Enrollment Automation

[smartspot2]

Thread for discussion/ideas/designs for enrollment automation. (Link to the current project board).

[smartspot2]

Some brief notes/ideas (from 11/15, and some of my thoughts):

Coordinators:

• Should be able to upload a list of mentors
  • Likely as a part of the coordinator interface
  • An issue here is that we'd need to modify the mentor model to also include the course; currently, the source of truth for the mentor's course is from mentor.section.course, which poses an issue if we were to create new unassigned mentors.
  • To avoid rewriting queries, we could leave mentor.section.course as a property that just looks at the mentor's course field, but this would cost more database queries.
• See when-to-meet kind of UI to give available section times
  • As a first POC, we could just have a manual input of times
• Coord puts list of times, sends to the backend, and makes mentor forms available to fill out
  • Maybe also have a way to check which mentors have filled out their preferences form, and see the current submitted data?
• Backend gives back various options from backend from LP
  • There'd be a few options for the LP costs; we can have an option where we aim for a uniform distribution, or for a more approx normal distribution, or some other configurable option by the coordinator
• Coord picks one output distribution, with the ability to perhaps manually edit the resulting data
• After confirming the assignments (through a list view or some other UI), have a button to create new sections for each of the mentors

Mentors:

• After time slots are confirmed, mentors should be able to access a form that allows them to rank their preferred time slots
  • Preferably also in calendar form, but for a POC we could just have some kind of dropdown
  • How many slots should we limit mentors to choose preferences for? Or should we not have any limits?

UI:

• Where should these forms/UIs be placed? Likely just in a new tab to the right of the current Resources tab, only viewable by mentors and coordinators?

LP:

• There is a past repo that does a similar kind of LP that we can take inspiration from (or steal): https://github.com/csmberkeley/matcher
  • This would add several more python dependencies though (networkx, numpy, and we'll likely not use pandas dataframes)