Fix #263, #417, #430 and #457.

[dzhuang]

#263 #417: (Bulk) Regrade/recacluate resulted in wrong gradechange state (percentage) when use_latest #430: Reopening latest session if use_latest should result in NONE, instead of the grades of that session or the grades of the second latest session. #457: Failure to stringify graded gradeChanges with state None.

• [ ] Naming of newly introduced field effective_time in GradeChange
• [ ] In terms of backward compatibility, is the new migration right?
• [x] More tests to cover changes

cc @inducer

[codecov[bot]]

Codecov Report

Merging #466 into master will increase coverage by 0.05%. The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #466      +/-   ##
==========================================
+ Coverage   96.36%   96.41%   +0.05%     
==========================================
  Files          45       45              
  Lines       11046    11100      +54     
  Branches     2055     2069      +14     
==========================================
+ Hits        10644    10702      +58     
+ Misses        342      340       -2     
+ Partials       60       58       -2

Impacted Files	Coverage Δ
course/grades.py	97.12% <ø> (+0.29%)	:arrow_up:
course/receivers.py	100% <100%> (ø)	:arrow_up:
course/flow.py	99.2% <100%> (+0.01%)	:arrow_up:
course/models.py	100% <100%> (+0.26%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update d0c89e4...58b3421. Read the comment docs.

[dzhuang]

@inducer Can you review this? Thanks. The two TODOs need your verfication, thanks.

[dzhuang]

@inducer Can you look at this PR and make sure the migration is correct? I wish to integrate this into my instance soon, new semester is coming. Thanks!

[dzhuang]

For record:

This might result in inconsistency when grade changes are ordered improperly.

grades accumulation strategy: use_latest A: grade time 2015 B: grade time 2016, flow completed 2014 C: grade time 2017 D: grade time 2018, flow completed 2013 Then A < B < C

But also D < B and C < D, which is a direct contradiction.

C's grade percentage should be the overall grade.

[dzhuang]

@inducer What about this? I wish to put the exempt issue later, since it's only debug-related (currently).

[inducer]

This is complicated (and terrifying) to review because of the long-term consequences. I just didn't have the time today. I'll try hard to get to it tomorrow night. Sorry!

[dzhuang]

NVM and Thanks alot.

[dzhuang]

Sorry for pushing you here @inducer. It seems I'll have to merge this into my instance tomorrow, if no progress is made here.

[dzhuang]

In cases when reopening a session, https://github.com/inducer/relate/blob/855801d463120e0fc74de08466e5d4ac39bac46d/course/flow.py#L510 ~~and https://github.com/inducer/relate/blob/855801d463120e0fc74de08466e5d4ac39bac46d/course/flow.py#L567~~ should also be adjusted.

Update: This seems won't be affected.

[dzhuang]

@inducer I've updated the code, looking forward to your review some day (not urging you to do it now, but hope that it will be merged someday).