app.py

from flask import Flask, render_template, request, jsonify, send_file, redirect, url_for, session
import validators
import os
import tempfile
import uuid
from urllib.parse import urlparse
from typing import Dict, Any, List

# Load environment variables from .env file
try:
    from dotenv import load_dotenv
    load_dotenv()
except ImportError:
    print("python-dotenv not installed. Using system environment variables only.")


from modules.technical_seo import TechnicalSEOModule
from modules.content_audit import ContentAuditModule
from modules.keywords import KeywordsModule
from modules.backlinks import BacklinksModule
from report_generator import ReportGenerator
from simple_pdf_generator import SimplePDFGenerator
from llm_recommendations import LLMRecommendations
from gsc_client import GSCClient
from utils import safe_pct
from benchmarks import BENCHMARKS, badge

app = Flask(__name__, static_folder='static')
app.secret_key = os.getenv('FLASK_SECRET_KEY', 'seo_report_generator_2024')


technical_module = TechnicalSEOModule(api_key=os.getenv('GOOGLE_API_KEY'))
content_module = ContentAuditModule()
keywords_module = KeywordsModule()
backlinks_module = BacklinksModule()
report_gen = ReportGenerator()
pdf_gen = SimplePDFGenerator()
llm_recommendations = LLMRecommendations()
try:
    gsc_client = GSCClient()
except ImportError as e:
    print(f"GSC client not available: {e}")
    gsc_client = None


reports_store = {}

def _transform_keywords_data(new_data: Dict[str, Any]) -> Dict[str, Any]:
    if not new_data or new_data.get('placeholder'):
        return {
            'placeholder': True,
            'message': 'No keyword data available',
            'total_keywords': 0,
            'position_distribution': {'top_3': 0, 'top_10': 0, 'top_50': 0, 'beyond_50': 0},
            'best_keywords': [],
            'opportunity_keywords': [],
            'data_source': 'Analysis failed'
        }
    

    totals = new_data.get('totals', {})
    distribution = new_data.get('distribution', {})
    movement = new_data.get('movement', {})
    best_keywords = new_data.get('best_keywords', [])
    declining_keywords = new_data.get('declining_keywords', [])
    opportunities = new_data.get('opportunities', [])
    data_sources = new_data.get('data_sources', {})
    

    pos_dist = {
        'top_3': distribution.get('top3', 0),
        'top_10': distribution.get('top10', 0),
        'top_50': distribution.get('top50', 0),
        'beyond_50': totals.get('keywords', 0) - distribution.get('top50', 0)
    }
    

    transformed_best_keywords = []
    for kw in best_keywords:
        transformed_best_keywords.append({
            'keyword': kw.get('keyword', ''),
            'position': kw.get('rank', 0),
            'clicks': 0,
            'impressions': kw.get('volume', 0),
            'url': kw.get('url', ''),
            'estimated_traffic': kw.get('estimated_traffic', 0),
            'trend': kw.get('trend', 'stable')
        })
    

    transformed_opportunities = []
    for opp in opportunities:
        transformed_opportunities.append({
            'keyword': opp.get('keyword', ''),
            'position': 0,
            'impressions': opp.get('volume', 0),
            'ctr': 0,
            'competitor_rank': opp.get('competitor_rank', 0),
            'priority_score': opp.get('priority_score', 0),
            'competitor_domain': opp.get('competitor_domain', '')
        })
    
    return {
        'total_keywords': totals.get('keywords', 0),
        'estimated_traffic': totals.get('estimated_traffic', 0),
        'position_distribution': pos_dist,
        'movement': movement,
        'best_keywords': transformed_best_keywords,
        'declining_keywords': declining_keywords,
        'opportunity_keywords': transformed_opportunities,
        'competitor_summary': new_data.get('competitor_summary', []),
        'data_source': f"{data_sources.get('positions', 'Unknown')} + {data_sources.get('volume', 'Unknown')}",
        'enrichment_rate': data_sources.get('enrichment_rate', 0),
        'meta': new_data.get('meta', {}),
        'placeholder': False
    }

@app.route('/')
def index():
    return render_template('index.html')

@app.route('/generate', methods=['POST'])
def generate_report():
    try:
        data = request.json
        url = data.get('url', '').strip()
        competitors = data.get('competitors', [])
        
        if not url:
            return jsonify({'error': 'Website URL is required'}), 400
            
        if not validators.url(url):
            return jsonify({'error': 'Please enter a valid URL'}), 400
        

        report_id = str(uuid.uuid4())
        

        competitor_domains = []
        competitor_list = []
        for comp in competitors:
            comp = comp.strip()
            if comp and validators.url(comp):
                competitor_list.append(comp)

                domain = urlparse(comp).netloc.replace('www.', '')
                competitor_domains.append(domain)
        

        technical_data = technical_module.analyze(url)
        

        content_data = content_module.analyze(url)
        

        # Check if GSC should be used
        use_gsc = False
        if gsc_client and 'gsc_tokens' in session and gsc_client.property_url:
            domain = urlparse(url).netloc.replace('www.', '')
            property_domain = urlparse(gsc_client.property_url).netloc.replace('www.', '')
            if domain == property_domain:
                use_gsc = True
        
        # Analyze keywords
        if use_gsc:
            keywords_result = app._analyze_with_gsc(url, competitor_domains)
        else:
            keywords_result = keywords_module.analyze(url, competitor_domains=competitor_domains)
        if not keywords_result.success:

            keywords_data = {
                'placeholder': True,
                'message': f'Keywords analysis failed: {keywords_result.error}',
                'total_keywords': 0,
                'position_distribution': {'top_3': 0, 'top_10': 0, 'top_50': 0, 'beyond_50': 0},
                'best_keywords': [],
                'opportunity_keywords': [],
                'data_source': 'Analysis failed'
            }
        else:

            keywords_data = _transform_keywords_data(keywords_result.data)
        

        print(f"DEBUG: Starting backlinks analysis for {url}")
        backlinks_result = backlinks_module.analyze(url)
        backlinks_data = backlinks_result.data
        print(f"DEBUG: Backlinks analysis result - Success: {backlinks_result.success}")
        print(f"DEBUG: Backlinks data keys: {list(backlinks_data.keys())}")
        if backlinks_data.get('total_backlinks'):
            print(f"DEBUG: Total backlinks found: {backlinks_data.get('total_backlinks')}")
        if backlinks_data.get('placeholder'):
            print(f"DEBUG: Using placeholder data: {backlinks_data.get('message')}")
        

        llm_rec_data = llm_recommendations.generate_recommendations(
            url, technical_data, content_data, keywords_data, backlinks_data
        )
        

        competitor_data = []
        for comp_url in competitor_list:
            comp_technical = technical_module.analyze(comp_url)
            comp_content = content_module.analyze(comp_url, quick_scan=True)
            

            comp_keywords_result = keywords_module.analyze(comp_url, competitor_domains=[], quick_scan=True)
            if comp_keywords_result.success:
                comp_keywords = _transform_keywords_data(comp_keywords_result.data)
            else:
                comp_keywords = {
                    'placeholder': True, 
                    'message': f'Keywords analysis failed: {comp_keywords_result.error}',
                    'total_keywords': 0,
                    'position_distribution': {'top_3': 0, 'top_10': 0, 'top_50': 0, 'beyond_50': 0},
                    'best_keywords': [],
                    'opportunity_keywords': [],
                    'data_source': 'Analysis failed'
                }
            

            comp_backlinks_result = backlinks_module.analyze(comp_url, quick_scan=True)
            comp_backlinks = comp_backlinks_result.data
            
            competitor_data.append({
                'url': comp_url,
                'technical': comp_technical,
                'content': comp_content,
                'keywords': comp_keywords,
                'backlinks': comp_backlinks
            })
        

        report_html = report_gen.generate_html_report(
            url=url,
            technical_data=technical_data,
            content_data=content_data,
            competitor_data=competitor_data,
            keywords_data=keywords_data,
            backlinks_data=backlinks_data,
            llm_recommendations=llm_rec_data,
            include_charts=True
        )
        

        reports_store[report_id] = {
            'url': url,
            'html': report_html,
            'technical_data': technical_data,
            'content_data': content_data,
            'keywords_data': keywords_data,
            'backlinks_data': backlinks_data,
            'llm_recommendations': llm_rec_data,
            'competitor_data': competitor_data
        }
        
        return jsonify({
            'success': True,
            'report_id': report_id,
            'redirect_url': f'/report/{report_id}'
        })
        
    except Exception as e:
        return jsonify({'error': f'Error generating report: {str(e)}'}), 500

@app.route('/report/<report_id>')
def view_report(report_id):
    if report_id not in reports_store:
        return redirect(url_for('index'))
    
    report_data = reports_store[report_id]
    return render_template('report.html', 
                         report_html=report_data['html'],
                         report_id=report_id,
                         url=report_data['url'])

@app.route('/download/<report_id>')
def download_html(report_id):
    if report_id not in reports_store:
        return jsonify({'error': 'Report not found'}), 404
    
    report_data = reports_store[report_id]
    
    with tempfile.NamedTemporaryFile(mode='w', suffix='.html', delete=False) as f:
        f.write(report_data['html'])
        temp_path = f.name
    
    filename = f"seo_report_{report_data['url'].replace('https://', '').replace('http://', '').replace('/', '_')}.html"
    
    return send_file(temp_path, as_attachment=True, download_name=filename, mimetype='text/html')

# PDF Download removed - now using browser print functionality

def _analyze_with_gsc(url: str, competitor_domains: List[str]):
    """Analyze keywords using GSC as primary source"""
    try:
        gsc_tokens = session.get('gsc_tokens', {})
        
        if not gsc_tokens.get('access_token'):
            return keywords_module.analyze(url, competitor_domains=competitor_domains)
        
        # Fetch GSC data using the updated method
        gsc_data = gsc_client.get_search_analytics(gsc_tokens)
        transformed_data = gsc_client.transform_gsc_data(gsc_data, urlparse(url).netloc)
        
        # Update session with potentially refreshed tokens
        session['gsc_tokens'] = gsc_tokens
        
        from modules.keywords import ModuleResult
        return ModuleResult(success=True, data=transformed_data)
        
    except Exception as e:
        print(f"GSC analysis failed: {e}")
        return keywords_module.analyze(url, competitor_domains=competitor_domains)

app._analyze_with_gsc = _analyze_with_gsc

@app.route('/auth/gsc/start')
def gsc_auth_start():
    """Start GSC OAuth flow"""
    if not gsc_client:
        return jsonify({'error': 'Google Search Console integration not available. Install: pip install google-api-python-client google-auth-oauthlib google-auth'}), 500
    
    try:
        auth_url = gsc_client.get_auth_url()
        return redirect(auth_url)
    except Exception as e:
        return jsonify({'error': f'OAuth setup failed: {str(e)}'}), 500

@app.route('/auth/gsc/callback')
def gsc_auth_callback():
    """Handle GSC OAuth callback"""
    auth_code = request.args.get('code')
    error = request.args.get('error')
    
    if error:
        return redirect(url_for('index', error=f'OAuth error: {error}'))
    
    if not auth_code:
        return redirect(url_for('index', error='No authorization code received'))
    
    try:
        tokens = gsc_client.exchange_code(auth_code)
        session['gsc_tokens'] = tokens
        return redirect(url_for('index', success='Google Search Console connected successfully'))
    except Exception as e:
        return redirect(url_for('index', error=f'Token exchange failed: {str(e)}'))

@app.route('/auth/gsc/status')
def gsc_auth_status():
    """Check GSC authentication status"""
    has_tokens = 'gsc_tokens' in session
    property_url = gsc_client.property_url
    
    return jsonify({
        'authenticated': has_tokens,
        'property_url': property_url,
        'client_configured': bool(gsc_client.client_id and gsc_client.client_secret)
    })

if __name__ == '__main__':
    app.run(debug=False, host='0.0.0.0', port=7860)