def acoustic_config(graph_db, textgrid_test_dir):
config = CorpusConfig('acoustic', **graph_db)
acoustic_path = os.path.join(textgrid_test_dir, 'acoustic_corpus.TextGrid')
with CorpusContext(config) as c:
c.reset()
parser = inspect_textgrid(acoustic_path)
c.load(parser, acoustic_path)
config.pitch_algorithm = 'acousticsim'
config.formant_source = 'acousticsim'
return config
def overlapped_config(graph_db, textgrid_test_dir, acoustic_syllabics):
config = CorpusConfig('overlapped', **graph_db)
acoustic_path = os.path.join(textgrid_test_dir, 'overlapped_speech')
with CorpusContext(config) as c:
c.reset()
parser = inspect_mfa(acoustic_path)
c.load(parser, acoustic_path)
c.encode_pauses(['sil'])
c.encode_utterances(min_pause_length=0)
c.encode_syllabic_segments(acoustic_syllabics)
c.encode_syllables()
config.pitch_algorithm = 'acousticsim'
config.formant_source = 'acousticsim'
return config
if args.corpus_name not in directories:
print(
'The corpus {0} does not have a directory (available: {1}). Please make it with a {0}.yaml file inside.'
.format(args.corpus_name, ', '.join(directories)))
sys.exit(1)
corpus_conf = load_config(corpus_name)
print('Processing...')
#Connect to local database at 8080
with ensure_local_database_running(corpus_name,
port=8080,
token=common.load_token()) as params:
#Load corpus context and config info
config = CorpusConfig(corpus_name, **params)
config.formant_source = 'praat'
# Common set up
if reset:
with CorpusContext(config) as c:
print("Resetting the corpus.")
c.reset()
common.loading(config, corpus_conf['corpus_directory'],
corpus_conf['input_format'])
common.lexicon_enrichment(config,
corpus_conf['unisyn_spade_directory'],
corpus_conf['dialect_code'])
common.speaker_enrichment(config,
corpus_conf['speaker_enrichment_file'])
common.basic_enrichment(
config, corpus_conf['vowel_inventory'] +
corpus_conf['extra_syllabic_segments'], corpus_conf['pauses'])
q = q.columns(c.phone.speaker.name.column_name('speaker'),
c.phone.discourse.name.column_name('discourse'),
c.phone.id.column_name('phone_id'),
c.phone.label.column_name('phone_label'),
c.phone.begin.column_name('begin'),
c.phone.end.column_name('end'),
c.phone.following.label.column_name('following_phone'),
c.phone.previous.label.column_name('previous_phone'),
c.phone.word.label.column_name('word'),
c.phone.cog.column_name('cog'),
c.phone.peak.column_name('peak'),
c.phone.slope.column_name('slope'),
c.phone.spread.column_name('spread'))
q.to_csv(output_path_word_initial)
print("Results for sibilants written to " + output_path + " and " +
output_path_word_initial)
if __name__ == '__main__':
with ensure_local_database_running('database') as config:
conf = CorpusConfig(corpus_name, **config)
conf.pitch_source = 'praat'
# config.pitch_algorithm = 'base'
conf.formant_source = 'praat'
conf.intensity_source = 'praat'
conf.praat_path = praat_path
if reset:
loading(conf)
acoustics(conf)
analysis(conf)
