forked from nfrechette/acl
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathgen_summary_stats.py
124 lines (102 loc) · 5.43 KB
/
gen_summary_stats.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
import numpy
import os
import sys
# This script depends on a SJSON parsing package:
# https://pypi.python.org/pypi/SJSON/1.1.0
# https://shelter13.net/projects/SJSON/
# https://bitbucket.org/Anteru/sjson/src
import sjson
if __name__ == "__main__":
if len(sys.argv) != 2:
print('Usage: python gen_summary_stats.py <path/to/input_file.sjson>')
sys.exit(1)
input_sjson_file = sys.argv[1]
if not input_sjson_file.endswith('.sjson'):
print('Expected SJSON input file, found: {}'.format(input_sjson_file))
sys.exit(1)
if not os.path.exists(input_sjson_file):
print('Input file not found: {}'.format(input_sjson_file))
sys.exit(1)
with open(input_sjson_file, 'r') as file:
input_sjson_data = sjson.loads(file.read())
input_data_type_def = {
'names': ('algorithm_names', 'raw_sizes', 'compression_ratios', 'durations', 'max_errors'),
'formats': ('S128', 'f4', 'f4', 'f4', 'f4')
}
columns_to_extract_acl_pre_v06 = (0, 1, 3, 5, 7)
columns_to_extract_acl_post_v06 = (1, 2, 4, 6, 8)
columns_to_extract_ue4 = (0, 1, 3, 4, 5)
output_csv_file_path_ratios = 'D:\\acl-dev\\tools\\graph_generation\\compression_ratios.csv'
output_csv_file_path_ratios_by_raw_size = 'D:\\acl-dev\\tools\\graph_generation\\compression_ratios_by_raw_size.csv'
output_csv_file_path_max_errors = 'D:\\acl-dev\\tools\\graph_generation\\max_errors.csv'
output_csv_file_path_max_errors_by_raw_size = 'D:\\acl-dev\\tools\\graph_generation\\max_errors_by_raw_size.csv'
output_csv_file_path_ratio_vs_max_error = 'D:\\acl-dev\\tools\\graph_generation\\ratio_vs_max_error.csv'
output_csv_file_path_durations = 'D:\\acl-dev\\tools\\graph_generation\\durations.csv'
output_csv_data_ratios = []
output_csv_data_ratios_by_raw_size = []
output_csv_data_max_errors = []
output_csv_data_max_errors_by_raw_size = []
output_csv_data_ratio_vs_max_error = []
output_csv_data_durations = []
output_csv_headers = []
for entry in input_sjson_data['inputs']:
if entry['version'] >= 0.6:
if entry['source'] == 'acl':
columns_to_extract = columns_to_extract_acl_post_v06
else:
columns_to_extract = columns_to_extract_ue4
else:
if entry['source'] == 'acl':
columns_to_extract = columns_to_extract_acl_pre_v06
else:
columns_to_extract = columns_to_extract_ue4
print('Parsing {} ...'.format(entry['header']))
csv_data = numpy.loadtxt(entry['file'], delimiter=',', dtype=input_data_type_def, skiprows=1, usecols=columns_to_extract)
filter = entry.get('filter', None)
if filter != None:
best_variable_data_mask = csv_data['algorithm_names'] == bytes(entry['filter'], encoding = 'utf-8')
csv_data = csv_data[best_variable_data_mask]
csv_data_ratios = numpy.sort(csv_data, order='compression_ratios')['compression_ratios']
csv_data_ratios_by_raw_size = numpy.sort(csv_data, order='raw_sizes')['compression_ratios']
csv_data_max_errors = numpy.sort(csv_data, order='max_errors')['max_errors']
csv_data_max_errors_by_raw_size = numpy.sort(csv_data, order='raw_sizes')['max_errors']
csv_data_ratio_by_max_error = numpy.sort(csv_data, order='max_errors')['compression_ratios']
csv_data_durations = numpy.sort(csv_data, order='durations')['durations']
output_csv_data_ratios.append(csv_data_ratios)
output_csv_data_ratios_by_raw_size.append(csv_data_ratios_by_raw_size)
output_csv_data_max_errors.append(csv_data_max_errors)
output_csv_data_max_errors_by_raw_size.append(csv_data_max_errors_by_raw_size)
output_csv_data_ratio_vs_max_error.append(csv_data_max_errors)
output_csv_data_ratio_vs_max_error.append(csv_data_ratio_by_max_error)
output_csv_data_durations.append(csv_data_durations)
output_csv_headers.append(entry['header'])
output_csv_data_ratios = numpy.column_stack(output_csv_data_ratios)
output_csv_data_ratios_by_raw_size = numpy.column_stack(output_csv_data_ratios_by_raw_size)
output_csv_data_max_errors = numpy.column_stack(output_csv_data_max_errors)
output_csv_data_max_errors_by_raw_size = numpy.column_stack(output_csv_data_max_errors_by_raw_size)
output_csv_data_ratio_vs_max_error = numpy.column_stack(output_csv_data_ratio_vs_max_error)
output_csv_data_durations = numpy.column_stack(output_csv_data_durations)
with open(output_csv_file_path_ratios, 'wb') as f:
header = bytes('{}\n'.format(','.join(output_csv_headers)), 'utf-8')
f.write(header)
numpy.savetxt(f, output_csv_data_ratios, delimiter=',', fmt=('%f'))
with open(output_csv_file_path_ratios_by_raw_size, 'wb') as f:
header = bytes('{}\n'.format(','.join(output_csv_headers)), 'utf-8')
f.write(header)
numpy.savetxt(f, output_csv_data_ratios_by_raw_size, delimiter=',', fmt=('%f'))
with open(output_csv_file_path_max_errors, 'wb') as f:
header = bytes('{}\n'.format(','.join(output_csv_headers)), 'utf-8')
f.write(header)
numpy.savetxt(f, output_csv_data_max_errors, delimiter=',', fmt=('%f'))
with open(output_csv_file_path_max_errors_by_raw_size, 'wb') as f:
header = bytes('{}\n'.format(','.join(output_csv_headers)), 'utf-8')
f.write(header)
numpy.savetxt(f, output_csv_data_max_errors_by_raw_size, delimiter=',', fmt=('%f'))
with open(output_csv_file_path_ratio_vs_max_error, 'wb') as f:
header = bytes('{}\n'.format(','.join(output_csv_headers)), 'utf-8')
f.write(header)
numpy.savetxt(f, output_csv_data_ratio_vs_max_error, delimiter=',', fmt=('%f'))
with open(output_csv_file_path_durations, 'wb') as f:
header = bytes('{}\n'.format(','.join(output_csv_headers)), 'utf-8')
f.write(header)
numpy.savetxt(f, output_csv_data_durations, delimiter=',', fmt=('%f'))