-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathcollect_results.py
138 lines (117 loc) · 5.54 KB
/
collect_results.py
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
# This program collects results from all the metrics across different 5 runs and does average.
import argparse
import numpy as np
if __name__=='__main__':
parser = argparse.ArgumentParser()
parser.add_argument('-res', '--result', help='Result File', default='out_fb_top_5_1000_con_tail_rel_5_times')
args = parser.parse_args()
result_file = args.result
# Read file
res = open(result_file, 'r').read().split('\n')[:-1]
res_model_runs_5 = []
# Metrics for one setup of CompGCN model, i.e., num_ins, num_context
train_acc = []
best_valid_acc = []
mrr_compgcn = []
hits1_compgcn = []
hits3_compgcn = []
hits10_compgcn = []
hits10_50_compgcn = []
# Metrics for one setup of a CompGCN_NBFNet, i.e., num_ins, num_context
mrr_compgcn_nbfnet = []
hits1_compgcn_nbfnet = []
hits3_compgcn_nbfnet = []
hits10_compgcn_nbfnet = []
hits10_50_compgcn_nbfnet = []
for i in range(len(res)):
if 'Epoch: 049' in res[i]:
t_acc = res[i].split(',')[1].split(':')[1]
val_acc = res[i].split(',')[3].split(':')[1]
train_acc.append(float(t_acc))
best_valid_acc.append(float(val_acc))
elif '>>>>>>>>>>>>' in res[i] and 'Early Stopping' not in res[i - 1]:
t_acc = res[i-1].split(',')[1].split(':')[1]
val_acc = res[i-1].split(',')[3].split(':')[1]
train_acc.append(float(t_acc))
best_valid_acc.append(float(val_acc))
elif 'MRR for CompGCN :' in res[i]:
mrr_compgcn.append(float(res[i].split(':')[1].split('(')[1][:-1]))
elif 'Hits@1 for CompGCN :' in res[i]:
hits1_compgcn.append(float(res[i].split(':')[1]))
elif 'Hits@3 for CompGCN :' in res[i]:
hits3_compgcn.append(float(res[i].split(':')[1]))
elif 'Hits@10 for CompGCN :' in res[i]:
hits10_compgcn.append(float(res[i].split(':')[1]))
elif 'Hits@10_50 for CompGCN :' in res[i]:
hits10_50_compgcn.append(float(res[i].split(':')[1].split('(')[1][:-1]))
elif 'MRR for CompGCN_NBFNet' in res[i]:
mrr_compgcn_nbfnet.append(float(res[i].split(':')[1].split('(')[1][:-1]))
elif 'Hits@1 for CompGCN_NBFNet' in res[i]:
hits1_compgcn_nbfnet.append(float(res[i].split(':')[1]))
elif 'Hits@3 for CompGCN_NBFNet' in res[i]:
hits3_compgcn_nbfnet.append(float(res[i].split(':')[1]))
elif 'Hits@10 for CompGCN_NBFNet' in res[i]:
hits10_compgcn_nbfnet.append(float(res[i].split(':')[1]))
elif 'Hits@10_50 for CompGCN_NBFNet' in res[i]:
hits10_50_compgcn_nbfnet.append(float(res[i].split(':')[1].split('(')[1][:-1]))
'''
print('mrr_compgcn', mrr_compgcn)
print('hits1_compgcn', hits1_compgcn)
print('hits3_compgcn', hits3_compgcn)
print('hits10_compgcn', hits10_compgcn)
print('hits10_50_compgcn', hits10_50_compgcn)
print('mrr_compgcn_nbfnet', mrr_compgcn_nbfnet)
print('hits1_compgcn_nbfnet', hits1_compgcn_nbfnet)
print('hits3_compgcn_nbfnet', hits3_compgcn_nbfnet)
print('hits10_compgcn_nbfnet', hits10_compgcn_nbfnet)
print('hits10_50_compgcn_nbfnet', hits10_50_compgcn_nbfnet)
print('==========================================================================================')
'''
print('mrr_compgcn', len(mrr_compgcn))
print(len(train_acc))
print(len(best_valid_acc))
# Average and output the results in the required format [MRR, Hits@1, Hits@3, Hits@10, Hits@10_50]
for i in range(len(mrr_compgcn)//5):
tr_acc = train_acc[i*5:(i*5+5)]
v_acc = best_valid_acc[i*5:(i*5+5)]
mrr = mrr_compgcn[i*5:(i*5+5)]
hits1 = hits1_compgcn[i*5:(i*5+5)]
hits3 = hits3_compgcn[i * 5:(i * 5 + 5)]
hits10 = hits10_compgcn[i * 5:(i * 5 + 5)]
hits10_50 = hits10_50_compgcn[i * 5:(i * 5 + 5)]
mrr_nbf = mrr_compgcn_nbfnet[i*25:(i*25+25)]
hits1_nbf = hits1_compgcn_nbfnet[i*25:(i*25+25)]
hits3_nbf = hits3_compgcn_nbfnet[i * 25:(i * 25 + 25)]
hits10_nbf = hits10_compgcn_nbfnet[i * 25:(i * 25 + 25)]
hits10_50_nbf = hits10_50_compgcn_nbfnet[i * 25:(i * 25 + 25)]
if np.inf in mrr:
#print('Handling Inf')
inf_index = set()
for j in range(len(mrr)):
if mrr[j] == np.inf:
inf_index.add(j)
#print(inf_index)
non_inf_index = set.difference(set(range(5)), inf_index)
#print(non_inf_index)
sum_ind = 0
for ind in non_inf_index:
sum_ind += mrr[ind]
mrr_compgcn_avg = sum_ind/len(non_inf_index)
else:
mrr_compgcn_avg = sum(mrr) / 5
tr_acc_avg = sum(tr_acc) / 5
v_acc_avg = sum(v_acc) / 5
hits1_avg = sum(hits1) / 5
hits3_avg = sum(hits3) / 5
hits10_avg = sum(hits10) / 5
hits10_50_avg = sum(hits10_50) / 5
mrr_nbf_avg = sum(mrr_nbf)/25
hits1_nbf_avg = sum(hits1_nbf) / 25
hits3_nbf_avg = sum(hits3_nbf) / 25
hits10_nbf_avg = sum(hits10_nbf) / 25
hits10_50_nbf_avg = sum(hits10_50_nbf) / 25
print('Model_%d_compgcn' % (i + 1), tr_acc_avg, v_acc_avg)
print('Model_%d_compgcn' %(i+1), mrr_compgcn_avg, hits1_avg, hits3_avg, hits10_avg, hits10_50_avg)
print('Model_%d_compgcn_nbfnet' % (i + 1), mrr_nbf_avg, hits1_nbf_avg, hits3_nbf_avg, hits10_nbf_avg,
hits10_50_nbf_avg)
print('==========================================================================================')