import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler
import numpy as np
import kaggle

kaggle.api.authenticate()

kaggle.api.dataset_download_files('andrewmvd/heart-failure-clinical-data', path='.', unzip=True)

results = pd.read_csv('heart_failure_clinical_records_dataset.csv')

#brak wierszy z NaN
results.dropna()

results = results.astype({"age": np.int64})


for col in results.columns:
    if results[col].dtype == np.float64:  # FLOATS TO VALUES IN [ 0, 1]
        dataReshaped = results[col].values.reshape(-1, 1)
        scaler = MinMaxScaler(feature_range=(0, 1))
        results[col] = scaler.fit_transform(dataReshaped)


# Podział zbioru 6:1:1
train, test = train_test_split(results, test_size= 1 - 0.6)

valid, test = train_test_split(test, test_size=0.5)

train.to_csv("train.csv", index=False)
valid.to_csv("valid.csv",index=False)
test.to_csv("test.csv",index=False)