DOLPHIN/nautilus_dolphin/test_5s_replay_parity.py

import sys
import time
from pathlib import Path
from datetime import datetime, timezone
import numpy as np
import pandas as pd

# Ensure we can import from the project root
sys.path.insert(0, str(Path(__file__).parent))

# Import Dolphin components
from nautilus_dolphin.nautilus.esf_alpha_orchestrator import NDAlphaEngine

# Mock configuration
CONFIG = {
    'engine': {
        'initial_capital': 25000.0,
        'vel_div_threshold': -0.02,
        'vel_div_extreme': -0.05,
        'max_leverage': 5.0,
        'use_alpha_layers': True,
        'seed': 42
    },
    'live_mode': False,
    'direction': 'short_only'
}

def test_parity():
    print("=== DOLPHIN-NAUTILUS PARITY TEST ===")
    
    # 1. Setup Data (load one sample day)
    vbt_dir = Path(r"C:\Users\Lenovo\Documents\- DOLPHIN NG HD HCM TSF Predict\vbt_cache")
    pq_files = sorted(vbt_dir.glob("*.parquet"))
    if not pq_files:
        print("Error: No parquet files found in vbt_cache")
        return
    
    sample_file = pq_files[0]
    date_str = sample_file.stem
    print(f"Testing for date: {date_str} using {sample_file.name}")
    
    df = pd.read_parquet(sample_file)
    asset_columns = [c for c in df.columns if c not in {
        'timestamp', 'scan_number', 'v50_lambda_max_velocity', 'v150_lambda_max_velocity',
        'v300_lambda_max_velocity', 'v750_lambda_max_velocity', 'vel_div',
        'instability_50', 'instability_150'
    }]
    
    # Pre-calculate vol_ok (p60 threshold from dynamic_beta_validate)
    vol_p60 = 0.000099 # Hardcoded champion p60
    bp = df['BTCUSDT'].values if 'BTCUSDT' in df.columns else None
    dvol = np.full(len(df), np.nan)
    if bp is not None:
        for i in range(50, len(bp)):
            seg = bp[max(0,i-50):i]
            if len(seg) >= 10:
                dvol[i] = float(np.std(np.diff(seg)/seg[:-1]))
    vol_ok_arr = np.where(np.isfinite(dvol), dvol > vol_p60, False)

    # 2. Path A: Legacy Batch (NDAlphaEngine.process_day)
    print("\nRunning Path A: Legacy Batch logic...")
    engine_a = NDAlphaEngine(**CONFIG['engine'])
    engine_a.set_acb(None) 
    
    batch_stats = engine_a.process_day(
        date_str=date_str,
        df=df,
        asset_columns=asset_columns,
        vol_regime_ok=vol_ok_arr,
        direction=-1
    )
    print(f"  Batch Result: PnL={batch_stats['pnl']:+.4f} Trades={batch_stats['trades']} Capital={batch_stats['capital']:.2f}")

    # 3. Path B: Modern Incremental (Direct NDAlphaEngine methods)
    print("\nRunning Path B: Modern Incremental logic (direct methods)...")
    engine_b = NDAlphaEngine(**CONFIG['engine'])
    engine_b.set_acb(None)
    
    # Orchestrate calls manually to match DolphinActor logic
    engine_b.begin_day(date_str, posture='APEX', direction=-1)
    
    for ri in range(len(df)):
        row = df.iloc[ri]
        vd = row.get('vel_div')
        if vd is None or not np.isfinite(float(vd)):
            engine_b._global_bar_idx += 1 # MUST increment to maintain parity with process_day logic
            continue

        v50_raw = row.get('v50_lambda_max_velocity')
        v750_raw = row.get('v750_lambda_max_velocity')
        v50_val = float(v50_raw) if (v50_raw is not None and np.isfinite(float(v50_raw))) else 0.0
        v750_val = float(v750_raw) if (v750_raw is not None and np.isfinite(float(v750_raw))) else 0.0

        prices = {}
        for ac in asset_columns:
            p = row.get(ac)
            if p is not None and p > 0 and np.isfinite(p):
                prices[ac] = float(p)
        
        if not prices:
             engine_b._global_bar_idx += 1 # MUST increment to maintain parity
             continue
             
        vol_ok = bool(vol_ok_arr[ri])
        
        # Step the bar (internally increments _global_bar_idx)
        engine_b.step_bar(
            bar_idx=ri,
            vel_div=float(vd),
            prices=prices,
            vol_regime_ok=vol_ok,
            v50_vel=v50_val,
            v750_vel=v750_val
        )

    incremental_stats = engine_b.end_day()
    print(f"  Incremental Result: PnL={incremental_stats['pnl']:+.4f} Trades={incremental_stats['trades']} Capital={incremental_stats['capital']:.2f}")

    # 4. Parity Check
    pnl_diff = abs(batch_stats['pnl'] - incremental_stats['pnl'])
    trade_diff = abs(batch_stats['trades'] - incremental_stats['trades'])
    
    print("\n=== PARITY ANALYSIS ===")
    print(f"  PnL Difference:    {pnl_diff:.8f}")
    print(f"  Trade Difference:  {trade_diff}")
    
    success = pnl_diff < 1e-6 and trade_diff == 0
    if success:
        print("\nSUCCESS: Batch and Incremental logic paths are identical.")
    else:
        print("\nFAILURE: Logic divergence detected.")
        sys.exit(1)

if __name__ == "__main__":
    test_parity()