First version of debugging notebook

[lsloan]

[I tried to add Peter (@knoopum) as a reviewer for this pull request, but GH won't let me add him. I've invited him to join this repo. Maybe he'll be able to review it after he accepts the invite.]

As I began debugging a recent problem with kartograafr/Canvas communications, I remembered Peter said he wanted to use kartograafr code from within Jupyter Notebook for support purposes. So I began using JN for debugging this problem, thinking it could become the basis for a support notebook. The document in this PR is the result.

As I put the notebook together, I realized for this to work really well, the kartograafr modules need to work more like an API than they do now. You'll notice comments in the notebook about this. I also stopped writing in the notebook when I'd gotten far enough to solve the problem at hand. (Actually, I went a little further than necessary when I added the AGOL queries. I was following the main application flow in the notebook.)

Prerequisites

If you want to run this notebook on your local machine, you will need:

• Proper configuration for authentication with Canvas and AGOL.
• In main.py, the line logger = None needs to be commented out. Not sure why, because it runs in production with that value.

Reviewing

• GitHub shows the raw JSON of the notebook file by default. To view a rendered version of it, click the "view" button on the bar to the right of the filename.
• To ensure no private personal information about students is saved to the repo, all output is removed from the notebook before committing it to GH. It makes the notebook a little more boring, but safer. If you want to see sample output, you'll need to run it yourself.

[lsloan]

@knoopum: See your email for an invitation to join the kartograafr repository. Then you can review this pull request and leave comments about my work here.