- added library documentation
- MLCitationMatcher by @mfilippo
- started to move away from brat standoff format as the default output
- removed the library
CRFPPas a dependency, and replaced with thesklearn- compatiblesklearn-crfsuite. - streamlined the installation process: the software is now installable via
pip(fixed bug #18)
- added support for CRFSuite, an implementation of CRF compatible with sklearn
- added test
tests/test_eval_ner.pyto run the evaluation of various ML models for the NER step
- substantial refactoring of
citation_extractor.ned.matchers.CitationMatcher - for the record: this was the version of the code used by M. Filipponi (@mfilippo)