composition.py

import json
from sklearn import tree
import numpy as np

from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn import svm
from sklearn.naive_bayes import GaussianNB
from sklearn.neighbors import KNeighborsClassifier


class Composition(object):
    def __init__(self):
        with open('history.json') as data_file:
            self.history = json.load(data_file)

        X = []
        y = []
        i = 0
        for game in self.history:
            winner_composition = self.composition_classe_to_number(game.get('winner').get('composition'))
            loser_composition = self.composition_classe_to_number(game.get('loser').get('composition'))

            if i == 1:
                i = 0
                X.append([winner_composition, loser_composition])
                y.append([1, len(game.get('winner').get('deaths'))])
            else:
                i = 1
                X.append([loser_composition, winner_composition])
                y.append([-1, - len(game.get('winner').get('deaths'))])

        X_fit = X[+100:]
        y_fit = y[+100:]

        X_test = X[:+100]
        y_test = y[:+100]
        X_fit = np.array(X_fit)
        y_fit = np.array(y_fit)
        nsamples, nx, ny = X_fit.shape
        X_fit = X_fit.reshape((nsamples, nx * ny))

        X_test = np.array(X_test)
        y_test = np.array(y_test)
        nsamples, nx, ny = X_test.shape
        X_test = X_test.reshape((nsamples, nx * ny))

        model = RandomForestClassifier()
        # Train the model using the training sets and check score
        model.fit(X_fit, y_fit)
        # #Equation coefficient and Intercept
        predict = model.predict(X_test)
        print(predict)
        print(y_test)
        win = 0
        deathdiff = 0

        i = 0
        comparison = []
        for prediction in predict:
            result = y_test[i]
            comparison.append([prediction[0] - result[0], prediction[1] - result[1]])

        for comp in comparison:
            if comp[0] == 0:
                win += 1
            deathdiff += abs(comp[1])

        print(win)
        print(deathdiff)

    # #
    # for prediction in predict - y_test:
    # 	if i == 0:
    # 		win = win + 1
    # 	else:
    # 		lose = lose + 1

    # print("wins : ", win)
    # print("loses : ", lose)
    def composition_classe_to_number(self, composition):
        number_composition = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
        for classe in composition:
            number_composition[self.classe_to_number(classe)] = 1

        return number_composition

    def composition_number_to_classe(self, composition):
        classe_composition = []
        for number in composition:
            classe_composition.append(self.number_to_classe(number))

        return classe_composition

    def classe_to_number(self, classe):

        return {
            'Osamodas': 0,
            'Sacrieur': 1,
            'Pandawa': 2,
            'Eniripsa': 3,
            'Eliotrope': 4,
            'Enutrof': 5,
            'Iop': 6,
            'Sadida': 7,
            'Sram': 8,
            'Feca': 9,
            'Ecaflip': 10,
            'Zobal': 11,
            'Cra': 12,
            'Steamer': 13,
            'Xelor': 14,
            'Roublard': 15
        }[classe]

    def number_to_classe(self, number):
        return {
            0: 'Osamodas',
            1: 'Sacrieur',
            2: 'Pandawa',
            3: 'Eniripsa',
            4: 'Eliotrope',
            5: 'Enutrof',
            6: 'Iop',
            7: 'Sadida',
            8: 'Sram',
            9: 'Feca',
            10: 'Ecaflip',
            11: 'Zobal',
            12: 'Cra',
            13: 'Steamer',
            14: 'Xelor',
            15: 'Roublard'
        }[number]


composition = Composition()