DEP: update version in deprecation warning of interpnd

[j-bowhay]

Closes #22097

It was reported that interpolate.interpnd.GradientEstimationWarning etc raised a warning saying they would be removed in 1.14. In fact they were only deprecated in this cycle and the version in the warning was outdated. This pr updates the version number so they can be removed in the next release.

[h-vetinari]

Thanks for tackling this. If it's urgent we can merge this as-is, though I'd like to fix this in a way that we don't forget it again so easily, see suggestion below.

[jorenham]

btw, there's a similar deprecation message issue in e.g. ttest_ind, caused by:

scipy/scipy/_lib/deprecation.py

Lines 220 to 222 in ba8553c

	message = (f"Arguments {intersection} are deprecated, whether passed "
	"by position or keyword. They will be removed in SciPy "
	"{version}. ")

(it's the missing f prefix on line 222)

which leads to e.g.

DeprecationWarning: Arguments {'permutations'} are deprecated, whether passed by position or keyword. They will be removed in SciPy {version}. Use ``method`` to perform a permutation test.
  scipy.stats.ttest_ind(permutations=1)

(i.e. the {version} isn't formatted)

[jorenham]

oh and the scipy.special.{lpn,lpmn,clpmn} don't warn about being deprecated when used, although they're deprecated in 1.15.0 (at least according to the docstrings)

[j-bowhay]

btw, there's a similar deprecation message issue in e.g. ttest_ind, caused by:

scipy/scipy/_lib/deprecation.py

Lines 220 to 222 in ba8553c

	message = (f"Arguments {intersection} are deprecated, whether passed "
	"by position or keyword. They will be removed in SciPy "
	"{version}. ")

(it's the missing f prefix on line 222)

which leads to e.g.

DeprecationWarning: Arguments {'permutations'} are deprecated, whether passed by position or keyword. They will be removed in SciPy {version}. Use ``method`` to perform a permutation test.
  scipy.stats.ttest_ind(permutations=1)

(i.e. the {version} isn't formatted)

Great spot thanks, has now been fixed.

oh and the scipy.special.{lpn,lpmn,clplm} don't warn about being deprecated when used, although they're deprecated in 1.15.0 (at least according to the docstrings)

@steppi what's the plan here?

[h-vetinari]

oh and the scipy.special.{lpn,lpmn,clplm} don't warn about being deprecated when used, although they're deprecated in 1.15.0 (at least according to the docstrings)

@steppi what's the plan here?

Could you or @jorenham add a warning? We shouldn't deprecate only in docstrings.

[h-vetinari]

I'd like to understand better why we have so many warnings that (before this PR) say that stuff will be removed in 1.14. That means either the warnings were all are wrong, or we missed to remove a lot of stuff.

[jorenham]

oh and the scipy.special.{lpn,lpmn,clplm} don't warn about being deprecated when used, although they're deprecated in 1.15.0 (at least according to the docstrings)

@steppi what's the plan here?

Could you or @jorenham add a warning? We shouldn't deprecate only in docstrings.

Sure, but I'm not clear which version I should mention that it'll be removed at.

[h-vetinari]

Look at when the deprecation in the docstring was added (1.15 IIUC?), then add two minor versions (=1.17)

[jorenham]

Look at when the deprecation in the docstring was added (1.15 IIUC?), then add two minor versions (=1.17)

Ah ok, I wasn't aware that the docstring was added in 1.15, because it looked like it was just forgotten to be removed 🤷🏻

[j-bowhay]

Thanks for the review @h-vetinari and all the eagle eyed spots @jorenham