Main block of transcribe_file_with_word_offset.py

Fig.1 Flowchart of Main Block of transcribe_file_with_word_time_offsets.py

Fig.1 呈現Main block的流程圖。 行155-175是External setup using os.environ and argparse.ArgumentParser區塊。 行176-209是Get Processing File List區塊。 行212-236是Set output_csv_filename and choices_list區塊。 行238-268是Batch-wise Speech Transcription區塊。

153 if __name__ == '__main__':
154     
155     os.environ["GOOGLE_APPLICATION_CREDENTIALS"]="GC-AI-Challenge-4dc21e90cad0.json"
156     parser = argparse.ArgumentParser(
157         description=__doc__,
158         formatter_class=argparse.RawDescriptionHelpFormatter)
159 
160     parser.add_argument('--beta', dest='beta', action='store_true')
161     parser.add_argument('--no-beta', dest='beta', action='store_false')
162     parser.add_argument('--nr', dest='nr', action='store_true')
163     parser.add_argument('--no-nr', dest='nr', action='store_false')
164     parser.add_argument('--reset', dest='reset', action='store_true')
165     parser.add_argument('--no-reset', dest='reset', action='store_false')
166     
167     parser.add_argument('-v', dest='verbose', action='store_true', help='Verbose mode')
168     
169     parser.add_argument('--batch_idx', dest='batch_idx', help='[1-60]')
170     parser.add_argument('--batch_size', dest='batch_idx', help='default 25')
171     parser.add_argument('--input_dir', dest='input_dir', help='input wav folder')
172     parser.set_defaults(beta=True, nr=True, reset=False, batch_idx='0', batch_size=25, input_dir='choices(entire)')
173     
174     
175     args = parser.parse_args()    
176     batch_size = args.batch_size
177     string_batch_range = args.batch_idx
178     input_dir = args.input_dir
179     
180     input_nr_dir = os.path.join(input_dir,'nr')
181     if not os.path.isdir(input_nr_dir):
182         os.makedirs(input_nr_dir)
183     
184     output_dir= 'result_'+ input_dir   
185     if not os.path.isdir(output_dir):
186         os.makedirs(output_dir)
187 
188     entireSpeechFileList = sorted(glob.glob(os.path.join(input_dir,'*.wav')))
189     #print('entire speech files in folder {}'.format(input_dir))
190     #print(entireSpeechFileList)
191     
192     if re.search('-', string_batch_range):
193          rr = re.match(r'([0-9]*)-([0-9]*)', string_batch_range)
194          if rr.group(1):
195              begin_idx = int(rr.group(1)) - 1
196          else:
197              begin_idx = 0
198          if rr.group(2):
199              end_idx = int(rr.group(2))
200          else:
201              end_idx = int(np.ceil(len(entireSpeechFileList)/batch_size))
202     else:
203         rr = re.match(r'([0-9]*)', string_batch_range)
204         if int(rr.group(1)) != 0:
205             begin_idx = int(rr.group(1))
206             end_idx = int(rr.group(1)) + 1
207         else:
208             begin_idx = 0
209             end_idx = int(np.ceil(len(entireSpeechFileList)/batch_size))
210     
211     
212     print('beta:', args.beta)
213     if args.beta:
214         suffix = 'beta'
215     else:
216         suffix = ''
217     print('suffix:', suffix)
218     
219     print('nr:', args.nr)
220     if args.nr:
221         suffix1 = 'nr'
222     else:
223         suffix1 = ''
224     print('suffix1:', suffix1)
225     
226 
227     output_csv_filename = os.path.join(output_dir,'transcribe_output_{}_{}.csv'.format(suffix,suffix1))
228     
229     '''check result csv exist or not'''
230     if os.path.isfile(output_csv_filename) and not(args.reset):
231         df = pd.read_csv(output_csv_filename,index_col=0)
232         existSpeechFileList = df.filename.tolist()
233         choices_list = df.sentence.tolist()
234     else:
235         choices_list = []
236         existSpeechFileList = []
237     
238     for b in range(begin_idx,end_idx):
239         
240         timestamp_xls_filename = os.path.join(output_dir,'transcribe_timestamp{}_{:02d}.xlsx'.format(suffix,b+1))
241         timestamp_writer = pd.ExcelWriter(timestamp_xls_filename)
242         
243         ''' check batch xlsx status 1. complete task 2. proccessing yet finished 3. new ''' 
244         titles = []
245         if os.path.isfile(timestamp_xls_filename):
246             print('load {} ...'.format(timestamp_xls_filename))
247             
248             book = load_workbook(timestamp_xls_filename)
249             
250             for ws in book.worksheets:
251                 mmm = re.match(r'(.+)\.wav_(.+)', ws.title)
252                 titles.append(os.path.join(input_dir,mmm.group(1)+ '.wav'))
253             titles = sorted(list(set(titles)))
254             print('number of sheets: {}'.format(len(titles)))
255             if len(titles) == batch_size or b == end_idx-1:
256                 continue
257             else:
258                 
259                 book = load_workbook(timestamp_xls_filename)
260                 timestamp_writer.book = book
261                 timestamp_writer.sheets = dict((ws.title, ws) for ws in book.worksheets)
262                 speechFileQueue = sorted(list(set(entireSpeechFileList[b*batch_size:(b+1)*batch_size]) - set(titles)))
263 
264                 batch_transcribe_speechFile(speechFileQueue,existSpeechFileList,choices_list,timestamp_writer,args.verbose,args.nr,beta=args.beta)
265         else:
266             print('create new {} !!!'.format(timestamp_xls_filename))
267             speechFileQueue = entireSpeechFileList[b*batch_size:(b+1)*batch_size]
268             
269             batch_transcribe_speechFile(speechFileQueue,existSpeechFileList,choices_list,timestamp_writer,args.verbose,args.nr,beta=args.beta)