Programmatic way to deal with greater numbers of documents and users

[saveli]

Is your feature request related to a problem? Please describe. I'm always frustrated when manually managing 20ish documents and users per semester I have to assign in a specific pattern, e.g.

• the workload matrix is huge and easy to confuse lines/columns when you have to scroll both vertically/horizontally
• bulk changes only count for rows, so selecting a couple documents and a couple users does affect all users, not just the selected ones
• creating and assigning each user as they change regularly (students in a course)
• managing/keeping track of what users do

Describe the solution you'd like Since the GUI is, in my opinion, optimised for few documents and users per project, it could stay that way, and projects with a huge number of documents/users and frequent changes should be managed using an API endpoint.

Describe alternatives you've considered I tried to look for existing APIs and found AERO and pycaprio. To my understanding, they lack the support for the aforementioned problems.

Additional context I'd be happy, if there are already existing other solutions.

[reckart]

Have you tried dynamic workload mode?

[reckart]

Oh, and have you had a look at the new functionality in the static workload mode to automatically pre-distribute documents to annotators based on a "x annotators per document" setting?

[saveli]

Sorry, forgot to mention this, I cannot use dynamic mode, since I have to assign in a specific (rule-based) pattern. This is also the root cause.

[reckart]

If you need a short-term solution, I would recommend you have a look at the database tables and write a script to inject your custom assignments there. The relevant table is annotation_document.