import spacy
# load a spacy language model
nlp = spacy.load("en_core_web_sm")
# add a custom component to language model
def custom_component(doc):
# do something with document
return doc
nlp.add_pipe(custom_component, name="custom")
# create flow for NLP
with Flow("Natural Language Processing") as flow:
# create a spaCy doc from text, the equivalent of nlp('This is some text')
doc = SpacyNLP(text="This is some text", nlp=nlp)
# extract default components from language model pipeline
tagger = SpacyTagger(nlp)
parser = SpacyParser(nlp)
ner = SpacyNER(nlp)
# extract a custom component from language model pipeline by name
custom_pipeline_component = SpacyComponent("custom")
flow.run()
def test_nlp_model_provided(self):
task = SpacyParser()
with pytest.raises(ValueError) as exc:
task.run()
assert "A spaCy pipeline must be provided" == str(exc.value)
def test_initialization(self):
task = SpacyParser()
assert task.nlp is None
def test_get_parser(self):
mock = Mock()
mock.parser = "parser"
task = SpacyParser(nlp=mock)
parser = task.run()
assert parser == "parser"
def test_nlp_model_provided(self):
task = SpacyParser()
with pytest.raises(ValueError,
match="A spaCy pipeline must be provided"):
task.run()