def run(self, articleText):
article = Article(articleId=articleText.articleTextId,
url=articleText.url,
publishDate=articleText.publishDate,
title=articleText.title,
text=articleText.text,
sentences=[])
for k, sentence in enumerate(self.spacyModel(articleText.text).sents):
tokenOffset = sentence[0].i
tokenCharOffset = sentence[0].idx
article.sentences.append(
Sentence(
startChar=sentence[0].idx,
endChar=sentence[-1].idx + len(str(sentence[-1])),
articleId=article.articleId,
sentenceId=k,
tokens=[
Token(i=token.i - tokenOffset,
startChar=token.idx - tokenCharOffset,
endChar=token.idx + len(str(token)) -
tokenCharOffset,
depType=token.dep_,
depHead=token.head.i - tokenOffset,
depLeftEdge=token.left_edge.i - tokenOffset,
depRightEdge=token.right_edge.i - tokenOffset,
posType=token.pos_,
posFineType=token.tag_,
lemma=token.lemma_,
text=token.text) for token in sentence
]))
self.output.append(article)
class TestSentence(TestCase):
def __init__(self, methodName):
super(TestSentence, self).__init__(methodName=methodName)
self.sentence = Sentence(startChar=0, endChar=0, articleId=0, sentenceId=0, tokens=None)
def setUp(self):
super().setUp()
def tearDown(self):
super().tearDown()
def test_sentence_test_get_hash_id(self):
self.assertEqual(self.sentence.test_get_hash_id(), True)
def test_sentence_test_to_data(self):
self.assertEqual(self.sentence.test_to_data(), True)
def test_sentence_test_from_data(self):
self.assertEqual(self.sentence.test_from_data(), True)
def get_outputs(sentences: List[pyDanticSentence]):
sentences = [
Sentence.from_data(json.loads(sentence.json()))
for sentence in sentences
]
tensorInputs = batchProcessor.process_input_batch(sentences)
tensorOutputs = model.get_outputs(**tensorInputs)
modelOutputs = batchProcessor.process_output_batch(tensorOutputs)
return [
pyDanticModelOutput.parse_raw(json.dumps(modelOutput.to_data()))
for modelOutput in modelOutputs
]
def __init__(self, methodName):
super(TestSentence, self).__init__(methodName=methodName)
self.sentence = Sentence(startChar=0,
endChar=0,
articleId=0,
sentenceId=0)
def from_data(cls, data):
data['publishDate'] = datetime.fromisoformat(data['publishDate'])
data['sentences'] = [Sentence.from_data(v) for v in data['sentences']]
return cls(**data)