'--export-dir',
dest='export_dir',
help=
'Export dir for nonsilence_phones.txt, silence_phones.txt and extra_questions.txt',
type=str,
default='data/local/dict/')
args = parser.parse_args()
if (args.singlefile != ''):
ids = [args.singlefile]
else:
if args.filelist == '':
print('No files specified for processing!')
sys.exit()
ids = common_utils.loadIdFile(args.filelist)
combinedDict = {}
for myid in ids:
print("I'm now opening ", myid)
importer = guessImportFunc(myid)
d = importer(myid)
combinedDict = merge_dicts(combinedDict, d)
variants = 0
for key in sorted(combinedDict.keys()):
#print 'Word:',key,combinedDict[key]
variants += len(combinedDict[key])
print('Dictionary size is ', len(combinedDict),
parser.add_argument('-p',
'--utterance-postfix-name',
dest='postfix',
help='--utterance-postfix-name',
type=str,
default='')
args = parser.parse_args()
if args.filelist == '':
print 'Corpus filelist is empty. Use -f to supply a filelist!'
else:
print 'Load ', args.filelist, ', ommit ', args.remove_extension
ids_raw = common_utils.loadIdFile(
args.filelist, remove_extension=args.remove_extension)
print 'I have', len(
ids_raw
), 'files. Some may have their audio missing, I\'ll check that for you...'
ids = []
#check for missing wav files:
omitted = 0
for myid in ids_raw:
check = myid + args.postfix + args.wav_extension
if os.path.isfile(check):
ids.append(myid)
elif os.path.isfile(myid + '_Kinect-Beam' + args.wav_extension):
ids.append(myid)
else:
print 'Warning, omitting', myid, 'because I can\'t find', check
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Prepares the files from the TUDA corpus (XML) into text transcriptions for KALDI')
parser.add_argument('-f', '--filelist', dest='filelist', help='process this file list', type=str, default = '')
parser.add_argument('-r', '--remove_extension', dest='remove_extension', help='remove this extension, to get plain file id', type=str, default='.xml')
parser.add_argument('-w', '--audio-file-extension', dest='wav_extension', help='extension for audio files', type=str, default='.wav')
parser.add_argument('-p', '--utterance-postfix-name', dest='postfix', help='--utterance-postfix-name', type=str, default='_Kinect-Beam')
args = parser.parse_args()
if args.filelist == '':
print 'Corpus filelist is empty. Use -f to supply a filelist!'
else:
print 'Load ', args.filelist, ', ommit ', args.remove_extension
ids_raw = common_utils.loadIdFile(args.filelist,remove_extension=args.remove_extension)
print 'I have',len(ids_raw),'files. Some may have their audio missing, I\'ll check that for you...'
ids = []
#check for missing wav files:
omitted = 0
for myid in ids_raw:
check = myid+args.postfix+args.wav_extension
if os.path.isfile(check):
ids.append(myid)
else:
print 'Warning, omitting',myid,'because I can\'t find',check
omitted += 1
print 'Found',len(ids),' wav files.'
print 'Omitted ',omitted,' xml transcription files (Some missing files is normal for the TUDA Kaldi corpus).'
parser = argparse.ArgumentParser(description='Prepares various sources of pronounciations and builds a lexicon that can be exported to KALDI')
parser.add_argument('-f', '--filelist', dest='filelist', help='Process this file list of lexicons', type=str, default = '')
parser.add_argument('-s', '--single-file', dest='singlefile', help='Process this single lexicon file', type=str, default = '')
parser.add_argument('-e', '--export-pickle', dest='export', help='Export pickle file of combined phoneme dictionary', type=str, default = '')
parser.add_argument('-d', '--export-dir', dest='export_dir', help='Export dir for nonsilence_phones.txt, silence_phones.txt and extra_questions.txt' , type=str, default='data/local/dict/')
args = parser.parse_args()
if(args.singlefile != ''):
ids = [args.singlefile]
else:
if args.filelist == '':
print 'No files specified for processing!'
sys.exit()
ids = common_utils.loadIdFile(args.filelist)
combinedDict = {}
for myid in ids:
print "I'm now opening ", myid
importer = guessImportFunc(myid)
d = importer(myid)
combinedDict = merge_dicts(combinedDict, d)
variants = 0
for key in sorted(combinedDict.iterkeys()):
#print 'Word:',key,combinedDict[key]
variants += len(combinedDict[key])
print 'Dictionary size is ', len(combinedDict), ' pronounciation variants ', variants