Closed j-carson closed 1 year ago
RectBivariateSpline
, one of the recommended replacements for interp2d
is already supported in class GeoidGPM
, see keyword argument kind
.
Should support for interp2d
be removed altogether or replaced with an other recommendation?
Honestly I am new to this package, so I was looking at all the kinds and discovered this one was based on deprecated functionality. It seems that the interp2d option should at least be deprecated in PyGeodesy? What happens when the function is removed from scipy is up to you as the maintainer, I think.
Thank you for reporting. Here the proposed resolution for the classes GeoidPGM
and GeoidG2012B
, forthcoming in PyGeodesy 23.10.06.
PyGeodesy
continues to support scipy.interpolate.interp2d
but issues a DeprecationWarning
when used with scipy 1.10
or later. Once interp2d
is missing in scipy.interpolate
, PyGeodesy
throws an error.
An enhancement may be, replace scipy.interpolate.interp2d
support in PyGeodesy
with something equivalent like -.RegularGridInterpolator
.
Lastly, see env variable PYGEODESY_WARNINGS
and function DeprecationWarnings
for more details.
Thank you! That sounds like a very reasonable resolution to me.
Great. Thank you again. Closing the issue.
Class GeoidPGM uses scipy interp2d function, which has been deprecated
Use of function https://github.com/mrJean1/PyGeodesy/blob/840f4ed496452e15e8eb2fe02f4a8d044f880c37/pygeodesy/geoids.py#L208C19-L208C19
scipy deprecation transition guide http://scipy.github.io/devdocs/notebooks/interp_transition_guide.html