app.py

import streamlit as st
from langchain_google_genai import ChatGoogleGenerativeAI
import re
from langchain_huggingface import HuggingFaceEndpoint
from langchain.chains import LLMChain
from langchain_core.prompts import PromptTemplate
import os

def generate_question(role, topic, difficulty_level):
    prompt = f"Generate an interview question for the role of {role} on the topic of {topic} with difficulty level {difficulty_level}. Just ask the question, that's it. No explanation is required."
    template = """Question: {question}
                  Answer: 
                """
    llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash-lite", google_api_key=st.secrets["GOOGLE_API_KEY"])
    response = llm.invoke(prompt)
    response = response.content
    return response
    

def evaluate_answer(question, user_answer):
    prompt = f"Question: {question}\nUser's Answer: {user_answer}\nEvaluate the answer, give a score out of 100, and provide feedback. Also, provide the best possible answer."
    llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash-lite", google_api_key=st.secrets["GOOGLE_API_KEY"])
    response = llm.invoke(prompt)
   
    evaluation = response.content
    # Extract score and feedback from the evaluation
     # Extract score using regular expressions
    score_match = re.search(r'(\d+)/100', evaluation)
    score = int(score_match.group(1)) if score_match else 0

    # Extract feedback
    feedback = evaluation.split('\n', 1)[1] if '\n' in evaluation else evaluation
    return score, feedback

def generate_report():
    st.write("### Interview Report")
    for i in range(st.session_state['total_questions']):
        st.write(f"**Question {i+1}:** {st.session_state['questions'][i]}")
        st.write(f"**Your Answer:** {st.session_state['answers'][i]}")
        st.write(f"**Score:** {st.session_state['scores'][i]}")
        st.write(f"**Feedback:** {st.session_state['feedback'][i]}")
        st.write("---")

# Initialize session state
if 'questions' not in st.session_state:
    st.session_state['questions'] = []
if 'answers' not in st.session_state:
    st.session_state['answers'] = []
if 'feedback' not in st.session_state:
    st.session_state['feedback'] = []
if 'scores' not in st.session_state:
    st.session_state['scores'] = []
if 'current_question' not in st.session_state:
    st.session_state['current_question'] = 0
if 'total_questions' not in st.session_state:
    st.session_state['total_questions'] = 10
if 'question_answered' not in st.session_state:
    st.session_state['question_answered'] = False
if 'interview_started' not in st.session_state:
    st.session_state['interview_started'] = False

st.title("Mock Interview Bot")

if not st.session_state['interview_started']:
    roles_and_topics = {      
  
    "Front-End Developer": ["HTML/CSS", "JavaScript and Frameworks (React, Angular, Vue.js)", "Responsive Design", "Browser Compatibility"],
    "Back-End Developer": ["Server-Side Languages (Node.js, Python, Ruby, PHP)", "Database Management (SQL, NoSQL)", "API Development", "Server and Hosting Management"],
    "Full-Stack Developer": ["Combination of Front-End and Back-End Topics", "Integration of Systems", "DevOps Basics"],
    "Mobile Developer": ["Android Development (Java, Kotlin)", "iOS Development (Swift, Objective-C)", "Cross-Platform Development (Flutter, React Native)"],
    "Data Scientist": ["Statistical Analysis", "Machine Learning Algorithms", "Data Wrangling and Cleaning", "Data Visualization"],
    "Data Analyst": ["Data Collection and Processing", "SQL and Database Querying", "Data Visualization Tools (Tableau, Power BI)", "Basic Statistics"],
    "Machine Learning Engineer": ["Supervised and Unsupervised Learning", "Model Deployment", "Deep Learning", "Natural Language Processing"],
    "DevOps Engineer": ["Continuous Integration/Continuous Deployment (CI/CD)", "Containerization (Docker, Kubernetes)", "Infrastructure as Code (Terraform, Ansible)", "Cloud Platforms (AWS, Azure, Google Cloud)"],
    "Cloud Engineer": ["Cloud Architecture", "Cloud Services (Compute, Storage, Networking)", "Security in the Cloud", "Cost Management"],
    "Cybersecurity Analyst": ["Threat Detection and Mitigation", "Security Protocols and Encryption", "Network Security", "Incident Response"],
    "Penetration Tester": ["Vulnerability Assessment", "Ethical Hacking Techniques", "Security Tools (Metasploit, Burp Suite)", "Report Writing and Documentation"],
    "Project Manager": ["Project Planning and Scheduling", "Risk Management", "Agile and Scrum Methodologies", "Stakeholder Communication"],
    "UX/UI Designer": ["User Research", "Wireframing and Prototyping", "Design Principles", "Usability Testing"],
    "Quality Assurance (QA) Engineer": ["Testing Methodologies", "Automation Testing", "Bug Tracking", "Performance Testing"],
    "Blockchain Developer": ["Blockchain Fundamentals", "Smart Contracts", "Cryptographic Algorithms", "Decentralized Applications (DApps)"],
    "Digital Marketing Specialist": ["SEO/SEM", "Social Media Marketing", "Content Marketing", "Analytics and Reporting"],
    "AI Research Scientist": ["AI Theory", "Algorithm Development", "Neural Networks", "Natural Language Processing"],
    "AI Engineer": ["AI Model Deployment", "Machine Learning Engineering", "Deep Learning", "AI Tools and Frameworks"],
    "Generative AI Specialist (GenAI)": ["Generative Models", "GANs (Generative Adversarial Networks)", "Creative AI Applications", "Ethics in AI"],
    "Generative Business Intelligence Specialist (GenBI)": ["Automated Data Analysis", "Business Intelligence Tools", "Predictive Analytics", "AI in Business Strategy"]


    }




    
    role = st.selectbox('Select Role', list(roles_and_topics.keys()))
    topic = st.selectbox('Select Topic', roles_and_topics[role])
    difficulty_level = st.selectbox("Select difficulty level:", ["Easy", "Medium", "Hard"])

    if st.button("Start Interview"):
        if role and topic and difficulty_level:
            st.session_state['questions'] = [generate_question(role, topic, difficulty_level) for _ in range(st.session_state['total_questions'])]
            st.session_state['current_question'] = 0
            st.session_state['interview_started'] = True
            st.session_state['question_answered'] = False

if st.session_state['interview_started']:
    current_question = st.session_state['current_question']
    if current_question < st.session_state['total_questions']:
        st.write(f"Question {current_question + 1}: {st.session_state['questions'][current_question]}")
        
        if not st.session_state['question_answered']:
            answer = st.text_area("Your Answer:", key=f"answer_{current_question}")
            if st.button("Submit Answer"):
                if answer:
                    st.session_state['answers'].append(answer)
                    score, feedback = evaluate_answer(st.session_state['questions'][current_question], answer)
                    st.session_state['scores'].append(score)
                    st.session_state['feedback'].append(feedback)
                    st.session_state['question_answered'] = True
                    st.write(f"Score: {score}")
                    st.write(f"Feedback: {feedback}")
        
        if st.session_state['question_answered']:
            if st.button("Next Question"):
                st.session_state['current_question'] += 1
                st.session_state['question_answered'] = False
    else:
        st.write("Interview Complete! Generating Report...")
        generate_report()