Update LSM module to convert the raw `.dat` files to OME-Zarr

[kabilar]

Currently the linc-convert LSM module converts raw TIFF files to OME-Zarr. By June 2025, LSM images from 3 brain sections will be uploaded to lincbrain.org/MIT. These data will not be converted to TIFF first but will rather be uploaded as raw .dat files. So we will need to update the linc-convert LSM module to load in these raw .dat files.

For reference, see the linc-lsm-private repository for Emine's MATLAB scripts for loading and converting the .dat files to .tiff. Some minor details are added in these pull requests: https://github.com/lincbrain/linc-lsm-private/pull/1, https://github.com/lincbrain/linc-lsm-private/pull/2.

cc @emineozeen @aaronkanzer @balbasty @calvinchai

[calvinchai]

When using read_spool from the repo to read the raw scan data, the result is very different from reading from tiff. I guess there is some issue with the function. I am taking a deeper look into it.

[calvinchai]

for reference, this is what we have from tiff

[kabilar]

Thanks @calvinchai. Emine just found out that Alan Watson updated the read_spool.py script today. The new version is here. The tool will load the data into memory as a numpy array. Could you please try out this new version and let us know what you find?

[calvinchai]

this reader works fine! thanks. Only issue I see is that it does not crop 2 rows like the old version. I am not sure if we want to hardcode this cropping. I will bring this to the slack channel discussion.