pub fn parse_ecb_plus_zip(data: &[u8]) -> Result<Vec<CorefDocument>>Expand description
Parse ECB+ from the real XML zip archive (ECB+.zip).
The zip contains files like ECB+/{topic}/{topic}_{doc}.xml.
Each XML file has:
<token t_id="N" sentence="S" number="P">text</token>elements<Markables>section with mention elements (ACTION_, HUMAN_, etc.)<Relations>section withCROSS_DOC_COREFlinking mentions to clusters