Update app.py

app.py

@@ -5,6 +5,7 @@ import torch
 import zipfile
 from TTS.api import TTS
 from pydub import AudioSegment
+import mecab
 
 os.environ["COQUI_TOS_AGREED"] = "1"
 
@@ -47,8 +48,15 @@ def convert_to_wav(input_audio_file):
         return "temp.wav"
     return input_audio_file
 
+def preprocess_text(txt, katsu):
+    if katsu:
+        tagger = mecab.Tagger()
+        txt = tagger.parse(txt).strip()
+    return txt
+
 def synthesize_text(text, input_audio_file, language):
     input_audio_file = convert_to_wav(input_audio_file)
+    text = preprocess_text(text, language == "ja")
     tts.tts_to_file(text=text, speaker_wav=input_audio_file, language=language, file_path="./output.wav")
     return "./output.wav"
 
@@ -73,15 +81,16 @@ def clone(text, input_file, language, url=None, use_url=False):
             input_audio_file = input_file.name
 
     output_file_path = synthesize_text(text, input_audio_file, language)
-    return output_file_path
+    returnoutput_file_path
 
 iface = gr.Interface(
     fn=clone,
-    inputs=["text", gr.File(label="Input File", file_types=[".zip", *AUDIO_FORMATS]), gr.Dropdown(choices=LANGUAGES, label="Language"), gr.Text(label="URL"), gr.Checkbox(label="Use URL", value=False)],
-    outputs=gr.Audio(type='filepath'),
-    title='Voice Clone',
-    description=""" by [Angetyde](https://youtube.com/@Angetyde?si=7nusP31nTumIkPTF) and [Tony Assi](https://www.tonyassi.com/ ) use this colab with caution <3. """,
-    theme=gr.themes.Base(primary_hue="teal", secondary_hue="teal", neutral_hue="slate")
+    inputs=["text", gr.File(label="Input File", file_types=[".zip", ".wav", ".mp3", ".flac", ".mp4"]), "language", "url", gr.Checkbox(label="Use URL")],
+    outputs="text",
+    title="Voice Cloning Application",
+    description="Enter text, select an audio file, choose a language, and optionally input a URL to clone a voice.",
+    theme="default",
+    allow_flagging="never"
 )
 
-iface.launch(share=True)
+iface.launch()