Update app.py

app.py

@@ -55,7 +55,6 @@ def text_to_speech(voice_id, text, session_id):
     if response.status_code != 200:
         return None
     
-    # Save temporary audio file with session ID
     audio_file_path = f'temp_voice_{session_id}.mp3'
     with open(audio_file_path, 'wb') as audio_file:
         audio_file.write(response.content)
@@ -135,30 +134,33 @@ def combine_audio_video(video_path, audio_path, output_path):
 
     subprocess.run(cmd, check=True)
 
-def process_video(voice_id, model, text, audio_file, input_type, progress=gr.Progress()):
+def process_input(voice, model, text, audio_file, progress=gr.Progress()):
     session_id = str(uuid.uuid4())
-    
-    # Handle audio source based on input type
-    if input_type == "text":
+    input_audio_path = None
+
+    if text and audio_file:
+        return None, "Please choose either text input or audio upload, not both."
+
+    if text:
         progress(0, desc="Generating speech...")
-        audio_path = text_to_speech(voice_id, text, session_id)
-        if not audio_path:
+        input_audio_path = text_to_speech(voice, text, session_id)
+        if not input_audio_path:
             return None, "Failed to generate speech audio."
-    else:  # audio upload
-        progress(0, desc="Processing uploaded audio...")
-        # Save uploaded audio to temporary file
-        temp_audio_path = f'temp_voice_{session_id}.mp3'
-        with open(temp_audio_path, 'wb') as f:
-            f.write(open(audio_file, 'rb').read())
-        audio_path = temp_audio_path
-    
+
+    elif audio_file:
+        progress(0, desc="Using uploaded audio...")
+        input_audio_path = audio_file.name
+
+    else:
+        return None, "Please provide either text or upload an audio file."
+
     progress(0.2, desc="Processing video...")
     video_path = os.path.join("models", model)
     
     try:
         progress(0.3, desc="Uploading files...")
         video_url = upload_file(video_path)
-        audio_url = upload_file(audio_path)
+        audio_url = upload_file(input_audio_path)
         
         if not video_url or not audio_url:
             raise Exception("Failed to upload files")
@@ -189,15 +191,14 @@ def process_video(voice_id, model, text, audio_file, input_type, progress=gr.Pro
         progress(0.8, desc="Falling back to simple combination...")
         try:
             output_path = f"output_{session_id}.mp4"
-            combine_audio_video(video_path, audio_path, output_path)
+            combine_audio_video(video_path, input_audio_path, output_path)
             progress(1.0, desc="Complete!")
             return output_path, f"Used fallback method. Original error: {str(e)}"
         except Exception as fallback_error:
             return None, f"All methods failed. Error: {str(fallback_error)}"
     finally:
-        # Cleanup
-        if os.path.exists(audio_path):
-            os.remove(audio_path)
+        if text and os.path.exists(input_audio_path):
+            os.remove(input_audio_path)
 
 def create_interface():
     voices = get_voices()
@@ -205,61 +206,48 @@ def create_interface():
     
     with gr.Blocks() as app:
         gr.Markdown("# JSON Train")
+
+        with gr.Row():
+            input_type = gr.Radio(choices=["Text", "Audio"], value="Text", label="Input Type")
+
         with gr.Row():
             with gr.Column():
-                input_type = gr.Radio(
-                    choices=["text", "audio"],
-                    label="Input Type",
-                    value="text"
-                )
+                voice_dropdown = gr.Dropdown(choices=[v[0] for v in voices], label="Select Voice", value=voices[0][0] if voices else None)
+                model_dropdown = gr.Dropdown(choices=models, label="Select Model", value=models[0] if models else None)
                 
-                with gr.Column(visible=True) as text_inputs:
-                    voice_dropdown = gr.Dropdown(
-                        choices=[v[0] for v in voices],
-                        label="Select Voice",
-                        value=voices[0][0] if voices else None
-                    )
-                    text_input = gr.Textbox(label="Enter text", lines=3)
-                
-                with gr.Column(visible=False) as audio_inputs:
-                    audio_upload = gr.Audio(label="Upload Audio", type="filepath")
-                
-                model_dropdown = gr.Dropdown(
-                    choices=models,
-                    label="Select Video Model",
-                    value=models[0] if models else None
+                text_input = gr.Textbox(label="Enter Text", lines=3)
+                audio_input = gr.Audio(label="Upload Audio", type="filepath")
+
+                text_input.show(True)
+                audio_input.show(False)
+
+                def toggle_input(input_type):
+                    if input_type == "Text":
+                        return gr.update(visible=True), gr.update(visible=False)
+                    else:
+                        return gr.update(visible=False), gr.update(visible=True)
+
+                input_type.change(
+                    fn=toggle_input,
+                    inputs=[input_type],
+                    outputs=[text_input, audio_input],
                 )
+                
                 generate_btn = gr.Button("Generate Video")
-            
+
             with gr.Column():
                 video_output = gr.Video(label="Generated Video")
                 status_output = gr.Textbox(label="Status", interactive=False)
         
-        def toggle_inputs(input_type):
-            return (
-                gr.Column.update(visible=(input_type == "text")),
-                gr.Column.update(visible=(input_type == "audio"))
-            )
-        
-        input_type.change(
-            fn=toggle_inputs,
-            inputs=[input_type],
-            outputs=[text_inputs, audio_inputs]
-        )
-        
-        def on_generate(voice_name, model_name, text, audio_file, input_type):
-            if input_type == "text":
-                voice_id = next((v[1] for v in voices if v[0] == voice_name), None)
-                if not voice_id:
-                    return None, "Invalid voice selected."
-            else:
-                voice_id = None  # Not needed for audio upload
-                
-            return process_video(voice_id, model_name, text, audio_file, input_type)
+        def on_generate(voice_name, model_name, text, audio_file):
+            voice_id = next((v[1] for v in voices if v[0] == voice_name), None)
+            if not voice_id:
+                return None, "Invalid voice selected."
+            return process_input(voice_id, model_name, text, audio_file)
         
         generate_btn.click(
             fn=on_generate,
-            inputs=[voice_dropdown, model_dropdown, text_input, audio_upload, input_type],
+            inputs=[voice_dropdown, model_dropdown, text_input, audio_input],
             outputs=[video_output, status_output]
         )