Deprecate image.coalignment

[wtbarnes]

image.coalignment has been moved to sunkit-image (See sunpy/sunkit-image#78).
This PR deprecates the coalignment module.
This deprecation will be included in 4.0 and thus we can remove the module in 4.1

Addresses #5425.

PR Checklist

• I have followed the guidelines in the Contributing document
• Changes follow the coding style of this project
• Changes have been formatted and linted
• Changes pass pytest style unit tests (and pytest passes).
• Changes include any required corresponding changes to the documentation
• Changes have been rebased and squashed to the minimal number (typically 1) of relevant commits
• Changes have a descriptive commit message with a short title
• I have added a Fixes #XXXX - or Closes #XXXX - comment to auto-close the issue that your PR addresses

[wtbarnes]

The doc fail is real. I've messed up the use of .. warning:: in the module docstring.

I take that back. It seems that sphinx does not like having the warning raised in coalignment.py.

[wtbarnes]

I just ended up slapping the deprecated decorator on every function.

[wtbarnes]

Well this has become slightly more complicated because of the use of apply_shifts in mapsequence_solar_derotate which I didn't realize until now,

sunpy/sunpy/physics/solar_rotation.py

Line 148 in a5c5be9

return apply_shifts(mc, yshift_keep, xshift_keep, clip=clip)

Three possible options:

• Do not deprecate apply_shifts or calculate_clipping
• Move apply_shifts and calculate_clipping to physics.solar_rotation
• Deprecate mapsequence_solar_derotate as well.

[dstansby]

I guess the pragmatic option is move the required code but make it private, and then we can make a decision on what to do about mapsequence_solar_derotate independently of this PR?

[wtbarnes]

I guess the pragmatic option is move the required code but make it private, and then we can make a decision on what to do about mapsequence_solar_derotate independently of this PR?

That seems sensible. I had kind of forgotten that physics.solar_rotation existed...

[dstansby]

One of the remote data test fails is real: https://github.com/sunpy/sunpy/runs/5557145861?check_suite_focus=true#step:7:5237 and I've just rebooted the figure tests to check them (I think the fail was the MPL issue that's now fixed)

[wtbarnes]

One of the remote data test fails is real: https://github.com/sunpy/sunpy/runs/5557145861?check_suite_focus=true#step:7:5237 and I've just rebooted the figure tests to check them (I think the fail was the MPL issue that's now fixed)

Ah thanks for catching that. I've ignored the warnings in the two doctests. We'll also need to remove that section of the guide once the coalignment stuff actually gets removed.