#!/usr/bin/env python3

import subprocess
import re
import csv
import os
from collections import defaultdict

def categorize_amex_transaction(description):
    description = description.lower()
    
    if any(keyword in description for keyword in ['carrefour', 'run market', 'intermarche']):
        return 'Groceries'
    if any(keyword in description for keyword in ['esko bar', 'le choka bleu', 'columbus cafe']):
        return 'Restaurants/Food'
    if any(keyword in description for keyword in ['openrouter', 'stripe-z.ai', 'claude.ai', 'ama eu sarl prime_new', 'scaleway', 'servperso* invoice pro']):
        return 'Online Services/Subscriptions'
    if any(keyword in description for keyword in ['air austral', 'run duty free', 'lm saint louis leroym4']):
        return 'Travel'
    if any(keyword in description for keyword in ['mon brico', 'sumup*kulture metisse', 'sumup*glamport', 'relay']):
        return 'Shopping'
        
    return 'Other'

def process_amex_files(file_list, output_csv=False, output_dir='../../output/csv'):
    expense_summary = defaultdict(float)
    total_expenses = 0
    all_transactions = []

    for file_path in file_list:
        try:
            result = subprocess.run(['pdftotext', '-layout', file_path, '-'], capture_output=True, text=True, check=True)
            content = result.stdout
        except (subprocess.CalledProcessError, FileNotFoundError) as e:
            print(f"Error processing {file_path}: {e}")
            continue

        # Regex for amex transactions
        transaction_regex = re.compile(r'(\d{1,2} \w{3})\s+\d{1,2} \w{3}\s+(.*?)\s+([\d,.]+)$(?<!CR$)', re.MULTILINE)
        
        lines = content.split('\n')
        for line in lines:
            # A simple heuristic to find transaction lines
            if re.match(r'\d{1,2} \w{3}', line) and not line.endswith('CR'):
                parts = line.split()
                if len(parts) > 3:
                    try:
                        date = parts[0] + ' ' + parts[1]
                        amount_str = parts[-1].replace(',', '.')
                        amount = float(amount_str)
                        description = ' '.join(parts[2:-1])
                        
                        category = categorize_amex_transaction(description)
                        expense_summary[category] += amount
                        total_expenses += amount
                        
                        # Store transaction for CSV output
                        all_transactions.append({
                            'Date': date,
                            'Description': description,
                            'Category': category,
                            'Amount': amount,
                            'Source': os.path.basename(file_path)
                        })
                    except (ValueError, IndexError):
                        continue
    
    # Output CSV if requested
    if output_csv and all_transactions:
        csv_file = os.path.join(output_dir, 'american_express_all_transactions.csv')
        os.makedirs(output_dir, exist_ok=True)
        with open(csv_file, 'w', newline='', encoding='utf-8') as csvfile:
            fieldnames = ['Date', 'Description', 'Category', 'Amount', 'Source']
            writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
            writer.writeheader()
            writer.writerows(all_transactions)
        print(f"\nTransaction data saved to {csv_file}")

    print("--- American Express Expense Summary for 2025 ---")
    print(f"Total Expenses Analyzed: €{total_expenses:,.2f}")
    print("\n--- Spending by Category ---")
    
    sorted_expenses = sorted(expense_summary.items(), key=lambda item: item[1], reverse=True)
    
    if total_expenses > 0:
        for category, total in sorted_expenses:
            percentage = (total / total_expenses) * 100
            print(f"{category:<25} €{total:9,.2f} ({percentage:5.2f}%)")
    else:
        print("No expenses found.")
    
    return all_transactions

if __name__ == "__main__":
    import argparse
    import glob
    
    parser = argparse.ArgumentParser(description='Process American Express statements')
    parser.add_argument('--pdf-dir', default='../data/pdf/american_express', 
                       help='Directory containing American Express PDF files')
    parser.add_argument('--output-dir', default='../../output/csv', 
                       help='Directory to save CSV output files')
    parser.add_argument('--csv', action='store_true', 
                       help='Output transaction data to CSV files')
    args = parser.parse_args()
    
    # Get all PDF files in the directory
    pdf_files = glob.glob(os.path.join(args.pdf_dir, "*.pdf"))
    
    # Sort files by date if possible
    pdf_files.sort()
    
    # Process all PDF files in the directory
    process_amex_files(pdf_files, args.csv, args.output_dir)