server/meta_generator.py

"""
Meta Description Generator
AI-powered meta description generation for SEO
"""

import os
import re
from typing import Dict, List

def generate_meta_description(page: Dict, api_keys: Dict = None) -> str:
    """
    Generate SEO-optimized meta description for a page
    Uses AI if available, falls back to smart extraction
    """
    title = page.get('title', '')
    text = page.get('text', '')
    headings = page.get('headings', [])
    url = page.get('url', '')
    
    # Try AI generation first
    ai_description = _generate_with_ai(title, text, headings, api_keys)
    if ai_description:
        return ai_description
    
    # Fallback to smart extraction
    return _generate_smart_description(title, text, headings)


def _generate_with_ai(title: str, text: str, headings: List, api_keys: Dict = None) -> str:
    """Generate meta description using AI (Groq or OpenAI)"""
    try:
        # Try Groq first (faster and free)
        groq_key = (api_keys or {}).get('groq') or os.getenv('GROQ_API_KEY')
        if groq_key:
            return _generate_with_groq(title, text, headings, groq_key)
        
        # Fallback to OpenAI
        openai_key = (api_keys or {}).get('openai') or os.getenv('OPENAI_API_KEY')
        if openai_key:
            return _generate_with_openai(title, text, headings, openai_key)
        
        return None
        
    except Exception as e:
        print(f"AI meta generation failed: {e}")
        return None


def _generate_with_groq(title: str, text: str, headings: List, api_key: str) -> str:
    """Generate using Groq API"""
    try:
        import requests
        
        # Extract key points from text
        text_sample = text[:1000] if text else ''
        headings_text = ' '.join([h.get('text', '') if isinstance(h, dict) else str(h) for h in headings[:5]])
        
        prompt = f"""اكتب وصف meta description مثالي لصفحة ويب بالمواصفات التالية:
- العنوان: {title}
- المحتوى: {text_sample}
- العناوين: {headings_text}

المتطلبات:
- الطول: 150-160 حرف بالضبط
- يجب أن يكون جذاب ومقنع
- يحتوي على كلمات مفتاحية
- يشجع على النقر
- بدون علامات تنصيص

اكتب الوصف فقط بدون أي نص إضافي:"""

        response = requests.post(
            'https://api.groq.com/openai/v1/chat/completions',
            headers={
                'Authorization': f'Bearer {api_key}',
                'Content-Type': 'application/json'
            },
            json={
                'model': 'llama-3.3-70b-versatile',
                'messages': [{'role': 'user', 'content': prompt}],
                'temperature': 0.7,
                'max_tokens': 100
            },
            timeout=15
        )
        
        if response.status_code == 200:
            data = response.json()
            description = data['choices'][0]['message']['content'].strip()
            # Clean up
            description = description.replace('"', '').replace("'", '').strip()
            # Ensure proper length
            if len(description) > 160:
                description = description[:157] + '...'
            return description
        
        return None
        
    except Exception as e:
        print(f"Groq meta generation error: {e}")
        return None


def _generate_with_openai(title: str, text: str, headings: List, api_key: str) -> str:
    """Generate using OpenAI API"""
    try:
        import requests
        
        text_sample = text[:1000] if text else ''
        headings_text = ' '.join([h.get('text', '') if isinstance(h, dict) else str(h) for h in headings[:5]])
        
        prompt = f"""Write a perfect meta description for a webpage with these details:
- Title: {title}
- Content: {text_sample}
- Headings: {headings_text}

Requirements:
- Exactly 150-160 characters
- Compelling and click-worthy
- Contains keywords
- No quotation marks

Write only the description:"""

        response = requests.post(
            'https://api.openai.com/v1/chat/completions',
            headers={
                'Authorization': f'Bearer {api_key}',
                'Content-Type': 'application/json'
            },
            json={
                'model': 'gpt-3.5-turbo',
                'messages': [{'role': 'user', 'content': prompt}],
                'temperature': 0.7,
                'max_tokens': 100
            },
            timeout=15
        )
        
        if response.status_code == 200:
            data = response.json()
            description = data['choices'][0]['message']['content'].strip()
            description = description.replace('"', '').replace("'", '').strip()
            if len(description) > 160:
                description = description[:157] + '...'
            return description
        
        return None
        
    except Exception as e:
        print(f"OpenAI meta generation error: {e}")
        return None


def _generate_smart_description(title: str, text: str, headings: List) -> str:
    """Generate meta description using smart extraction (no AI)"""
    # Clean text
    text = re.sub(r'\s+', ' ', text).strip()
    
    # Try to find first meaningful paragraph
    sentences = re.split(r'[.!?]+', text)
    
    description = ''
    for sentence in sentences:
        sentence = sentence.strip()
        if len(sentence) > 50 and len(sentence) < 200:
            description = sentence
            break
    
    # If no good sentence, use title + first heading
    if not description and headings:
        first_heading = headings[0]
        heading_text = first_heading.get('text', '') if isinstance(first_heading, dict) else str(first_heading)
        description = f"{title} - {heading_text}"
    
    # If still nothing, use title + generic text
    if not description:
        description = f"{title} - اكتشف خدماتنا ومنتجاتنا المتميزة"
    
    # Ensure proper length
    if len(description) > 160:
        description = description[:157] + '...'
    elif len(description) < 120:
        description += ' - تواصل معنا الآن للمزيد من المعلومات'
    
    return description


def generate_meta_descriptions_for_audit(audit: Dict, api_keys: Dict = None) -> List[Dict]:
    """Generate meta descriptions for all pages in audit"""
    pages = audit.get('pages', [])
    results = []
    
    for page in pages:
        url = page.get('url', '')
        title = page.get('title', '')
        
        # Check if page already has meta description
        html = page.get('html', '')
        existing_meta = re.search(r'<meta\s+name=["\']description["\']\s+content=["\']([^"\']+)["\']', html, re.IGNORECASE)
        
        has_meta = bool(existing_meta)
        existing_description = existing_meta.group(1) if existing_meta else None
        
        # Generate new description
        new_description = generate_meta_description(page, api_keys)
        
        # Analyze existing if present
        issues = []
        if has_meta:
            if len(existing_description) < 120:
                issues.append('قصير جداً (أقل من 120 حرف)')
            elif len(existing_description) > 160:
                issues.append('طويل جداً (أكثر من 160 حرف)')
            
            if not any(keyword in existing_description.lower() for keyword in title.lower().split()[:3]):
                issues.append('لا يحتوي على كلمات مفتاحية من العنوان')
        
        results.append({
            'url': url,
            'title': title,
            'has_meta': has_meta,
            'existing': existing_description,
            'suggested': new_description,
            'issues': issues,
            'status': 'جيد ✓' if has_meta and not issues else 'يحتاج تحسين ⚠️' if has_meta else 'مفقود ❌'
        })
    
    return results


def get_meta_description_recommendations(audit: Dict) -> Dict:
    """Get summary and recommendations for meta descriptions"""
    pages = audit.get('pages', [])
    
    total_pages = len(pages)
    pages_with_meta = 0
    pages_with_good_meta = 0
    pages_missing_meta = 0
    
    for page in pages:
        html = page.get('html', '')
        meta_match = re.search(r'<meta\s+name=["\']description["\']\s+content=["\']([^"\']+)["\']', html, re.IGNORECASE)
        
        if meta_match:
            pages_with_meta += 1
            description = meta_match.group(1)
            if 120 <= len(description) <= 160:
                pages_with_good_meta += 1
        else:
            pages_missing_meta += 1
    
    score = int((pages_with_good_meta / total_pages * 100)) if total_pages > 0 else 0
    
    recommendations = []
    if pages_missing_meta > 0:
        recommendations.append(f'أضف meta description لـ {pages_missing_meta} صفحة')
    
    if pages_with_meta - pages_with_good_meta > 0:
        recommendations.append(f'حسّن meta description لـ {pages_with_meta - pages_with_good_meta} صفحة')
    
    return {
        'score': score,
        'total_pages': total_pages,
        'pages_with_meta': pages_with_meta,
        'pages_with_good_meta': pages_with_good_meta,
        'pages_missing_meta': pages_missing_meta,
        'status': 'ممتاز ✅' if score >= 90 else 'جيد ✓' if score >= 70 else 'يحتاج تحسين ⚠️',
        'recommendations': recommendations
    }