DOLPHIN/nautilus_dolphin/test_forewarner_gamut.py

import os
import sys
import numpy as np
import pandas as pd
from collections import Counter
from pathlib import Path

# Adjust paths
PROJECT_ROOT = Path(__file__).resolve().parent
sys.path.insert(0, str(PROJECT_ROOT))

from mc.mc_ml import DolphinForewarner
from mc.mc_sampler import MCSampler

MODELS_DIR = PROJECT_ROOT / "mc_results" / "models"

def determine_risk_level(report):
    env = report.envelope_score
    cat = report.catastrophic_probability
    champ = report.champion_probability

    if cat > 0.25 or env < -1.0:
        return "RED"
    elif env < 0 or cat > 0.10:
        return "ORANGE"
    elif env > 0 and champ > 0.4:
        return "AMBER"
    elif env > 0.5 and champ > 0.6:
        return "GREEN"
    else:
        return "AMBER"

def run_gamut_test():
    print("======================================================================")
    print("MONTE CARLO FOREWARNER GAMUT TEST ('Montecarlo the Montecarlo')")
    print("======================================================================")
    
    if not MODELS_DIR.exists():
        print(f"[ERROR] Models directory not found at {MODELS_DIR}.")
        return

    print("[1/3] Initializing ML Models...")
    try:
        forewarner = DolphinForewarner(models_dir=str(MODELS_DIR))
    except Exception as e:
        print(f"[ERROR] Failed to load ML models: {e}")
        return
        
    print("[2/3] Generating 2,500 random parameter configurations across the manifold...")
    sampler = MCSampler(base_seed=1337)
    # Generate a large set of raw trials (they don't even need to be strictly 'valid' 
    # for the ML model to assess them, but we'll use the sampler to get structurally sound vectors)
    trials = sampler.generate_trials(n_samples_per_switch=10, max_trials=2500)
    print(f"      Generated {len(trials)} unique parameter vectors.")
    
    print("[3/3] Feeding gamut into the Forewarning Alert System...")
    
    results = []
    level_counts = Counter()
    
    for i, trial in enumerate(trials):
        config_dict = trial.to_dict()
        report = forewarner.assess_config_dict(config_dict)
        level = determine_risk_level(report)
        
        level_counts[level] += 1
        results.append({
            'trial_id': i,
            'level': level,
            'env_score': report.envelope_score,
            'cat_prob': report.catastrophic_probability,
            'champ_prob': report.champion_probability,
            'predicted_roi': report.predicted_roi,
            'config': config_dict
        })
        
        if (i + 1) % 500 == 0:
            print(f"      Assessed {i + 1} / {len(trials)} configs...")

    print("\n======================================================================")
    print("FOREWARNER RESPONSE DISTRIBUTION")
    print("======================================================================")
    
    total = len(trials)
    print(f"GREEN  (Safe / Optimal):      {level_counts['GREEN']:>4} ({level_counts['GREEN']/total*100:.1f}%)")
    print(f"AMBER  (Transitional):        {level_counts['AMBER']:>4} ({level_counts['AMBER']/total*100:.1f}%)")
    print(f"ORANGE (High Risk Warning):   {level_counts['ORANGE']:>4} ({level_counts['ORANGE']/total*100:.1f}%)")
    print(f"RED    (Catastrophic Alert):  {level_counts['RED']:>4} ({level_counts['RED']/total*100:.1f}%)")
    
    print("\n======================================================================")
    print("SAMPLE EXTREMUM PROFILES")
    print("======================================================================")
    
    # Sort by envelope score
    results_sorted = sorted(results, key=lambda x: x['env_score'])
    
    print("\\n--- WORST 'RED' CONFIGURATION DETECTED ---")
    worst = results_sorted[0]
    print(f"Envelope Score: {worst['env_score']:.3f} | Catastrophic Prob: {worst['cat_prob']:.1%} | Predicted ROI: {worst['predicted_roi']:.1f}%")
    print(f"Key Triggers: max_leverage={worst['config']['max_leverage']:.2f}, fraction={worst['config']['fraction']:.2f}, vel_div_extreme={worst['config']['vel_div_extreme']:.4f}")
    
    print("\n--- BEST 'GREEN' CONFIGURATION DETECTED ---")
    best = results_sorted[-1]
    print(f"Envelope Score: {best['env_score']:.3f} | Champion Prob: {best['champ_prob']:.1%} | Predicted ROI: {best['predicted_roi']:.1f}%")
    print(f"Key Triggers: max_leverage={best['config']['max_leverage']:.2f}, fraction={best['config']['fraction']:.2f}, vel_div_extreme={best['config']['vel_div_extreme']:.4f}")

if __name__ == "__main__":
    run_gamut_test()