Relion >= 3.1 offsets convention: _rlnOriginX(Y)Angst #1302

j-c-c · 2025-07-14T18:54:14Z

Adopt Relion's newer offsets metadata fields: #1118

adopt _rlnOriginX(Y)Angst in favor of _rlnOriginX(Y)
Still read in both conventions
Internally use new convention in metadata
Will save off in new convention

codecov · 2025-07-14T19:14:18Z

Codecov Report

❌ Patch coverage is 97.75281% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.65%. Comparing base (04948e6) to head (2cea2fc).
⚠️ Report is 45 commits behind head on develop.

Files with missing lines	Patch %	Lines
src/aspire/downloader/data_fetcher.py	0.00%	2 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1302      +/-   ##
===========================================
+ Coverage    90.61%   90.65%   +0.04%     
===========================================
  Files          133      133              
  Lines        14389    14433      +44     
===========================================
+ Hits         13038    13084      +46     
+ Misses        1351     1349       -2

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

garrettwrong · 2025-08-04T11:47:42Z

What's the update here? Thanks

j-c-c · 2025-08-06T18:11:46Z

What's the update here? Thanks

The update is I felt very unsatisfied with this implementation. It changes the behavior of our ImageSource classes and involved adapting a lot of tests. I've opened a PR (#1317) that maintains current behavior of our ImageSource classes and just updates the metadata convention for offsets. Going to leave both PRs until we have a chance to discuss which is preferred.

garrettwrong · 2025-09-25T15:04:38Z

Cool. This is still a fair amount of changes, but a lot of it is test code, so I think we can get it in this release if it can be made ready soon. Thanks

… fix pixel_size bug. Fix broken tests.

…el_size for CoordinateSource/ArrayImageSource. Fix downsample offsets.

garrettwrong

Sorry, almost there, one issue, otherwise minor tweaks and this'll be good enough to push forward. This source stuff gets dicey fast.

Thanks!

src/aspire/source/image.py

src/aspire/utils/misc.py

src/aspire/source/image.py

garrettwrong · 2025-10-08T13:18:26Z

tests/test_relion_source.py

-        rln_src_32.offsets, sim.offsets, rtol=0, atol=atol, strict=True
-    )
+    np.testing.assert_allclose(rln_src_64.offsets, sim.offsets, rtol=0, atol=atol)
+    np.testing.assert_allclose(rln_src_32.offsets, sim.offsets, rtol=0, atol=atol)


cool thanks

…oherent between source and metadata.

j-c-c · 2025-10-08T20:09:20Z

@garrettwrong Ok, sorry that took until end of day. Here you go.

src/aspire/utils/misc.py

garrettwrong

Great, thanks!

janden

Sorry for the delay on this.

Looks great! Just a few things.

janden · 2025-10-17T09:23:33Z

src/aspire/image/xform.py

            centered_fft=self.centered_fft,
        )

+        return Image(im_ds, pixel_size=im.pixel_size).stack_reshape(


So we are bypassing the standard Image.downsample method in order to keep the pixel_size the same. Why?

By the time this generation pipeline is kicked off by requesting images the pixel_size has already been adjusted as an image attribute and in metadata in the ImageSource.downsample method. If we use Image.downsample here the pixel size will get adjusted twice.

src/aspire/source/coordinates.py

src/aspire/source/image.py

src/aspire/utils/misc.py

janden · 2025-10-17T09:56:02Z

src/aspire/utils/misc.py

+    if not close:
+        warnings.warn(
+            f"User provided pixel_size: {reference_pixel_size} angstrom, does not match"
+            f" pixel_size found in metadata: {pixel_sizes} angstrom. Setting"


This is a little confusing, since I don't think we always check the return value and act on it. Should this be updated?

j-c-c self-assigned this Jul 14, 2025

j-c-c added cleanup extern Relating to external changes labels Jul 14, 2025

j-c-c mentioned this pull request Aug 6, 2025

Angstrom offsets #1317

Closed

j-c-c mentioned this pull request Aug 7, 2025

Downsample pixel_size bug #1318

Closed

j-c-c force-pushed the rlnIriginAngst branch from a7af69f to e0687f5 Compare September 16, 2025 14:20

j-c-c changed the base branch from develop to pxsz September 16, 2025 15:05

garrettwrong force-pushed the pxsz branch from 113a619 to 89b7388 Compare September 24, 2025 16:52

Base automatically changed from pxsz to develop September 24, 2025 18:05

j-c-c force-pushed the rlnIriginAngst branch from b279d25 to 0182f7c Compare September 25, 2025 14:43

j-c-c added 16 commits September 29, 2025 14:37

Simulation pixel_size default=1.0. Private/public Image.downsample to…

9e5eb62

… fix pixel_size bug. Fix broken tests.

Fix gallery pixel_size

3a2a851

Adopt Relion 3.1 offset convention (_rlnOriginX(Y)Angst). Default pix…

7688e41

…el_size for CoordinateSource/ArrayImageSource. Fix downsample offsets.

Require pixel_size for all ImageSource objects

5737bd6

tox

94a556b

cleanup Simulation pixel_size behavior. Fix docstring. Add test.

fc0a059

Resolve Errors and failing tests

2588e71

Fix galleries

b4fc52c

fix param in gallery

e001c01

more rebase cleanup

137ddd4

Update Simulation pixel_size behavior. Warn on mismatch.

f740d88

_populate_pixel_size. Handle mismatches. tests

6ee4315

Fix failing tests

8849024

pixel_size setter/getter.

2d00e25

add offset tests

099bb38

test src.downsample.pixel_size == src.img.downsample.pixel_size

73fc3a8

garrettwrong mentioned this pull request Oct 7, 2025

Metadata Dtypes #1328

Open

j-c-c added 2 commits October 7, 2025 11:51

offsets stored in doubles. pixels stored as float.

8deee0b

remove strict

f89a5b5

j-c-c requested a review from garrettwrong October 8, 2025 12:32

garrettwrong requested changes Oct 8, 2025

View reviewed changes

j-c-c added 4 commits October 8, 2025 13:02

pixel_size logic

3dee8c3

update check_pixel_size

a9d2d41

cast pixel_size as float correctly. Add test to check pixel_size is c…

fdeec0f

…oherent between source and metadata.

check for scalar pixel_size. Add test.

341d44e

j-c-c requested a review from garrettwrong October 8, 2025 20:09

garrettwrong requested changes Oct 9, 2025

View reviewed changes

src/aspire/utils/misc.py Outdated Show resolved Hide resolved

change argument names

a3c6f4a

j-c-c requested a review from garrettwrong October 9, 2025 12:46

garrettwrong previously approved these changes Oct 9, 2025

View reviewed changes

j-c-c marked this pull request as ready for review October 9, 2025 12:57

j-c-c requested a review from janden as a code owner October 9, 2025 12:57

j-c-c mentioned this pull request Oct 9, 2025

ImageSource.save() does not write Relion >=3.1 starfiles #1329

Open

janden requested changes Oct 17, 2025

View reviewed changes

j-c-c added 2 commits October 17, 2025 08:34

Remove None default for CoordinateSource pixel_size

f58eab7

typos

1bab70c

janden previously approved these changes Oct 17, 2025

View reviewed changes

j-c-c added 2 commits October 17, 2025 10:33

clarifying pixel_size comment in Downsample xform

4680fc0

update check_pixel_size warning message. fix broken test.

7d10379

j-c-c dismissed stale reviews from janden and garrettwrong via 7d10379 October 17, 2025 14:52

j-c-c added 2 commits October 17, 2025 11:19

fix gallery

339fef0

update extract_particles CLI

2cea2fc

j-c-c merged commit 38069c0 into develop Oct 17, 2025
35 checks passed

j-c-c mentioned this pull request Oct 17, 2025

Relion now uses rlnOriginX(Y)Angstrom instead of rlnOriginX(Y) #1118

Closed

Relion >= 3.1 offsets convention: _rlnOriginX(Y)Angst #1302

Relion >= 3.1 offsets convention: _rlnOriginX(Y)Angst #1302

Uh oh!

Conversation

j-c-c commented Jul 14, 2025

Uh oh!

codecov bot commented Jul 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

garrettwrong commented Aug 4, 2025

Uh oh!

j-c-c commented Aug 6, 2025

Uh oh!

garrettwrong commented Sep 25, 2025

Uh oh!

garrettwrong left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

garrettwrong Oct 8, 2025

Choose a reason for hiding this comment

Uh oh!

j-c-c commented Oct 8, 2025

Uh oh!

Uh oh!

garrettwrong left a comment

Choose a reason for hiding this comment

Uh oh!

janden left a comment

Choose a reason for hiding this comment

Uh oh!

janden Oct 17, 2025

Choose a reason for hiding this comment

Uh oh!

j-c-c Oct 17, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

janden Oct 17, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

codecov bot commented Jul 14, 2025 •

edited

Loading