plebo/plot_results_summary.py at main · sighellan/plebo · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
from plot_cluster_grouped_results import plot_clustered_subfig, groups
from optimisation import plot_order, opt_names

import matplotlib.pyplot as plt
import pickle
import numpy as np
plt.rcParams["font.family"] = "Times New Roman"

experiment_ids = {
    'synth': '835893-840977',
    'strong': '837796',
    'selection': '838921'
}

rows = {
    'Direct': 1,
    'Prior': 2,
    'No transfer': 0
}

columns = {
    'synth': 0,
    'strong': 1,
    'selection': 2
}

titles = ['Synthetic', 'Strong', 'Selection']

def dict_key(dd, val):
    for kk, vv in dd.items():
        if vv == val:
            return kk
    return None


scale = 0.8

fig, axes = plt.subplots(3, 3, figsize=(scale*10,scale*6))

for exp_label in experiment_ids.keys():
    exp_id = experiment_ids[exp_label]
    results = pickle.load(open('Figures/results_plots/%s.p' %exp_id, 'rb'))
    N_tasks, N_its = np.shape(results['HBO_numpyro'])
    for subset_label in rows.keys():
        ax = axes[rows[subset_label], columns[exp_label]]
        # Plot groups separately to ease interpretation
        plot_subset = groups[subset_label]
        plot_clustered_subfig(ax, plot_subset, results, N_tasks, N_its, mute=True, alpha=0.1)

for ii in range(2):
    for jj in range(3):
        axes[ii,jj].set_xticks([])

for ii in range(3):
    for jj in range(3):
        axes[ii,jj].spines['top'].set_visible(False)
        axes[ii,jj].spines['right'].set_visible(False)

for ii in range(3):
    ymin, ymax = np.inf, -np.inf
    for jj in range(3):
        ymin = min(ymin, axes[ii,jj].get_ylim()[0])
        ymax = max(ymax, axes[ii,jj].get_ylim()[1])

    for jj in range(3):
        axes[ii,jj].set_ylim([ymin, ymax])

        if jj != 0:
            axes[ii,jj].set_yticks([])

axes[2,1].set_xlabel('Iteration')
axes[1,0].set_ylabel("Perf. compared to PLeBO")

for jj in range(3):
    axes[0,jj].set_title(titles[jj])

handles_all, labels_all = [], []
for jj in range(3):
    handles, labels = axes[jj,0].get_legend_handles_labels()
    handles_all += handles
    labels_all += labels

filtered_handles, filtered_labels = [], []
for jj in range(len(labels_all)):
    if labels_all[jj] not in filtered_labels:
        filtered_labels.append(labels_all[jj])
        filtered_handles.append(handles_all[jj])

order_list = [plot_order[dict_key(opt_names, ff)] for ff in filtered_labels]

sorted_handles = [xx for _, xx in sorted(zip(order_list, filtered_handles))]
sorted_labels = [xx for _, xx in sorted(zip(order_list, filtered_labels))]

fig.legend(sorted_handles, sorted_labels, loc='lower center', ncols=5, frameon=False,
          bbox_to_anchor=(0.5, -0.1))

plt.savefig('Figures/results_plots/subsets/overview_results.png',
            bbox_inches='tight', dpi=400)