gwac/plot_longrange.py at master · lukasjf/gwac · 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
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
from os import listdir
from collections import defaultdict
from ast import literal_eval
import torch
import numpy as np
import matplotlib.pyplot as plt

from labellines import labelLine, labelLines

results = {'oddeven':defaultdict(list), 'multioddeven':defaultdict(list)}

for f in listdir('longrange_results'):
    if not f.startswith('size_wise_acc'):
        continue
    print(f)
    with open('longrange_results/' + f, "r") as res:
        for line in res.readlines():
            model, dataset, seed, size, accuracy, nodes = line.split(',')
            size, accuracy, nodes = int(size), float(accuracy), int(nodes)
            if not dataset in results:
                results[dataset] = {}
            if not model in results[dataset]:
                results[dataset][model] = {}
            if not size in results[dataset][model]:
                results[dataset][model][size] = []
            results[dataset][model][size].append(accuracy)

model_order = ['gin', 'gcn', 'gat', 'neg', 'universal', 'itergnn', 'gwacs', 'gwaciter', 'gwacact']
for dataset in results:
    size_order = sorted(results[dataset][model_order[0]])
    break

print(model_order)
print(size_order)

for model in model_order:
    for dataset in results:
        print(model)
        if not model in results[dataset]:
            continue
        if dataset != "oddeven":
            continue
        accs = [np.mean(results[dataset][model][size]) for size in size_order]
        stds = [np.std(results[dataset][model][size]) for size in size_order]
        accs = [str(acc)[:4] for acc in accs]
        stds = [str(std)[:4] for std in stds]
        accs = ['\makebox{{{}\\rpm{}}}'.format(accs[i], stds[i]) for i in range(len(accs))]
        print(model, dataset)
        print(' & '.join(accs))

"""##############################################################################"""

str_for_model = {"gin": "GIN", "gat": "GAT", "gcn": "GCN", "neg":"NEG", "universal":"UT",
                 "itergnn":"IterGNN", "gwacs":"GwAC-S", "gwacgru":"GwAC-GRU",
                 "gwaclstm":"GwAC-LSTM", "gwacact": "GwAC-UT", "gwaciter": "GwAC-Iter",
                 "gwacatt": "GwAC-ATT"}

results = {}

for f in listdir('longrange_results'):
    if not f.startswith('distance_wise_acc'):
        continue
    with open('longrange_results/' + f, "r") as res:
        for line in res.readlines():
            model, dataset, seed, distance, accuracy, nodes = line.split(',')
            distance, accuracy, nodes = int(distance), float(accuracy), int(nodes)
            if not dataset in results:
                results[dataset] = {}
            if not model in results[dataset]:
                results[dataset][model] = {}
            if not distance in results[dataset][model]:
                results[dataset][model][distance] = []
            results[dataset][model][distance].append(accuracy)

dataset = 'oddeven'
plt.figure(figsize=(9, 4.5))
for m in model_order:
    xs = [d for d in results[dataset][m]]
    xs = xs[1::2]
    xs = [x for x in xs if x < 19]
    print(m, results[dataset][m])
    ys = [np.array(results[dataset][m][d-1]) + np.array(results[dataset][m][d]) for d in xs]
    stds = np.array([np.std(y) for y in ys])
    ys = np.array([np.mean(y)/2 for y in ys])
    xs = xs
    plt.plot(xs, ys, label="{}".format(str_for_model[m]))
plt.xticks([0] + list(range(1, 19, 2)), [0]+["{}-{}".format(i, i+1) for i in range(1, 19,2)])
plt.xlabel("Node distance")
plt.ylabel("Accuracy")
    #plt.fill_between(xs, ys-stds, ys+stds, alpha=0.3)
labelLines(plt.gca().get_lines(), align=True, xvals=[2, 4.5, 1.1, 2., 3.5, 4.5, 7.5, 9.5, 8.5], fontsize=14)
#plt.legend()
plt.savefig('underreaching.pdf')
#plt.show()
plt.clf()

print("###################UNDERREACHING########")

print(xs)
for d in xs:
    numbers= []
    for m in model_order:
        ys = np.array(results[dataset][m][d - 1]) + np.array(results[dataset][m][d])
        std = np.std(ys/4)
        ys = np.mean(ys/2)
        numbers.append('\makebox{{{}\\rpm{}}}'.format(str(ys)[:4], str(std)[:4]))
    print("{}-{} & ".format(d-1, d) + " & ".join(numbers) + "\\\\")


"""##############################################################################"""

results = {}

for f in listdir('longrange_results'):
    if not f.startswith('training_range_accs'):
        continue
    with open('longrange_results/' + f, "r") as res:
        for line in res.readlines():
            model, dataset, seed, size, accuracy = line.split(',')
            size, accuracy, nodes = int(size), float(accuracy), int(nodes)
            if not dataset in results:
                results[dataset] = {}
            if not model in results[dataset]:
                results[dataset][model] = {}
            if not size in results[dataset][model]:
                results[dataset][model][size] = []
            results[dataset][model][size].append(accuracy)

dataset = 'oddeven'
plt.figure(figsize=(9, 4.5))
for m in model_order:
    xs = [size for size in results[dataset][m]]
    xvalues = range(len(xs))
    ys = np.array([np.mean(results[dataset][m][d]) for d in xs])
    stds = np.array([np.std(results[dataset][m][d]) for d in xs])
    plt.plot(xvalues, ys, label="{}".format(str_for_model[m]))
    #plt.fill_between(xvalues, ys -stds, ys + stds,alpha=0.3)
    plt.xticks(xvalues, xs)
plt.xlabel("Graph size")
plt.ylabel("Accuracy")
labelLines(plt.gca().get_lines(), align=True, xvals=[1.6, 1.1, 0.1, 0.5, 2.5, 4.5, 6.2, 4.5, 6, 0.3], fontsize=14)
plt.savefig('oversmoothing.pdf')
#plt.show()
plt.clf()

print("###################OVERSMOOTHING########")

for model in model_order:
    for dataset in results:
        if dataset != "oddeven":
            continue
        print(size_order)
        accs = [np.mean(results[dataset][model][size]) for size in size_order]
        stds = [np.std(results[dataset][model][size]) for size in size_order]
        accs = [str(acc)[:4] for acc in accs]
        stds = [str(std)[:4] for std in stds]
        accs = ['\makebox{{{}\\rpm{}}}'.format(accs[i], stds[i]) for i in range(len(accs))]
        print(model, dataset)
        print(' & '.join(accs))


"""##############################################################################"""