add function for audio combination

app.py

@@ -106,7 +106,7 @@ def infer(audio_input_path):
     if result.returncode == 0:
         print("Command executed successfully.")
         concat_out, file_list = concatenate_audio()
-        return concat_out, gr.update(choices=file_list, value=file_list[0], visible=True), gr.update(visible=True), gr.update(value=file_list, visible=True)
+        return concat_out, gr.update(choices=file_list, value=file_list[0], visible=True), gr.update(visible=True), gr.update(value=file_list, visible=True), gr.update(visible=True)
     else:
         print("Error executing command.")
         raise gr.Error("Error executing command")
@@ -114,8 +114,33 @@ def infer(audio_input_path):
 def load_chosen_audio(audio_path):
     return audio_path
 
+def overlay_audio(original_mp3: str, translated_wav: str) -> str:
+    # Load original MP3 and convert to WAV
+    original = AudioSegment.from_mp3(original_mp3).set_frame_rate(16000).set_channels(1)
+    
+    # Lower the volume (reduce by 6 dB, which is half)
+    original = original - 6
+    
+    # Load the translated WAV
+    translated = AudioSegment.from_wav(translated_wav).set_frame_rate(16000).set_channels(1)
+    
+    # Overlay the translated speech over the original
+    combined = original.overlay(translated)
+    
+    # Create a temporary file to save the output
+    temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
+    combined.export(temp_file.name, format="wav")
+    
+    print(f"Final audio saved at: {temp_file.name}")
+    return temp_file.name  # Return the temporary file path
+
+def process_final_combination(audio_in, chosen_translated):
+    audio_in = process_audio(audio_in)
+    temp_output_path = overlay_audio(audio_in, chosen_translated)
+    return gr.update(value=temp_output_path, visible=True)
+
 def hide_previous():
-    return gr.update(visible=False), gr.update(visible=False), gr.update(visible=False)
+    return gr.update(visible=False), gr.update(visible=False), gr.update(visible=False), gr.update(visible=False), gr.update(visible=False)
 
 css="""
 div#col-container{
@@ -131,11 +156,16 @@ with gr.Blocks(css=css) as demo:
         audio_input = gr.Audio(label="Audio IN", type="filepath")
         submit_btn = gr.Button("Submit")
         output_result = gr.Audio(label="Translated result")
-        dropdown_wav_selector = gr.Dropdown(
-            label="Pick a generated audio to load",
-            value = None,
-            visible=False
-        )
+
+        with gr.Row():
+            dropdown_wav_selector = gr.Dropdown(
+                label="Pick a generated audio to load",
+                value = None,
+                visible=False,
+                scale=2
+            )
+            choose_this_btn = gr.Button("Use this one", scale=1)
+        combined_output = gr.Audio("Combined Outpu", visible=False)
         with gr.Accordion("Downloadable audio Output list", open=False, visible=False) as result_accordion:
             wav_list = gr.Files(label="Output Audio List", visible=False)
 
@@ -155,14 +185,20 @@ with gr.Blocks(css=css) as demo:
         queue = False
     )
 
+    choose_this_btn.click(
+        fn = process_final_combination,
+        inputs = [audio_input, dropdown_wav_selector],
+        outputs = [combined_output]
+    )
+
     submit_btn.click(
         fn = hide_previous,
         inputs = None,
-        outputs = [dropdown_wav_selector, result_accordion,  wav_list]
+        outputs = [dropdown_wav_selector, result_accordion,  wav_list, choose_this_btn, combined_output]
     ).then(
         fn = infer,
         inputs = [audio_input],
-        outputs = [output_result, dropdown_wav_selector, result_accordion, wav_list]
+        outputs = [output_result, dropdown_wav_selector, result_accordion, wav_list, choose_this_btn]
     )
 
 demo.queue().launch(show_api=False, show_error=True)