Update app.py

app.py

@@ -65,22 +65,20 @@ def save_uploaded_audio(audio_file, session_id):
     if audio_file is None:
         return None
     
-    # Get the file extension from the original filename
-    _, ext = os.path.splitext(audio_file.name)
-    if not ext:
-        ext = '.mp3'  # Default extension if none is found
-    
-    # Save the uploaded audio file with session ID
-    audio_file_path = f'temp_voice_{session_id}{ext}'
-    with open(audio_file_path, 'wb') as f:
-        if isinstance(audio_file, str):  # If it's a file path
-            with open(audio_file, 'rb') as source:
-                f.write(source.read())
-        else:  # If it's a file object
-            audio_file.seek(0)
-            f.write(audio_file.read())
+    # If audio_file is already a path, just copy it
+    if isinstance(audio_file, str):
+        ext = os.path.splitext(audio_file)[1]
+        if not ext:
+            ext = '.mp3'
+        output_path = f'temp_voice_{session_id}{ext}'
+        
+        # Copy the file to our temporary location
+        with open(audio_file, 'rb') as source:
+            with open(output_path, 'wb') as dest:
+                dest.write(source.read())
+        return output_path
     
-    return audio_file_path
+    return None
 
 def upload_file(file_path):
     with open(file_path, 'rb') as file:
@@ -113,7 +111,7 @@ def lipsync_api_call(video_url, audio_url):
 
 def check_job_status(job_id):
     headers = {"x-api-key": B_KEY}
-    max_attempts = 30  # Limit the number of attempts
+    max_attempts = 30
     
     for _ in range(max_attempts):
         response = requests.get(f"{API_URL}/{job_id}", headers=headers)
@@ -224,42 +222,42 @@ def create_interface():
     
     with gr.Blocks() as app:
         gr.Markdown("# JSON Train")
-        with gr.Row():
-            with gr.Column():
-                input_type = gr.Radio(
-                    choices=["text", "audio"],
-                    label="Input Type",
-                    value="text"
-                )
-                
-                # Text-to-speech inputs
-                with gr.Group() 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)
-                
-                # Audio upload input
-                with gr.Group() 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
+        
+        input_type = gr.Radio(
+            choices=["text", "audio"],
+            label="Input Type",
+            value="text"
+        )
+        
+        with gr.Column():
+            # Text-to-speech inputs
+            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
                 )
-                generate_btn = gr.Button("Generate Video")
+                text_input = gr.Textbox(label="Enter text", lines=3)
+            
+            # Audio upload input
+            with gr.Column(visible=False) as audio_inputs:
+                audio_upload = gr.Audio(label="Upload Audio", type="filepath")
             
-            with gr.Column():
-                video_output = gr.Video(label="Generated Video")
-                status_output = gr.Textbox(label="Status", interactive=False)
+            model_dropdown = gr.Dropdown(
+                choices=models,
+                label="Select Video Model",
+                value=models[0] if models else None
+            )
+            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.Group.update(visible=(input_type == "text")),
-                gr.Group.update(visible=(input_type == "audio"))
+                gr.Column.update(visible=(input_type == "text")),
+                gr.Column.update(visible=(input_type == "audio"))
             )
         
         input_type.change(