Update app.py

app.py

@@ -34,16 +34,25 @@ st.write("Analyze all 27 emotions from uploaded audio with enhanced detection to
 # Audio Preprocessing
 def make_audio_scarier(audio_path, output_path):
     try:
-        commands = [
-            f"ffmpeg -i {audio_path} -af 'asetrate=44100*0.8,aresample=44100' temp1.wav",
-            f"ffmpeg -i temp1.wav -af 'reverb=0.8:0.2:0.5:0.5:0.5:0.5' temp2.wav",
-            f"ffmpeg -i temp2.wav -af 'atempo=1.2' {output_path}"
-        ]
-        for cmd in commands:
-            subprocess.run(cmd, shell=True, check=True)
+        # Step 1: Adjust pitch (slower rate for scarier effect)
+        cmd1 = f"ffmpeg -i {audio_path} -af 'asetrate=44100*0.8,aresample=44100' temp1.wav"
+        subprocess.run(cmd1, shell=True, check=True, stderr=subprocess.PIPE, text=True)
+        
+        # Step 2: Apply reverb with adjusted parameters
+        cmd2 = f"ffmpeg -i temp1.wav -af 'reverb=0.4:0.7:0.5:0.5:0.5:0.02' temp2.wav"
+        subprocess.run(cmd2, shell=True, check=True, stderr=subprocess.PIPE, text=True)
+        
+        # Step 3: Adjust tempo
+        cmd3 = f"ffmpeg -i temp2.wav -af 'atempo=1.2' {output_path}"
+        subprocess.run(cmd3, shell=True, check=True, stderr=subprocess.PIPE, text=True)
+        
+        # Clean up temporary files
         for temp_file in ["temp1.wav", "temp2.wav"]:
             if os.path.exists(temp_file):
                 os.remove(temp_file)
+    except subprocess.CalledProcessError as e:
+        st.error(f"Audio processing failed: {str(e)} - Command: {e.cmd}, Output: {e.stderr}")
+        raise
     except Exception as e:
         st.error(f"Audio processing failed: {str(e)}")
         raise
@@ -83,9 +92,9 @@ def perform_audio_emotion_detection(audio_path):
         # Enhanced boosting based on audio features
         features = extract_audio_features(audio_path)
         if features.get("pitch_mean", 0) < 200 and features.get("energy_mean", 0) > 0.1 and features.get("zcr_mean", 0) > 0.1:
-            emotion_dict["fearful"] = min(1.0, emotion_dict.get("fearful", 0) + 0.4)  # Increased boost
+            emotion_dict["fearful"] = min(1.0, emotion_dict.get("fearful", 0) + 0.4)
             top_emotion = "fearful" if emotion_dict["fearful"] > emotion_dict[top_emotion] else top_emotion
-        elif features.get("energy_mean", 0) > 0.25:  # Stricter threshold
+        elif features.get("energy_mean", 0) > 0.25:
             emotion_dict["angry"] = min(1.0, emotion_dict.get("angry", 0) + 0.35)
             top_emotion = "angry" if emotion_dict["angry"] > emotion_dict[top_emotion] else top_emotion
         elif features.get("pitch_mean", 0) > 500 and features.get("energy_mean", 0) < 0.05:
@@ -99,7 +108,7 @@ def perform_audio_emotion_detection(audio_path):
             top_emotion = "surprise" if emotion_dict["surprise"] > emotion_dict[top_emotion] else top_emotion
         # Fallback to avoid neutral if score is low
         if emotion_dict["neutral"] > 0.5 and max([v for k, v in emotion_dict.items() if k != "neutral"]) > 0.3:
-            emotion_dict["neutral"] = max(0.0, emotion_dict["neutral"] - 0.2)  # Reduce neutral weight
+            emotion_dict["neutral"] = max(0.0, emotion_dict["neutral"] - 0.2)
             top_emotion = max(emotion_dict, key=emotion_dict.get)
         return emotion_dict, top_emotion
     except Exception as e:
@@ -124,7 +133,6 @@ def perform_text_emotion_detection(text):
                     "pride", "realization", "relief", "remorse", "sadness", "surprise", "neutral"]
         emotions_dict = {result['label']: result['score'] for result in results if result['label'] in emotions}
         top_emotion = max(emotions_dict, key=emotions_dict.get)
-        # Reduce neutral influence if other emotions are strong
         if emotions_dict.get("neutral", 0) > 0.5 and max([v for k, v in emotions_dict.items() if k != "neutral"]) > 0.4:
             emotions_dict["neutral"] = max(0.0, emotions_dict["neutral"] - 0.15)
             top_emotion = max(emotions_dict, key=emotions_dict.get)
@@ -237,7 +245,7 @@ def display_analysis_results(audio_path):
     # Avoid neutral if other emotions are competitive
     top_emotion = max(combined_emotions, key=combined_emotions.get)
     if combined_emotions["neutral"] > 0.5 and max([v for k, v in combined_emotions.items() if k != "neutral"]) > 0.4:
-        combined_emotions["neutral"] = max(0.0, combined_emotions["neutral"] - 0.25)  # Stronger reduction
+        combined_emotions["neutral"] = max(0.0, combined_emotions["neutral"] - 0.25)
         top_emotion = max(combined_emotions, key=combined_emotions.get)
 
     sentiment = "POSITIVE" if top_emotion in ["admiration", "amusement", "approval", "caring", "desire", "excitement", 
@@ -254,7 +262,7 @@ def display_analysis_results(audio_path):
         st.markdown(f"**{sentiment_icon} {sentiment.capitalize()}** (Based on {top_emotion})")
         st.subheader("Sarcasm")
         sarcasm_icon = "😏" if is_sarcastic else "😐"
-        st.markdown(f"**{sarcasm_icon} {'Detected' if is_sarcastic else 'Not Detected'}** (Score: {sarcasm_score:.3f})")
+        st.markdown(f"**{sarc_icon} {'Detected' if is_sarcastic else 'Not Detected'}** (Score: {sarcasm_score:.3f})")
 
     with col2:
         st.subheader("Emotion Distribution")
@@ -304,4 +312,5 @@ def main():
     """)
 
 if __name__ == "__main__":
-    main()
+    main()
+