Fix dragging annotation into a different orga’s dataset

[fm3]

When drag’n’dropping an annotation into a dataset in view-mode, the annotation is uploaded into that dataset, regardless of what dataset is reference din the NML tags.

This PR now makes this available if the organization name has changed as well.

URL of deployed dev instance (used for testing):

• https://uploadannotationorgaspecific.webknossos.xyz

Steps to test:

• Create annotation, download
• In the NML file, edit the organization name in the experiment tag to some non-existing orga, or remove it completely
• Uploading that annotation by dragging it into a dataset view mode should still work, and create an annotation for that dataset.

Issues:

• fixes Uploading Annotation into dataset should also include overwritingOrganizationName #7681

---

• Updated changelog
• Considered common edge cases

[MichaelBuessemeyer]

Thanks for this improvement 🙏

I only found one potential improvement of the changes. But as I am unsure whether my suggestion keeps the same functionality, feel free to ignore it and merge

[MichaelBuessemeyer]

Shouldn't the if-check check for overwritingOrganizationName being defined and not for overwritingDatasetName? Or is there some other place where it is ensured that in case overwritingDatasetName is defined, then overwritingOrganizationName is also defined?

And maybe the same syntax as in the line above can be used here as well, as imo the code is more readable this way. Or is there a semantic difference between my suggestion and your code?

Suggested change

	organizationName = if (overwritingDatasetName.isDefined) overwritingOrganizationName
	else parseOrganizationName(parameters \ "experiment")
	organizationName = overwritingOrganizationName.getOrElse(parseOrganizationName(parameters \ "experiment"))

[fm3]

Yes, there is in fact a difference. I want the overwritingOrganizationName, even if it is None, if the overwritingDatasetName is defined.

This case will probably not happen, because the frontend sends both or neither. But from the backend point of view, the organizationName from the NML becomes invalid once the datasetName is overwritten (this was already the case before this PR). So in that moment I want to yield the None, so that it will later be filled by the disambiguation logic via dataset name, and not the organizationName from the NML.

[MichaelBuessemeyer]

Please wait before merging: I am running into CSP errors in Chrome (version 125.0.6422.60) & Ubuntu:

Let's tackle this in this PR as well, as the zip uploading in an annotation does not work on the master as well (with the chrome version)

[MichaelBuessemeyer]

Found a CSP fix 🎉