Add a warning if all levels of a categorical response are the same

[gabrielodom]

If the user supplies a data frame with all the same label (for instance, all female), and attempts to run AES-PCA on this with the single-level factor as a response, two things will happen:

1. the code will take 2-3 times longer to execute (no idea why), and
2. all the p-values will be 0 (is this bad?)

Normally, this would not happen, as predicting the outcome of a variable that's always TRUE doesn't seem like a reasonable thing to do, so we should add a warning.