app.py

import os
import io
import streamlit as st
import docx
import docx2txt
import tempfile
import time
import re
import concurrent.futures
from functools import lru_cache
from transformers import pipeline

# ... [Keep your existing configurations and constants] ...

#####################################
# Preload Models - Optimized with DistilBART
#####################################
@st.cache_resource(show_spinner=True)
def load_models():
    """Load optimized models at startup"""
    with st.spinner("Loading AI models..."):
        # Use smaller, faster model
        models = {
            'summarizer': pipeline(
                "summarization", 
                model="distilbart-base-cs",  # Faster than BART
                max_length=300,              # Reduced context window
                truncation=True,
                num_return_sequences=1
            )
        }
        return models

models = load_models()

#####################################
# Function: Extract Text from File - Optimized
#####################################
@lru_cache(maxsize=16, typed=False)
def extract_text_from_file(file_obj):
    """Optimized text extraction with early exit"""
    filename = file_obj.name
    ext = os.path.splitext(filename)[1].lower()
    text = ""
    MAX_TEXT = 15000  # Reduced processing limit
    
    try:
        if ext == ".docx":
            doc = docx.Document(file_obj)
            # Only process first 50 paragraphs (approx 10 pages)
            text = "\n".join(para.text for para in doc.paragraphs[:50] if para.text.strip())[:MAX_TEXT]
        elif ext == ".doc":
            # Direct conversion using docx2txt
            text = docx2txt.process(file_obj.stream.read())[:MAX_TEXT]
        elif ext == ".txt":
            text = file_obj.read().decode("utf-8")[:MAX_TEXT]
    except Exception as e:
        text = f"Error: {str(e)}"
    
    return text

# ... [Keep your existing extraction functions] ...

#####################################
# Optimized Summarization
#####################################
def summarize_resume_text(resume_text):
    """Faster summarization with input truncation"""
    start_time = time.time()
    
    # Truncate text for summarization
    text_to_summarize = resume_text[:1024]
    base_summary = models['summarizer'](
        text_to_summarize, 
        max_length=150,  # Smaller summary
        truncation=True
    )[0]['summary_text']
    
    # Parallel extraction with thread pool
    with concurrent.futures.ThreadPoolExecutor() as executor:
        # Reduced number of parallel tasks
        name_future = executor.submit(extract_name, resume_text[:200])
        age_future = executor.submit(extract_age, resume_text)
        industry_future = executor.submit(extract_industry, resume_text, base_summary)
        
        # Get results
        name = name_future.result()
        age = age_future.result()
        industry = industry_future.result()
        skills, work = extract_skills_and_work(resume_text)  # Sequential
    
    # Format summary (simplified)
    return f"**Name**: {name}\n**Age**: {age}\n**Industry**: {industry}\n\n{base_summary}", 0.1

# ... [Keep your scoring and feedback functions] ...

#####################################
# Optimized Streamlit Interface
#####################################
st.title("Google Resume Analyzer")

# Initialize session state properly
if 'progress' not in st.session_state:
    st.session_state['progress'] = 0
if 'last_update' not in st.session_state:
    st.session_state['last_update'] = time.time()

uploaded_file = st.file_uploader("Upload your resume", type=["docx", "doc", "txt"])

if uploaded_file and st.button("Analyze"):
    # Use exponential backoff for progress updates
    progress_interval = 0.1
    max_retries = 10
    
    with st.spinner():
        # Step 1: Text extraction
        st.session_state['progress'] = 33
        resume_text = extract_text_from_file(uploaded_file)
        if "Error" in resume_text:
            st.error(resume_text)
            st.session_state['progress'] = 100
            continue
        
        # Step 2: Summarization
        st.session_state['progress'] = 66
        summary, _ = summarize_resume_text(resume_text)
        
        # Step 3: Scoring
        st.session_state['progress'] = 100
        
        # Display results
        st.subheader("Analysis Complete!")
        st.markdown(summary)
        
        # Display scores
        overall_score, category_scores, score_breakdown = calculate_google_match_score(summary)
        show_score(overall_score)
        
        # Display feedback
        feedback, _ = generate_template_feedback(category_scores)
        st.markdown(feedback)

# Progress bar implementation
if st.session_state['progress'] < 100:
    st.progress(st.session_state['progress'], 100)
    time.sleep(progress_interval)

def show_score(score):
    """Display score with appropriate formatting"""
    score_percent = int(score * 100)
    if score >= 0.85:
        st.success(f"**Match Score**: {score_percent}% 🌟")
    elif score >= 0.70:
        st.success(f"**Match Score**: {score_percent}% ✅")
    elif score >= 0.50:
        st.warning(f"**Match Score**: {score_percent}% ⚠️")
    else:
        st.error(f"**Match Score**: {score_percent}% 🔍")

# ... [Keep your remaining functions] ...