Thursday, 2024/02/15
10:00 am ET
Terrell Russell, Kory Draughn, Josh Moore (GerBI / OME), Joost Willemse (Leiden University), Ander Astudillo (SURF), Alice Stuart-Lee (SURF), Guillaume Gay (France Bioimaging), Marjolijn Mertz (NKI - Netherlands Cancer Institute), Nirav Merchant (CyVerse / UArizona), Rolf Harkes (NKI - Netherlands Cancer Institute)
- Serialization of OMERO to iRODS via PRC
- AA) Save (As)
- Josh added target data object location for saving into iRODS namespace
- No longer just named 'transfer.tar'
- Leaves naming decision to the human
- They have to know where they can save a datafile
- Overwriting will happen today without first checking the target
- What does error handling look like? Bad path / bad permissions?
- Josh - any errors will be shown in the OMERO UI
- Currently hard-coded user/pass combo to talk to iRODS
- OMERO could hold an annotation for username
- But nowhere to get passwords at this time
- OIDC would be favorite if possible
- Only OMERO-web viable for now
- Other clients don't know how to do this fancy web stuff (yet)
- Inital work/success will just be 'service account'
- Assume iRODS admins will handle the 'archive' side
- Mapping file of usernames generated by admins
- iRODS rodsadmin session can proxy as the correct user
- Current OMERO LDAP integration uses OMERO email/username OR an LDAP query
- mapping to iRODS could be similar, local or external
- https://azure.github.io/Avere/legacy/ops_guide/4_6/html/username_map_file.html
- Josh added target data object location for saving into iRODS namespace
- BB) Load
- Could/Should just be a flag on the same script to read/load?
- Should it be able to overwrite in OMERO?
- Would have to be a delete flag? Maybe doesn't make sense.
- Would always upload/create a new ID in OMERO
- AA) Save (As)
- Looking for guinea pigs
- If you'd like to archive OMERO information into elsewhere… this script provides a means to do that
- Nirav - has a use case, NIH data management plans - if people ask for access to your data, this could provide a means to bundle a study to share with another OMERO user or elsewhere. The storage system can be exposed for sharing without exposing the OMERO instance to others.
- This is a vote for Save and Load
- Joost - so this would not reuse the same ID in OMERO? … correct
- What happens if OMERO data is deleted?
- The archive/saved version is independent / decoupled
- What happens if OMERO data is deleted?
- Alice - Dutch Bioimaging also interested in testing this
- Guillaume - If you delete the raw-data in OMERO the annotations and thumbnails stay so you could have a way to keep a 'trace' of archived data (out of scope for this group)
- Consortium
- Check for overwrite/existence before sending/saving
- Confirm errors are sent back to OMERO
- Consider bidirectional save/load in the same script
- Think about mapping file
- What happens when no user found…
- Next Meeting
- March 2024