You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Note: this is an RFC with a suggested enhancement, it is not 100% we want this yet. More consideration is needed before starting implementation.
It would be useful to annotate certain elements of a report in order to come back to them later. For example: with programming exercises throughout the semester, one similar pair is often not enough to build a case. By flagging these submissions, teachers can go back to a report later to be reminded of what elements contributed to the suspicion.
How it would look like
I think we would like two types of annotations:
Tags (e.g. suspicious, AI-generated, innocent, ...) that allow easy searching, flagging and grouping of submissions
Initially I think we could keep this per submissions. One could also tag a pair by tagging one (or both) submissions of that pair.
There should also be support for bulk tagging using the tables. This would allow tagging a whole cluster as "plagiarism" at once.
Annotations on submissions. These would preferably support markdown to allow for proper formatting.
Annotating a whole submission
Annotating a specific line (or section) on a single submission
Annotating lines (or sections) in a pair of two submissions (e.g. "This function is copied")
In addition it would be useful to have an overview of all annotations with a quick link to the submission, pair or cluster.
Changes needed
This requires the annotations to be persistent in a way. As it might also be beneficial to also export these annotations elsewhere, we would need to expose this through some kind of interface.
API endpoints need to be added (/reports/{id}/annotations and /reports/{id}/tags) to create, read and edit annotations.
The UI needs to be updated to be able to add these annotations and send them to the API
Specifically the annotations on pairs should be designed with care: showing this properly feels like a challenge
The data tables should be able to search for tags, as well as search for, remove and bulk edit tags.
These features should only be enabled when running in server mode (not when running standalone)
Note: this is an RFC with a suggested enhancement, it is not 100% we want this yet. More consideration is needed before starting implementation.
It would be useful to annotate certain elements of a report in order to come back to them later. For example: with programming exercises throughout the semester, one similar pair is often not enough to build a case. By flagging these submissions, teachers can go back to a report later to be reminded of what elements contributed to the suspicion.
How it would look like
I think we would like two types of annotations:
suspicious
,AI-generated
,innocent
, ...) that allow easy searching, flagging and grouping of submissionsIn addition it would be useful to have an overview of all annotations with a quick link to the submission, pair or cluster.
Changes needed
This requires the annotations to be persistent in a way. As it might also be beneficial to also export these annotations elsewhere, we would need to expose this through some kind of interface.
/reports/{id}/annotations
and/reports/{id}/tags
) to create, read and edit annotations.Related issue: #552
The text was updated successfully, but these errors were encountered: