"""
Convert 1-minute data to higher timeframes (5m, 15m, 1h, 4h, daily).

This allows testing the same strategy on different timeframes to find
the optimal trading interval.
"""

import pandas as pd
import os
import sys

sys.path.append(os.path.dirname(os.path.abspath(__file__)))
import config


def resample_to_timeframe(df, timeframe):
    """
    Resample 1-minute OHLCV data to a higher timeframe.
    
    Parameters
    ----------
    df : pd.DataFrame
        1-minute OHLCV data with DatetimeIndex
    timeframe : str
        Pandas resample frequency: '5min', '15min', '1H', '4H', '1D'
        
    Returns
    -------
    pd.DataFrame
        Resampled OHLCV data
    """
    # Ensure index is datetime
    if not isinstance(df.index, pd.DatetimeIndex):
        df = df.set_index('timestamp')
    
    # Resample OHLCV data
    resampled = df.resample(timeframe).agg({
        'open': 'first',
        'high': 'max',
        'low': 'min',
        'close': 'last',
        'volume': 'sum'
    }).dropna()
    
    return resampled


def create_all_timeframes(source_file):
    """
    Create datasets for all timeframes from 1-minute source data.
    
    Parameters
    ----------
    source_file : str
        Path to 1-minute CSV file
        
    Returns
    -------
    dict
        Dictionary of {timeframe: output_file_path}
    """
    print("\n" + "="*70)
    print("CREATING MULTI-TIMEFRAME DATASETS")
    print("="*70)
    
    # Load source data
    print(f"\n📥 Loading source data: {source_file}")
    df = pd.read_csv(source_file)
    df['timestamp'] = pd.to_datetime(df['timestamp'])
    df = df.set_index('timestamp')
    
    print(f"   Loaded {len(df):,} 1-minute candles")
    print(f"   Period: {df.index[0]} to {df.index[-1]}")
    
    # Define timeframes to create
    timeframes = {
        '5min': '5-minute',
        '15min': '15-minute',
        '30min': '30-minute',
        '1H': '1-hour',
        '4H': '4-hour',
        '1D': 'daily'
    }
    
    output_files = {}
    
    for tf_code, tf_name in timeframes.items():
        print(f"\n🔄 Creating {tf_name} dataset...")
        
        # Resample
        resampled = resample_to_timeframe(df, tf_code)
        
        # Save
        output_file = os.path.join(
            config.DATA_DIR, 
            f"btc_{tf_code.lower()}_binance.csv"
        )
        
        # Reset index to have timestamp as column
        resampled = resampled.reset_index()
        resampled.to_csv(output_file, index=False)
        
        output_files[tf_name] = output_file
        
        print(f"   ✓ Created {len(resampled):,} {tf_name} candles")
        print(f"   ✓ Saved to: {output_file}")
    
    return output_files


def estimate_trades_per_timeframe():
    """Display estimated trades per day for each timeframe."""
    print("\n" + "="*70)
    print("ESTIMATED TRADING CHARACTERISTICS BY TIMEFRAME")
    print("="*70)
    
    print("\nBased on current parameters (MA 15/60, ~2 trades/day on 1-min):\n")
    
    timeframes = [
        ("1-minute", 1, "~2 trades/day", "High frequency, many signals"),
        ("5-minute", 5, "~8-12 trades/week", "Active intraday"),
        ("15-minute", 15, "~2-4 trades/week", "Swing trading"),
        ("30-minute", 30, "~1-2 trades/week", "Position trading"),
        ("1-hour", 60, "~5-10 trades/month", "Low frequency"),
        ("4-hour", 240, "~1-3 trades/month", "Very selective"),
        ("Daily", 1440, "~2-6 trades/year", "Long-term holds")
    ]
    
    print(f"{'Timeframe':<15} {'Candle (min)':<15} {'Frequency':<20} {'Style'}")
    print("-" * 70)
    for tf, mins, freq, style in timeframes:
        print(f"{tf:<15} {mins:<15} {freq:<20} {style}")
    
    print("\n💡 Hypothesis:")
    print("   - Longer timeframes = Less noise, fewer false signals")
    print("   - Better win rate but fewer opportunities")
    print("   - May need to adjust MA periods (15/60 → 10/30 or 20/80)")


def main():
    """Main execution."""
    print("\n🌍 MULTI-TIMEFRAME DATA CREATOR")
    
    # Use the 2-year Binance data as source
    source_file = os.path.join(config.DATA_DIR, "btc_1m_binance_730d.csv")
    
    if not os.path.exists(source_file):
        print(f"\n✗ Source file not found: {source_file}")
        print(f"   Please ensure you have 1-minute data available.")
        return
    
    # Create all timeframes
    output_files = create_all_timeframes(source_file)
    
    # Show characteristics
    estimate_trades_per_timeframe()
    
    print("\n" + "="*70)
    print("✅ MULTI-TIMEFRAME DATASETS CREATED")
    print("="*70)
    
    print("\n📁 Created files:")
    for tf_name, filepath in output_files.items():
        print(f"   - {tf_name}: {filepath}")
    
    print("\n💡 Next steps:")
    print("  1. Test current parameters on each timeframe:")
    print("     - Update config.py to use each timeframe file")
    print("     - Run: python run_backtest.py")
    print("     - Compare results")
    print("\n  2. Optimize parameters per timeframe:")
    print("     - Run: python optimize_parameters.py (for each TF)")
    print("     - Find best settings for each interval")
    print("\n  3. Compare performance:")
    print("     - Which timeframe gives best Sharpe ratio?")
    print("     - Which has best win rate?")
    print("     - Which has lowest drawdown?")
    
    print("="*70 + "\n")


if __name__ == "__main__":
    main()