"""
Convert comprehensive knowledge base TXT to ALU Brain JSON format
This will make your knowledge base actually usable by the chatbot!
"""

import json
from pathlib import Path
import re

def parse_comprehensive_kb():
    """Parse the comprehensive knowledge base text file"""
    
    kb_file = Path("data/alu_knowledge/alu_ultimate_knowledge_base.txt")
    
    if not kb_file.exists():
        print(f"❌ Knowledge base file not found: {kb_file}")
        return None
    
    with open(kb_file, 'r', encoding='utf-8') as f:
        content = f.read()
    
    # Split by major sections (## headers)
    sections = re.split(r'\n##\s+', content)
    
    entries = []
    entry_id = 1
    
    for section in sections:
        if not section.strip():
            continue
            
        lines = section.split('\n')
        title = lines[0].strip()
        
        # Get the content
        content_lines = []
        current_subsection = ""
        
        for line in lines[1:]:
            if line.strip():
                if line.startswith('###'):  # Subsection
                    current_subsection = line.replace('###', '').strip()
                elif line.startswith('-') or line.startswith('*'):  # List item
                    content_lines.append(line.strip())
                else:
                    content_lines.append(line.strip())
        
        full_content = '\n'.join(content_lines)
        
        # Create entry
        entry = {
            "id": f"comprehensive_kb_{entry_id}",
            "question": f"Tell me about {title}",
            "answer": full_content,
            "type": "text",
            "keywords": [word.lower() for word in title.split() if len(word) > 3],
            "category": "Comprehensive Knowledge Base",
            "source": "ALU Comprehensive KB"
        }
        
        entries.append(entry)
        entry_id += 1
        
        # Also create entries for common questions
        if "club" in title.lower():
            entries.append({
                "id": f"comprehensive_kb_{entry_id}",
                "question": "How to create a club at ALU?",
                "answer": full_content,
                "type": "text",
                "keywords": ["club", "create", "start", "society", "organization"],
                "category": "Campus Life",
                "source": "ALU Comprehensive KB"
            })
            entry_id += 1
        
        if "founder" in title.lower() or "leadership" in title.lower():
            entries.append({
                "id": f"comprehensive_kb_{entry_id}",
                "question": "Who founded ALU?",
                "answer": full_content,
                "type": "text",
                "keywords": ["founder", "fred", "swaniker", "leadership", "history"],
                "category": "About ALU",
                "source": "ALU Comprehensive KB"
            })
            entry_id += 1
    
    return entries

def create_brain_json():
    """Create the ALU Brain JSON file"""
    
    print("============================================================")
    print("🧠 CONVERTING COMPREHENSIVE KB TO ALU BRAIN FORMAT")
    print("============================================================\n")
    
    entries = parse_comprehensive_kb()
    
    if not entries:
        print("❌ No entries created")
        return False
    
    # Create the brain JSON structure
    brain_data = {
        "category": "comprehensive_knowledge",
        "description": "Comprehensive ALU Knowledge Base - All ALU information",
        "version": "1.0",
        "last_updated": "2024-11-20",
        "entries": entries
    }
    
    # Save to alu_brain directory
    output_file = Path("alu_brain/comprehensive_knowledge.json")
    with open(output_file, 'w', encoding='utf-8') as f:
        json.dump(brain_data, f, indent=2, ensure_ascii=False)
    
    print(f"✅ Created {len(entries)} entries")
    print(f"📁 Saved to: {output_file}")
    print(f"\n📝 Sample entries:")
    for entry in entries[:3]:
        print(f"  - {entry['question']}")
    
    print("\n============================================================")
    print("✅ CONVERSION COMPLETE!")
    print("============================================================")
    print("\nNext steps:")
    print("1. Commit and push this file to Hugging Face")
    print("2. Restart the Space")
    print("3. The chatbot will now use your comprehensive knowledge!")
    
    return True

if __name__ == "__main__":
    create_brain_json()