Update app.py

app.py

@@ -1,39 +1,53 @@
 import gradio as gr
-from transformers import pipeline
+from gtts import gTTS
+from pydub import AudioSegment
+from io import BytesIO
 
-# Hugging Face TTS pipelines
-korean_tts = pipeline("text-to-speech", model="facebook/mms-tts-ko")
-us_english_tts = pipeline("text-to-speech", model="facebook/mms-tts-en-US")
-uk_english_tts = pipeline("text-to-speech", model="facebook/mms-tts-en-GB")
+# Step 1: Define a function to generate and merge TTS audio for multiple languages
+def multilingual_tts(korean_text, british_english_text, american_english_text):
+    # Language mapping
+    texts = {
+        "ko": korean_text,
+        "en-gb": british_english_text,  # British English
+        "en-us": american_english_text,  # American English
+    }
+    
+    combined_audio = AudioSegment.silent(duration=0)  # Empty audio to start
+    
+    for lang, text in texts.items():
+        if text.strip():  # Process only if text is provided
+            tld = 'co.uk' if lang == "en-gb" else 'com'
+            tts = gTTS(text, lang="en" if lang.startswith("en") else lang, tld=tld)
+            audio_file = BytesIO()
+            tts.write_to_fp(audio_file)
+            audio_file.seek(0)
+            tts_audio = AudioSegment.from_file(audio_file, format="mp3")
+            combined_audio += tts_audio + AudioSegment.silent(duration=500)  # Add silence between languages
+    
+    # Save combined audio to a file
+    output_file = "combined_output.mp3"
+    combined_audio.export(output_file, format="mp3")
+    
+    return output_file
 
-# Function to generate audio from text
-def generate_audio(text, language):
-    if language == "Korean":
-        audio = korean_tts(text)
-    elif language == "US English":
-        audio = us_english_tts(text)
-    elif language == "UK English":
-        audio = uk_english_tts(text)
-    else:
-        return "Invalid language selection."
-    return audio["audio"]
-
-# Gradio Interface
+# Step 2: Create Gradio interface
 with gr.Blocks() as demo:
-    gr.Markdown("## Multilingual TTS Application")
-
+    gr.Markdown("## Multilingual TTS: Generate a Single Audio File")
+    
     with gr.Row():
-        text_input = gr.Textbox(label="Enter text to convert to speech")
-        language_selector = gr.Dropdown(
-            choices=["Korean", "US English", "UK English"], label="Select language"
-        )
-
-    audio_output = gr.Audio(label="Generated Speech")
-    generate_button = gr.Button("Generate")
-
+        korean_input = gr.Textbox(label="Enter Korean Text:", placeholder="안녕하세요")
+        british_english_input = gr.Textbox(label="Enter British English Text:", placeholder="Hello (British)")
+        american_english_input = gr.Textbox(label="Enter American English Text:", placeholder="Hello (American)")
+    
+    output_audio = gr.Audio(label="Generated Speech", type="filepath")
+    generate_button = gr.Button("Generate Speech")
+    
     generate_button.click(
-        fn=generate_audio, inputs=[text_input, language_selector], outputs=audio_output
+        multilingual_tts, 
+        inputs=[korean_input, british_english_input, american_english_input], 
+        outputs=output_audio
     )
 
-# Launch the app
-demo.launch()
+# Run the app
+if __name__ == "__main__":
+    demo.launch()