app_new.py

"""
Web Interface for Job Application AI Agent - React Frontend Version

This module provides a Flask web application for the job application AI agent
with a modern React frontend and REST API endpoints.
"""
# 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
import os
import logging
import json
import uuid
import time
import sys
import importlib
from datetime import datetime
from pathlib import Path
from flask import Flask, render_template, request, redirect, url_for, flash, send_file, session, jsonify
from flask_cors import CORS
import pandas as pd
import zipfile
import io

from job_apply_ai.scraper.linkedin import LinkedInScraper
from job_apply_ai.cv_modifier.cv_analyzer import CVAnalyzer, CVModifier, batch_process_jobs
from job_apply_ai.utils.helpers import ensure_directory_exists
from CV_Tailor.ai_cv_tailor import AILangGraphTailor

# CV Analyzer Module
from job_apply_ai.analyzer.ats_evaluator import ATSEvaluator
import docx

# Changes made to add the Apify Actor
# Added Dependencies
import urllib.parse
from apify_client import ApifyClient

# Configure logging
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)

# Initialize Flask app
app = Flask(__name__, static_folder='static', static_url_path='/static', template_folder='templates')
app.secret_key = os.environ.get('SECRET_KEY', 'dev_key_for_testing')
app.config['SESSION_COOKIE_NAME'] = f"job_apply_ai_session_{int(time.time())}"
app.config['JSON_SORT_KEYS'] = False

# Enable CORS for API endpoints
CORS(app, resources={r"/api/*": {"origins": "*"}})

# Keep runtime files local to the project unless overridden by env var.
runtime_root = os.environ.get('JOB_APPLY_AI_DATA_DIR', os.path.join(os.getcwd(), '.runtime'))
if not os.path.isabs(runtime_root):
    runtime_root = os.path.abspath(runtime_root)
app.config['UPLOAD_FOLDER'] = os.path.join(runtime_root, 'uploads')
ensure_directory_exists(app.config['UPLOAD_FOLDER'])

# Create output directories
app.config['CV_OUTPUT_DIR'] = os.path.join(app.config['UPLOAD_FOLDER'], 'cvs')
app.config['JOBS_OUTPUT_DIR'] = os.path.join(app.config['UPLOAD_FOLDER'], 'jobs')
app.config['STATE_DIR'] = os.path.join(app.config['UPLOAD_FOLDER'], 'session_state')
ensure_directory_exists(app.config['CV_OUTPUT_DIR'])
ensure_directory_exists(app.config['JOBS_OUTPUT_DIR'])
ensure_directory_exists(app.config['STATE_DIR'])

# Define where the system's Jinja ATS template lives
app.config['ATS_TEMPLATE_PATH'] = os.path.join(os.getcwd(), 'data', 'ats_template.docx')

app.config['SESSION_TYPE'] = 'filesystem'

SUPPORTED_TAILORING_MODES = {'local', 'api'}
SUPPORTED_LLM_PROVIDERS = {'ollama', 'groq', 'grok', 'openai'}


def _parse_bool(value, default=False):
    if value is None:
        return default
    if isinstance(value, bool):
        return value
    return str(value).strip().lower() in {'1', 'true', 'yes', 'on'}

def _session_state_path(state_id):
    return os.path.join(app.config['STATE_DIR'], f"{state_id}.json")

def _get_tailoring_mode():
    """Resolve active tailoring mode from session, then environment."""
    mode = (session.get('tailoring_mode') or '').strip().lower()
    if mode in SUPPORTED_TAILORING_MODES:
        return mode
    env_mode = (os.environ.get('CV_TAILORING_MODE', 'local') or 'local').strip().lower()
    if env_mode not in SUPPORTED_TAILORING_MODES:
        env_mode = 'local'
    return env_mode

def _set_tailoring_mode(mode):
    mode = (mode or '').strip().lower()
    if mode in SUPPORTED_TAILORING_MODES:
        session['tailoring_mode'] = mode
        return mode
    return _get_tailoring_mode()


def _get_llm_provider():
    provider = (session.get('llm_provider') or '').strip().lower()
    if provider in SUPPORTED_LLM_PROVIDERS:
        return provider
    env_provider = (os.environ.get('LLM_PROVIDER', 'ollama') or 'ollama').strip().lower()
    if env_provider not in SUPPORTED_LLM_PROVIDERS:
        env_provider = 'ollama'
    return env_provider


def _set_llm_provider(provider):
    provider = (provider or '').strip().lower()
    if provider not in SUPPORTED_LLM_PROVIDERS:
        provider = _get_llm_provider()
    session['llm_provider'] = provider
    # Keep compatibility with integrations that read environment variables.
    os.environ['LLM_PROVIDER'] = provider
    return provider


def _is_summary_tailoring_enabled():
    if 'enable_professional_summary' in session:
        return _parse_bool(session.get('enable_professional_summary'))
    return _parse_bool(os.environ.get('CV_ENABLE_SUMMARY_TAILORING', '0'))


def _set_summary_tailoring_enabled(enabled):
    enabled_bool = _parse_bool(enabled)
    session['enable_professional_summary'] = enabled_bool
    os.environ['CV_ENABLE_SUMMARY_TAILORING'] = '1' if enabled_bool else '0'
    return enabled_bool


def _is_cover_letter_enabled():
    if 'include_cover_letters' in session:
        return _parse_bool(session.get('include_cover_letters'))
    return _parse_bool(os.environ.get('API_INCLUDE_COVER_LETTERS', '0'))


def _set_cover_letter_enabled(enabled):
    enabled_bool = _parse_bool(enabled)
    session['include_cover_letters'] = enabled_bool
    os.environ['API_INCLUDE_COVER_LETTERS'] = '1' if enabled_bool else '0'
    return enabled_bool

def _sanitize_filename(text):
    """Sanitize text to be safe for use in filenames."""
    if not text:
        return 'file'
    # Replace problematic characters with underscores
    sanitized = text.replace('/', '_').replace('\\', '_').replace(' ', '_')
    # Remove any other special characters that could cause issues
    sanitized = ''.join(c for c in sanitized if c.isalnum() or c == '_')
    # Limit length
    return sanitized[:20]

def _clear_job_context(keep_cv_template=True):
    """Clear prior search/CV generation state for a fresh workflow."""
    state_id = session.pop('jobs_state_id', None)
    if state_id:
        state_path = _session_state_path(state_id)
        if os.path.exists(state_path):
            try:
                os.remove(state_path)
            except OSError:
                pass

    for key in [
        'jobs_file', 'excel_filename', 'generated_cvs', 'successful_jobs',
        'failed_jobs', 'current_cv', 'current_cv_filename',
    ]:
        session.pop(key, None)

    if not keep_cv_template:
        session.pop('cv_template', None)

def _save_processed_jobs(processed_jobs):
    state_id = str(uuid.uuid4())
    state_path = _session_state_path(state_id)
    with open(state_path, 'w', encoding='utf-8') as f:
        json.dump(processed_jobs, f, ensure_ascii=False)
    session['jobs_state_id'] = state_id
    return state_id

def _load_processed_jobs():
    state_id = session.get('jobs_state_id')
    if not state_id:
        return []
    state_path = _session_state_path(state_id)
    if not os.path.exists(state_path):
        return []
    try:
        with open(state_path, 'r', encoding='utf-8') as f:
            data = json.load(f)
            return data if isinstance(data, list) else []
    except Exception as e:
        logger.error(f"Failed to load session job state: {str(e)}")
        return []

def _build_professional_summary(job, matched_categories):
    """Build a concise, professional summary tailored to role and extracted skills."""
    job_title = (job.get('title') or 'Professional').strip()
    company = (job.get('company') or 'your target company').strip()
    flat_skills = []
    for skills in (matched_categories or {}).values():
        for s in skills or []:
            skill = str(s).strip()
            if skill:
                flat_skills.append(skill)
    deduped = []
    seen = set()
    for skill in flat_skills:
        key = skill.lower()
        if key not in seen:
            seen.add(key)
            deduped.append(skill)
    top_skills = deduped[:5]
    if top_skills:
        skills_text = ", ".join(top_skills)
        return (
            f"{job_title} professional with hands-on experience in {skills_text}. "
            f"Delivers reliable, scalable outcomes through strong collaboration, ownership, "
            f"and structured problem-solving. Ready to contribute immediate impact at {company}."
        )
    return (
        f"{job_title} professional focused on delivering measurable outcomes through "
        f"technical execution, collaboration, and continuous improvement. "
        f"Motivated to contribute meaningful impact at {company}."
    )

# Addition to optimize CV Analyzer
def _enrich_job_with_skills(job, analyzer=None):
    """
    Just-In-Time (JIT) processing: Runs the CVAnalyzer only if skills are missing.
    """
    # If it already has categories, skip the heavy NLP processing
    if 'matched_categories' in job and job['matched_categories']:
        return job
        
    if not analyzer:
        analyzer = CVAnalyzer()
        
    description = job.get('description', '')
    if description and description != "Description empty.":
        logger.info(f"JIT Analysis running for job: {job.get('title')}")
        matched_skills, matched_requirements, matched_categories = analyzer.extract_skills_from_description(description)
        job['matched_skills'] = matched_skills
        job['matched_requirements'] = matched_requirements
        job['matched_categories'] = matched_categories
    else:
        logger.warning(f"Skipping JIT Analysis for {job.get('title')} - No description.")
        job['matched_skills'] = []
        job['matched_requirements'] = []
        job['matched_categories'] = {}
        
    return job

# ==================== React Frontend Routes ====================

@app.route('/app')
@app.route('/app/<path:path>')
def react_app(path=''):
    """Serve React app for all routes under /app"""
    react_build_path = os.path.join(os.path.dirname(__file__), '../../frontend/dist')
    if os.path.exists(react_build_path):
        index_path = os.path.join(react_build_path, 'index.html')
        if os.path.exists(index_path):
            return send_file(index_path)
    # Fallback to template if React build doesn't exist
    return render_template('index.html')

@app.route('/')
def index():
    """Render the home page."""
    return render_template('index.html')

# ==================== REST API Endpoints for React ====================

@app.route('/api/health', methods=['GET'])
def api_health():
    """Health check endpoint."""
    return jsonify({'status': 'ok', 'version': '2.0.0'})

@app.route('/api/config', methods=['GET'])
def api_config():
    """Get current app configuration."""
    return jsonify({
        'tailoring_mode': _get_tailoring_mode(),
        'llm_provider': _get_llm_provider(),
        'enable_professional_summary': _is_summary_tailoring_enabled(),
        'include_cover_letters': _is_cover_letter_enabled(),
        'supported_modes': list(SUPPORTED_TAILORING_MODES),
        'supported_providers': list(SUPPORTED_LLM_PROVIDERS),
    })

@app.route('/api/set-tailoring-mode', methods=['POST'])
def api_set_tailoring_mode():
    """Set tailoring mode via API."""
    data = request.get_json()
    mode = data.get('mode', 'local')
    new_mode = _set_tailoring_mode(mode)
    return jsonify({'success': True, 'mode': new_mode})


@app.route('/api/settings', methods=['POST'])
def api_save_settings():
    """Save runtime settings from frontend modal."""
    try:
        data = request.get_json() or {}

        mode = _set_tailoring_mode(data.get('tailoring_mode', _get_tailoring_mode()))
        provider = _set_llm_provider(data.get('llm_provider', _get_llm_provider()))
        summary_enabled = _set_summary_tailoring_enabled(
            data.get('enable_professional_summary', _is_summary_tailoring_enabled())
        )
        cover_letters_enabled = _set_cover_letter_enabled(
            data.get('include_cover_letters', _is_cover_letter_enabled())
        )

        return jsonify({
            'success': True,
            'settings': {
                'tailoring_mode': mode,
                'llm_provider': provider,
                'enable_professional_summary': summary_enabled,
                'include_cover_letters': cover_letters_enabled,
            },
        })
    except Exception as e:
        logger.error(f"Error saving settings: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

@app.route('/api/upload-cv', methods=['POST'])
def api_upload_cv():
    """Upload CV template via API."""
    try:
        if 'file' not in request.files:
            return jsonify({'success': False, 'error': 'No file provided'}), 400
        
        file = request.files['file']
        if file.filename == '':
            return jsonify({'success': False, 'error': 'No file selected'}), 400
        
        if not file.filename.endswith('.docx'):
            return jsonify({'success': False, 'error': 'Only .docx files are supported'}), 400
        
        filename = f"cv_template_{int(time.time())}.docx"
        filepath = os.path.join(app.config['UPLOAD_FOLDER'], filename)
        file.save(filepath)
        
        session['cv_template'] = filepath
        logger.info(f"CV template uploaded: {filename}")
        
        return jsonify({
            'success': True,
            'filename': file.filename,
            'message': 'CV template uploaded successfully'
        })
    except Exception as e:
        logger.error(f"Error uploading CV: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

@app.route('/api/user-profile', methods=['GET', 'POST'])
def api_user_profile():
    """Get or save the user's base profile data into the session."""
    try:
        if request.method == 'POST':
            data = request.get_json() or {}
            # Save the incoming data to the Flask session
            session['user_profile'] = {
                'first_name': data.get('first_name', ''),
                'last_name': data.get('last_name', ''),
                'email': data.get('email', ''),
                'phone': data.get('phone', ''),
                'linkedin': data.get('linkedin', ''),
                'github': data.get('github', '')
            }
            logger.info("User profile saved to session.")
            return jsonify({'success': True, 'message': 'Profile saved successfully.'})
            
        # If GET request, return whatever is currently in the session
        profile = session.get('user_profile', {})
        return jsonify({'success': True, 'profile': profile})
        
    except Exception as e:
        logger.error(f"Error handling user profile: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

# Added Apify Actor Job Search
@app.route('/api/apify-search', methods=['POST'])
def api_apify_search():
    """Standalone endpoint to search for jobs via Apify and save results to JSON."""
    """Standalone endpoint to search via Apify, fetch descriptions locally, and save to JSON."""
    try:
        data = request.get_json()
        keyword = data.get('keyword', '').strip()
        location = data.get('location', '').strip()
        max_jobs = int(data.get('max_jobs', 10))
        max_days_old = int(data.get('max_days_old', 14))
        
        if not keyword or not location:
            return jsonify({'success': False, 'error': 'Keyword and location are required'}), 400
        
        logger.info(f"Starting dedicated Apify search: {keyword} in {location}")
        
        # ==========================================
        # 1. URL Construction & Apify Call
        # ==========================================
        base_url = "https://www.linkedin.com/jobs/search?"
        params = {
            "keywords": keyword,
            "location": location,
        }

        if max_days_old:
            seconds = max_days_old * 24 * 60 * 60
            params["f_TPR"] = f"r{seconds}"

        final_url = base_url + urllib.parse.urlencode(params, safe="%2C")
        
        apify_token = os.environ.get("APIFY_API_TOKEN")
        if not apify_token:
            return jsonify({'success': False, 'error': 'Apify API token missing'}), 500

        client = ApifyClient(apify_token)
        
        # Force Apify to scrape at least 10 to satisfy its internal requirements
        apify_count = max(10, max_jobs) 
        run_input = {
            "urls": [final_url],
            "count": apify_count
        }

        app_env = os.environ.get("ENVIOR")

        if app_env == "PROD":

            logger.info("Running in Production")
            logger.info("Executing Apify Actor to get job list...")
            run = client.actor("curious_coder/linkedin-jobs-scraper").call(run_input=run_input)
            dataset = client.dataset(run["defaultDatasetId"]).list_items().items

            
            if not dataset:
                return jsonify({
                    'success': True,
                    'jobs': [],
                    'message': 'No jobs found on LinkedIn for these parameters.'
                })

            # ==========================================
            # 2. Map the Data 
            # ==========================================
            jobs = []
            for item in dataset[:max_jobs]:
                jobs.append({
                    'id': str(item.get('id')),
                    'title': item.get('title', 'Unknown Title'),
                    'company': item.get('companyName', item.get('company', 'Unknown Company')),
                    'link': item.get('link', ''),
                    'apply_url': item.get('applyUrl', ''),
                    'description': item.get('descriptionText', 'No Description')
                })
        else:
            # ==========================================
            # Simulate Delay & Load Mock Data
            # ==========================================
            logger.info("Running in Development")
            logger.info("Simulating Apify scraping delay (5 seconds)...")
            time.sleep(5)
            
            # Target the specific JSON file you saved previously
            mock_file_path = ".runtime/uploads/jobs/apify_jobs_20260424_184439.json"
            
            if not os.path.exists(mock_file_path):
                return jsonify({'success': False, 'error': f'Mock file not found: {mock_file_path}'}), 500
                
            with open(mock_file_path, 'r', encoding='utf-8') as f:
                all_mock_jobs = json.load(f)
                
            # Enforce the max_jobs limit requested by the client to mimic actual behavior
            jobs = all_mock_jobs[:max_jobs]

        # ==========================================
        # 3. Save Results to a JSON File
        # ==========================================
        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
        filename = f"apify_jobs_mocked{timestamp}.json"
        
        filepath = os.path.join(app.config['JOBS_OUTPUT_DIR'], filename)

        with open(filepath, 'w', encoding='utf-8') as f:
            json.dump(jobs, f, indent=4, ensure_ascii=False)
            
        logger.info(f"Successfully saved {len(jobs)} jobs with descriptions to {filename}")

        _save_processed_jobs(jobs)
        
        return jsonify({
            'success': True,
            'jobs': jobs,
            'json_file': filename,
            'message': f'Found and saved {len(jobs)} jobs to JSON.'
        })
        
    except Exception as e:
        logger.error(f"Error in Apify search endpoint: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500


# CV Analyzer
@app.route('/api/evaluate-fit', methods=['POST'])
def api_evaluate_fit():
    """Run heavy semantic analysis to get a fit score before generating a CV."""
    try:
        data = request.get_json()
        job_id = data.get('job_id')
        
        if not job_id:
            return jsonify({'success': False, 'error': 'No job ID provided'}), 400
            
        # 1. Ensure the user has uploaded their base CV
        user_resume_path = session.get('cv_template')
        if not user_resume_path or not os.path.exists(user_resume_path):
            return jsonify({'success': False, 'error': 'Please upload a base CV first.'}), 400
            
        # 2. Extract raw text from the base CV
        try:
            doc = docx.Document(user_resume_path)
            resume_text = "\n".join([para.text for para in doc.paragraphs if para.text.strip()])
        except Exception as e:
            return jsonify({'success': False, 'error': f'Failed to read CV: {str(e)}'}), 500
            
        # 3. Find the specific Job Description from the session
        jobs = _load_processed_jobs()
        job = next((j for j in jobs if j.get('id') == job_id or str(jobs.index(j)) == str(job_id)), None)
        
        if not job or not job.get('description'):
            return jsonify({'success': False, 'error': 'Job description not found.'}), 404
            
        # 4. Run the Heavy Evaluator (Lazy Loads SBERT if it's the first time!)
        logger.info(f"Running deep semantic evaluation for job: {job.get('title')}")
        evaluator = ATSEvaluator()
        result = evaluator.evaluate_fit(resume_text, job['description'])
        
        return jsonify({
            'success': True,
            'match_score': result['match_score'],
            'matched_skills': result['matched_skills'],
            'missing_skills': result['missing_skills']
        })
        
    except Exception as e:
        logger.error(f"Error in evaluate-fit: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

@app.route('/api/search', methods=['POST'])
def api_search_jobs():
    """Search for jobs via API."""
    try:
        data = request.get_json()
        keyword = data.get('keyword', '').strip()
        location = data.get('location', '').strip()
        max_jobs = int(data.get('max_jobs', 10))
        max_days_old = int(data.get('max_days_old', 14))
        
        if not keyword or not location:
            return jsonify({'success': False, 'error': 'Keyword and location are required'}), 400
        
        _set_tailoring_mode(data.get('tailoring_mode', _get_tailoring_mode()))
        _clear_job_context(keep_cv_template=True)
        
        logger.info(f"Searching jobs: {keyword} in {location}")
        scraper = LinkedInScraper(headless=True)
        jobs = scraper.scrape_job_listings(keyword, location, max_jobs=max_jobs, max_days_old=max_days_old)
        
        if not jobs:
            return jsonify({
                'success': True,
                'jobs': [],
                'message': 'No jobs found'
            })
        
        # Fetch descriptions for each job
        for i, job in enumerate(jobs):
            try:
                logger.info(f"Fetching description {i+1}/{len(jobs)}")
                title, company, description = scraper.fetch_job_description(job['link'])
                jobs[i]['description'] = description
            except Exception as e:
                logger.warning(f"Failed to fetch description for {job.get('title')}: {str(e)}")
                jobs[i]['description'] = ''
        
        # Save to Excel
        today_date = datetime.today().strftime("%Y-%m-%d")
        filename = f"linkedin_jobs_{today_date}_{int(time.time())}.xlsx"
        filepath = os.path.join(app.config['JOBS_OUTPUT_DIR'], filename)
        df = pd.DataFrame(jobs)
        df.to_excel(filepath, index=False)
        session['jobs_file'] = filepath
        session['excel_filename'] = filename
        
        # Extract skills from job descriptions
        analyzer = CVAnalyzer()
        processed_jobs = []
        
        for i, job in enumerate(jobs):
            if job.get('description'):
                try:
                    matched_skills, matched_requirements, matched_categories = analyzer.extract_skills_from_description(job['description'])
                    job['matched_skills'] = matched_skills
                    job['matched_categories'] = matched_categories
                    # Add unique ID
                    job['id'] = str(i)
                    processed_jobs.append(job)
                except Exception as e:
                    logger.warning(f"Failed to analyze skills for {job.get('title')}: {str(e)}")
        
        _save_processed_jobs(processed_jobs)
        
        return jsonify({
            'success': True,
            'jobs': processed_jobs,
            'excel_file': filename,
            'message': f'Found {len(processed_jobs)} jobs'
        })
    except Exception as e:
        logger.error(f"Error searching jobs: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

@app.route('/api/jobs', methods=['GET'])
def api_get_jobs():
    """Get stored jobs from session."""
    try:
        jobs = _load_processed_jobs()
        return jsonify({
            'success': True,
            'jobs': jobs,
            'count': len(jobs)
        })
    except Exception as e:
        logger.error(f"Error getting jobs: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

# @app.route('/api/generate-cv/<job_index>', methods=['POST'])
@app.route('/api/generate-cv/<job_id>', methods=['POST'])
def api_generate_cv(job_id):
    """Generate CV for a single job with JIT Analysis."""
    try:
        jobs = _load_processed_jobs()

        # job_idx = int(job_index)

        # 1. Find the actual job index using the UUID (with legacy fallback)
        job_idx = -1
        for i, j in enumerate(jobs):
            if j.get('id') == job_id or str(i) == str(job_id):
                job_idx = i
                break
        
        # if job_idx < 0 or job_idx >= len(jobs):
        if job_idx == -1:
            return jsonify({'success': False, 'error': 'Invalid job index'}), 400
        
        job = jobs[job_idx]

        # 2. JIT Analysis: Enrich the job with skills if it hasn't been done yet
        job = _enrich_job_with_skills(job)

        # 3. Save the enriched job back to the session so we don't re-analyze if they click again
        jobs[job_idx] = job
        _save_processed_jobs(jobs)

        # cv_template_path = session.get('cv_template')
        
        # if not cv_template_path or not os.path.exists(cv_template_path):
        #     return jsonify({'success': False, 'error': 'CV template not found'}), 400

        user_resume_path = session.get('cv_template')
        ats_template_path = app.config.get('ATS_TEMPLATE_PATH')
        
        if not user_resume_path or not os.path.exists(user_resume_path):
            return jsonify({'success': False, 'error': 'User resume not found. Please upload a resume first.'}), 400
            
        if not ats_template_path or not os.path.exists(ats_template_path):
            return jsonify({'success': False, 'error': 'System ATS template not found on server.'}), 500
        
        # tailoring_mode = _get_tailoring_mode()
        # matched_categories = job.get('matched_categories', {})
        # professional_summary = _build_professional_summary(job, matched_categories)
        
        # Generate output filename
        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
        company_slug = _sanitize_filename(job.get('company', 'Company'))
        title_slug = _sanitize_filename(job.get('title', 'Position'))
        output_filename = f"CV_{timestamp}_{company_slug}_{title_slug}.docx"
        output_path = os.path.join(app.config['CV_OUTPUT_DIR'], output_filename)

        # ==========================================
        # EXECUTE LANGGRAPH AI TAILOR
        # ==========================================
        logger.info(f"Starting LangGraph Tailor for {job.get('title')}")
        ai_tailor = AILangGraphTailor()
        
        # Extract raw text from the USER'S uploaded resume
        original_resume_text = ai_tailor.extract_text_from_docx(user_resume_path)
        
        # Run the Graph to get the tailored JSON
        final_state = ai_tailor.run_pipeline(original_resume_text, job.get('description', ''))
        
        # Map the JSON to the SYSTEM'S ATS Jinja template and save it
        ai_tailor.generate_final_docx(
            tailored_dict=final_state["tailored_resume"],
            template_path=ats_template_path, 
            output_path=output_path
        )
        
        
        # Modify CV
        # modifier = CVModifier(cv_template_path)
        
        # generated_cover_letter_filename = None

        # if tailoring_mode == 'api':
        #     try:
        #         generation_result = _generate_cv_with_api_tailoring(job, cv_template_path, output_path)
        #         provider = generation_result.get('provider', 'unknown')
        #         generated_cover_letter_filename = generation_result.get('cover_letter_filename')
        #         logger.info(f"API tailoring completed using {provider}")
        #     except Exception as e:
        #         # Fallback to local tailoring
        #         logger.warning(f"API tailoring failed, falling back to local: {str(e)}")
        #         if _is_summary_tailoring_enabled():
        #             modifier.update_profile_summary(professional_summary)
        #         modifier.update_skills_section(matched_categories)
        #         modifier.save_modified_cv(output_path)
        # else:
        #     # Local tailoring mode
        #     if _is_summary_tailoring_enabled():
        #         modifier.update_profile_summary(professional_summary)
        #     modifier.update_skills_section(matched_categories)
        #     modifier.save_modified_cv(output_path)
        
        # return jsonify({
        #     'success': True,
        #     'filename': output_filename,
        #     'cover_letter_filename': generated_cover_letter_filename,
        #     'job_title': job.get('title'),
        #     'company': job.get('company'),
        #     'message': f"CV generated for {job.get('title')} at {job.get('company')}"
        # })
        return jsonify({
            'success': True,
            'filename': output_filename,
            'job_title': job.get('title'),
            'company': job.get('company'),
            'job': job, # Keeps UI updated
            'message': f"CV successfully tailored for {job.get('company')}"
        })
    except Exception as e:
        logger.error(f"Error generating CV: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

@app.route('/api/generate-all-cvs', methods=['POST'])
def api_generate_all_cvs():
    """Generate CVs for multiple jobs with JIT Analysis."""
    try:
        data = request.get_json()
        # job_indices = data.get('job_indices', [])
        requested_ids = data.get('job_indices', [])
        jobs = _load_processed_jobs()
        
        # if not job_indices:
        #     job_indices = list(range(len(jobs)))

        # 1. Map the requested UUIDs back to their actual list indices
        job_indices = []
        if requested_ids:
            for req_id in requested_ids:
                for i, j in enumerate(jobs):
                    if j.get('id') == req_id or str(i) == str(req_id):
                        job_indices.append(i)
                        break
        else:
            # Fallback if no IDs are passed
            job_indices = list(range(len(jobs)))
        
        # if not session.get('cv_template') or not os.path.exists(session.get('cv_template')):
        #     return jsonify({'success': False, 'error': 'CV template not found'}), 400
        
        # cv_template_path = session.get('cv_template')

        user_resume_path = session.get('cv_template')
        ats_template_path = app.config.get('ATS_TEMPLATE_PATH')
        
        if not user_resume_path or not os.path.exists(user_resume_path):
            return jsonify({'success': False, 'error': 'User resume not found. Please upload a resume first.'}), 400
            
        if not ats_template_path or not os.path.exists(ats_template_path):
            return jsonify({'success': False, 'error': 'System ATS template not found on server.'}), 500

        tailoring_mode = _get_tailoring_mode()
        successful = []
        failed = []

        # Instantiate the Analyzer ONCE for the entire batch to save CPU
        analyzer = CVAnalyzer()
        jobs_modified = False
        
        for idx in job_indices:
            try:
                if idx < 0 or idx >= len(jobs):
                    continue

                # JIT Analysis for each job in the batch
                job = _enrich_job_with_skills(jobs[idx], analyzer)
                jobs[idx] = job
                jobs_modified = True

                # Now the data is ready for your custom Tailor
                matched_categories = job.get('matched_categories', {})
                professional_summary = _build_professional_summary(job, matched_categories)
                
                timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
                company_slug = _sanitize_filename(job.get('company', 'Company'))
                title_slug = _sanitize_filename(job.get('title', 'Position'))
                output_filename = f"CV_{timestamp}_{company_slug}_{title_slug}.docx"
                output_path = os.path.join(app.config['CV_OUTPUT_DIR'], output_filename)

                # ==========================================
                # EXECUTE LANGGRAPH AI TAILOR
                # ==========================================
                logger.info(f"Starting LangGraph Tailor for {job.get('title')}")
                ai_tailor = AILangGraphTailor()
                
                # Extract text from the user's uploaded resume
                original_resume_text = ai_tailor.extract_text_from_docx(user_resume_path)
                
                # Run the Graph to get the tailored JSON
                final_state = ai_tailor.run_pipeline(original_resume_text, job.get('description', ''))
                
                # Map the JSON to the system ATS template and save it
                ai_tailor.generate_final_docx(
                    tailored_dict=final_state["tailored_resume"],
                    template_path=ats_template_path, 
                    output_path=output_path
                )              
                
                # modifier = CVModifier(cv_template_path)
                
                # cover_letter_filename = None

                # if tailoring_mode == 'api':
                #     try:
                #         generation_result = _generate_cv_with_api_tailoring(job, cv_template_path, output_path)
                #         cover_letter_filename = generation_result.get('cover_letter_filename')
                #     except Exception as e:
                #         if _is_summary_tailoring_enabled():
                #             modifier.update_profile_summary(professional_summary)
                #         modifier.update_skills_section(matched_categories)
                #         modifier.save_modified_cv(output_path)
                # else:
                #     if _is_summary_tailoring_enabled():
                #         modifier.update_profile_summary(professional_summary)
                #     modifier.update_skills_section(matched_categories)
                #     modifier.save_modified_cv(output_path)
                
                # successful.append({
                #     'job_index': idx,
                #     'filename': output_filename,
                #     'cover_letter_filename': cover_letter_filename,
                #     'job_title': job.get('title'),
                #     'company': job.get('company'),
                # })
                successful.append({
                    'job_index': idx,
                    'filename': output_filename,
                    'cover_letter_filename': None, # Adjust if adding cover letter logic back
                    'job_title': job.get('title'),
                    'company': job.get('company'),
                })
            except Exception as e:
                logger.error(f"Error generating CV for job {idx}: {str(e)}")
                failed.append({
                    'job_index': idx,
                    'error': str(e),
                    'job_title': jobs[idx].get('title') if idx < len(jobs) else 'Unknown',
                })

        if jobs_modified:
            _save_processed_jobs(jobs)
        
        # Create ZIP file
        zip_buffer = io.BytesIO()
        with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
            for cv in successful:
                file_path = os.path.join(app.config['CV_OUTPUT_DIR'], cv['filename'])
                if os.path.exists(file_path):
                    zip_file.write(file_path, arcname=cv['filename'])
                cover_letter_filename = cv.get('cover_letter_filename')
                if cover_letter_filename:
                    cover_letter_path = os.path.join(app.config['CV_OUTPUT_DIR'], cover_letter_filename)
                    if os.path.exists(cover_letter_path):
                        zip_file.write(cover_letter_path, arcname=cover_letter_filename)
        
        zip_buffer.seek(0)
        zip_filename = f"CVs_{datetime.now().strftime('%Y%m%d_%H%M%S')}.zip"
        zip_path = os.path.join(app.config['CV_OUTPUT_DIR'], zip_filename)
        
        with open(zip_path, 'wb') as f:
            f.write(zip_buffer.getvalue())
        
        return jsonify({
            'success': True,
            'successful': successful,
            'failed': failed,
            'zip_filename': zip_filename,
            'total_generated': len(successful),
            'total_failed': len(failed),
        })
    except Exception as e:
        logger.error(f"Error generating all CVs: {str(e)}")
        return jsonify({'success': False, 'error': str(e)}), 500

@app.route('/api/download/<filename>', methods=['GET'])
def api_download_file(filename):
    """Download a file."""
    try:
        import urllib.parse
        # Decode URL-encoded filename
        decoded_filename = urllib.parse.unquote(filename)
        
        # Security check: prevent directory traversal
        if '..' in decoded_filename or '/' in decoded_filename or '\\' in decoded_filename:
            return jsonify({'error': 'Invalid filename'}), 400
        
        file_path = os.path.join(app.config['CV_OUTPUT_DIR'], decoded_filename)
        if not os.path.exists(file_path):
            file_path = os.path.join(app.config['JOBS_OUTPUT_DIR'], decoded_filename)
        
        if not os.path.exists(file_path):
            logger.warning(f"File not found: {decoded_filename}")
            return jsonify({'error': 'File not found'}), 404
        
        logger.info(f"Downloading file: {decoded_filename}")
        return send_file(file_path, as_attachment=True, download_name=decoded_filename)
    except Exception as e:
        logger.error(f"Error downloading file: {str(e)}")
        return jsonify({'error': str(e)}), 500

# Communication endpoint with the Browser Extension (Auto-Apply)
@app.route('/api/latest-application-data', methods=['GET'])
def api_latest_application_data():
    """Endpoint for the Chrome Extension to fetch real user data for autofill."""
    
    # 1. Grab the user profile from the session
    user_profile = session.get('user_profile')
    
    if not user_profile:
        return jsonify({
            'success': False, 
            'error': 'No user profile found. Please save your profile in the React app first!'
        }), 400

    # 2. Grab the latest generated CV from the session
    latest_cv = session.get('latest_cv_filename')
    cv_url = None
    
    if latest_cv:
        import urllib.parse
        # Dynamically build the full download URL based on the server's current host
        safe_filename = urllib.parse.quote(latest_cv)
        cv_url = f"{request.host_url}api/download/{safe_filename}"

    return jsonify({
        'success': True,
        'user': user_profile,
        # 'cv_url': cv_url,
        'message': 'Real session data retrieved successfully'
    })

def _generate_cv_with_api_tailoring(job, cv_template, output_path):
    """Use API subproject updater to generate a tailored CV and optional cover letter."""
    api_project_root = os.path.join(os.getcwd(), "Automatic CV and Cover Letter with API")
    if not os.path.exists(api_project_root):
        raise FileNotFoundError("API subproject folder not found")

    if api_project_root not in sys.path:
        sys.path.append(api_project_root)

    try:
        # Ensure integrations that depend on env vars see the active provider.
        os.environ['LLM_PROVIDER'] = _get_llm_provider()
        APIIntegration = importlib.import_module('src.utils.openai_integration').OpenAIIntegration
        DocumentUpdater = importlib.import_module('src.updaters.document_updater').DocumentUpdater

        cover_letter_template = os.environ.get(
            'API_COVER_LETTER_TEMPLATE_PATH',
            os.path.join(api_project_root, 'data', 'Cover Letter_Imon .docx')
        )
        if not os.path.isabs(cover_letter_template):
            cover_letter_template = os.path.abspath(cover_letter_template)

        if not os.path.exists(cover_letter_template):
            if _is_cover_letter_enabled():
                raise FileNotFoundError("Cover letter template not found")
            # Cover letters are disabled, so reuse CV template path as a safe placeholder.
            cover_letter_template = cv_template

        description = (job.get('description') or '').strip()
        if not description:
            raise ValueError("Job description is empty")

        llm_integration = APIIntegration()
        provider = getattr(llm_integration, 'provider', 'unknown')
        if provider in ('openai', 'grok', 'groq') and not llm_integration.is_api_key_set():
            raise ValueError(f"API key not configured for provider '{provider}'")

        updater = DocumentUpdater(cv_template, cover_letter_template, llm_integration)
        updater.update_cv(description, output_path)

        generated_cover_letter_filename = None
        if _is_cover_letter_enabled():
            cv_basename = os.path.basename(output_path)
            if cv_basename.startswith('CV_'):
                cover_letter_filename = f"CoverLetter_{cv_basename[3:]}"
            else:
                cover_letter_filename = f"CoverLetter_{cv_basename}"
            cover_letter_output_path = os.path.join(os.path.dirname(output_path), cover_letter_filename)
            updater.update_cover_letter(description, cover_letter_output_path)
            generated_cover_letter_filename = cover_letter_filename

        return {
            'provider': provider,
            'cover_letter_filename': generated_cover_letter_filename,
        }
    except Exception as e:
        raise Exception(f"API tailoring error: {str(e)}")

# ==================== Legacy Routes (keep for backward compatibility) ====================

@app.route('/search', methods=['GET', 'POST'])
def search_jobs():
    """Legacy search route - redirects to React app"""
    if request.method == 'POST':
        keyword = request.form.get('keyword', '')
        location = request.form.get('location', '')
        max_jobs = int(request.form.get('max_jobs', 10))
        
        _set_tailoring_mode(request.form.get('tailoring_mode', _get_tailoring_mode()))
        
        if not keyword or not location:
            flash('Please enter both job title and location', 'error')
            return redirect(url_for('index'))
        
        try:
            scraper = LinkedInScraper(headless=True)
            jobs = scraper.scrape_job_listings(keyword, location, max_jobs=max_jobs)
            
            if not jobs:
                flash('No jobs found. Try different search terms.', 'warning')
                return redirect(url_for('index'))
            
            for i, job in enumerate(jobs):
                logger.info(f"Fetching description for job {i+1}/{len(jobs)}: {job['title']}")
                title, company, description = scraper.fetch_job_description(job['link'])
                jobs[i]['description'] = description
            
            today_date = datetime.today().strftime("%Y-%m-%d")
            filename = f"linkedin_jobs_{today_date}.xlsx"
            filepath = os.path.join(app.config['JOBS_OUTPUT_DIR'], filename)
            
            df = pd.DataFrame(jobs)
            df.to_excel(filepath, index=False)
            session['jobs_file'] = filepath
            session['excel_filename'] = filename
            
            analyzer = CVAnalyzer()
            processed_jobs = []
            
            for job in jobs:
                if job.get('description'):
                    matched_skills, matched_requirements, matched_categories = analyzer.extract_skills_from_description(job['description'])
                    job['matched_skills'] = matched_skills
                    job['matched_categories'] = matched_categories
                    processed_jobs.append(job)
            
            _save_processed_jobs(processed_jobs)
            
            return render_template('job_list.html', 
                                  jobs=processed_jobs, 
                                  excel_file=filename,
                                  excel_path=filepath,
                                  tailoring_mode=_get_tailoring_mode(),
                                  llm_provider=_get_llm_provider())
            
        except Exception as e:
            logger.error(f"Error during job search: {str(e)}")
            flash(f'An error occurred: {str(e)}', 'error')
            return redirect(url_for('index'))
    
    return redirect(url_for('index'))

@app.route('/upload_cv', methods=['GET', 'POST'])
def upload_cv():
    """Legacy CV upload route"""
    if request.method == 'POST':
        if 'cv_file' not in request.files:
            flash('No file part', 'error')
            return redirect(request.url)
        
        file = request.files['cv_file']
        if file.filename == '':
            flash('No selected file', 'error')
            return redirect(request.url)
        
        if not file.filename.endswith('.docx'):
            flash('Only .docx files are supported', 'error')
            return redirect(request.url)
        
        try:
            filename = f"cv_template_{int(time.time())}.docx"
            filepath = os.path.join(app.config['UPLOAD_FOLDER'], filename)
            file.save(filepath)
            session['cv_template'] = filepath
            flash('CV template uploaded successfully', 'success')
            return redirect(url_for('index'))
        except Exception as e:
            logger.error(f"Error uploading CV: {str(e)}")
            flash(f'Error uploading CV: {str(e)}', 'error')
            return redirect(request.url)
    
    return render_template('upload_cv.html')

@app.errorhandler(404)
def not_found(error):
    """Handle 404 errors"""
    return render_template('404.html'), 404

@app.errorhandler(500)
def internal_error(error):
    """Handle 500 errors"""
    logger.error(f"Internal server error: {str(error)}")
    return render_template('500.html'), 500

if __name__ == '__main__':
    port = int(os.environ.get('PORT', 5050))
    debug = os.environ.get('FLASK_DEBUG', 'False').lower() == 'true'
    app.run(host='0.0.0.0', port=port, debug=debug)