def test_semantic_role_labeling(self):
predictor = pretrained.srl_with_elmo_luheng_2018()
sentence = "If you liked the music we were playing last night, you will absolutely love what we're playing tomorrow!"
result = predictor.predict_json({"sentence": sentence})
assert result["words"] == [
"If", "you", "liked", "the", "music", "we", "were", "playing", "last", "night", ",",
"you", "will", "absolutely", "love", "what", "we", "'re", "playing", "tomorrow", "!"
]
assert result["verbs"] == [
{"verb": "liked",
"description": "If [ARG0: you] [V: liked] [ARG1: the music we were playing last night] , you will absolutely love what we 're playing tomorrow !",
"tags": ["O", "B-ARG0", "B-V", "B-ARG1", "I-ARG1", "I-ARG1", "I-ARG1", "I-ARG1", "I-ARG1", "I-ARG1", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O"]},
{"verb": "were",
"description": "If you liked the music we [V: were] playing last night , you will absolutely love what we 're playing tomorrow !",
"tags": ["O", "O", "O", "O", "O", "O", "B-V", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O"]},
{"verb": "playing",
"description": "If you liked [ARG1: the music] [ARG0: we] were [V: playing] [ARGM-TMP: last] night , you will absolutely love what we 're playing tomorrow !",
"tags": ["O", "O", "O", "B-ARG1", "I-ARG1", "B-ARG0", "O", "B-V", "B-ARGM-TMP", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O"]},
{"verb": "will",
"description": "If you liked the music we were playing last night , you [V: will] absolutely love what we 're playing tomorrow !",
"tags": ["O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "B-V", "O", "O", "O", "O", "O", "O", "O", "O"]},
{"verb": "love",
"description": "[ARGM-ADV: If you liked the music we were playing last night] , [ARG0: you] [ARGM-MOD: will] [ARGM-ADV: absolutely] [V: love] [ARG1: what we 're playing tomorrow] !",
"tags": ["B-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "I-ARGM-ADV", "O", "B-ARG0", "B-ARGM-MOD", "B-ARGM-ADV", "B-V", "B-ARG1", "I-ARG1", "I-ARG1", "I-ARG1", "I-ARG1", "O"]},
{"verb": "'re",
"description": "If you liked the music we were playing last night , you will absolutely love what we [V: 're] playing tomorrow !",
"tags": ["O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "B-V", "O", "O", "O"]},
{"verb": "playing",
"description": "If you liked the music we were playing last night , you will absolutely love [ARG1: what] [ARG0: we] 're [V: playing] [ARGM-TMP: tomorrow] !",
"tags": ["O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "B-ARG1", "B-ARG0", "O", "B-V", "B-ARGM-TMP", "O"]}
]
def _get_predictor():
return [
pretrained.srl_with_elmo_luheng_2018(),
biaffine_parser_stanford_dependencies_todzat_2017(),
named_entity_recognition_with_elmo_peters_2018(),
lemma(),
pretrained.neural_coreference_resolution_lee_2017(),
fine_grained_named_entity_recognition_with_elmo_peters_2018()
]
def main(args):
if args.tagger == "srl":
model = srl_with_elmo_luheng_2018()
tagger_function = model.predict
elif args.tagger == "oie":
model = open_information_extraction_stanovsky_2018()
tagger_function = lambda sentence: model.predict_json(
{"sentence": sentence})
elif args.tagger == "dep":
model = biaffine_parser_stanford_dependencies_todzat_2017()
tagger_function = lambda sentence: get_verb_info_from_graph(
get_nx_graph_from_dep(model.predict(sentence)))
else:
raise RuntimeError(f"Unknown tagger type: {args.tagger}")
if args.include_coref:
coref_model = neural_coreference_resolution_lee_2017()
else:
coref_model = None
if args.verbose:
logging.basicConfig(level=logging.INFO)
make_files_for_semparse(args.data_files_path, args.output_path,
tagger_function, coref_model)