search

mobie / mobie.github.io

1 stars 3 forks source link

ome-zarr mobie samples #98

Open constantinpape opened 1 year ago

constantinpape commented 1 year ago

Here is an overview of current MoBIE projects, which have ome.zarr data and samples:

Spatial Transcriptomics Data

https://github.com/mobie/mouse-embryo-spatial-transcriptomics-project

This contains 40 or so positions, each with image data (stored as ome.zarr on the ebi-s3) and transcriptomics data stored as csv. The data for the first position is available here:

The image data has 4 channels.

Covid HTM Data

https://github.com/mobie/covid-if-project

This contains data from a HTM screen for covid antibodies and contains a few hundred image positions. Here are the addresses for one of the positions:

Platynereis Data

The raw data and cell segmentation from https://github.com/mobie/platybrowser-project is also available as ome.zarr (the data for the project itself is stored as n5):

will-moore commented 1 year ago

Hi @constantinpape - I tried to open a few of those Covid HTM samples in ome-ngff-validator, and in neuroglancer but I'm getting CORs errors. E.g. https://ome.github.io/ome-ngff-validator/?source=https://s3.embl.de/covid-if-project/20200406_164555_328/images/ome-zarr/nuclei_WellE06_0000.ome.zarr The Spatial Transcriptomics are working 👍

constantinpape commented 1 year ago

I tried to open a few of those Covid HTM samples in ome-ngff-validator, and in neuroglancer but I'm getting CORs errors.

Yes, some option is not enabled by default on the EMBL S3 and we have run into this issue before. @tischi do you remember how this was fixed by Josep? I think you also solved this with him recently.

tischi commented 1 year ago

Which bucket? What is the exact error message? I think I have to ask Josep to fix this.

will-moore commented 1 year ago

https://ome.github.io/ome-ngff-validator/?source=https://s3.embl.de/covid-if-project/20200406_164555_328/images/ome-zarr/nuclei_WellE06_0000.ome.zarr fails to load https://s3.embl.de/covid-if-project/20200406_164555_328/images/ome-zarr/nuclei_WellE06_0000.ome.zarr/.zattrs

Access to fetch at 'https://s3.embl.de/covid-if-project/20200406_164555_328/images/ome-zarr/nuclei_WellE06_0000.ome.zarr/.zattrs' from origin 'https://ome.github.io' has been blocked by CORS policy: No 'Access-Control-Allow-Origin' header is present on the requested resource.

Same for https://s3.embl.de/i2k-2020/platy-raw.ome.zarr/.zattrs

tischi commented 1 year ago

@will-moore could you please try again?

will-moore commented 1 year ago

Looks good 👍 thanks. However, the validator is reporting Dimension separator must be / for version 0.4. https://ome.github.io/ome-ngff-validator/?source=https://s3.embl.de/covid-if-project/20200406_164555_328/images/ome-zarr/nuclei_WellE06_0000.ome.zarr I believe this will be relaxed in future versions (https://github.com/ome/ngff/pull/161) but currently that won't be read by ome-zarr-py (see fix in https://github.com/ome/ome-zarr-py/pull/235)

tischi commented 1 year ago

Wow, I am surprised that this data is so ancient that we still have . separator...

tischi commented 1 year ago

@constantinpape once you have more time again, maybe good to double check this.

constantinpape commented 1 year ago

Yes, I can confirm: 

$ mc ls embl/covid-if-project/20200406_164555_328/images/ome-zarr/nuclei_WellE06_0000.ome.zarr/s0
[2021-11-19 14:35:32 CET]   271B STANDARD .zarray
[2021-11-19 14:35:32 CET]    31B STANDARD .zattrs
[2021-11-19 14:35:32 CET] 856KiB STANDARD 0.0
[2021-11-19 14:35:32 CET] 859KiB STANDARD 0.1
[2021-11-19 14:35:32 CET] 705KiB STANDARD 1.0
[2021-11-19 14:35:32 CET] 700KiB STANDARD 1.1

I think that stems from a bug when creating this data that has since been fixed. (And technically the current data is not valid; it specifies format version 0.4, but uses .). -> This should be fixed and it's in principle easy to write something that does that on the fly, but I don't have time for that right now, maybe next week.

titansmc commented 1 year ago

The CORS error is gone now, we had one migration done and I have dissallowed the ome.github.io domain. It is back now. Cheers.