Data hosted on cellpainting-gallery s3
can be browsed at https://idr.github.io/idr0125-way-cellpainting/.
Use scripts/s3_sync.py
to generate aws s3 sync
commands to download all OME-NGFF plates as metadata ONLY (no zarr chunks)
Import of metadata-only NGFF plates requires ZarrReader with --depth=100
:
$ omero import --depth=100 --bulk screenA/idr0125-screenA-bulk.yml --file /tmp/idr0125.log --errs /tmp/idr0125.err
Need to view an image from each plate OR omero render set Image:ID rdef.yml
for each plate, to avoid subsequent ResourceError after symlinking. See scripts/render_set_cmd.py
Run scripts/get_import_paths.py
to create imported_paths.txt
with the managed repo path for each imported plate.
Run scripts/symlink_cmd.py
which consumes imported_paths.txt
to output the commands to replace each plate in the managed repo with a symlink to the corresponding plate on a mounted s3 bucket. These can be saved in scripts/symlinks.bash
. Need to run sudo chmod +x symlinks.bash
then sudo -u omero-server -s
and bash symlinks.bash
.
Run omero render set Plate:ID
for every plate. NB: takes ~12hrs per plate? TBD: split this job between several parallel processes.