Skip to content

Remove explicitly setting frames per chunk to 1 - #661

Merged
yousefmoazzam merged 1 commit into
mainfrom
simplify-frames-per-chunk-mutation
Dec 10, 2025
Merged

yousefmoazzam merged 1 commit into
mainfrom
simplify-frames-per-chunk-mutation

Conversation

@yousefmoazzam

@yousefmoazzam yousefmoazzam commented Dec 8, 2025 •

Copy link
Copy Markdown
Collaborator

Fixes #659
Fixes IMGDA-674

Checklist

  • I have added tests that prove my fix is effective or that my feature works
  • New and existing unit tests pass locally with my changes
  • I have made corresponding changes to the documentation

@yousefmoazzam
yousefmoazzam force-pushed the simplify-frames-per-chunk-mutation branch from e6f5f77 to 3c51d04 Compare December 9, 2025 10:59
@yousefmoazzam
yousefmoazzam requested a review from dkazanc December 9, 2025 10:59
The frames per chunk need not be set to 1 unless the minimum block
length is 1. So, the `if` statement that remains, which sets the frames
per chunk to be the same as the minimum block length under the
appropriate condition, is sufficient to ensure a frames per chunk value
that will work across all blocks within a section.
@yousefmoazzam
yousefmoazzam force-pushed the simplify-frames-per-chunk-mutation branch from 3c51d04 to 871f40b Compare December 9, 2025 14:14
@yousefmoazzam

Copy link
Copy Markdown
Collaborator Author

It was discussed in sprint planning that this was good to go.

The IRIS CI is currently failing due to a lack of a release of httomolib that contains the new httomolib.misc.utils module. This issue is unrelated to these changes to the frames per chunk calculation, so I'm going to go ahead and merge anyway.

@yousefmoazzam
yousefmoazzam merged commit b9d1b30 into main Dec 10, 2025
5 of 6 checks passed
@yousefmoazzam
yousefmoazzam deleted the simplify-frames-per-chunk-mutation branch December 10, 2025 16:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Fragile intermediate data saver's frames-per-chunk calculation in multi-process runs can cause hanging

1 participant