Add examples and counter_examples field to regex filter entries

[tazlin]

With the growing list of regex filter entries, it would benefit moderation (particularly from a regression testing and documentation point of view) to have two extra fields:

• examples
  • A (comma delimited?) string of positive examples that the regex should catch.
• counter_examples
  • A (comma delimited?) string of negative examples that the regex should not catch, such as substrings of larger words that should not count for detection. (See https://en.wikipedia.org/wiki/Scunthorpe_problem for a treatise on the underlying reason this field is needed).