Florian Bourgey
Florian Bourgey
It seems that Sphinx does not support `.pyi` files (see [here](https://github.com/sphinx-doc/sphinx/pull/4824) and [here](https://github.com/sphinx-doc/sphinx/issues/7630) and also [here](https://github.com/lakehq/sail/pull/123)) `pandas.DataFrame.columns` is a `pandas._libs.properties.AxisProperty` located [here](https://github.com/pandas-dev/pandas/blob/main/pandas/_libs/properties.pyi#L18C1-L27C19) The same thing exists for `pandas.Interval` for which...
Nonetheless, I slightly updated the `pandas.DataFrame.columns` doc (added a Return and See also section)
@steppi, yes, I've updated the `legendre.h` file. Sorry, first time modifying such files in scipy. Below is what I have updated Can you advise on what the best procedure is...
I have created https://github.com/scipy/xsf/pull/78
@lucascolley @steppi https://github.com/scipy/xsf/pull/78 fixes the issue https://github.com/scipy/scipy/issues/24099 I had added a test in this scipy PR. How do you think I should proceed here?
Following your comment https://github.com/scipy/scipy/pull/23622#pullrequestreview-3235489898, it appears that ```python from scipy.stats import levy_stable levy_stable.parameterization = "S1" levy_stable(1.006, beta=1).pdf(0.1) ``` raises an error ```python ValueError: The function value at x=0.009368566561400948 is NaN;...
Thanks for all those tests. > Maybe better transition logic is based on the observed behavior of the plots - SciPy's calculation of the PDF is too high in the...
This would still depend on `cutoff_exponent` and an additional parameter for testing how *much larger* the scipy PDF is compared to the approximation. What would be the much larger? I...
When after the `cutoff_exponent` and the relative error is above 10%, we return the asymptotic PDF. It might be useful to also add more tests when far in the tail?...