Description:

spanextructure 0.1.1

Span Extructure

You might think the name is mispelled but it ain't. It is a word play on spaCy's Span, extract and structure. span_exctructure is a spaCy component that builds upon SpanRuler and regex to extract structured information, e.g. dates, amounts with currency and multipliers etc.
Installation
pip install span_extructure

Usage
import spacy

nlp = spacy.blank("en")

# Optionally add config if varying from default values
config = {
"overwrite": False, # default: False
"rules": [
{
"patterns": [[{"SHAPE": "dd.dd.dddd"}]],
"extruct": r"(?P<day>[0-3]\d).(?P<month>0[1-9]|1[0-2]).(?P<year>20[0-5]\d|19\d\d)",
"label": "DATE",
}
]
}
nlp.add_pipe("span_extructure", config=config)

doc = nlp("This date 21.04.1986 will be a DATE entity while the structured information will be extracted to `Span._.extructure`")
for e in doc.ents:
print(f"{e.text}\t{e.label_}\t{e._.extructure}")

>>> 21.04.1986 DATE {'day': '21', 'month': '04', 'year': '1986'}

License

For personal and professional use. You cannot resell or redistribute these repositories in their original state.

Share

• Share on Facebook
• Share on Twitter