Merge branch 'master' of ssh://invmac/~/Public/repos/speech-scoring

speech_samplegen.py

@@ -198,9 +198,10 @@ def synth_generator():
                 prog = tqdm(words)
                 prog.set_postfix(variant=v,voice=s.name,rate=s.rate)
                 for w in tqdm(words):
-                    prog.set_postfix(word=w)
+                    prog.set_description('Synthesizing text:"{}"'.format(w))
                     synthed = s.generate_audio(w, v)
                     writer(synthed)
+                prog.close()
         end_time = time.time()
         time_str = hms_string(end_time - start_time)
         print("It took {} to synthsize all variants.".format(time_str))
@@ -243,20 +244,21 @@ def generate_audio_for_stories():
     # story_file = './inputs/all_stories_hs.json'
     story_file = './inputs/all_stories.json'
     stories_data = json.load(open(story_file))
-    # text_list = [t[0] for i in stories_data.values() for t in i]
-    text_list = [i for g in stories_data.values() for i in g]
-    generate_audio_for_text_list(text_list)
+    # text_list_dup = [t[0] for i in stories_data.values() for t in i]
+    text_list_dup = [t for i in stories_data.values() for t in i]
+    text_list = sorted(list(set(text_list_dup)))
+    generate_audio_for_text_list(text_list[:5])
 
 def generate_test_audio_for_stories():
     story_file = './inputs/all_stories_hs.json'
     # story_file = './inputs/all_stories.json'
     stories_data = json.load(open(story_file))
-    text_list = [t[0] for i in stories_data.values() for t in i]
+    text_list_dup = [t[0] for i in stories_data.values() for t in i]
+    text_list = sorted(list(set(text_list_dup)))
     # text_list = [i.replace('-','') for g in stories_data.values() for i in g]
     word_list = [i.strip('\n_') for i in open('./inputs/wordlist.txt','r').readlines()]
     text_set = set(text_list)
     new_word_list = [i for i in word_list if i not in text_set and len(i) > 4]
-    # len(new_word_list)
     test_words = new_word_list[:int(len(text_list)/5+1)]
     generate_audio_for_text_list(test_words)