Update app.py

app.py

@@ -61,29 +61,18 @@ def generate_script(user_prompt: str, model_id: str, token: str, duration: int):
         return f"Error generating script: {e}", "", ""
 
 # ---------------------------------------------------------------------
-# Voice-Over Generation Function
+# Voice-Over Generation Function (Inactive)
 # ---------------------------------------------------------------------
 @spaces.GPU(duration=300)
 def generate_voice(script: str, speaker: str = "default"):
     try:
-        # Load TTS model
-        tts_model_path = "tts_models/en/ljspeech/tacotron2-DDC"
-        vocoder_model_path = "vocoder_models/en/ljspeech/hifigan_v2"
-        synthesizer = Synthesizer(tts_model_path, vocoder_model_path)
-
-        # Generate audio
-        wav = synthesizer.tts(script)
-
-        # Save output to a file
-        output_path = f"{tempfile.gettempdir()}/generated_voice.wav"
-        synthesizer.save_wav(wav, output_path)
-
-        return output_path
+        # Placeholder for inactive state
+        return "Voice-over generation is currently inactive."
     except Exception as e:
-        return f"Error generating voice-over: {e}"
+        return f"Error: {e}"
 
 # ---------------------------------------------------------------------
-# Music Generation Function (using facebook/musicgen-medium)
+# Music Generation Function (facebook/musicgen-medium)
 # ---------------------------------------------------------------------
 @spaces.GPU(duration=300)
 def generate_music(prompt: str, audio_length: int):
@@ -117,23 +106,15 @@ def generate_music(prompt: str, audio_length: int):
 
 
 # ---------------------------------------------------------------------
-# Audio Blending Function with Ducking
+# Audio Blending Function with Ducking (Inactive)
 # ---------------------------------------------------------------------
 def blend_audio(voice_path: str, music_path: str, ducking: bool):
     try:
-        voice = AudioSegment.from_file(voice_path)
-        music = AudioSegment.from_file(music_path)
-
-        if ducking:
-            music = music - 10  # Lower music volume for ducking
-
-        combined = music.overlay(voice)
-        output_path = f"{tempfile.gettempdir()}/final_promo.wav"
-        combined.export(output_path, format="wav")
-
-        return output_path
+        # Placeholder for inactive state
+        return "Audio blending functionality is currently inactive."
     except Exception as e:
-        return f"Error blending audio: {e}"
+        return f"Error: {e}"
+
 
 # ---------------------------------------------------------------------
 # Gradio Interface
@@ -145,23 +126,23 @@ with gr.Blocks() as demo:
     """)
 
     with gr.Tabs():
-        # Step 1: Generate Script
-        with gr.Tab("Step 1: Generate Script"):
-            with gr.Row():
-                user_prompt = gr.Textbox(label="Promo Idea", placeholder="E.g., A 30-second promo for a morning show.")
-                llama_model_id = gr.Textbox(label="Llama Model ID", value="meta-llama/Meta-Llama-3-8B-Instruct")
-                duration = gr.Slider(label="Duration (seconds)", minimum=15, maximum=60, step=15, value=30)
-
-            generate_script_button = gr.Button("Generate Script")
-            script_output = gr.Textbox(label="Generated Script", lines=5)
-            sound_design_output = gr.Textbox(label="Sound Design", lines=3)
-            music_suggestion_output = gr.Textbox(label="Music Suggestions", lines=3)
-
-            generate_script_button.click(
-                fn=lambda user_prompt, model_id, duration: generate_script(user_prompt, model_id, hf_token, duration),
-                inputs=[user_prompt, llama_model_id, duration],
-                outputs=[script_output, sound_design_output, music_suggestion_output],
-            )
+# Step 1: Generate Script
+with gr.Tab("Step 1: Generate Script"):
+    with gr.Row():
+        user_prompt = gr.Textbox(label="Promo Idea", placeholder="E.g., A 30-second promo for a morning show.")
+        llama_model_id = gr.Textbox(label="Llama Model ID", value="meta-llama/Meta-Llama-3-8B-Instruct")
+        duration = gr.Slider(label="Duration (seconds)", minimum=15, maximum=60, step=15, value=30)
+
+    generate_script_button = gr.Button("Generate Script")
+    script_output = gr.Textbox(label="Generated Voice-Over Script", lines=5)
+    sound_design_output = gr.Textbox(label="Sound Design Suggestions", lines=3)
+    music_suggestion_output = gr.Textbox(label="Music Suggestions", lines=3)
+
+    generate_script_button.click(
+        fn=lambda user_prompt, model_id, duration: generate_script(user_prompt, model_id, hf_token, duration),
+        inputs=[user_prompt, llama_model_id, duration],
+        outputs=[script_output, sound_design_output, music_suggestion_output],
+    )
 
         # Step 2: Generate Voice
         with gr.Tab("Step 2: Generate Voice"):
@@ -178,30 +159,19 @@ with gr.Blocks() as demo:
             )
 
         # Step 3: Generate Music
-        with gr.Tab("Step 3: Generate Music"):
-            with gr.Row():
-                audio_length = gr.Slider(label="Music Length (tokens)", minimum=128, maximum=1024, step=64, value=512)
-                music_prompt_display = gr.Textbox(
-                    label="Music Suggestions from Step 1", 
-                    lines=3, 
-                    interactive=False
-                )
-
-            generate_music_button = gr.Button("Generate Music")
-            music_output = gr.Audio(label="Generated Music", type="filepath")
-
-            # Pass music suggestions from Step 1 to Step 3
-            generate_script_button.click(
-                fn=lambda _, __, ___: None,  # Dummy function to update the display
-                inputs=[],
-                outputs=[music_suggestion_output],
-            )
-
-            generate_music_button.click(
-                fn=lambda prompt, audio_length: generate_music(prompt, audio_length),
-                inputs=[music_suggestion_output, audio_length],
-                outputs=[music_output],
-            )
+with gr.Tab("Step 3: Generate Music"):
+    with gr.Row():
+        audio_length = gr.Slider(label="Music Length (tokens)", minimum=128, maximum=1024, step=64, value=512)
+
+    generate_music_button = gr.Button("Generate Music")
+    music_output = gr.Audio(label="Generated Music", type="filepath")
+
+    # Pass music suggestions from Step 1 to Step 3
+    generate_music_button.click(
+        fn=lambda music_suggestion, audio_length: generate_music(music_suggestion, audio_length),
+        inputs=[music_suggestion_output, audio_length],
+        outputs=[music_output],
+    )
 
         # Step 4: Blend Audio
         with gr.Tab("Step 4: Blend Audio"):