poker_hand_prediction.py

'''
Contains things common to all models implemented here; such as imports, configs, etc.
'''

# -------------------- Imports -------------------- #
import numpy as np, pandas as pd, os
from prettytable import PrettyTable


# -------------------- Globals and Configs -------------------- #
feature_names = list()
for index in range(1, 6):
    feature_names.extend(["Suit"+str(index), "Rank"+str(index)])

feature_names.append('class')


training_input_file = os.path.abspath('../datasets/csv/train.csv')
testing_input_file = os.path.abspath('../datasets/csv/test.csv')

np.random.seed(666)     # seed for reproducible results

# To store configs
class myConfigs:
    features = 0
    classes = 0


config = myConfigs()


# -------------------- Data -------------------- #

train_data = pd.read_csv(training_input_file, names=feature_names)
test_data = pd.read_csv(testing_input_file, names=feature_names)

# Get features of data
config.features = len(train_data.columns) - 1
config.classes = len(set(train_data['class']))

# Shuffle training data
train_data = train_data.sample(frac=1).reset_index(drop=True)

# Seperate data and classes
train_y = np.array(train_data['class'])
train_x = np.array(train_data.drop('class', 1))

test_y = np.array(test_data['class'])
test_x = np.array(test_data.drop('class', 1))

if __name__ == '__main__':
    tab = PrettyTable(['Config', 'Value'])
    configs = vars(config)

    for key in configs:
        tab.add_row([key, configs[key]])
    print(tab)