Brightside

Python library for plotting common plots and formatting text and tables in Jupyter Notebook.

Examples

Loading the libraries:

%matplotlib inline

import brightside as bs
import matplotlib
from matplotlib import pyplot as plt
from scipy import stats
import numpy as np

Loss and Accuracy plots

Loading the log data into numpy arrays.

epochs, train_loss, train_acc, val_loss, val_acc = np.loadtxt('training.log', delimiter=' ', 
                                                              usecols=(1, 3, 5, 7, 9), unpack=True)
labels = ['Train', 'Validation']

Plotting the loss

fig, ax = bs.plot_loss([train_loss,val_loss], labels, epochs=epochs, figsize=(10,3))
plt.savefig('loss_plot.png', format='png', dpi=300)

Plotting the accuracy

fig, ax = bs.plot_accuracy([train_acc,val_acc], labels, epochs=epochs, figsize=(5,5))
plt.savefig('acc_plot.png', format='png', dpi=300)

Confusion matrix plot

Loading predictions and groundtruth into numpy arrays.

predictions, groundtruth = np.loadtxt('predictions.csv', delimiter=' ', unpack=True)
labels = np.genfromtxt('labels.txt',dtype='str')

Plotting the confusion matrix

fig, ax = bs.show_confusion_matrix(predictions, groundtruth, labels,
                                   figsize=(8,8), annot=False,
                                   linewidths=0.005, linecolor='gray',
                                   square=True, show_xticks=True,
                                   cbar_kws={"shrink": 0.75})
plt.savefig('cm.png', format='png', dpi=300)

Printing a table on Jupyter

Making an array of random numbers.

np.random.seed(42)
rand_numbers = np.random.uniform(0,1,100)

Showing the table

bs.print_table('Statistics', [('Min:', rand_numbers.min()),
('Max:', rand_numbers.max()),
('Mean:', rand_numbers.mean()),
('Median:', np.median(rand_numbers)),
('StdDev:', rand_numbers.std()),
('Mode:', stats.mode(rand_numbers).mode[0]),
('Distinct:', np.size(np.unique(rand_numbers)))])

Statistics

Min:	0.005522117123602399
Max:	0.9868869366005173
Mean:	0.47018074337820936
Median:	0.4641424546894926
StdDev:	0.29599822663249037
Mode:	0.005522117123602399
Distinct:	100

Plotting a sequence

Loading the sequential data

sequences = np.genfromtxt('sequences.csv', delimiter=' ', dtype='str')
labels = np.genfromtxt('activity_labels.txt', delimiter=',', dtype='str')

Preparing the sequences

import itertools
import pandas as pd

def time2ind(time):
    hour, minute, second = [int(time[i:i + 2]) for i in range(0, 6, 2)]
    index = 120 * hour + 2 * minute + (1 if second > 30 else 0)
    return index

days = -np.ones((2, 2880), np.int)
dates = ['2019-05-20', '2019-05-21']
for date, time, label in sequences:
    index = time2ind(time)
    i = 0 if date == '2019-05-20' else 1
    days[i, index] = int(label)

times = list()
for time in itertools.product(range(24), range(60), range(0, 60, 30)):
    times.append(':'.join([str(t).zfill(2) for t in time]))

colors = sns.color_palette("hls", len(labels))
labels_colors = {i: (label, colors[i]) for i, label in enumerate(labels)}

dataframe = pd.DataFrame(data=days, index=dates, columns=times)
fig, ax, leg = bs.show_sequences(dataframe, labels_colors, figsize=(8, 1.5), xticklabels=100, mask_value=-1,
                                 leg_square_size=7, show_box=True, plot_xlabel='Time', plot_ylabel='Date')
plt.savefig('sequences.png', format='png', dpi=300)