OMERO Perkin Elmer Operetta Import CLI vs Insight

Dear all,
I tried to import a Perkin Elmer Operetta Screening into Omero Server. My preferred option is an inplace import with ln_s. The Operetta Export contains a xml file as well as all the individual tif files from the screen. When I try to import only the xml file to avoid duplicates from command line using:

omero import --transfer=ln_s '/pathTo/Index.idx.xml' --skip upgrade

I get an error:

[      main] ERROR     ome.formats.importer.cli.ErrorHandler - UNREADABLE_FILE: /pathTo/Images/Index.idx.xml

When I alternatively use the Insight client and select the xml file, I get the complete screen imported correctly.
Is there anything I am doing wrong?
Thanks for your help,
Alex

Hi Alex,

A few questions:

  • Are insight and the cli running on the same machine?
  • If yes, are they running as the same user?
  • If not, is /pathTo mounted the same way on both?

UNREADABLE_FILE points most likely to a pure access issue rather than anything specific about the application or the file type.

~Josh

1 Like

Hi Josh,
thanks for the answer.
So the CLI is run via ssh terminal directly on the server using the recommended inplace import user. The insight is run on a client Mac with the real “owner”/user of the file. The path is identical and pointing to the same file on a file server. I checked that the file is readable from the ssh terminal for both the omero system user as well as for the inplace user. I imported successfully using the inplace user for the same user using the CLI e.g. tif files or lif files.
An other question: Can I also use the CLI to attach files (e.g. results tables) and leave them at their original location on the file server? If yo I would appreciate a small example.
Thanks
Alex

Hi Josh,
I could solve the problem, there was a space in the pathTo, that I overlocked, so everything is working as expected.
One more question, Can a screen be placed inside a project or dataset?
Thanks Alex

whew. Glad to hear it!

No, it cannot. The two hierarchies are independent of one another.

~Josh

Hi Josh,
thanks for the answer. I think I made a mistake. I meant a plate. Can a plat ebe imported into a dataset or project?
Interestingly when I use the bulk import method and specify for a high content screening plate a target in the form of “Dataset:name” in the bulk import file Omero performs the import and ends up with all the images stored in the Dataset defined in the bulk import file and additionally the Plates on the top level. Here is a small example:
bulk.csv

Dataset:name:Fig3_d,Fig3d_WD_MaSat488_60_1,pathToscreen index file

results in the following import:
Image
Is this still ok?
Can I somehow further structure this, so that the plate would belong to a project folder for the given manuscript?
Thanks Alex

Also no. :slight_smile: Sorry.

Interesting. That’s a surprise. It’s certainly permissible, though there may be some odd edge cases if you try to move or delete the images from their dataset location.

Unfortunately not.

~Josh

Sorry, I dropped the ball on one of your questions. To attach a file with the CLI this snippet would work:

IMAGE=$(omero import a.fake)
FILE=$(omero upload a.fake)
ANN=$(omero obj new FileAnnotation file=$FILE
omero obj new ImageAnnotationLink parent=$IMAGE child=$ANN

but it does copy the file. https://github.com/ome/omero-upload provides a function for doing this without uploading the file, but it will need some scripting to work in your case. You can see an example at https://github.com/IDR/idr0052-walther-condensinmap/blob/master/scripts/upload_and_create_rois.py

All the best,
~Josh