theislab / spatialdata-db

BSD 3-Clause "New" or "Revised" License
1 stars 0 forks source link

Merscope Data/Reader mismatch #9

Open Farbodch opened 5 months ago

Farbodch commented 5 months ago

The merscope reader-function requires parquet-formatted data for the "cell_boundaries" input, but 16 out of the 17 merscope databases contain purely hdf5-formatted files in their cell_boundaries subfolder. For reference, the only merscope without this issue is "ush6cMERSCOPEHumanUterineCancerPatient2-RACostain", which does have a .parquet cell_boundaries file in the main directory. This db also happens to be the only one with the cell_boundaries data situated in its main directory and not in a subfolder in the main directory.

LucaMarconato commented 5 months ago

Please see some code to make it work here: https://github.com/scverse/spatialdata-io/issues/71.

It would be nice to have this functionality available in spatialdata-io, if any of you end up using the code from Can Ergen I can put you two in contact via Zulip and coordinate a joint PR.