from pathlib import Path
import os
if __name__ == '__main__':
opts = docopt(__doc__)
# load model
filename = opts['-i']
f = open(filename, 'rb')
model = pickle.load(f)
f.close()
# load evaluation corpus
corpus = opts['-c']
if opts['--final']:
reader = InterTASSReader(
corpus, res_filename="InterTASS/ES/TASS2017_T1_test_res.qrel")
else:
reader = InterTASSReader(corpus)
X, y_true = list(reader.X()), list(reader.y())
# normalize
#X = model.normalize(X)
# classify
y_pred = model.predict(X)
# evaluate and print
evaluator = Evaluator()
evaluator.evaluate(y_true, y_pred)
evaluator.print_results()
stats.py [options] -f <folder>
stats.py -h | --help
Options:
-f <folder> Folder containing TASS corpora.
-h --help Show this screen.
"""
from docopt import docopt
from os import path
from sentiment.tass import InterTASSReader
from collections import Counter
if __name__ == '__main__':
opts = docopt(__doc__)
# load corpora
filenames = [
f'intertass-{location}-train-tagged.xml'
for location in ['ES', 'CR', 'PE']
]
folder = opts['-f']
for filename in filenames:
filepath = path.join(folder, filename)
reader = InterTASSReader(filepath)
y = list(reader.y())
print('***')
print(filename)
print('tweets:', len(y))
print(Counter(y))
from sentiment.evaluator import Evaluator
from sentiment.tass import InterTASSReader
if __name__ == '__main__':
opts = docopt(__doc__)
# load model
filename = opts['-i']
f = open(filename, 'rb')
model = pickle.load(f)
f.close()
# load corpus
if not opts['--final']:
reader = InterTASSReader('TASS/InterTASS/TASS2017_T1_development.xml')
else:
reader = InterTASSReader('TASS/InterTASS/TASS2017_T1_test.xml',
'TASS/InterTASS/TASS2017_T1_test_res.qrel')
X, y_true = list(reader.X()), list(reader.y())
# classify
y_pred = model.predict(X)
# evaluate and print
evaluator = Evaluator()
evaluator.evaluate(y_true, y_pred)
evaluator.print_results()
evaluator.print_confusion_matrix()
# detailed confusion matrix, for result analysis
from sentiment.evaluator import Evaluator
from sentiment.tass import InterTASSReader
if __name__ == '__main__':
opts = docopt(__doc__)
# load model
filename = opts['-i']
f = open(filename, 'rb')
model = pickle.load(f)
f.close()
# load evaluation corpus
corpus = opts['-c']
final = opts.get('-f', None)
reader = InterTASSReader(corpus, final)
X, y_true = list(reader.X()), list(reader.y())
# classify
y_pred = model.predict(X)
# evaluate and print
evaluator = Evaluator()
evaluator.evaluate(y_true, y_pred)
evaluator.print_results()
evaluator.print_confusion_matrix()
# detailed confusion matrix, for result analysis
cm_items = defaultdict(list)
for i, (true, pred) in enumerate(zip(y_true, y_pred)):
def count_tweets(path):
reader = InterTASSReader(path)
dist = Counter(reader.y())
return dict(dist), sum(dist.values())
from sentiment.tass import InterTASSReader
from sentiment.baselines import MostFrequent
from sentiment.classifier import SentimentClassifier
models = {
'basemf': MostFrequent,
'clf': SentimentClassifier,
}
if __name__ == '__main__':
opts = docopt(__doc__)
# load corpora
corpus = opts['-i']
reader = InterTASSReader(corpus)
X, y = list(reader.X()), list(reader.y())
development = opts['-d']
reader = InterTASSReader(development)
X_dev, y_dev = list(reader.X()), list(reader.y())
# train model
model_type = opts['-m']
if model_type == 'clf':
model = models[model_type](clf=opts['-c'])
else:
model = models[model_type]() # baseline
model.fit(X, y, X_dev, y_dev)
# save model
from sentiment.baselines import MostFrequent
from sentiment.classifier import SentimentClassifier
models = {
'basemf': MostFrequent,
'clf': SentimentClassifier,
}
if __name__ == '__main__':
opts = docopt(__doc__)
# load corpora
corpus = opts['-i']
reader = InterTASSReader(corpus)
X, y = list(reader.X()), list(reader.y())
# train model
model_type = opts['-m']
if model_type == 'clf':
model = models[model_type](clf=opts['-c'])
else:
model = models[model_type]() # baseline
if opts['--train']:
#print("lalolanda")
model.fit(X, y)
else:
#print('otra cosa')
model.cross_validation(X, y)
import pickle
from sentiment.tass import InterTASSReader, GeneralTASSReader
from sentiment.baselines import MostFrequent
from sentiment.classifier import SentimentClassifier
models = {
'basemf': MostFrequent,
'clf': SentimentClassifier,
}
if __name__ == '__main__':
opts = docopt(__doc__)
# load corpora
reader1 = InterTASSReader('TASS/InterTASS/tw_faces4tassTrain1000rc.xml')
X1, y1 = list(reader1.X()), list(reader1.y())
reader2 = GeneralTASSReader(
'TASS/GeneralTASS/general-tweets-train-tagged.xml', simple=True)
X2, y2 = list(reader2.X()), list(reader2.y())
X, y = X1 + X2, y1 + y2
# train model
model_type = opts['-m']
if model_type == 'clf':
model = models[model_type](clf=opts['-c'])
else:
model = models[model_type]() # baseline
model.fit(X, y)
matplotlib.use('TkAgg')
import matplotlib.pyplot as plt
models = {
'basemf': MostFrequent,
'clf': SentimentClassifier,
}
if __name__ == '__main__':
opts = docopt(__doc__)
# load training corpus
reader1 = InterTASSReader(cfg.tweets['InterTASS']['train']['path'])
X1, y1 = list(reader1.X()), list(reader1.y())
reader2 = GeneralTASSReader(
cfg.tweets['GeneralTASS']['train']['path'], simple=True)
X2, y2 = list(reader2.X()), list(reader2.y())
X, y = X1 + X2, y1 + y2
# load development corpus (for evaluation)
reader = InterTASSReader(cfg.tweets['InterTASS']['development']['path'])
Xdev, y_true = list(reader.X()), list(reader.y())
# load model if given
if opts['-i']:
filename = opts['-i']
f = open(filename, 'rb')