def test_deletion_only_reversed(self):
ali = [
(['begin'], ['begin']),
(['a', 'b'], ['b']),
(['end'], ['end']),
]
expectation = [
(['a'], []),
]
self.assertEqual(extract_mismatch(ali), expectation)
def test_substitution_with_deletion(self):
ali = [
(['begin'], ['begin']),
(['a', 'b'], ['c']),
(['end'], ['end']),
]
expectation = [
(['a', 'b'], ['c']),
]
self.assertEqual(extract_mismatch(ali), expectation)
def test_insertion_only(self):
ali = [
(['begin'], ['begin']),
(['a'], ['a', 'b']),
(['end'], ['end']),
]
expectation = [
([], ['b']),
]
self.assertEqual(extract_mismatch(ali), expectation)
def test_double_substitution(self):
ali = [
(['a'], ['a']),
(['a'], ['b']),
(['a'], ['b']),
(['a'], ['a']),
(['a'], ['b']),
]
expectation = [(['a', 'a'], ['b', 'b']), (['a'], ['b'])]
self.assertEqual(extract_mismatch(ali), expectation)
def test_trivial(self):
ali = [(['a'], ['a'])]
expectation = []
self.assertEqual(extract_mismatch(ali), expectation)
references = {}
with open(args.text_references) as f:
for line in f:
fields = line.split()
references[fields[0]] = fields[1:]
candidate_possible_words = []
for line in sys.stdin:
fields = line.split()
_, utt_id, _, _, _ = parse_oov_id(fields[0])
candidate_line = fields[1:]
reference_line = references[utt_id]
alignment = align(reference_line, candidate_line)
mismatches = extract_mismatch(alignment)
oov_mismatch = find_in_mismatches(mismatches, "<UNK-OI>")
total_ref_len += len(reference_line)
total_nb_errors += number_of_errors(mismatches)
matching_oovs = intersection(oov_list, oov_mismatch[0])
if len(matching_oovs) in oov_hits:
oov_hits[len(matching_oovs)].append((utt_id, matching_oovs))
else:
oov_hits[len(matching_oovs)] = [(utt_id, matching_oovs)]
candidate_possible_words.append(oov_mismatch[0])
print(fields[0], oov_mismatch[0], '--', oov_mismatch[1])
with open(args.reference_file, 'w') as f: