snorkelflow.studio.spacy_labeling_function

class snorkelflow.studio.spacy_labeling_function(name=None, resources=None)

Bases: resources_fn_labeling_function

Convenience decorator for spacy based labeling functions.

The decorated function must have an additional nlp argument which represents the spacy en-core-web-sm model.

Parameters:

• name (Optional[str], default: None) – Name of the LF

• resources (Optional[Mapping[str, Any]], default: None) – Labeling resources passed in to f via kwargs

Examples

>>> from snorkelflow.studio import spacy_labeling_function
>>> @spacy_labeling_function(name="my_spacy_lf")
... def many_entities(x, nlp) -> int:
...     return 0 if len(nlp(x.txt).ents) > 5 else -1

__init__(name=None, resources=None)

Methods

__init__([name, resources])