forecaster_engine.py

import os
import json
import pandas as pd
import numpy as np
from datetime import datetime

DATA_FILE = os.path.join(os.path.dirname(__file__), "data", "nifty50_daily.parquet")
PREDICTIONS_FILE = os.path.join(os.path.dirname(__file__), "predictions.json")

def generate_predictions():
    if not os.path.exists(DATA_FILE):
        print(f"Data file missing: {DATA_FILE}")
        return
        
    df_all = pd.read_parquet(DATA_FILE)
    tickers = df_all['ticker'].unique()
    
    predictions = {}
    forecast_date = None
    
    for ticker in tickers:
        df = df_all[df_all['ticker'] == ticker].copy()
        df.sort_values('date', inplace=True)
        df.set_index('date', inplace=True)
        
        if len(df) < 130:
            continue
            
        forecast_date_ts = df.index[-1] + pd.Timedelta(days=1)
        # Advance past weekends roughly for display
        if forecast_date_ts.weekday() >= 5:
            forecast_date_ts += pd.Timedelta(days=(7 - forecast_date_ts.weekday()))
            
        if forecast_date is None:
            forecast_date = forecast_date_ts.strftime('%Y-%m-%d')
            
        daily_close = df['close']
        
        # Features
        delta = daily_close.diff()
        gain = (delta.where(delta > 0, 0)).rolling(window=2).mean()
        loss = (-delta.where(delta < 0, 0)).rolling(window=2).mean()
        rs = gain / loss
        rsi_2 = 100 - (100 / (1 + rs))
        
        sma_3 = daily_close.rolling(window=3).mean()
        dist_sma3 = daily_close / sma_3
        
        sma_5 = daily_close.rolling(window=5).mean()
        dist_sma5 = daily_close / sma_5
        
        df_eval = pd.DataFrame({
            'close': daily_close,
            'rsi_2': rsi_2,
            'dist_sma3': dist_sma3,
            'dist_sma5': dist_sma5,
            '1d_ret': daily_close.pct_change()
        }).dropna()
        
        # Target for historical testing
        df_eval['actual_dir'] = np.where(df_eval['close'].shift(-1) > df_eval['close'], 1, -1)
        
        # The last row is TODAY. We don't have tomorrow's close, so actual_dir is wrong for the last row.
        # We test on the 120 days BEFORE today
        test_set = df_eval.iloc[-121:-1]
        today_data = df_eval.iloc[-1]
        
        best_acc = 0
        best_rule = None
        
        # 1. RSI-2 threshold
        for thresh in [10, 20, 30, 40, 50, 60, 70, 80, 90]:
            for op in ['<', '>']:
                sig = np.where(test_set['rsi_2'] < thresh if op == '<' else test_set['rsi_2'] > thresh, 1, -1)
                acc = (sig == test_set['actual_dir']).mean()
                if acc > best_acc:
                    best_acc = acc
                    best_rule = ('rsi_2', thresh, op)
                    
        # 2. SMA distance threshold
        for feature in ['dist_sma3', 'dist_sma5']:
            for thresh in [0.95, 0.98, 1.0, 1.02, 1.05]:
                sig = np.where(test_set[feature] < thresh, 1, -1)
                acc = (sig == test_set['actual_dir']).mean()
                if acc > best_acc:
                    best_acc = acc
                    best_rule = (feature, thresh, '<')
                    
        # 3. 1d return
        sig = np.where(test_set['1d_ret'] < 0, 1, -1)
        acc = (sig == test_set['actual_dir']).mean()
        if acc > best_acc:
            best_acc = acc
            best_rule = ('1d_ret', 0, '<')
            
        feature, thresh, op = best_rule
        val = today_data[feature]
        
        if op == '<':
            prediction = 1 if val < thresh else -1
        else:
            prediction = 1 if val > thresh else -1
            
        predictions[ticker] = {
            "prediction": "UP" if prediction == 1 else "DOWN",
            "probability": round(best_acc * 100, 2),
            "rule_used": f"{feature} {op} {thresh}"
        }
        
    output = {
        "generated_at": datetime.now().isoformat(),
        "forecast_date": forecast_date,
        "predictions": predictions
    }
    
    with open(PREDICTIONS_FILE, "w") as f:
        json.dump(output, f, indent=4)
        
    print(f"Generated predictions for {forecast_date}. Saved to {PREDICTIONS_FILE}")
    return output

if __name__ == "__main__":
    generate_predictions()