ium_444452/Scripts/train_neural_network.py

#!/usr/bin/python
import datetime
import os
import sys
import pandas as pd
from keras.models import Sequential
from keras import layers
from keras.preprocessing.text import Tokenizer
from keras.preprocessing.sequence import pad_sequences
import logging

logging.getLogger("tensorflow").setLevel(logging.ERROR)

data_path = ''
num_words = 0
epochs = 0
batch_size = 0
pad_length = 0


def tokenize(x, x_train):
    global pad_length, num_words
    tokenizer = Tokenizer(num_words=num_words)
    tokenizer.fit_on_texts(x)
    train_x = tokenizer.texts_to_sequences(x_train)
    vocabulary_length = len(tokenizer.word_index) + 1
    train_x = pad_sequences(train_x, padding='post', maxlen=pad_length)
    return train_x, vocabulary_length


def save_model(model):
    # model_name = 'neural_net_' + datetime.datetime.today().strftime('%d-%b-%Y-%H:%M:%S')
    model_name = 'neural_net'
    model.save(os.path.join(os.getcwd(), 'model', model_name), save_format='h5', overwrite=True)


def train_model(model, x_train, y_train):
    global epochs, batch_size
    model.fit(x_train, y_train, epochs=epochs, verbose=False, batch_size=batch_size)


def get_model(vocabulary_length):
    global pad_length, batch_size
    model = Sequential()
    model.add(layers.Embedding(input_dim=vocabulary_length,
                               output_dim=batch_size,
                               input_length=pad_length))
    model.add(layers.Flatten())
    model.add(layers.Dense(10, activation='relu'))
    model.add(layers.Dense(1, activation='sigmoid'))
    model.compile(optimizer='adam',
                  loss='binary_crossentropy',
                  metrics=['accuracy'])
    return model


def split_data(data):
    x = data['tokens']
    y = data['fraudulent']
    return x, y


def load_data(data_path, filename) -> pd.DataFrame:
    return pd.read_csv(os.path.join(data_path, filename))


def read_params():
    global data_path, num_words, epochs, batch_size, pad_length
    data_path, num_words, epochs, batch_size, pad_length = sys.argv[1].split(',')
    num_words = int(num_words)
    epochs = int(epochs)
    batch_size = int(batch_size)
    pad_length = int(pad_length)


def main():
    read_params()
    global data_path
    abs_data_path = os.path.abspath(data_path)
    train_data = load_data(abs_data_path, 'train_data.csv')
    test_data = load_data(abs_data_path, 'test_data.csv')
    x_train, y_train = split_data(train_data)
    x_test, _ = split_data(test_data)
    x_train, vocab_size = tokenize(pd.concat([x_train, x_test]), x_train)
    model = get_model(vocab_size)
    train_model(model, x_train, y_train)
    save_model(model)


if __name__ == '__main__':
    main()