def reset(corpus_name): """Remove the database files produced from import.""" with ensure_local_database_running(corpus_name, port=8080, ip=server_ip, token=load_token()) as params: config = CorpusConfig(corpus_name, **params) with CorpusContext(config) as c: print('Resetting the corpus.') c.reset()
print("wrote to ", dest) if __name__ == '__main__': parser = argparse.ArgumentParser() parser.add_argument("timit_path", help="path to timit corpus") parser.add_argument("dest_path", help="path to destination") parser.add_argument("--reset", help="set to true to reset corpus", default=False) parser.add_argument("--convert", help="set to true if converting mfccs", default=False) args = parser.parse_args() corpus_name = "TIMIT" with ensure_local_database_running('database') as config: conf = CorpusConfig(corpus_name, **config) if args.reset: loading(conf, args.timit_path) if args.convert: filename_to_path = {} for root, dirs, files in os.walk(args.timit_path): for file in files: if re.match(".*\.[Ww][Aa][Vv]", file) is not None: src_filename = re.sub("\.[Ww][Aa][Vv]", "", file) path = os.path.join(root, file) filename_to_path[src_filename] = path export_textgrid(conf, args.dest_path, filename_to_path)
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 = common.load_config(corpus_name) print('Processing...') ignored_speakers = corpus_conf.get('ignore_speakers', []) if reset: common.reset(corpus_name) ip = common.server_ip if docker: ip = common.docker_ip with ensure_local_database_running(corpus_name, ip=ip, port=common.server_port, token=common.load_token()) as params: print(params) config = CorpusConfig(corpus_name, **params) config.formant_source = 'praat' # Common set up 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'])
directories = [ x for x in os.listdir(base_dir) if os.path.isdir(x) and x != 'Common' ] 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,
args = parser.parse_args() corpus_name = args.corpus_name reset = args.reset directories = [ x for x in os.listdir(base_dir) if os.path.isdir(x) and x != 'Common' ] 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 = common.load_config(corpus_name) print('Processing...') with ensure_local_database_running(corpus_name) as params: print(params) config = CorpusConfig(corpus_name, **params) config.formant_source = 'praat' with CorpusContext(config) as c: print(c.hierarchy) # Common set up if reset: common.reset(config) 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,