graf.py

# coding=utf-8

# Guardo métodos generales de visualización
# -----------------------------------------------------------------------------

import matplotlib.pyplot as plt
import matplotlib.mlab as mlab
import numpy as np


def graf_GP(x, mu, std):
    plt.plot(x, mu, color="#4682b4", alpha=0.3)
    plt.fill(np.concatenate([x, x[::-1]]), np.concatenate([mu - 1.9600 * std,
             (mu + 1.9600 * std)[::-1]]), alpha=.5, fc='#C0C0C0', ec='None',
             label='95% confidence interval')

def graf_hist(values, label):
    # std = np.std(values)
    # mean = np.mean(values)
    # x = np.linspace(mean - 4 * std, mean + 4 * std, 100)
    # plt.plot(x, mlab.normpdf(x, mean, std), 'k--')

    n, bins, patches = plt.hist(values, 60, normed=False, histtype='bar', alpha=0.6,
                                label=label)

def sample_hist(values, real = None, feature_name='Feature', save_path='', show=False):
    """

    Parameters
    ----------
    values: Valores con los que generar el histograma (features sampleadas)
    real: Valor real de la feature
    feature_name: Nombre de la feature
    """
    mean = np.mean(values)
    std = np.std(values)

    x = np.linspace(mean - 4*std,mean +4*std,100)

    plt.figure(figsize=(6*3.13,9))
    plt.plot(x,mlab.normpdf(x,mean,std), 'k--')

    n, bins, patches = plt.hist(values, 30, normed=1, histtype='bar', color = 'c', alpha=0.6)

    plt.axvline(x= real, color = 'r', label=u'Real value')

    plt.title(feature_name, size=18)
    plt.xlabel('Feature Value', size=14)
    #plt.ylabel('Probability density', rotation = 'horizontal', labelpad=60, size=14)


    plt.legend(loc='upper right')

    if show:
        plt.show()
        
    if save_path != '':
        plt.savefig(save_path + feature_name + '.png')

    plt.close()