def __init__(self, lexicon_name='', *args, **kwargs):
super().__init__(*args, **kwargs)
self.lexicon_name = lexicon_name
model0 = emb.get_custom0()
self.embedding = emb.get_custom3(initial_model=model0,
lexicon_name=self.lexicon_name,
suffix='model0_' + self.lexicon_name)
def main():
parser = argparse.ArgumentParser(
description='Build inequalities.',
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('output', help='Output file')
args = parser.parse_args()
model = emb.get_custom0()
lexicon = read_bing_liu(res.bing_liu_lexicon_path)
print('Save inequalities to %s', args.output)
feat.find_ineq(model, lexicon, args.output)
def load_resources(self):
super().load_resources()
self.word2vec = emb.get_custom0(word2vec_param=self.word2vec_param)
# anew_french = reader.read_anew(res.anew_french_lexicon_path)
# lexicons = [('lidilem_adjectifs', lidilem_adjectifs_lexicon),
# ('lidilem_noms', lidilem_noms_lexicon),
# ('lidilem_verbes', lidilem_verbes_lexicon),
# ('blogoscopie', blogoscopie_lexicon)]
lexicons = [('lidilem', lidilem_lexicon),
('blogoscopie', blogoscopie_lexicon)]
for (lex_name, lex) in lexicons:
utils.remove_multi_words_in_lexicon(lex)
model_name = 'model0'
logger.info('Processing %s', model_name)
model = emb.get_custom0(train_path=train_path)
for (lex_test_name, lex_test) in lexicons:
logger.info('Compare %s on %s', model_name, lex_test_name)
emb.compare_model_with_lexicon(model, lex_test)
emb.compare_model_with_lexicon_class(model, lex_test)
model_name = 'model1'
logger.info('Processing %s', model_name)
for lex_train_name, lex_train in lexicons:
logger.info('Training %s with %s', model_name, lex_train_name)
model = emb.get_custom1(train_path=train_path, lexicon=lex_train, suffix=lex_train_name)
for lex_test_name, lex_test in lexicons:
# Skip test lexicon if it's the same as the training lexicon
if lex_test_name == lex_train_name:
continue
logger.info('Compare %s on %s', model_name, lex_test_name)
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.embedding = emb.get_custom0()
# -*- coding: utf-8 -*-
import logging
import numpy as np
import embeddings as emb
import resources as res
import reader
import utils
logger = logging.getLogger(__name__)
models = []
models.append(('model0', emb.get_custom0()))
models.append(('model1', emb.get_custom1()))
models.append(('modelGnews', emb.get_gnews()))
model0 = emb.get_custom0()
# model1 = emb.get_custom1()
modelGnews = emb.get_gnews()
topn = 1000
sample_size = 5000
bing_liu_lexicon = reader.read_bing_liu(res.bing_liu_lexicon_path)
nrc_emotion_lexicon = reader.read_nrc_emotion(res.nrc_emotion_lexicon_path)
nrc_emotions_lexicon = reader.read_nrc_emotions(res.nrc_emotion_lexicon_path)
lexicons = [('bing_liu_lexicon', bing_liu_lexicon),
('nrc_emotion_lexicon', nrc_emotion_lexicon),
