def dataset_modification(config_path, output_path):
data_config_path = '{}/data.yaml'.format(config_path)
data = Utility.yaml_load(data_config_path)
dataset = data['subset']
print dataset
for i in xrange(5, 20, 4):
if training_set[i-1] == 'i':
continue
subset = training_set[0:i]
# print subset
# print len(subset), len(subset)*50
dataset[ '{}-{}'.format(subset[0], subset[len(subset)-1]) ] = subset
# print dataset
# print data
Utility.yaml_save('{}/data.yaml'.format(output_path), data)
pass
def change_exp(config_path, output_path, subset):
exp = Utility.yaml_load( '{}/experiment.yaml'.format(config_path) )
exp['data_set']['optim_dur'] = 'a-{}'.format(subset[len(subset)-1])
exp['data_set']['train'] = 'a-{}'.format(subset[len(subset)-1])
print exp
Utility.yaml_save('{}/experiment.yaml'.format(output_path), exp)
pass
Utility.make_directory(set_out_path)
if Utility.is_dir_exists(set_stress_path) & Utility.is_dir_exists(set_utt_base_path):
print ch
for i in xrange(1, 51):
name = 'tscsd{}{}'.format(ch, Utility.fill_zero(i, 2))
yaml_filename = '{}/{}.utt.yaml'.format(set_utt_base_path, name )
if not Utility.is_file_exist(yaml_filename):
continue
full_file = '{}/{}.lab'.format(set_syllable_full_path, name)
count = [0]
yaml = Utility.yaml_load(yaml_filename)
add_stress(yaml, count, name)
if not (len(Utility.read_file_line_by_line(full_file)) == count[0] ):
print 'Not equal'
print name, len(Utility.read_file_line_by_line(full_file)), count[0]
out_file = '{}/{}.utt.yaml'.format(set_out_path, name)
Utility.yaml_save(out_file, yaml)
print 'all no {} names '.format(coun_no)
pass
def gen_utt(file_path, out_file):
out_list = []
phrase, word, syllable, phone = collections.OrderedDict(
), collections.OrderedDict(), collections.OrderedDict(
), collections.OrderedDict()
word_index, syllable_index, phone_index = 0, 0, 0
output_array = []
lines = Utility.read_file_line_by_line(file_path)
for line in lines:
#regex
pat = re.compile(
r"""
^(?P<prev_phoneme>.+)-(?P<cur_phoneme>.+)\+(?P<next_phoneme>.+)
/A:(?P<prev_phone_position>.+)_.+-(?P<cur_phone_position>.+)_.+\+(?P<next_phone_position>.+)_.+
/B:(?P<prev_tone>.+)-(?P<cur_tone>.+)\+(?P<next_tone>.+)
/C:(?P<prev_syllable_position>.+)_.+-(?P<cur_syllable_position>.+)_.+\+(?P<next_syllable_position>.+)_.+
/D:(?P<prev_number_of_phone>.+)-(?P<cur_number_of_phone>.+)\+(?P<next_number_of_phone>.+)
/E:(?P<prev_word_position>.+)-(?P<cur_word_position>.+)\+(?P<next_word_position>.+)
/F:.+_(?P<prev_number_of_syllable>.+)-.+_(?P<cur_number_of_syllable>.+)\+.+_(?P<next_number_of_syllable>.+)
/G:.+_(?P<number_of_syllable_in_sentence>.+)_(?P<number_of_word_in_sentence>.+)
/H:(?P<prev_word_partofspeech>.+)_.+-(?P<cur_word_partofspeech>.+)_.+\+(?P<next_word_partofspeech>.+)_.+
/I:(?P<cur_stress>.+)
?
""", re.VERBOSE)
match = re.match(pat, line)
if match == None:
raise (UtteranceException('Unmatched context: {:}'.format(line)))
#Check if pau or sil add new entity
if (match.group('cur_phoneme') == 'sil'):
#Phone unit
od = collections.OrderedDict()
od['unit'] = 'phone'
od['entity'] = match.group('cur_phoneme')
output_array.append(od)
elif (match.group('cur_phoneme') == 'pau'):
#Phone unit
od = collections.OrderedDict()
od['unit'] = 'phone'
od['entity'] = match.group('cur_phoneme')
output_array[len(output_array) - 1]['inners'].append(od)
else:
# Check if Start of utterance, add new utterance
if (match.group('cur_word_position')
== '1') & (match.group('cur_syllable_position') == '1') & (
match.group('cur_phone_position') == '1'):
#Utterance unit
od = collections.OrderedDict()
od['unit'] = 'utterance'
od['inners'] = []
output_array.append(od)
#print output_array
# Check if Start of word, add new word
if (match.group('cur_syllable_position')
== '1') & (match.group('cur_phone_position') == '1'):
#Word unit
od = collections.OrderedDict()
od['unit'] = 'word'
word_index += 1
od['word_index'] = word_index
od['part_of_speech'] = match.group('cur_word_partofspeech')
od['inners'] = []
output_array[len(output_array) - 1]['inners'].append(od)
#print "output_array"
# Check if Start of syllable, add new syllable
if (match.group('cur_phone_position') == '1'):
# Tone entity
od = collections.OrderedDict()
od['unit'] = 'syllable'
syllable_index += 1
od['syllable_index'] = syllable_index
od['tone_type'] = int(match.group('cur_tone'))
od['stress'] = match.group('cur_stress')
od['inners'] = []
word_index = len(output_array[len(output_array) - 1]['inners'])
output_array[len(output_array) -
1]['inners'][word_index - 1]['inners'].append(od)
#print output_array
# Add new phone
od = collections.OrderedDict()
od['unit'] = 'phone'
phone_index += 1
od['phone_index'] = phone_index
od['entity'] = match.group('cur_phoneme')
word_index = len(output_array[len(output_array) - 1]['inners'])
syllable_index = len(output_array[len(output_array) -
1]['inners'][word_index -
1]['inners'])
output_array[len(output_array) -
1]['inners'][word_index -
1]['inners'][syllable_index -
1]['inners'].append(od)
# print output_array
Utility.yaml_save(out_file, output_array)
pass
