Add checks for logistic regression assumptions

[craig-parylo]

There are six assumptions underlying logistic regression:

1. The outcome is binary,
2. Observations are independent,
3. No multicollinearity among predictors,
4. No extreme outliers,
5. Linear relationship between predictors and logic of outcome,
6. Sample size must be sufficiently large.

https://r4ds.github.io/bookclub-islr/addendum---logistic-regression-assumptions.html

It would be helpful if there were automated checks of the input glm model to ensure it passes these tests (where applicable).