negspacy

Version 1.0 is a major version update providing support for spaCy 3.0's new interface for adding pipeline components. As a result, it is not backwards compatible with previous versions of negspacy.

If your project uses spaCy 2.3.5 or earlier, you will need to use version 0.1.9. See See archived readme.

• Big thanks to @lcreteig for pull request #20, which allows a user to update the extension name, thus being able to use multiple instances of negspacy in a pipeline:

negex = Negex(nlp, extension_name="myownname")

• Small modifications to make it easier to users to add or remove patterns on the fly:

nlp = spacy.load("en_core_web_sm")
negex = Negex(nlp)
patterns = negex.get_patterns()
negex.add_patterns(
        pseudo_negations=["my favorite pattern"],
        termination=["these are", "great patterns"],
        preceding_negations=["wow a negation"],
        following_negations=["extra negation"],
    )
negex.remove_patterns(termination = "but", preceding_negations=["wow a negation"])

• fixes issue #11
• modify termsets to distinguish generic and clinical phrases, issue #10
• confirm support for spaCy 2.2.4

Not publishing to pypi

Adds ability for user to consider first tokens of noun chunks as negation terms.

• #7 Bug fix to fix issue in spacy 2.1.8 where using an entity ruler and not finding any entities would cause doc.is_nered to return false. This would make negspacy believe named entity recognition was not performed, throwing an error, incorrectly.
• moved patterns to termset lookup file, making it easier to support multiple languages. #6
• add additional patterns (contractions)

• Users can specify negation terminology
• Improved error handling