DEP: extend some announced deprecations due to out-of-band 1.13 release

[h-vetinari]

Implement the plan to delay some deprecations, due to the out-of-band release of 1.13 that was not accounted for when these deprecation warnings were written.

I've also moved the deprecation sections away from the top line of docstrings, because that is read/displayed by lots of tooling (e.g. IDEs), and should still contain the description of the function (even though we want to place the deprecation warning prominently).

[h-vetinari]

To summarize, this extends the following deprecation announcements, while keeping the originally intended ~1 year window between deprecation and removal.

• 1.13 -> 1.14:
  • follow-up to Clean up sparse array API #18507 (not done on main yet)
  • DEP: replace scipy.interpolate.interp2d by a stub #18583 (has PR, but already delayed to 1.14)
  • follow-up to DEP: signal: deprecate using medfilt and order_filter with float128 and object dtypes #18341 (has PR)
• 1.14 -> 1.15:
  • follow-up to DEP: signal: deprecate cmplx_sort #18587
  • follow-up to DEP: integrate: Deprecate initial values other than None or 0 in cumulative_trapezoid #18219
  • deprecate scipy.stats.rvs_ratio_uniforms (follow-up to MAINT: Move rvs_ratio_uniforms to sampling #18622)
  • follow-up to DEP: deprecate scipy.signal wavelets #19340
  • follow-up to DEP: integrate: deprecate romberg and quadrature #19510

In the future we may choose to completely avoid announcing specific versions for removal, but for now I just did a minimal/mechanical update here.

[j-bowhay]

Thanks @h-vetinari. Looks like the test failure here are real but the rest seems reasonable.

I have been very busy recently so haven't kept up with all the discussions however this seems to match the conclusion of the linked discussion.

I think we should also consider bumping:

scipy/scipy/_lib/deprecation.py

Lines 57 to 63 in 2353b15

	message = (
	f"`scipy.{sub_package}.{module}.{attribute}` is deprecated along with "
	f"the `scipy.{sub_package}.{module}` namespace. "
	f"`scipy.{sub_package}.{module}.{attribute}` will be removed "
	f"in SciPy 1.13.0, and the `scipy.{sub_package}.{module}` namespace "
	f"will be removed in SciPy 2.0.0."
	)

[h-vetinari]

I think we should also consider bumping:

I did this now, even though I think it's kinda a borderline case in terms of utility IMO 🤷

[j-bowhay]

Thanks, all seems in order. Are you ok updating the tracker too?

[h-vetinari]

Thanks, all seems in order. Are you ok updating the tracker too?

Yes, was planning to do that

[tylerjereddy]

@h-vetinari thanks for doing this so quickly, and apologies for not having clearer/louder messaging on the timeline