edsnlp.pipelines.core.terminology.factory

DEFAULT_CONFIG = dict(terms=None, regex=None, attr='TEXT', ignore_excluded=False, term_matcher='exact', term_matcher_config={}) module-attribute

create_component(nlp, name, label, terms, attr, regex, ignore_excluded, term_matcher, term_matcher_config)

Source code in edsnlp/pipelines/core/terminology/factory.py

@Language.factory(
    "eds.terminology",
    default_config=DEFAULT_CONFIG,
    assigns=["doc.ents", "doc.spans"],
)
def create_component(
    nlp: Language,
    name: str,
    label: str,
    terms: Optional[Dict[str, Union[str, List[str]]]],
    attr: Union[str, Dict[str, str]],
    regex: Optional[Dict[str, Union[str, List[str]]]],
    ignore_excluded: bool,
    term_matcher: TerminologyTermMatcher,
    term_matcher_config: Dict[str, Any],
):
    assert not (terms is None and regex is None)

    if terms is None:
        terms = dict()
    if regex is None:
        regex = dict()

    return TerminologyMatcher(
        nlp,
        label=label,
        terms=terms,
        attr=attr,
        regex=regex,
        ignore_excluded=ignore_excluded,
        term_matcher=term_matcher,
        term_matcher_config=term_matcher_config,
    )