preProcess.py

import pandas as pd
from sklearn.preprocessing import StandardScaler, MinMaxScaler
import pickle
import numpy as np


def scaled(train_data):
    
    feature_scaler = StandardScaler()
    
    feature_columns = ['Duration 1', 'Duration 2', 'Sensor Value', 'Sensor Value 2', 'Consumable Life']
    
    train_data[feature_columns] = train_data[feature_columns].astype(np.float32)

    train_data[feature_columns] = feature_scaler.fit_transform(train_data[feature_columns])

    with open('feature_scaler.pkl', 'wb') as f:
        pickle.dump(feature_scaler, f)
    
    return train_data

def split(data):
    all_run_ids = data['Run ID'].unique()  
    total = len(all_run_ids)

    train_end = int(0.7 * total)
    val_end = int(0.85 * total)

    # train_end = int(1)
    # val_end = int(2)

    train_ids = all_run_ids[:train_end]
    val_ids = all_run_ids[train_end:val_end]
    test_ids = all_run_ids[val_end:]

    train_data = data[data['Run ID'].isin(train_ids)]
    val_data = data[data['Run ID'].isin(val_ids)]
    test_data = data[data['Run ID'].isin(test_ids)]
    print("Train size:", len(train_data))
    print("Val size:", len(val_data))
    print("Test size:", len(test_data))

    return train_data.reset_index(drop=True), val_data.reset_index(drop=True), test_data.reset_index(drop=True)