Update app.py

app.py

@@ -2,7 +2,7 @@ import gradio as gr
 from huggingface_hub import InferenceClient
 import re
 
-# ---- Load and parse questions from knowledge.txt ----
+# ---- Load and parse questions ----
 def load_questions(file_path):
     with open(file_path, 'r') as f:
         data = f.read()
@@ -20,88 +20,104 @@ def load_questions(file_path):
 
 all_questions = load_questions('knowledge.txt')
 
-# ---- Tagging interview questions ----
+# ---- Categorize questions ----
 questions_by_type = {
     'Technical': [q for q in all_questions if any(keyword in q['question'].lower() for keyword in [
-    'function', 'linked list', 'url', 'rest', 'graphql', 'garbage', 'cap theorem', 'sql', 'hash table',
-    'stack', 'queue', 'recursion', 'reverse', 'bfs', 'dfs', 'time complexity', 'binary search tree',
-    'web application', 'chat system', 'load balancing', 'caching', 'normalization', 'acid', 'indexing',
-    'sql injection', 'https', 'xss', 'hash', 'vulnerabilities'])],
-    
+        'function', 'linked list', 'url', 'rest', 'graphql', 'garbage', 'cap theorem', 'sql', 'hash table',
+        'stack', 'queue', 'recursion', 'reverse', 'bfs', 'dfs', 'time complexity', 'binary search tree',
+        'web application', 'chat system', 'load balancing', 'caching', 'normalization', 'acid', 'indexing',
+        'sql injection', 'https', 'xss', 'hash', 'vulnerabilities'])],
+
     'Competency-Based Interview': [q for q in all_questions if any(keyword in q['question'].lower() for keyword in [
-    "Debugging","Learning Fast","Deadlines","Teamwork","Leadership","Mistake Recovery","Conflict Management","Decision Making"])],
-    
+        "Debugging","Learning Fast","Deadlines","Teamwork","Leadership","Mistake Recovery","Conflict Management","Decision Making"])],
+
     'Case': [q for q in all_questions if any(keyword in q['question'].lower() for keyword in [
-    "A/B Testing","Financial Modeling","Automation","Data Analysis","Regression","Business Opportunity","Stakeholder Alignment"])] 
+        "A/B Testing","Financial Modeling","Automation","Data Analysis","Regression","Business Opportunity","Stakeholder Alignment"])]
 }
 
-# ---- Hugging Face Client ----
+# ---- HF Client ----
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
-# ---- Set interview type ----
+# ---- Interview type selection ----
 def set_type(choice, user_profile):
-    user_profile["interview_type"] = choice
+    user_profile.update({
+        "interview_type": choice,
+        "state": "idle"
+    })
     return "Great! What’s your background and what field/role are you aiming for?", user_profile
 
-# ---- Save background ----
+# ---- Background ----
 def save_background(info, user_profile):
     user_profile["field"] = info
-    return "Awesome! Type 'start' below to begin your interview.", user_profile
+    return "Awesome! Type 'start' below when you're ready to begin your interview.", user_profile
 
-# ---- Main respond logic ----
+# ---- Main Respond Logic ----
 def respond(message, chat_history, user_profile):
     message = message.strip().lower()
 
+    # Safety: Ensure interview type + field are set
     if not user_profile.get("interview_type") or not user_profile.get("field"):
-        bot_msg = "Please finish steps 1 and 2 before starting the interview."
-        chat_history.append((message, bot_msg))
+        chat_history.append((message, "Please finish steps 1 and 2 first."))
         return chat_history
 
-    # Start interview
-    if message == 'start':
+    state = user_profile.get("state", "idle")
+
+    # Handle start
+    if message == 'start' and state == 'idle':
         interview_type = user_profile['interview_type']
         selected_questions = questions_by_type.get(interview_type, [])
-        user_profile['questions'] = selected_questions
-        user_profile['current_q'] = 0
-        user_profile['user_answers'] = []
+
         if not selected_questions:
-            bot_msg = "No questions available for this interview type."
-        else:
-            bot_msg = f"First question: {selected_questions[0]['question']}"
-        chat_history.append((message, bot_msg))
+            chat_history.append((message, "No questions found for this interview type."))
+            return chat_history
+
+        user_profile.update({
+            'questions': selected_questions,
+            'current_q': 0,
+            'user_answers': [],
+            'state': 'running'
+        })
+
+        intro_msg = f"Great, let's begin your {interview_type} interview for the {user_profile['field']} role.\nI’ll be asking you a few questions to assess your knowledge. Take your time!"
+        first_q = selected_questions[0]['question']
+        chat_history.append((message, intro_msg))
+        chat_history.append(("", f"First question: {first_q}"))
         return chat_history
 
-    # Stop interview early
-    if message == 'stop':
-        user_profile['questions'] = []
-        bot_msg = "Interview stopped. Type 'start' to begin again or 'feedback' to get feedback."
-        chat_history.append((message, bot_msg))
+    # Handle stop
+    if message == 'stop' and state == 'running':
+        user_profile["state"] = "stopped"
+        chat_history.append((message, "Interview stopped. Type 'start' to begin again or 'feedback' if you'd like an analysis (if available)."))
         return chat_history
 
     # Handle feedback
     if message == 'feedback':
-        if not user_profile.get("questions"):
-            chat_history.append((message, "No completed interview found. Please complete the interview first."))
+        if state != 'completed':
+            chat_history.append((message, "Feedback is only available after completing the interview."))
             return chat_history
+
         feedback = generate_feedback(user_profile)
         chat_history.append((message, feedback))
         return chat_history
 
-    # Handle interview questions
-    if user_profile.get("questions"):
+    # Handle interview question flow
+    if state == 'running':
         q_index = user_profile['current_q']
-        if q_index < len(user_profile['questions']):
+        questions = user_profile['questions']
+
+        if q_index < len(questions):
             user_profile['user_answers'].append(message)
             user_profile['current_q'] += 1
 
-            if user_profile['current_q'] < len(user_profile['questions']):
-                next_question = user_profile['questions'][user_profile['current_q']]['question']
+            if user_profile['current_q'] < len(questions):
+                next_question = questions[user_profile['current_q']]['question']
                 chat_history.append((message, f"Next question: {next_question}"))
             else:
-                chat_history.append((message, "Interview complete! Type 'feedback' if you'd like an analysis."))
+                user_profile['state'] = 'completed'
+                chat_history.append((message, "✅ Interview complete! You can type 'feedback' to receive an analysis."))
             return chat_history
 
-    # General fallback chat using LLM if no interview ongoing
+    # Handle idle/stopped fallback small talk using LLM
     messages = [
         {"role": "system", "content": f"You are a professional interviewer conducting a {user_profile['interview_type']} interview for a candidate in the {user_profile['field']} field."}
     ]
@@ -115,7 +131,7 @@ def respond(message, chat_history, user_profile):
     chat_history.append((message, bot_msg))
     return chat_history
 
-# ---- Feedback generator ----
+# ---- Feedback ----
 def generate_feedback(user_profile):
     feedback = []
     questions = user_profile.get('questions', [])
@@ -136,9 +152,9 @@ def generate_feedback(user_profile):
         feedback.append(fb)
     return "\n".join(feedback)
 
-# ---- Gradio interface ----
+# ---- Gradio UI ----
 with gr.Blocks() as demo:
-    user_profile = gr.State({"interview_type": "", "field": ""})
+    user_profile = gr.State({"interview_type": "", "field": "", "state": "idle"})
     chat_history = gr.State([])
 
     gr.Markdown("# 🎤 Welcome to Intervu")