def test_verb_extraction_passive(mocker):
verb_dict = {
'pos_': ['VERB', 'VERB'],
'initial_value': 'was eaten',
'aux': False,
'replacement_value': 'was eaten',
'indexes': (2, 4),
'dep_': ['auxpass', 'ROOT'],
'active': True
}
verbs_extractor_instance = VerbsExtractor()
mocker.patch.object(textacy.extract,
"pos_regex_matches",
return_value=["test_element"])
mocker.patch.object(VerbsExtractor, "filtering_verbs")
mocker.patch.object(VerbsExtractor,
"clean_init_verb",
return_value=["", verb_dict])
mocker.patch.object(VerbsExtractor, "clean_aux_passive")
verbs_extractor_instance.verb_extraction([{
"doc": "doc_example",
"context": {
"obj_type": ""
}
}])
assert VerbsExtractor.filtering_verbs.call_count == 1
assert VerbsExtractor.clean_init_verb.call_count == 1
assert VerbsExtractor.clean_aux_passive.call_count == 1
def test_cut_verbs_empty():
doc = get_doc_example("He")
span = doc[2:4]
verbs_extractor_instance = VerbsExtractor()
cut_verbs_result = verbs_extractor_instance.cut_verbs({}, span, doc)
assert not cut_verbs_result
def test_cut_verbs_dict():
doc = get_doc_example("He is a well trained athlete.")
span = doc[3:5]
verbs_extractor_instance = VerbsExtractor()
cut_verbs_result = verbs_extractor_instance.cut_verbs({}, span, doc)
assert isinstance(cut_verbs_result, dict)
def test_extract(mocker):
verbs_extractor_instance = VerbsExtractor()
mocker.patch.object(VerbsExtractor, "verb_extraction")
text_information = mocker.patch.object(context_extractor,
'TextInformation')
verbs_extractor_instance.extract(text_information)
assert VerbsExtractor.verb_extraction.call_count == 1
def test_clean_init_verb():
pos_to_remove = ['CCONJ', 'PART']
doc = get_doc_example("He is a well trained athlete.")
span = doc[3:5]
verbs_extractor_instance = VerbsExtractor()
span_return, verb_dict_return = verbs_extractor_instance.clean_init_verb(
pos_to_remove, doc, span)
assert span_return.text == "well trained"
assert verb_dict_return["indexes"] == (3, 5)
def test_clean_aux_passive_true():
doc = get_doc_example("The food was eaten by the dogs.")
span = doc[2:3]
verb_dict = {
'pos_': ['VERB', 'VERB'],
'initial_value': 'was eaten',
'aux': False,
'replacement_value': 'was eaten',
'indexes': (2, 4),
'dep_': ['auxpass', 'ROOT'],
'active': True
}
verbs_extractor_instance = VerbsExtractor()
clean_aux_passive_result = verbs_extractor_instance.clean_aux_passive(
span, verb_dict)
assert not clean_aux_passive_result['active']
def test_filtering_verbs_not_cut_verbs(mocker):
verb_dict = {
'replacement_value': 'well trained',
'pos_': ['ADV', 'VERB'],
'aux': False,
'indexes': (3, 5),
'dep_': ['advmod', 'amod'],
'initial_value': 'well trained',
'active': True
}
verbs_extractor_instance = VerbsExtractor()
mocker.patch.object(VerbsExtractor, "cut_verbs")
filtering_verbs_result = verbs_extractor_instance.filtering_verbs(
verb_dict, [1], "")
assert VerbsExtractor.cut_verbs.call_count == 0
assert not filtering_verbs_result
def create_pipeline(self):
"""Create a pipeline with multiple steps, each step will be a class used to extract a part of the context
Returns
-------
list
a list of classes, each of them has a method called "extract" to extract the corresponding part of the context,
that method will be used to execute the list as a pipeline
"""
pipeline = [
VerbsExtractor(),
AgentsExtractor(),
IndirectObjectExtractor(),
AdverbModExtractor(),
DirectObjectExtractor(),
SubjectExtractor(),
ActionExtractor()
]
return pipeline