Codestin Search App

ayshih · 2024-03-12T02:04:08Z

Description

#14819 moved the IERS files to a separate package (astropy-iers-data), and now IERS-A is essentially bundled with astropy. However, there persist some remnants of the old logic, where IERS-A is available only if the download were triggered. This results in the peculiar behavior that if automatic downloading of IERS-A files is disabled (iers.conf.auto_download = False) or fails, the bundled IERS-A file is ignored even though it is present (see #13227 (comment)).

This PR fixes the logic so that the bundled IERS-A file is used if iers.conf.auto_download = False, with corresponding fixes in documentation and tests.

This PR does not attempt to fix the logic so that a previously downloaded IERS-A file is used if iers.conf.auto_download is subsequently set to False. I didn't want to wrap my head around whether the downloaded file should have priority over the bundled file, given that astropy-iers-data could also be upgraded.

Fixes #13227: With astropy 6.0, the response to the report would be to upgrade astropy-iers-data to obtain a recent IERS-A table. This PR then fixes the residual bug where setting iers.conf.auto_download = False would ignore that bundled IERS-A table.

Fixes #16128: Because the bundled IERS-A is now no longer ignored by the default IERS_Auto when downloading is blocked, it's now typically impossible to trigger the "degraded accuracy" error/warning unless the user expressly forces the use of the IERS-B data alone (e.g., through the earth_orientation_table ScienceState).

Other related issues:

By checking this box, the PR author has requested that maintainers do NOT use the "Squash and Merge" button. Maintainers should respect this when possible; however, the final decision is at the discretion of the maintainer that merges the PR.

github-actions · 2024-03-12T02:04:36Z

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.

github-actions · 2024-03-12T02:04:45Z

👋 Thank you for your draft pull request! Do you know that you can use [ci skip] or [skip ci] in your commit messages to skip running continuous integration tests until you are ready?

neutrinoceros · 2024-03-12T07:39:43Z

Just a heads-up that this patch seems to touch some of the same lines than #16070, so merge conflicts might ensue. That said, my own PR is currently stuck so I don't have a problem with this one going in first, mind you. I'm also happy to serve as a reviewer if needed !

ayshih · 2024-03-12T13:31:36Z

I think #16070 will be sufficiently orthogonal – improving how IERS-A and IERS-B are combined versus merely ensuring that the combination even happens – that any merge conflicts should be trivial to resolve.

pllim

Thanks! Does the content of https://docs.astropy.org/en/stable/utils/data.html also need updating?

ayshih · 2024-03-12T18:14:46Z

Ah, indeed, that page should be edited too

mhvk

Thanks so much! I have one question on the implementation, whether it might not be even simpler to just preserve whatever has been downloaded (which in itself would seem more logical). What do you think? I am happy to go with what you have as well, as it resolves the biggest issues.

Beyond that, only a nitpick on the changelog entry...

mhvk · 2024-03-12T19:49:25Z

@@ -0,0 +1 @@
+The return type of ``IERS_Auto.open()`` is no longer ``IERS_B`` when automatic updating of the IERS-A file is disabled or the downloading of the new file fails.


I think we should add what it is now! Maybe

IERS_Auto.open() now defaults to the combination of IERS_A and B data bundled in the astropy-iers-data dependency rather than to IERS_B only when automatic ...

Overhauled text

mhvk · 2024-03-12T19:57:34Z

@@ -786,7 +790,7 @@ def open(cls):

        """
        if not conf.auto_download:


I guess the alternative path, which would preserve anything downloaded, would simply be to remove this stanza and in the else clause of the for loop warn only if conf.auto_download is set.

Am I missing something?

Discussion concluded we should leave this for follow-up

ayshih · 2024-03-13T15:43:29Z

I have one question on the implementation, whether it might not be even simpler to just preserve whatever has been downloaded (which in itself would seem more logical). What do you think?

Yes, that could be done, but the edge case that worries me is a user who does not want automatic downloading but at some point in the past happened to download IERS-A. Once that file is in the cache, that stale IERS-A file would continue to be used (by default) despite any subsequent updates of astropy-iers-data.

Of course, the natural solution would be to inspect both the IERS-A file from astropy-iers-data and any downloaded IERS-A file in the cache, and use whichever one has more recent non-predicted values, since that presumably means it is the more recent file. What do you think?

mhvk · 2024-03-13T17:41:09Z

@ayshih - yes, good point. So, I guess to get this to work one would have to also change _refresh_table_as_needed and in the download failure path check whether IERS_A_FILE is more up to date then the current one: effectively, for the no download case, one could do new_table = cls.read() there and fall through to the part that updates self as needed.

p.s. We can obviously move this to a new issue/PR!

ayshih · 2024-03-14T18:30:25Z

Let's get this PR in to solve the easy-to-encounter bug, and defer the improved handling of two IERS-A files to a future PR. I've updated the documentation more, including https://docs.astropy.org/en/stable/utils/data.html

mhvk

Looks good! (modulo the failure of course!)

mhvk · 2024-03-14T19:52:32Z

 therefore recommend the following guidelines:

- * Do one of the following:
+ * Set ``astropy.utils.iers.conf.auto_download = False`` in your Astropy config


I like this change to have just one recommended way!

mhvk

@ayshih - I have a few more nitpicky comments on docstrings, to join those of @eerovaher. Will merge once those are done.

Thanks!

mhvk · 2024-03-20T16:40:27Z

@@ -786,7 +790,7 @@ def open(cls):

        """
        if not conf.auto_download:


Discussion concluded we should leave this for follow-up

neutrinoceros

Thank you so much for doing this. Not only is it fixing a subtle bug that has already confused several users, it's also making the underlying logic easier to reason about ! I only have a couple small suggestions and questions, but overall LGTM

ayshih · 2024-07-08T19:56:06Z

I totally forgot about this PR. I've addressed the comments from @neutrinoceros and rebased the PR.

neutrinoceros

LGTM now, thanks !

mhvk

OK, let's get this in. Thanks, @ayshih!

github-actions Bot added Docs utils.iers labels Mar 12, 2024

ayshih force-pushed the iers_a branch 2 times, most recently from 7c17771 to 7ea06a4 Compare March 12, 2024 03:23

ayshih marked this pull request as ready for review March 12, 2024 04:31

pllim added this to the v6.1.0 milestone Mar 12, 2024

pllim added Bug API change PRs and issues that change an existing API, possibly requiring a deprecation period labels Mar 12, 2024

pllim reviewed Mar 12, 2024

View reviewed changes

pllim requested review from astrofrog, mhvk and taldcroft March 12, 2024 15:25

mhvk reviewed Mar 12, 2024

View reviewed changes

mhvk reviewed Mar 14, 2024

View reviewed changes

ayshih force-pushed the iers_a branch from e298744 to 9f1bd1c Compare March 14, 2024 22:31

eerovaher reviewed Mar 18, 2024

View reviewed changes

Comment thread docs/utils/data.rst Outdated

eerovaher reviewed Mar 18, 2024

View reviewed changes

Comment thread docs/changes/utils/16187.api.rst Outdated

mhvk reviewed Mar 20, 2024

View reviewed changes

ayshih mentioned this pull request Mar 21, 2024

Add GONG H-Alpha map source sunpy/sunpy#7451

Merged

5 tasks

neutrinoceros reviewed Mar 29, 2024

View reviewed changes

neutrinoceros mentioned this pull request Apr 3, 2024

ENH: include exclusive values from IERS_B in IERS_Auto (blocked by #16725) #16070

Draft

1 task

astrofrog removed this from the v6.1.0 milestone Apr 4, 2024

astrofrog added this to the v7.0.0 milestone Apr 4, 2024

ayshih added 3 commits July 8, 2024 15:54

Fix IERS_Auto logic now that IERS-A is bundled

2da17a6

Documentation updates

dbea96f

Addressed comments

30fc213

ayshih force-pushed the iers_a branch from 9435353 to 30eab3f Compare July 8, 2024 19:55

ayshih force-pushed the iers_a branch from 30eab3f to fe708ef Compare July 8, 2024 19:57

Addressed more comments

b484bde

ayshih force-pushed the iers_a branch from fe708ef to b484bde Compare July 8, 2024 20:01

neutrinoceros approved these changes Jul 9, 2024

View reviewed changes

mhvk approved these changes Jul 9, 2024

View reviewed changes

mhvk merged commit 3ebe9b8 into astropy:main Jul 9, 2024

This was referenced Jul 10, 2024

BUG/PERF: defects introduced in #16187 #16693

Closed

Revert "Fix IERS_Auto logic now that IERS-A is bundled" #16695

Closed

ayshih mentioned this pull request Jul 10, 2024

Fix the caching of the table constructed using bundled IERS-A and IERS-B files #16697

Merged

1 task

neutrinoceros mentioned this pull request Jul 15, 2024

TST: regression in weekly cron #16724

Closed

olebole mentioned this pull request Nov 10, 2024

[7.0.0rc1] Mark some "future" test in coordinates as remote_data #17359

Merged

1 task

ayshih mentioned this pull request Nov 13, 2024

Even if IERS_Auto predictive values are stale, do not download IERS-A if downloading is disabled #17387

Merged

ayshih deleted the iers_a branch June 3, 2025 16:30

pllim mentioned this pull request May 28, 2026

IERS-A data emitted IERSDegradedAccuracyWarning in astropy 5.3.3 but throws ValueError in astropy 7 or later #19823

Closed

		@@ -0,0 +1 @@
		The return type of ``IERS_Auto.open()`` is no longer ``IERS_B`` when automatic updating of the IERS-A file is disabled or the downloading of the new file fails.

		@@ -786,7 +790,7 @@ def open(cls):

		"""
		if not conf.auto_download:

Uh oh!

Uh oh!

Conversation

ayshih commented Mar 12, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Uh oh!

github-actions Bot commented Mar 12, 2024

Uh oh!

github-actions Bot commented Mar 12, 2024

Uh oh!

neutrinoceros commented Mar 12, 2024

Uh oh!

ayshih commented Mar 12, 2024

Uh oh!

pllim left a comment

Choose a reason for hiding this comment

Uh oh!

ayshih commented Mar 12, 2024

Uh oh!

mhvk left a comment

Choose a reason for hiding this comment

Uh oh!

mhvk Mar 12, 2024

Choose a reason for hiding this comment

Uh oh!

ayshih Mar 14, 2024

Choose a reason for hiding this comment

Uh oh!

mhvk Mar 12, 2024

Choose a reason for hiding this comment

Uh oh!

mhvk Mar 20, 2024

Choose a reason for hiding this comment

Uh oh!

ayshih commented Mar 13, 2024

Uh oh!

mhvk commented Mar 13, 2024

Uh oh!

ayshih commented Mar 14, 2024

Uh oh!

mhvk left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

mhvk Mar 14, 2024

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

mhvk left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

mhvk Mar 20, 2024

Choose a reason for hiding this comment

Uh oh!

neutrinoceros left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ayshih commented Jul 8, 2024

Uh oh!

neutrinoceros left a comment

Choose a reason for hiding this comment

Uh oh!

mhvk left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

ayshih commented Mar 12, 2024 •

edited

Loading