Update app.py

app.py

@@ -166,9 +166,14 @@ def is_square_video(video_path):
     return gr.update(visible=True)
 
 
-def txt_to_driving_video(input_text, video_type):
-    audio_outpath = gpu_wrapped_elevenlabs_pipeline_generate_voice(text=input_text, voice=None)
-    video_outpath = gpu_wrapped_stf_pipeline_execute(audio_outpath, video_type)
+def txt_to_driving_video(input_text, audio_file, video_type):
+    
+    if audio_file is None and input_text is not None:
+        audio_file = gpu_wrapped_elevenlabs_pipeline_generate_voice(text=input_text, voice=None)
+
+    # 오디오파일이 있으면 텍스트보다 오디오를 우선적으로 처리
+    video_outpath = gpu_wrapped_stf_pipeline_execute(audio_file, video_type)
+    
     return video_outpath
 
 
@@ -208,15 +213,16 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown("**Text to talking video generation tool**\n\n")
 
     #gr.Markdown("### 1. Text to audio")
-    gr.Markdown("### 1. Text to Driving-Video")
+    gr.Markdown("### 1. Text/Audio to Driving-Video")
     with gr.Row():
-        
-        script_txt = gr.Text()
+        with gr.Column(scale=1):
+            script_txt = gr.Text(label="Textbox(Enter text OR upload an audio file.)")
+            audio_file = gr.Audio(label="Upload audio(Enter text OR upload an audio file.)", type="file")
         # audio_gen_button = gr.Button("Audio generation", variant="primary")
         # with gr.Column():
         #     txt2video_gen_button = gr.Button("txt2video generation", variant="primary")
         video_type = gr.Radio(choices=["Female video", "Male video"], label="Select video type", value="Female video")
-        txt2video_gen_button = gr.Button("txt2video generation", variant="primary")
+        txt2video_gen_button = gr.Button("Txt/Audio to video generation", variant="primary")
 
         #with gr.Column():
             #audio_gen_button = gr.Button("Audio generation", variant="primary")
@@ -326,7 +332,7 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     txt2video_gen_button.click(
         fn=txt_to_driving_video,
         inputs=[
-            script_txt, video_type
+            script_txt, audio_file, video_type
         ],
         outputs=[video_input],
         show_progress=True