def test_dev(self):
fname = 'tests/mockdata/dev.phn'
corpus1 = tokenlists_to_corpus(load_annotation(fname))
corpus2 = load_corpus_txt(fname)
assert (corpus1 == corpus2)
def _load_corpus(fname):
return load_corpus_txt(fname)
def _load_corpus(fname):
return load_corpus_txt(fname)
with open('tests/mockdata/dev.phn.annotation.pkl', 'rb') as fid:
corpus = pickle.load(fid)
assert (corpus == load_annotation('tests/mockdata/dev.phn'))
def test_badline(self):
with pytest.raises(ReadError):
load_annotation('tests/mockdata/mockcorpus_small_badline.phn')
def test_badinterval(self):
with pytest.raises(ReadError):
load_annotation('tests/mockdata/mockcorpus_small_badinterval.phn')
def test_badfloat(self):
with pytest.raises(ReadError):
load_annotation('tests/mockdata/mockcorpus_small_badfloat.phn')
class TestCorpusAnnotationFromPhoneFile(object):
def test_dev(self):
fname = 'tests/mockdata/dev.phn'
corpus1 = tokenlists_to_corpus(load_annotation(fname))
corpus2 = load_corpus_txt(fname)
assert (corpus1 == corpus2)
if __name__ == '__main__':
fname = 'tests/mockdata/mockcorpus_small.phn'
corpus1 = tokenlists_to_corpus(load_annotation(fname))
corpus2 = load_corpus_txt(fname)
assert (corpus1 == corpus2)
corpus = args['trs'][1]
# if corpus is "other", change resource_dir to get the transcriptions/vad
phn_corpus_file = path.join(resource_dir, '{}.phn'.format(corpus))
wrd_corpus_file = path.join(resource_dir, '{}.wrd'.format(corpus))
vad_file = path.join(resource_dir, '{}.vad'.format(corpus))
print vad_file
if verbose:
print banner('LOADING FILES')
# load gold phones and gold words
with verb_print(' loading word corpus file',
verbose, True, True, True):
wrd_corpus = load_corpus_txt(wrd_corpus_file)
with verb_print(' loading phone corpus file',
verbose, True, True, True):
phn_corpus = load_corpus_txt(phn_corpus_file)
# load across and withing folds
with verb_print(' loading folds cross',
verbose, True, True, True):
#fragments_cross = load_split(folds_cross_file,
# multiple=False)
intervals_vad = [load_split(vad_file,
multiple=False)]
# get list of file names from vad:
# names = load_names(vad_file)
try:
def test_collapse():
is1 = [Interval(0, 1), Interval(2,3),
Interval(4, 5)]
assert(collapse(is1) == is1)
is2 = [Interval(0, 1), Interval(1,2)]
assert(collapse(is2) == [Interval(0, 2)])
is3 = [Interval(0, 1), Interval(2, 3),
Interval(0, 1.5)]
assert(collapse(is3) == [Interval(0, 1.5),
Interval(2,3)])
is4 = [Interval(0, 1), Interval(2, 3),
Interval(0, 10)]
assert(collapse(is4) == [Interval(0.,10)])
corpus = load_corpus_txt('tests/mockdata/tiny.phn')
disc_clsdict = load_classes_txt('tests/mockdata/tiny.classes', corpus)
gold_clsdict = load_classes_txt('tests/mockdata/tiny.classes', corpus)
def test_cover():
assert(cover(disc_clsdict) == 8.0)
assert(cover(gold_clsdict) == 8.0)
def test_coverage():
assert(coverage(disc_clsdict, gold_clsdict) == 1.0)
def test_ued():
assert(ued('kitten', 'sitting') == 3)
def test_ned():
assert(ned('kitten', 'sitting') == 3/7)
def test_dev(self):
fname = 'tests/mockdata/dev.phn'
corpus1 = tokenlists_to_corpus(load_annotation(fname))
corpus2 = load_corpus_txt(fname)
assert (corpus1 == corpus2)
assert (corpus == load_annotation('tests/mockdata/dev.phn'))
def test_badline(self):
with pytest.raises(ReadError):
load_annotation('tests/mockdata/mockcorpus_small_badline.phn')
def test_badinterval(self):
with pytest.raises(ReadError):
load_annotation('tests/mockdata/mockcorpus_small_badinterval.phn')
def test_badfloat(self):
with pytest.raises(ReadError):
load_annotation('tests/mockdata/mockcorpus_small_badfloat.phn')
class TestCorpusAnnotationFromPhoneFile(object):
def test_dev(self):
fname = 'tests/mockdata/dev.phn'
corpus1 = tokenlists_to_corpus(load_annotation(fname))
corpus2 = load_corpus_txt(fname)
assert (corpus1 == corpus2)
if __name__ == '__main__':
fname = 'tests/mockdata/mockcorpus_small.phn'
corpus1 = tokenlists_to_corpus(load_annotation(fname))
corpus2 = load_corpus_txt(fname)
assert (corpus1 == corpus2)
from tde.measures.nlp import collapse, cover, coverage, ued, ned, NED
from tde.util.reader import load_corpus_txt, load_classes_txt
def test_collapse():
is1 = [Interval(0, 1), Interval(2, 3), Interval(4, 5)]
assert (collapse(is1) == is1)
is2 = [Interval(0, 1), Interval(1, 2)]
assert (collapse(is2) == [Interval(0, 2)])
is3 = [Interval(0, 1), Interval(2, 3), Interval(0, 1.5)]
assert (collapse(is3) == [Interval(0, 1.5), Interval(2, 3)])
is4 = [Interval(0, 1), Interval(2, 3), Interval(0, 10)]
assert (collapse(is4) == [Interval(0., 10)])
corpus = load_corpus_txt('tests/mockdata/tiny.phn')
disc_clsdict = load_classes_txt('tests/mockdata/tiny.classes', corpus)
gold_clsdict = load_classes_txt('tests/mockdata/tiny.classes', corpus)
def test_cover():
assert (cover(disc_clsdict) == 8.0)
assert (cover(gold_clsdict) == 8.0)
def test_coverage():
assert (coverage(disc_clsdict, gold_clsdict) == 1.0)
def test_ued():
assert (ued('kitten', 'sitting') == 3)
