Update app.py

app.py

@@ -12,7 +12,7 @@ DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
 print(f"Using device: {DEVICE}")
 
 class EmotionAwareTranscriber:
-    def __init__(self, model_size="base"):  # Fixed method name from _init_ to __init__
+    def __init__(self, model_size="base"):
         print("Initializing models...")
 
         # Initialize Whisper
@@ -157,37 +157,26 @@ class EmotionAwareTranscriber:
                 "audio": None
             }
 
-# Add installation code for Google Colab
-def install_dependencies():
-    print("Installing required packages...")
-    import subprocess
-    
-    # Install required packages
-    subprocess.run(["pip", "install", "gradio", "torch", "transformers", "librosa", "gtts", "numpy"])
-    
-    # Check if ffmpeg is installed, and install if needed
-    try:
-        import ffmpeg
-    except ImportError:
-        print("Installing ffmpeg...")
-        subprocess.run(["apt-get", "update", "-qq"])
-        subprocess.run(["apt-get", "install", "-y", "-qq", "ffmpeg"])
-    
-    print("Dependencies installed successfully.")
+# Initialize the transcriber first
+transcriber = EmotionAwareTranscriber()
 
-# Initialize the transcriber
-process_audio_wrapper.last_audio = None  # Initialize the class attribute
+# Define a global variable to store the last audio file path
+last_audio_file = None
 
+# Define the process_audio_wrapper function AFTER initializing the variable
 def process_audio_wrapper(audio_path, style):
+    global last_audio_file
+    
     result = transcriber.process_audio(audio_path, style)
 
     # Clean up previous audio files
-    if hasattr(process_audio_wrapper, "last_audio") and process_audio_wrapper.last_audio:
+    if last_audio_file and os.path.exists(last_audio_file):
         try:
-            os.unlink(process_audio_wrapper.last_audio)
-        except:
-            pass
-    process_audio_wrapper.last_audio = result["audio"]
+            os.unlink(last_audio_file)
+        except Exception as e:
+            print(f"Error cleaning up audio file: {e}")
+    
+    last_audio_file = result["audio"]
 
     return (
         result["transcription"],
@@ -196,46 +185,35 @@ def process_audio_wrapper(audio_path, style):
         result["audio"] if result["audio"] else None
     )
 
-# Main execution with proper checks for Colab
-if __name__ == "__main__":  # Fixed double underscore
-    # Check if running in Colab
-    try:
-        import google.colab
-        IN_COLAB = True
-    except:
-        IN_COLAB = False
-    
-    if IN_COLAB:
-        install_dependencies()
-    
-    # Initialize transcriber after dependencies are installed
-    transcriber = EmotionAwareTranscriber()
-    
-    # Gradio interface
-    with gr.Blocks(title="Emotion-Aware Audio Transcriber") as demo:
-        gr.Markdown("# 🎤 Emotion-Aware Audio Transcriber")
-        gr.Markdown("Upload an audio file to get a transcription with emotional analysis and response")
-
-        with gr.Row():
-            audio_input = gr.Audio(label="Upload Audio", type="filepath")
-            style_selector = gr.Radio(
-                ["motivational", "calm", "energetic", "angry"],
-                label="Response Style",
-                value="motivational"
-            )
-            submit_btn = gr.Button("Process", variant="primary")
-
-        with gr.Column():
-            transcription_output = gr.Textbox(label="Transcription")
-            emotion_output = gr.Textbox(label="Detected Emotion")
-            response_output = gr.Textbox(label="Generated Response")
-            audio_output = gr.Audio(label="Spoken Response")
-
-        submit_btn.click(
-            fn=process_audio_wrapper,
-            inputs=[audio_input, style_selector],
-            outputs=[transcription_output, emotion_output, response_output, audio_output]
+# Gradio interface
+with gr.Blocks(title="Emotion-Aware Audio Transcriber") as demo:
+    gr.Markdown("# 🎤 Emotion-Aware Audio Transcriber")
+    gr.Markdown("Upload an audio file to get a transcription with emotional analysis and response")
+
+    with gr.Row():
+        audio_input = gr.Audio(label="Upload Audio", type="filepath")
+        style_selector = gr.Radio(
+            ["motivational", "calm", "energetic", "angry"],
+            label="Response Style",
+            value="motivational"
         )
+        submit_btn = gr.Button("Process", variant="primary")
+
+    with gr.Column():
+        transcription_output = gr.Textbox(label="Transcription")
+        emotion_output = gr.Textbox(label="Detected Emotion")
+        response_output = gr.Textbox(label="Generated Response")
+        audio_output = gr.Audio(label="Spoken Response")
+
+    submit_btn.click(
+        fn=process_audio_wrapper,
+        inputs=[audio_input, style_selector],
+        outputs=[transcription_output, emotion_output, response_output, audio_output]
+    )
 
-    # Launch with share=True for Colab to generate a public URL
-    demo.launch(debug=True, share=True)
+# Launch the app
+if __name__ == "__main__":
+    demo.launch()
+else:
+    # This part is crucial for HuggingFace Spaces deployment
+    app = demo