medcat.ner.vocab_based_annotator
I would just ignore this whole class, it’s just a lot of rules that work nicely for CDB once the software is trained the main thing are the context vectors.
Module Contents
Functions
|
Given a name it will check should it be annotated based on config rules. If yes |
Attributes
- medcat.ner.vocab_based_annotator.logger
- medcat.ner.vocab_based_annotator.maybe_annotate_name(name, tkns, doc, cdb, config, label='concept')
Given a name it will check should it be annotated based on config rules. If yes the annotation will be added to the doc._.ents array.
- Parameters:
name (str) – The name found in the text of the document.
tkns (List[spacy.tokens.Token]) – Tokens that belong to this name in the spacy document.
doc (Doc) – Spacy document to be annotated with named entities.
cdb (CDB) – Concept database.
config (Config) – Global config for medcat.
label (str) – Label for this name (usually concept if we are using a vocab based approach).
- Returns:
Optional[Span] – The Span, if relevant.
- Return type:
Optional[spacy.tokens.Span]