def test_ngrams(self):
ngrams = generate_ngrams(2, self.corpus)
for first, second in [ngrams[i:i+2] for i in range(len(ngrams) - 2)]:
self.assertEquals(first.split()[-1], second.split()[0])
from glob import glob
from os.path import join as pjoin
from sys import argv, exit
from dump_speaker_data import extract_feature_utterances
from sample_size_test import generate_ngrams
if __name__ == "__main__":
try:
files = glob(pjoin(argv[1], "*.xml"))
except IndexError:
print "You must include the directory containing xml files"
exit()
utterances = extract_feature_utterances(files, "pos", "MOT")
bigrams = [generate_ngrams(2, ut) for ut in utterances if len(ut) > 1]
bigrams = reduce(list.__add__, bigrams, [])
poses = set([word for ut in utterances for word in ut])
print "# Utterances:", len(utterances)
print "# Bigrams:", len(bigrams)
print "# Bigrams Types", len(set(bigrams))
print "# of parts of speech tags:", len(poses)
"""
WORDS
# Utterances: 10445
# Bigrams: 36079
# Bigrams Types 10162
POS
def test_statistic_funs(self):
ngrams = generate_ngrams(3, self.corpus)
self.assertEqual(dice_stat(ngrams, ngrams), 1)
self.assertEqual(dice_stat(ngrams, []), 0)
