Test with a real HCS OME-TIFF dataset

[ziw-liu]

          I'm looking at this now, I can make you a small test dataset

EDIT: Check if this fits the bill: cm.mantis/2023_07_21_mm_hsc_dataset/mm_hsc_dataset_2

Originally posted by @ieivanov in #159 (comment)

[ziw-liu]

@ieivanov thanks for making this! I think this is a nice example to test against. One very minor point is that it would be nice to have non-single digit column numbers (e.g. 12) to prevent future regressions.

[ieivanov]

Yup, I can make that happen. I'll have to re-acquire this dataset anyway as we saw that the data is of the wrong file size due to a MM bug somewhere (micro-manager/micro-manager#1746)

[ieivanov]

One wrinkle we missed yesterday - an important feature of the iohub readers is adding singleton dimensions where they may be missing. In the test data we should include acquisitions with fewer than 5 dimensions to test that iohub correctly turns those into TCZYX datasets

[ziw-liu]

In the test data we should include acquisitions with fewer than 5 dimensions to test that iohub correctly turns those into TCZYX datasets

This is already tested with the older (2020-gamma) MM datasets. I don't think we need to acquire the full suite of the combinations unless we aim to replace the old ones completely.

[ieivanov]

Ah, right, OK!