Re: Do special functions need to have all kwargs/attributes/methods of ufuncs?
Albert Steppi <[email protected]> Thu, 21 Sep 2023 11:27:25 -0400
| Newsgroups | gmane.comp.python.scientific.devel |
|---|---|
| Message-ID | <CAPceVpWKrL+YNeDfYTY=Lr90quxibTtat005Uz-+NM+EpMpJNA@mail.gmail.com> |
> > But it works today? > > ``` > In [1]: import numpy as np > > In [2]: import scipy.special as special > > In [3]: isinstance(special.erf, np.ufunc) > Out[3]: True > > In [4]: isinstance(special.ndtr, np.ufunc) > Out[4]: True > Sorry, you need to set `export SCIPY_ARRAY_API=1`, to put the dispatch to alternative backends in place. As things currently stand, ndtr, erf, and some others, will be replaced with functions which dispatch to a given backend depending on the type of the input. On Thu, Sep 21, 2023 at 03:35 Evgeni Burovski <[email protected]> wrote: > > On Wed, Sep 20, 2023 at 10:33 PM Albert Steppi <[email protected]> > wrote: > >> > >> I think making something like isinstance(special.ndtr, np.ufunc) work > would require some pretty ugly hackery, so I'd prefer to at least break > this. > > But it works today? > > ``` > In [1]: import numpy as np > > In [2]: import scipy.special as special > > In [3]: isinstance(special.erf, np.ufunc) > Out[3]: True > > In [4]: isinstance(special.ndtr, np.ufunc) > Out[4]: True > ``` > _______________________________________________ > SciPy-Dev mailing list -- [email protected] > To unsubscribe send an email to [email protected] > https://mail.python.org/mailman3/lists/scipy-dev.python.org/ > Member address: [email protected] > _______________________________________________ SciPy-Dev mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3/lists/scipy-dev.python.org/ Member address: [email protected]