Update data_per_streamline capabilities

[skoudoro]

The goal of this PR is to allow inhomogeneous array for data_per_streamlines in a Tractogram.

I encounter this case when I wanted to save the labels of a cluster using Quickbundles from DIPY.

All saved centroids have a different number of streamlines, which make the array inhomogeneous.

data_per_streamline = {} data_per_streamline['clusters_labels'] = np.array([c.indices for c in clusters], dtype=object) data_per_streamline['clusters_sizes'] = clusters.clusters_sizes() 

See the following tutorial: https://docs.dipy.org/stable/examples_built/segmentation/segment_quickbundles.html

[codecov]

Codecov Report

Attention: Patch coverage is 84.61538% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 92.26%. Comparing base (0e925ab) to head (c907b45).

Files	Patch %	Lines
nibabel/streamlines/tractogram.py	84.61%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@ Coverage Diff @@ ## master #1291 +/- ## ========================================== - Coverage 92.27% 92.26% -0.01%  ========================================== Files 99 99 Lines 12460 12470 +10 Branches 2561 2565 +4 ========================================== + Hits 11497 11505 +8  - Misses 641 642 +1  - Partials 322 323 +1 

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

[effigies]

Is it possible to test this case? I think passing something like np.array([1, np.array([0, 1])], dtype=object)?

[effigies]

Hi Serge, thanks, and apologies for the slow response. I'm going to look into what's going on with the pre-release tests.

[skoudoro]

Hi @effigies,

Thank you for the review. no worries.

I will come back on this next week. What is the ETA of the next nibabel release?

[effigies]

ETA April. Might aim for late March or early May, as April will be busy for me.