I've tried to follow the pattern of MAINT: special: Add von Mises CDF from xsf #25186 though here this is a change to the public api.
Added a post in the development forum
AI Generation Disclosure
Codex helped me with the move and the doc as a draft. I've then updated the code, and added the examples.