def load_model(filename='model_cnn_', size='sm', path='/models', type_= '.bin'):
root_path = utils.get_path_to_project_dir(os.getcwd())
full_path = root_path + path + '/' + filename + size + type_
if (os.path.isfile(full_path)):
nlp = utils.read_obj(filename + size, path='/models', type_=type_)
else:
nlp = spacy.load("pt_core_news_" + size)
utils.write_obj(nlp, filename + size, path='/models',type_=type_)
return nlp
# get df_testemunhas by doc
df = utils.get_df_testemunhas_by_doc(list_docs_testemunhas)
utils.write_obj(df,
'df_excerpts',
path=self.data_interim,
type_=self.type)
# get df with union of testemunhas by doc
df_ = utils.df_union_testemunhas(df)
utils.write_obj(df_,
'df_witnesses',
path=self.data_interim,
type_=self.type)
return df_
if __name__ == "__main__":
pred = Predictor(size='sm', type_='.bin', sample_size=6)
pred.predict()
print(f'tipo do obj:{type(pred)}')
# print df_excerpts
df_ = utils.read_obj('df_excerpts', path='data/interim', type_='.bin')
print(df_.shape)
print(df_)
# print df_witnesses
df_wit = utils.read_obj('df_witnesses', path='data/interim', type_='.bin')
print(df_wit.shape)
print(df_wit)
def test_read_obj():
lista = ['a', 'b']
utils.write_obj(lista, "lista")
lista2 = utils.read_obj("lista")
assert lista == lista2
def test_predict():
pred = Predictor(sample_size=6)
df_wit = pred.predict()
df_test2 = utils.read_obj('df_witnesses', '/test/core', type_='.bin')
assert df_wit.columns == df_test2.columns
def test_load_model():
pred = Predictor(sample_size=6)
ml = pred.load_model()
ml2 = utils.read_obj('model_cnn_sm', '/test/core', type_='.bin')
assert type(ml) == type(ml2)