Update app.py

app.py

@@ -2,123 +2,101 @@ import streamlit as st
 from transformers import pipeline
 import re
 
-# Load HuggingFace models
-emotion_classifier = pipeline("text-classification", model="bhadresh-savani/distilbert-base-uncased-emotion", top_k=3)
+# Load models
+emotion_classifier = pipeline("text-classification", model="bhadresh-savani/distilbert-base-uncased-emotion", top_k=1)
 intent_classifier = pipeline("zero-shot-classification", model="facebook/bart-large-mnli")
+text_generator = pipeline("text2text-generation", model="google/flan-t5-base", max_new_tokens=100)
 
 # Candidate intents
 candidate_tasks = [
     "change mobile plan",
-    "upgrade phone",
     "top up balance",
     "report service outage",
     "ask for billing support",
-    "request human support",
     "reactivate service",
     "cancel subscription",
-    "check account status"
+    "check account status",
+    "upgrade device"
 ]
 
-# Define emotion priority rules
+# Emotion scoring
 urgent_emotions = {"anger", "annoyance", "disgust", "frustration", "sadness"}
 moderate_emotions = {"confusion", "concern", "nervousness", "fear"}
 
-# Helper: Calculate support score
-def calculate_support_score(text, emotion):
-    score = 0.0
-
-    # Emotion-based scoring
+def get_emotion_score(emotion):
     if emotion in urgent_emotions:
-        score += 0.5
+        return 1.0
     elif emotion in moderate_emotions:
-        score += 0.3
-
-    # Symbol analysis
-    if "!" in text:
-        score += 0.1 * min(text.count("!"), 3)  # max +0.3
-
-    # Keyword triggers
-    urgent_keywords = ["not working", "out of service", "no signal", "urgent", "immediately", "stopped"]
-    for kw in urgent_keywords:
-        if kw in text.lower():
-            score += 0.2
-            break
+        return 0.6
+    else:
+        return 0.2
 
-    return min(score, 1.0)
+def get_content_score(text, top_intents):
+    score = 0.0
+    trigger_words = ["out of service", "urgent", "not working", "stopped", "can't", "immediately"]
+    if any(kw in text.lower() for kw in trigger_words):
+        score += 0.3
+    if any(label in ["top up balance", "reactivate service", "report service outage"] for label in top_intents):
+        score += 0.4
+    if any(label in ["change mobile plan", "ask for billing support"] for label in top_intents):
+        score += 0.2
+    return min(score + 0.1, 1.0)
 
-# Helper: Generate auto-response
-def generate_auto_reply(top_intent):
-    intent_responses = {
-        "change mobile plan": "You can change your mobile plan by visiting your account dashboard or exploring our latest offers here: [link].",
-        "top up balance": "To restore service, please top up your balance through our payment portal: [link].",
-        "check account status": "You can check your account status and active services via our self-service portal: [link].",
-    }
-    return intent_responses.get(top_intent.lower(), "We have received your request and are processing it.")
+def generate_reply(input_text, intent):
+    prompt = f"Generate a polite and helpful customer service reply to this intent: '{intent}'. Customer said: '{input_text}'"
+    return text_generator(prompt)[0]['generated_text']
 
-# App UI
-st.set_page_config(page_title="AI Customer Support Decision Engine", layout="centered")
-st.title("📞 AI Customer Emotion, Intent & Support Decision Analyzer")
+# UI
+st.set_page_config(page_title="Customer Support Assistant", layout="centered")
+st.title("📞 Smart Customer Support Assistant (for Agents Only)")
 
-user_input = st.text_area("Please enter the customer's message or conversation:", height=150)
+user_input = st.text_area("Enter customer's message or complaint:", height=150)
 
-if st.button("Analyze"):
+if st.button("Analyze Message"):
     if user_input.strip() == "":
-        st.warning("Please enter a message to analyze.")
+        st.warning("Please enter a customer message.")
     else:
-        with st.spinner("Analyzing..."):
-
-            # 1. Emotion Analysis
-            emotion_results = emotion_classifier(user_input)
-            top_emotion = emotion_results[0][0]
-            emotion_label = top_emotion['label']
-            emotion_score = top_emotion['score']
+        with st.spinner("Processing..."):
 
-            # 2. Support Score
-            support_score = calculate_support_score(user_input, emotion_label)
+            # Emotion detection
+            emotion_result = emotion_classifier(user_input)
+            emotion_label = emotion_result[0]['label']
+            emotion_score = get_emotion_score(emotion_label)
 
-            # 3. Intent Detection
+            # Intent detection
             intent_result = intent_classifier(user_input, candidate_tasks)
-            top_labels = intent_result['labels']
-            top_scores = intent_result['scores']
-            intent_pairs = list(zip(top_labels, top_scores))
-            filtered_intents = [(label, score) for label, score in intent_pairs if score > 0.15]
-
-            # Display emotion
-            st.subheader("📌 Emotion Analysis")
-            st.write(f"**Primary Emotion**: {emotion_label}")
-            st.write(f"**Confidence Score**: {emotion_score:.2f}")
-
-            # Display support recommendation
-            st.subheader("🛎️ Human Support Priority")
-            st.write(f"**Support Score**: {support_score:.2f}")
-            if support_score >= 0.7:
-                st.error("🔴 Strongly Recommend Human Support")
-            elif support_score >= 0.4:
-                st.warning("🟠 Consider Human Support")
-            else:
-                st.success("🟢 Automated Response is Sufficient")
-
-            # Output decision
-            st.subheader("🧠 System Recommendation")
-            if support_score < 0.4:
-                top_intent = top_labels[0]
-                st.markdown(f"🤖 **Auto-reply Recommendation**: _{top_intent}_")
-                st.info(generate_auto_reply(top_intent))
-            else:
-                st.markdown("👩‍💼 **Customer Intent Summary for Human Agent:**")
+            top_intents = [label for label, score in zip(intent_result['labels'], intent_result['scores']) if score > 0.2]
 
-                if filtered_intents:
-                    st.markdown("### 🔺 Priority Pyramid")
-                    st.markdown(f"- **Primary Intent**: {filtered_intents[0][0]} (confidence: {filtered_intents[0][1]:.2f})")
+            # Content score
+            content_score = get_content_score(user_input, top_intents)
 
-                    if len(filtered_intents) > 1:
-                        st.markdown("### 🔸 Secondary Intents")
-                        for label, score in filtered_intents[1:]:
-                            st.markdown(f"- {label} (confidence: {score:.2f})")
+            # Final decision score
+            final_score = (0.5 * emotion_score) + (0.5 * content_score)
 
-                    st.markdown("### 🔹 Additional Clues")
-                    if "!" in user_input:
-                        st.write("- Urgent tone detected (multiple '!')")
-                    st.write(f"- Detected Emotion: {emotion_label}")
+            st.subheader("🧾 System Summary")
+
+            if final_score < 0.5:
+                st.markdown("### 🟢 This message was handled automatically.")
+                if top_intents:
+                    auto_intent = top_intents[0]
+                    auto_reply = generate_reply(user_input, auto_intent)
+                    st.markdown("#### 🤖 Auto-Response Sent to User:")
+                    st.success(auto_reply)
+                else:
+                    st.info("No clear intent detected. A general auto-reply was used.")
+            else:
+                st.markdown("### 🔴 Human Support Required")
+
+                # Customer Profile Summary
+                st.markdown("#### 👤 Customer Status:")
+                st.write(f"- **Emotion detected**: {emotion_label.capitalize()}")
+                st.write(f"- **Tone**: {'Urgent' if emotion_score > 0.8 else 'Concerned' if emotion_score > 0.5 else 'Calm'}")
+
+                if top_intents:
+                    st.markdown("#### 🧩 Detected Customer Needs:")
+                    for intent in top_intents:
+                        reply = generate_reply(user_input, intent)
+                        st.markdown(f"**• {intent.capitalize()}**")
+                        st.write(f"Recommended Action: {reply}")
                 else:
-                    st.info("No clear customer intents detected.")
+                    st.warning("No clear intent detected. Manual review recommended.")