Update app.py

app.py

@@ -29,39 +29,63 @@ if uploaded_file:
     audio.export(audio_path, format="wav")
 
     try:
-        pass  # Placeholder to avoid empty try
-    except Exception as e:
-        st.error(f"❌ Processing failed: {e}")
-        tab1, tab2, tab3, tab4 = st.tabs(["📝 Transcript", "📋 Summary", "💬 Emotions", "📈 Trends"])
-
-        with tab1:
-            st.subheader("📝 Speaker-Simulated Transcript")
-            st.markdown(diarized_transcript, unsafe_allow_html=True)
-
-        with tab2:
-            st.subheader("📋 Contextual Summary")
-            st.write(summary[0]["summary_text"])
-
-        with tab3:
-            st.subheader("💬 Emotional Insights (Overall)")
-            if 'emotion_scores' in locals():
-                for emo in emotion_scores[0]:
-                    st.write(f"{emo['label']}: {round(emo['score']*100, 2)}%")
-            else:
-                st.write("No emotional data to display.")
+        # Transcribe
+        st.info("🔄 Transcribing with Whisper (mixed-language support)...")
+        asr = pipeline("automatic-speech-recognition", model="openai/whisper-large")
+        result = asr(audio_path, return_timestamps=True, generate_kwargs={"language": "<|en|>"})
+        raw_transcript = result.get("text", "")
+
+        if not raw_transcript:
+            st.error("❌ Could not generate a transcript. Please try a different audio.")
+        else:
+            # Simulated Speaker Diarization
+            st.info("🗣️ Simulating speaker separation...")
+            sentences = re.split(r'(?<=[.?!])\s+', raw_transcript)
+            diarized_transcript = ""
+            for idx, sentence in enumerate(sentences):
+                speaker = "Speaker 1" if idx % 2 == 0 else "Speaker 2"
+                diarized_transcript += f"{speaker}: {sentence}\n\n"
+
+            # Summarization
+            st.info("📋 Summarizing conversation...")
+            summarizer = pipeline("summarization", model="philschmid/bart-large-cnn-samsum")
+            summary = summarizer(raw_transcript, max_length=256, min_length=60, do_sample=False)
+
+            # Emotion tagging
+            st.info("🎭 Extracting emotional tones...")
+            emotion_model = pipeline("text-classification", model="j-hartmann/emotion-english-distilroberta-base", return_all_scores=True)
+            emotion_scores = emotion_model(raw_transcript)
+
+            # Layout with Tabs
+
+            with tab1:
+                st.subheader("📝 Speaker-Simulated Transcript")
+                st.markdown(diarized_transcript, unsafe_allow_html=True)
+
+            with tab2:
+                st.subheader("📋 Contextual Summary")
+                st.write(summary[0]["summary_text"])
+
+            with tab3:
+                st.subheader("💬 Emotional Insights (Overall)")
+                if 'emotion_scores' in locals():
+                    for emo in emotion_scores[0]:
+                        st.write(f"{emo['label']}: {round(emo['score']*100, 2)}%")
+                else:
+                    st.write("No emotional data to display.")
+                
+            with tab4:
+                st.subheader("📈 Emotional Trends Over Time")
+
+                session_dates = ["2024-04-01", "2024-04-08", "2024-04-15", "2024-04-22"]
+                anxiety_scores = [70, 65, 55, 40]
+                sadness_scores = [30, 20, 25, 15]
             
-        with tab4:
-            st.subheader("📈 Emotional Trends Over Time")
-
-            session_dates = ["2024-04-01", "2024-04-08", "2024-04-15", "2024-04-22"]
-            anxiety_scores = [70, 65, 55, 40]
-            sadness_scores = [30, 20, 25, 15]
-        
-            fig = go.Figure()
-            fig.add_trace(go.Scatter(x=session_dates, y=anxiety_scores, mode='lines+markers', name='Anxiety'))
-            fig.add_trace(go.Scatter(x=session_dates, y=sadness_scores, mode='lines+markers', name='Sadness'))
-            fig.update_layout(title='Emotional Trends', xaxis_title='Date', yaxis_title='Score (%)')
-            st.plotly_chart(fig)
+                fig = go.Figure()
+                fig.add_trace(go.Scatter(x=session_dates, y=anxiety_scores, mode='lines+markers', name='Anxiety'))
+                fig.add_trace(go.Scatter(x=session_dates, y=sadness_scores, mode='lines+markers', name='Sadness'))
+                fig.update_layout(title='Emotional Trends', xaxis_title='Date', yaxis_title='Score (%)')
+                st.plotly_chart(fig)
 
             # Export Button
             st.subheader("📥 Export Session Report")
@@ -116,3 +140,39 @@ if uploaded_file:
     finally:
         if os.path.exists(audio_path):
             os.remove(audio_path)
+
+tab1, tab2, tab3, tab4 = st.tabs(["📝 Transcript", "📋 Summary", "💬 Emotions", "📈 Trends"])
+
+with tab1:
+    st.subheader("📝 Speaker-Simulated Transcript")
+    if 'diarized_transcript' in locals():
+        st.markdown(diarized_transcript, unsafe_allow_html=True)
+    else:
+        st.warning("Transcript not available.")
+
+with tab2:
+    st.subheader("📋 Contextual Summary")
+    if 'summary' in locals():
+        st.write(summary[0]["summary_text"])
+    else:
+        st.warning("Summary not available.")
+
+with tab3:
+    st.subheader("💬 Emotional Insights (Overall)")
+    if 'emotion_scores' in locals():
+        for emo in emotion_scores[0]:
+            st.write(f"{emo['label']}: {round(emo['score']*100, 2)}%")
+    else:
+        st.warning("No emotional data to display.")
+
+with tab4:
+    st.subheader("📈 Emotional Trends Over Time")
+    session_dates = ["2024-04-01", "2024-04-08", "2024-04-15", "2024-04-22"]
+    anxiety_scores = [70, 65, 55, 40]
+    sadness_scores = [30, 20, 25, 15]
+
+    fig = go.Figure()
+    fig.add_trace(go.Scatter(x=session_dates, y=anxiety_scores, mode='lines+markers', name='Anxiety'))
+    fig.add_trace(go.Scatter(x=session_dates, y=sadness_scores, mode='lines+markers', name='Sadness'))
+    fig.update_layout(title='Emotional Trends', xaxis_title='Date', yaxis_title='Score (%)')
+    st.plotly_chart(fig)