inference.py

"""
Nifty 50 Ensemble Inference Script
Usage:
    python inference.py --data data.csv --model models_v2_1m --horizon 1
"""
import argparse
import pandas as pd
import numpy as np
from nifty_ensemble_v2 import NiftyEnsembleV2

def main():
    parser = argparse.ArgumentParser(description='Nifty 50 Directional Prediction')
    parser.add_argument('--data', required=True, help='CSV with OHLCV columns (Datetime,Open,High,Low,Close,Volume)')
    parser.add_argument('--model', required=True, help='Model directory (e.g., models_v2_1m or models_v2_5m)')
    parser.add_argument('--horizon', type=int, default=1, help='Prediction horizon in minutes')
    parser.add_argument('--output', default='predictions.csv', help='Output CSV path')
    args = parser.parse_args()
    
    # Load data
    df = pd.read_csv(args.data, index_col='Datetime', parse_dates=True)
    print(f"[Inference] Loaded {len(df)} rows from {args.data}")
    
    # Load model
    pipe = NiftyEnsembleV2()
    pipe.load(args.model)
    
    # Predict
    results = pipe.predict(df)
    df_out = results['df'].copy()
    df_out['pred_up'] = results['pred']
    df_out['proba_up'] = results['proba'][:, 1]
    df_out['regime'] = results['regime']
    
    # Regime labels
    regime_labels = {0: 'Bull', 1: 'Bear', 2: 'Transitional', 3: 'High-Vol', 4: 'Low-Vol'}
    df_out['regime_label'] = df_out['regime'].map(lambda x: regime_labels.get(x, f'Regime_{x}'))
    
    # Signal strength
    df_out['signal'] = np.where(df_out['proba_up'] > 0.6, 'STRONG_UP',
                         np.where(df_out['proba_up'] < 0.4, 'STRONG_DOWN', 'NEUTRAL'))
    
    # Keep relevant columns
    out_cols = ['Open','High','Low','Close','target','pred_up','proba_up',
                'regime','regime_label','signal']
    available = [c for c in out_cols if c in df_out.columns]
    df_out[available].to_csv(args.output)
    
    print(f"[Inference] Saved predictions to {args.output}")
    print(f"  Up predictions: {df_out['pred_up'].sum()} ({df_out['pred_up'].mean()*100:.1f}%)")
    print(f"  Mean up-probability: {df_out['proba_up'].mean():.3f}")
    print(f"  Regime distribution:\n{df_out['regime_label'].value_counts().to_string()}")

if __name__ == '__main__':
    main()