Duplicate grade_ids

[tuncbkose]

Here is a proposal on how to deal with duplicate grade_ids during autograding. Feedback is very welcome. Note that three of these commits might be worth considering merging independent of this PR:

• 3c51d4d fixes an old testing bug unrelated to this PR (I can make a pr for this separately if requested).
• c28a52a changes CheckCellMetadata error message to add duplicate cells as a possible reason why the error is encountered. This might be desirable as it would allow users to notice the duplication problem during validation and try to fix it before submission.
• bac7c76 tweaks the error handling in the base converter.py to allow continuing with the other submitted notebooks (from the same student) in an assignment even if one of them is problematic. This way, for example, if an assignment had multiple notebooks and only one contained a duplicated cell, the rest would still get autograded.

Dealing with duplicate grade_ids

The main idea of this PR is to add a duplicate flag to cells that were observed to have duplicate grade_ids during autograding, so that they can be ignored and the rest of the notebook can be autograded. At the end of autograding, the duplicate flag is caught in order to raise an error and let the instructor know that some cells were ignored. These ignored cells are also marked with a comment for easy identification, however, the problem of the instructor editing the notebook remains unfixed (as it breaks the feedback workflow, #1202).

3b3c672 changes the order cells are processed by `DeduplicateIds´ with the idea that often if one is copy-pasting a cell, the pasted cell goes below the original one. This makes it more likely for the original to be more important for the grading, implying that we should be ignoring the cells lower in the notebook.

[github-actions]

👈 Launch a Binder on branch AaltoSciComp/nbgrader/duplicate-2