LLM Autograding manual graded questions

[Victorsss-Orz]

This is a part of the SIIP grant for Automatic Short Answer Grading using Large Language Models. The first PR works on the real basic functions, and more features will be added later.
We added a grade button in the manual grading assessment question page that utilizes large language models to autograde all problems, and update the scores and feedbacks.

[codecov]

Codecov Report

Attention: Patch coverage is 18.48341% with 172 lines in your changes missing coverage. Please review.

Project coverage is 69.01%. Comparing base (dfa97d9) to head (83e9b51).

❗ Current head 83e9b51 differs from pull request most recent head 100d658

Please upload reports for the commit 100d658 to get more accurate results.

Files	Patch %	Lines
apps/prairielearn/src/lib/bot-grading.ts	16.12%	156 Missing ⚠️
...alGrading/assessmentQuestion/assessmentQuestion.ts	15.78%	16 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #9887      +/-   ##
==========================================
- Coverage   69.23%   69.01%   -0.23%     
==========================================
  Files         484      486       +2     
  Lines       76262    76380     +118     
  Branches     6447     6439       -8     
==========================================
- Hits        52799    52710      -89     
- Misses      23012    23219     +207     
  Partials      451      451              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jonatanschroeder]

This is by no means a complete or formal review, I'm only providing some initial thoughts.

[SethPoulsen]

I've been working closely with Victor throughout this PR, so most of the comments and changes, I've just shared with him directly offline.

There is obviously a lot more work to do in order to make these features work well, but I think this represents a good MVP for something that works minimally enough so that we should merge it and make additional changes incrementally.

[SethPoulsen]

@Victorsss-Orz it looks like you have conflicts with the yarn.lock file. One way to fix this would be to re-check out the version of yarn.lock from the master branch, and then re-run your yarn add openai command. There might be a faster way than that, I don't know though.

[Victorsss-Orz]

@Victorsss-Orz it looks like you have conflicts with the yarn.lock file. One way to fix this would be to re-check out the version of yarn.lock from the master branch, and then re-run your yarn add openai command. There might be a faster way than that, I don't know though.

Done

[nwalters512]

A few more comments, then this should be good to merge.

[Victorsss-Orz]

Looking forward to getting this merged