build: pandas min version

[nstarman]

See #15849 (comment)

[github-actions]

Thank you for your contribution to Astropy! 🌌 This checklist is meant to remind the package maintainers who will review this pull request of some common things to look for.

• Do the proposed changes actually accomplish desired goals?
• Do the proposed changes follow the Astropy coding guidelines?
• Are tests added/updated as required? If so, do they follow the Astropy testing guidelines?
• Are docs added/updated as required? If so, do they follow the Astropy documentation guidelines?
• Is rebase and/or squash necessary? If so, please provide the author with appropriate instructions. Also see instructions for rebase and squash.
• Did the CI pass? If no, are the failures related? If you need to run daily and weekly cron jobs as part of the PR, please apply the "Extra CI" label. Codestyle issues can be fixed by the bot.
• Is a change log needed? If yes, did the change log check pass? If no, add the "no-changelog-entry-needed" label. If this is a manual backport, use the "skip-changelog-checks" label unless special changelog handling is necessary.
• Is this a big PR that makes a "What's new?" entry worthwhile and if so, is (1) a "what's new" entry included in this PR and (2) the "whatsnew-needed" label applied?
• At the time of adding the milestone, if the milestone set requires a backport to release branch(es), apply the appropriate "backport-X.Y.x" label(s) before merge.

[pllim]

With v6.1rc out, we don't backport to v6.0.x anymore, so I adjusted the milestone. FYI.

Also, need to update this:

astropy/tox.ini

Line 80 in 7542bc9

oldestdeps: pandas==1.4.*

[pllim]

Actually the tox.ini setup suggested that we are supposed to support pandas>=1.4 so maybe we should not backport this.

[nstarman]

It would be nice if we could tie oldest-deps in tox to our pyproject.

[pllim]

It would be nice if we could tie oldest-deps in tox to our pyproject.

@WilliamJamieson might have wrote something up for such things at one point but I don't remember where. I vaguely remember @jdavies-st had ideas too. 🤞

[pllim]

From a quick grep, we also don't need this anymore:

astropy/astropy/table/table.py

Line 95 in 7542bc9

__doctest_requires__ = {"*pandas": ["pandas>=1.1"]}

Also the blurb about "Version 0.14" in docs/install.rst under pandas.

Maybe can also get rid of this but @taldcroft should confirm:

astropy/astropy/table/table.py

Lines 4147 to 4154 in 7542bc9

	# If pandas is older than 0.24 the type may have turned to float
	if column.dtype.kind != out[name].dtype.kind:
	warnings.warn(
	f"converted column '{name}' from {column.dtype} to"
	f" {out[name].dtype}",
	TableReplaceWarning,
	stacklevel=3,
	)

[jdavies-st]

It would be nice if we could tie oldest-deps in tox to our pyproject.

@WilliamJamieson might have wrote something up for such things at one point but I don't remember where. I vaguely remember @jdavies-st had ideas too. 🤞

https://github.com/spacetelescope/minimum_dependencies is the package you want. It was originally a script within jwst, but William pulled it out into its own package. It parses the minimum versions of packages specified in pyproject.toml and then writes out a requirements.txt file with these versions pinned. The CI job can then install these versions first, then the package without deps, and see whether the tests pass.

[nstarman]

Thanks @jdavies-st! Perfect for a follow up PR simplifying our tox.ini.

[taldcroft]

Fine for table

[pllim]

But @taldcroft , can we remove the snippet mentioned here?

#16308 (comment)

[taldcroft]

But @taldcroft , can we remove the snippet mentioned here?

#16308 (comment)

Yes, that can go since pandas >= 0.24 is always true. The one thing is that this removes the warning if the user specifies use_nullable_int = False in the call to to_pandas(). I think this is reasonable and a good thing -- the user should not to be warned about something they specifically opted into. The doc strings needs to be updated.

        use_nullable_int : bool, default=True
            Convert integer MaskedColumn to pandas nullable integer type.
            If ``use_nullable_int=False`` and there are masked elements then
            the column is converted to float with NaN for missing elements.

This is a minor API change but a good cleanup for 7.0. In a perfect world this change will break a test (for the warning) but I'm not sure off-hand.

[nstarman]

pre-commit autofix

[nstarman]

pre-commit.ci autofix

[nstarman]

Hm. Appears that pandas v1.5.3 doesn't raise the same error as later versions, instead emitting a warning.