V4.2.0 - Final release of GUM series 4
Final release of GUM series 4:
- Added s_type="multiple" for sentences containing multiple types (previously under "other")
- Standardized some @rend from "italics" to always "italic"
- Standardized hyphens/dashes in number ranges to have POS tag 'TO' (e.g. in ranges of years), matching the syntactic analysis
- Changed some inconsistent POS tags for IPA name pronunciation from FW to NP
- Added better imperative mood labeling to CoreNLP UD morph features based on manual s_type annotations
- Removed spurious spans in RST files and fixed some segmentations not conforming to guidelines
- Numerous assorted error corrections