Severity Text - Option to apply standard processing

[djaglowski]

Building upon #181, will users have expectations about exactly how severity text is captured? in other words, should users be able to apply standard processing rules to the text that is retained? For example:

severity:
  parse_from: some_field
  text_action: none # "none" is default, but also accepts "lower", "upper", "title", "known_only", etc

none - preserve severity text exactly
lower - lowercase severity text
upper - uppercase severity text
known_only - if matches a known level, use that text's predefined level string. otherwise, use "unknown" or similar