Florian M
Florian M
We should have a look end-to-end at what is causing the long layer names when uploading tiffs which are then converted by the worker convert_to_wkw job. I think wk moves...
### URL of deployed dev instance (used for testing): - https://___.webknossos.xyz ### Steps to test: - abc ### TODOs: - [x] evolution - [x] determine properties map from AnnotationWithTracings object...
Datasets not hosted on our site (nor our cloud storage) could theoretically still be read by the worker. Wk would need to supply the remote credentials to the worker.
As discussed in https://scm.slack.com/archives/C5AKLAV0B/p1764149494650819?thread_ts=1764071136.741679&cid=C5AKLAV0B The evolution should just take the value of the first matching user. We can check manually first if any users have ambiguous names and fix that...
Currently there is no way to change a user’s name. There should be a way to do that for the user themselves. We could also consider allowing it for organization...
The command palette already improves keyboard navigation a lot, by allowing to jump between pages. Once there, however, I almost always have to use the mouse again. It would be...
To account for accidental dataset deletions, e.g. from multi-delete, see #8817, a trash should be added that keeps the dataset for a certain number of days and allows restoring.
An OME-Zarr dataset was uploaded and the heuristics in processing + exploring after the dataset upload failed to interpret it correctly. Compare https://scm.slack.com/archives/CMBMU5684/p1762760956386899?thread_ts=1762444741.596309&cid=CMBMU5684 for more context and example data.
On instances with deletion disabled and uploadToPaths configured to local disk, it would be nice to offer a measure to prevent accidental deletion. When finishUploadToPaths is called, wk could run...
Could the upload logic be simplified by using the datasetId directly in place of the uploadId? I think it’s a 1:1 relation already.