minor bug fixes to outlier detection

[elray1]

• add validation that x does not contain duplicate values to detect_outlr. This resolves the part of Issue validation for correct grouping structure #26 that relates to this function.
• prevent error when calling detect_outlr without having explicitly loaded the tibble package by using tibble::... when constructing a default argument value
• take more care when standardizing names of components returned by feasts::STL; hopefully will be more robust across versions of feasts (name of the seasonal column changed between the time I wrote this and now).
• prevent error in dplyr::case_when with disagreeing data types
• update outliers vignette so that the seasonal STL method uses seasonality; it got lost at some point
• add some (very) minimal unit tests for expected errors from detect_outlr

[ryantibs]

Looks good, thanks! I notice that you still are working in your old epitools repo clone. I will go ahead and merge.

Git is repointing this to the newly renamed epiprocess repo, but I think it would be safer (fewer potential problems downstream) if you just nuked and re-cloned epiprocess, assuming you don't have un-pushed work.