FormatROD: support TY5 compression

[biochem-fan]

This also adds support for gzip/bzipped files, non-unity gain and binning.
This solves issue #878.

@dagewa What is the right order to add tests? Necessary images are in the following PRs:

• Add two images written by CrysAlisPro dials/data-files#71
• Add two images written by CrysAlisPro dials/data#534

Once test images are merged, I can add concrete tests (e.g. testing the gain). But if we merge test data before this dxtbx PR, the existing "all imports work" test will fail and block other commits.

[dagewa]

I will merge the image data first. I don't think the "all imports work" test is run by CI any more. We might get a failure on the DLS Jenkins job, but if you add the concrete tests soon this isn't a problem.

[biochem-fan]

@dagewa Thank you very much for merging test images.
I confirmed that this works:

dials_data("image_examples") / "Hypix6000-monoclinic_lysozyme1_1_1.rodhypix.bz2"

I am writing my own tests in tests/format. How can I actually execute it, ideally only it, not others? The document mentions libtbx.pytest but my development build does not have it.

[dagewa]

If you have built DIALS using its own bootstrap.py (in installer/) then you should be able to run

pytest --regression tests/format/test_module.py::test_name

[biochem-fan]

@dagewa Thanks.

I added tests. In my local computer,

pytest --regression tests/format/test_FormatROD.py -v

passed. However, test runs on GitHub failed to find the newly added test images.

This is strange, because both PRs into data-files and data have been merged.

dials-data in PyPI has been automatically updated to 2.4.140 yesterday. However, one in conda-forge remains 2.4.139. The GitHub action used the old one (probably from conda-forge).

+ echo DIALS_DATA_VERSION_FULL=version.full=2.4.139

https://github.com/cctbx/dxtbx/actions/runs/22379175084/job/64775933987?pr=879#step:8:24

How do I update them?

[dagewa]

The conda-forge package updated now. We might need to force it to use that version, by setting

dials-data>=2.4.140

in the files in .conda-envs/

[biochem-fan]

@dagewa Thanks! Tests are running now. If nothing fails, this PR is ready to be merged.

[dagewa]

Thanks for adding this. I tested with a TY6 data set to make sure I could still read the images properly and confirmed it was fine.