Initial RFC-3 support

[melissalinkert]

Companion to glencoesoftware/bioformats2raw#330.

Draft for now, as this still needs tests. I kind of suspect there is a dimension ordering issue (in particular with FLIM-ModuloAlongC.ome.tiff), but needs more investigation to determine which PR needs work (if either).

[melissalinkert]

I kind of suspect there is a dimension ordering issue (in particular with FLIM-ModuloAlongC.ome.tiff), but needs more investigation to determine which PR needs work (if either).

glencoesoftware/bioformats2raw@29867c6 is expected to fix the ordering issue, so this PR just needs some tests.

[melissalinkert]

Tests added in 506a0fe depend upon the current state of glencoesoftware/bioformats2raw#330. I tested this locally by checking out the rfc-3 branch that corresponds to the bioformats2raw PR, then running ./gradlew clean build publishToMavenLocal. Build is expected to fail here, but taking out of draft since I have no more planned changes.

Proposed next steps:

• review Report Modulo dimensions as separate axes (RFC-3) bioformats2raw#330 and merge once we're happy with it (using changes here to assist with testing if needed)
• RC of bioformats2raw
• update this PR to consume RC (build should then pass)
• review this PR and merge once we're happy
• RC of raw2ometiff

[melissalinkert]

Tests should now cover the case in glencoesoftware/bioformats2raw#330 (comment), and are passing locally. The build failure is still expected as this is dependent upon the current state of glencoesoftware/bioformats2raw#330 being installed locally.

[sbesson]

I applied the latest changes of this PR to the set of OME-Zarr samples generated as per glencoesoftware/bioformats2raw#330 (comment).

I encountered similar failures for the ICS and OIR samples

+ ./raw2ometiff-0.11.0-SNAPSHOT/bin/raw2ometiff /Users/sbesson/Desktop/RFC-3/ome-zarr/0.9.dev1/default/ics/20240612_training_test-tiles_01_stitched.zarr /Users/sbesson/Desktop/RFC-3/ome-tiff/raw2ometiff/0.9.dev1-default/ics/20240612_training_test-tiles_01_stitched.ome.tiff -p
Exception in thread "main" picocli.CommandLine$ExecutionException: Error while calling command (com.glencoesoftware.pyramid.PyramidFromDirectoryWriter@41294f8): java.lang.RuntimeException: loci.formats.FormatException: Mismatch on dimension C; expected 1, got 4

+ ./raw2ometiff-0.11.0-SNAPSHOT/bin/raw2ometiff /Users/sbesson/Desktop/RFC-3/ome-zarr/0.9.dev1/default/oir/DAPI_mCherry_4T_5Lambda-420-630-w10nm-s50nm.zarr /Users/sbesson/Desktop/RFC-3/ome-tiff/raw2ometiff/0.9.dev1-default/oir/DAPI_mCherry_4T_5Lambda-420-630-w10nm-s50nm.ome.tiff -p
Exception in thread "main" picocli.CommandLine$ExecutionException: Error while calling command (com.glencoesoftware.pyramid.PyramidFromDirectoryWriter@41294f8): java.lang.RuntimeException: loci.formats.FormatException: Mismatch on dimension C; expected 1, got 5

I suspect we need to decide on how to resolve glencoesoftware/bioformats2raw#330 (comment) in order to unblock this

[melissalinkert]

With 648f08c and the current state of glencoesoftware/bioformats2raw#330, I was able to successfully convert the .ics data without error (I haven't tried the .oir one yet). I was initially very concerned that data was being converted incorrectly when looking at the output OME-TIFF (since large areas are blank in many planes), but comparing the original .ics and converted OME-TIFF plane-by-plane in ImageJ confirmed that this is expected.