import pandas as pd
import tensorflow as tf
from keras.models import Sequential
from keras.layers import Dense
from keras import utils
import os

EPOCHS = int(os.environ['EPOCHS'])
if EPOCHS <= 0:
    EPOCHS = 1000

X_train = pd.read_csv('./X_train.csv',
                              engine = 'python',
                              encoding = 'ISO-8859-1',
                              sep=',')

X_val = pd.read_csv('./X_val.csv',
                              engine = 'python',
                              encoding = 'ISO-8859-1',
                              sep=',')

Y_train = pd.read_csv('./Y_train.csv',
                              engine = 'python',
                              encoding = 'ISO-8859-1',
                              sep=',')

Y_val = pd.read_csv('./Y_val.csv',
                              engine = 'python',
                              encoding = 'ISO-8859-1',
                              sep=',')


Y_train = utils.to_categorical(Y_train)
Y_val = utils.to_categorical(Y_val)

model = Sequential(
     [               
         Dense(100, input_dim=X_train.shape[1], activation='relu'),
         Dense(70, activation='relu'),
         Dense(50, activation='relu'),
         Dense(4, activation='softmax')
     ], name = "Powerlifters_model"
)

model.compile(
    loss=tf.keras.losses.CategoricalCrossentropy(),
    optimizer=tf.keras.optimizers.Adam(),
    metrics=['accuracy'])

model.fit(
  X_train,Y_train,
    epochs = EPOCHS,
    validation_data=(X_val, Y_val)
)

model.save('model')