Update app.py

app.py

@@ -38,19 +38,18 @@ def parse_multilingual_text(input_text):
     return [(lang, text.strip()) for lang, text in segments if lang in LANGUAGES.keys()]
 
 @spaces.GPU
-def generate_segment_audio(text, lang, speaker_url, pipe):
-
-    # Ensure text is a string
+def generate_segment_audio(text, lang, speaker_audio, pipe):
     if not isinstance(text, str):
         text = text.decode("utf-8") if isinstance(text, bytes) else str(text)
-
-    # Check and process speaker_audio if it is a string (path to file)
-    if isinstance(speaker_audio, str):
-        print("Speaker Audio Type:", type(speaker_audio))
-        with open(speaker_audio, 'rb') as file:
-            speaker_audio = file.read()
-
-    audio_data = pipe.generate(text, speaker_audio, lang)
+    if speaker_audio is not None:
+        if isinstance(speaker_audio, str):
+            with open(speaker_audio, 'rb') as file:
+                speaker_audio_data = file.read()
+        else:
+            speaker_audio_data = speaker_audio
+    else:
+        speaker_audio_data = None
+    audio_data = pipe.generate(text, speaker_audio_data, lang)
     resample_audio = resampler(newsr=24000)
     audio_data_resampled = next(resample_audio([{'sample_rate': 24000, 'samples': audio_data.cpu()}]))['samples_24k']
     audio_np = audio_data_resampled.cpu().numpy()