BF: Ensure little-endian whatever the platform. trx should works on s…

[skoudoro]

fixes #83

This pull request introduces a better handling of endianness to ensure cross-platform compatibility when reading and writing TRX files. The changes add utility functions to enforce little-endian byte order, update all relevant file writing operations to use these utilities, and include extensive tests to verify correct behavior for various data types and scenarios.

• Added _get_dtype_little_endian and _ensure_little_endian utility functions in trx_file_memmap.py to convert data types and arrays to little-endian format, ensuring all data written to disk conforms to the TRX file specification for cross-platform compatibility.
• Updated all file writing operations in TrxFile.deepcopy and related functions to use _ensure_little_endian, ensuring that positions, offsets, and all per-vertex, per-streamline, and per-group data are saved in little-endian byte order. [1] [2] [3] [4] [5] [6]
• Modified the _create_memmap function to always use little-endian dtypes when mapping files, and updated the test for memmap creation accordingly. [1] [2]
• Updated trx/workflows.py to ensure all streamline and data arrays are written in little-endian order using the new utility functions, covering positions, offsets, and additional data arrays. [1] [2]
• Added comprehensive tests in test_memmap.py to verify correct dtype conversion, array onversion, and roundtrip data integrity for various data types and endianness scenarios.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (master@43edfbb). Learn more about missing BASE report.

Additional details and impacted files

@@            Coverage Diff            @@
##             master      #89   +/-   ##
=========================================
  Coverage          ?   59.76%           
=========================================
  Files             ?       12           
  Lines             ?     2411           
  Branches          ?        0           
=========================================
  Hits              ?     1441           
  Misses            ?      970           
  Partials          ?        0           

Flag	Coverage Δ
unittests	59.76% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.

[skoudoro]

I just tested on Fedora machine and it seems all good. see the screenshot below:

ready to be merged @arokem

[skoudoro]

Also, Can you accept the request for codecov @arokem or @frheault ?

Thank you

[arokem]

I believe I just approved the codecov request.

[skoudoro]

I believe I just approved the codecov request.

I confirm, and I do not have the permission to configure it.

So I let you do it and add the secret to make it work. Thanks @arokem

[frheault]

@skoudoro Do you have a branch on the Dipy repo using this PR content to ensure that these changes are not breaking outside of the repo. This could avoid surprises down the line.

[skoudoro]

Hi @frheault,

I tried with the last release, no issue. but let me connect to fedora and try with master branch

[arokem]

Given it works with the released DIPY, let's go ahead and merge this. We can keep an eye on this with future DIPY versions as well.