Update app.py

app.py

@@ -4,6 +4,8 @@ import torch
 from TTS.api import TTS
 import os
 import json
+import scipy.io.wavfile as wavfile
+import numpy as np
 
 os.environ["COQUI_TOS_AGREED"] = "1"
 
@@ -16,19 +18,26 @@ def clone(text, audio):
     # Generowanie mowy
     wav = tts.tts(text=text, speaker_wav=audio, language="pl")
     
-    # Zapisywanie pliku audio
-    tts.save_wav(wav, file_path="./output.wav")
+    # Konwersja do numpy array i zapisanie jako plik WAV
+    wav_np = np.array(wav)
+    wavfile.write("./output.wav", 24000, (wav_np * 32767).astype(np.int16))
     
     # Uzyskanie informacji o fonemach
     phonemes_info = tts.synthesizer.get_phonemes(text, language="pl")
     
     # Przygotowanie informacji o fonemach
     phonemes_data = []
+    cumulative_duration = 0
     for phoneme, duration in phonemes_info:
+        start_time = cumulative_duration
+        end_time = start_time + duration
         phonemes_data.append({
             "phoneme": phoneme,
+            "start": float(start_time),
+            "end": float(end_time),
             "duration": float(duration)
         })
+        cumulative_duration = end_time
     
     # Zapisywanie informacji o fonemach do pliku JSON
     with open("./phonemes_info.json", "w", encoding="utf-8") as f: