Update app.py

app.py

@@ -13,6 +13,7 @@ import time
 import base64
 import io
 import streamlit.components.v1 as components
+import numpy as np
 
 # Suppress warnings for a clean console
 logging.getLogger("torch").setLevel(logging.CRITICAL)
@@ -20,6 +21,14 @@ logging.getLogger("transformers").setLevel(logging.CRITICAL)
 warnings.filterwarnings("ignore")
 os.environ["TOKENIZERS_PARALLELISM"] = "false"
 
+# Check if NumPy is available
+try:
+    test_array = np.array([1, 2, 3])
+    torch.from_numpy(test_array)
+except Exception as e:
+    st.error(f"NumPy is not available or incompatible with PyTorch: {str(e)}. Ensure 'numpy' is in requirements.txt and reinstall dependencies.")
+    st.stop()
+
 # Check if CUDA is available, otherwise use CPU
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 print(f"Using device: {device}")
@@ -173,7 +182,7 @@ def process_uploaded_audio(audio_file):
         st.error(f"Error processing uploaded audio: {str(e)}")
         return None
     finally:
-        if temp_file_path and os.path.exists(temp_file_path) and not st.session_state.get('keep_temp', False):
+        if temp_file_path and os.path.exists(temp_file_path):
             os.remove(temp_file_path)
 
 # Show model information
@@ -221,6 +230,19 @@ def custom_audio_recorder():
     }
     </script>
     <button id="record-btn" onclick="toggleRecording()">Start Recording</button>
+    <style>
+    #record-btn {
+        background-color: #f63366;
+        color: white;
+        border: none;
+        padding: 10px 20px;
+        border-radius: 5px;
+        cursor: pointer;
+    }
+    #record-btn:hover {
+        background-color: #ff0000;
+    }
+    </style>
     """
     return components.html(audio_recorder_html, height=100)
 
@@ -228,18 +250,25 @@ def custom_audio_recorder():
 def display_analysis_results(transcribed_text):
     emotions_dict, top_emotion, emotion_map, sentiment = perform_emotion_detection(transcribed_text)
     is_sarcastic, sarcasm_score = perform_sarcasm_detection(transcribed_text)
-    st.header("Results")
+    st.header("Analysis Results")
     st.text_area("Transcribed Text", transcribed_text, height=100, disabled=True)
-    col1, col2 = st.columns(2)
+    col1, col2 = st.columns([1, 2])
     with col1:
         st.subheader("Sentiment")
-        st.write(f"{sentiment} ({top_emotion})")
-    with col2:
+        sentiment_icon = "👍" if sentiment == "POSITIVE" else "👎" if sentiment == "NEGATIVE" else "😐"
+        st.markdown(f"{sentiment_icon} {sentiment} (Based on {top_emotion})")
         st.subheader("Sarcasm")
-        st.write(f"{'Detected' if is_sarcastic else 'Not Detected'} (Score: {sarcasm_score:.2f})")
-    if emotions_dict:
-        fig = px.bar(x=list(emotions_dict.keys()), y=list(emotions_dict.values()), labels={'x': 'Emotion', 'y': 'Score'})
-        st.plotly_chart(fig)
+        sarcasm_icon = "😏" if is_sarcastic else "😐"
+        st.markdown(f"{sarcasm_icon} {'Detected' if is_sarcastic else 'Not Detected'} (Score: {sarcasm_score:.2f})")
+    with col2:
+        st.subheader("Emotions")
+        if emotions_dict:
+            st.markdown(f"*Dominant:* {emotion_map.get(top_emotion, '❓')} {top_emotion.capitalize()} (Score: {emotions_dict[top_emotion]:.3f})")
+            fig = px.bar(x=list(emotions_dict.keys()), y=list(emotions_dict.values()),
+                         labels={'x': 'Emotion', 'y': 'Score'}, title="Emotion Distribution")
+            st.plotly_chart(fig, use_container_width=True)
+        else:
+            st.write("No emotions detected.")
 
 # Process base64 audio
 def process_base64_audio(base64_data):
@@ -261,27 +290,40 @@ def process_base64_audio(base64_data):
 
 # Main App Logic
 def main():
-    tab1, tab2 = st.tabs(["Upload Audio", "Record Audio"])
+    if 'debug_info' not in st.session_state:
+        st.session_state.debug_info = []
+    tab1, tab2 = st.tabs(["📁 Upload Audio", "🎙 Record Audio"])
     with tab1:
-        audio_file = st.file_uploader("Upload Audio", type=["wav", "mp3", "ogg"])
+        st.header("Upload an Audio File")
+        audio_file = st.file_uploader("Choose an audio file", type=["wav", "mp3", "ogg"])
         if audio_file:
-            st.audio(audio_file)
-            if st.button("Analyze Uploaded Audio"):
-                with st.spinner("Analyzing..."):
-                    temp_path = process_uploaded_audio(audio_file)
-                    if temp_path:
-                        text = transcribe_audio(temp_path)
-                        if text:
-                            display_analysis_results(text)
+            st.audio(audio_file.getvalue())
+            if st.button("Analyze Upload", key="analyze_upload"):
+                with st.spinner("Analyzing audio..."):
+                    temp_audio_path = process_uploaded_audio(audio_file)
+                    if temp_audio_path:
+                        transcribed_text = transcribe_audio(temp_audio_path)
+                        if transcribed_text:
+                            display_analysis_results(transcribed_text)
+                        else:
+                            st.error("Could not transcribe audio. Try clearer audio.")
     with tab2:
+        st.header("Record Your Voice")
+        st.subheader("Browser-Based Recorder")
         audio_data = custom_audio_recorder()
-        if audio_data and st.button("Analyze Recorded Audio"):
-            with st.spinner("Analyzing..."):
-                temp_path = process_base64_audio(audio_data)
-                if temp_path:
-                    text = transcribe_audio(temp_path)
-                    if text:
-                        display_analysis_results(text)
+        if audio_data and st.button("Analyze Recording", key="analyze_rec"):
+            with st.spinner("Processing recording..."):
+                temp_audio_path = process_base64_audio(audio_data)
+                if temp_audio_path:
+                    transcribed_text = transcribe_audio(temp_audio_path)
+                    if transcribed_text:
+                        display_analysis_results(transcribed_text)
+                    else:
+                        st.error("Could not transcribe audio. Speak clearly.")
+        st.subheader("Manual Text Input")
+        manual_text = st.text_area("Enter text to analyze:", placeholder="Type your text...")
+        if st.button("Analyze Text", key="analyze_manual") and manual_text:
+            display_analysis_results(manual_text)
     show_model_info()
 
 if __name__ == "__main__":