Model report - used to find issues in models or questions

[johnjosephhorton]

I created a report that is generated by iterating through all models and administers a survey consisting of all of our example questions.

model_report.txt

This can be created by running:

make model-report

When you are running locally, if there are exceptions that are not QuestionAnswerValidation exceptions, it will generate a report.

Uses

1. What models cannot/do not return a generated token string at all?
2. What models / questions have responses that cannot be repaired (but probably could be) with smarter 'fixes'?
3. What models do we not have prices for?

For 1 - if something seems 'wrong' with the model, we should probably remove it from available models For 2 - What might a fix look like? You can see how a question answer gets fixed in the edsl/questions/QuestionX.py For 3 - Can we add that model to the spreadsheet?