plot_results.py

import seaborn as sns
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import argparse

parser = argparse.ArgumentParser(description='run training loop')
parser.add_argument('csv')
args = parser.parse_args()


solution_length = pd.read_csv(args.csv)
solution_length_filtered = solution_length[solution_length['DeepQube'] != 9999]
success = pd.read_csv(args.csv)
success['DeepQube'] = success['DeepQube'].apply(
    lambda x: 'Solved' if x != 9999 else 'Fail')

num_samples = len(solution_length_filtered['DeepQube'])

gods_number_data = []
gods_number_frequency = num_samples*0.01*np.array([0.015, 0.061, 0.24, 0.90,
                                                   3.11, 9.81, 25.33, 36.77,
                                                   21.3, 2.46, 0.0075])

for i in range(4, 15):
    for _ in range(round(gods_number_frequency[i-4])):
        if len(gods_number_data) < num_samples:
            gods_number_data.append(i)

solution_length_filtered.insert(2, "GodsNumber", gods_number_data, True)

plt.figure(1, figsize=(16, 9))
plot1 = sns.histplot(data=solution_length_filtered, binwidth=1, legend=False)
plt.xticks(fontsize='x-large')
plt.yticks(fontsize='x-large')
plt.legend(('Optimal Solver', 'Baseline', 'DeepQube'), fontsize='xx-large')
plt.xlabel('Solution move count (quarter turn metric)', fontsize='xx-large')
plt.ylabel('Count', fontsize='xx-large')
plt.ylim(bottom=0)

plt.savefig('histogram.png', dpi=300)

plt.figure(2)
plot2 = sns.countplot(x='DeepQube', data=success, edgecolor=(
    0, 0, 0), palette=['limegreen', 'orangered'])
plot2.set(xlabel='', ylabel='Count')
i = 0
for p in plot2.patches:
    h = p.get_height()

    count = success['DeepQube'].value_counts()[i]

    plot2.text(p.get_x() + p.get_width()/2.0, h + 50,
               f'{count} ({count/5000 * 100}%)', ha='center')
    i += 1

plt.savefig('count_plot.png', dpi=300)