app.py

import os
import gradio as gr
from llama_index.readers.file import PDFReader
from llama_index.core import VectorStoreIndex
from llama_index.llms.openai import OpenAI


# Set API key (best practice: use HF secrets in actual deployment)
os.environ['OPENAI_API_KEY'] = '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'

# Global vars
query_engine = None
interview_questions = []
resume_summary = ""

# Step 1: Load Resume
def load_resume(file):
    global query_engine, interview_questions, resume_summary

    reader = PDFReader()
    documents = reader.load_data(file=file.name)

    index = VectorStoreIndex.from_documents(documents)
    query_engine = index.as_query_engine()

    # Summarize resume
    resume_summary = query_engine.query("Summarize the key points in this resume")

    # Generate interview questions
    q_prompt = "Generate 5 interview questions based on this resume:\n" + str(resume_summary)
    interview_questions.clear()
    for i in range(5):
        q = query_engine.query(q_prompt + f"\nQuestion {i+1}")
        interview_questions.append(str(q))

    return f"✅ Resume uploaded.\n\n📝 Summary:\n{resume_summary}"

# Step 2: Show Questions
def show_questions():
    if not interview_questions:
        return "❌ Please upload and analyze a resume first."
    return "\n".join([f"{i+1}. {q}" for i, q in enumerate(interview_questions)])

# Step 3: Evaluate Answer
def evaluate_answer(answer):
    if not answer.strip():
        return "⚠️ Please provide an answer."
    word_count = len(answer.split())
    score = min(word_count // 10, 5)
    stars = "⭐" * score
    return f"✅ Answer received.\nScore: {stars} ({score}/5)"

# Step 4: Rate Resume
def rate_resume():
    if not resume_summary:
        return "❌ Upload a resume first."
    rating = query_engine.query("Evaluate and rate the quality of this resume from 1 to 10. Only return the number.")
    return f"📊 Resume Rating: {rating}/10"

# Interface layout
with gr.Blocks() as demo:
    gr.Markdown("# 🤖 Resume Interview Bot\nUpload your resume, get interview questions, answer them, and get feedback!")

    with gr.Row():
        resume_input = gr.File(label="📄 Upload Resume (.pdf)", file_types=[".pdf"])
        resume_status = gr.Textbox(label="Resume Summary", lines=6)
    gr.Button("Analyze Resume").click(load_resume, inputs=resume_input, outputs=resume_status)

    gr.Markdown("### 🎯 Generated Interview Questions")
    question_box = gr.Textbox(label="Questions", lines=7)
    gr.Button("Get Questions").click(show_questions, outputs=question_box)

    gr.Markdown("### 🗣️ Answer a Question")
    answer_input = gr.Textbox(label="Your Answer")
    answer_result = gr.Textbox(label="Feedback", interactive=False)
    gr.Button("Submit Answer").click(evaluate_answer, inputs=answer_input, outputs=answer_result)

    gr.Markdown("### 📈 Resume Quality Rating")
    rating_output = gr.Textbox(label="Rating", interactive=False)
    gr.Button("Rate Resume").click(rate_resume, outputs=rating_output)

demo.launch()