1
|
|
|
import json |
2
|
|
|
import os |
3
|
|
|
from pathlib import Path |
4
|
|
|
|
5
|
|
|
from ..utils import commonpath |
6
|
|
|
from ..utils import safe_dumps |
7
|
|
|
from ..utils import short_filename |
8
|
|
|
|
9
|
|
|
|
10
|
|
|
class FileStorage(object): |
11
|
|
|
def __init__(self, path, logger, default_machine_id=None): |
12
|
|
|
self.path = Path(path) |
13
|
|
|
self.default_machine_id = default_machine_id |
14
|
|
|
if not self.path.exists(): |
15
|
|
|
self.path.mkdir(parents=True) |
16
|
|
|
self.path = self.path.resolve() |
17
|
|
|
self.logger = logger |
18
|
|
|
self._cache = {} |
19
|
|
|
|
20
|
|
|
def __str__(self): |
21
|
|
|
return str(self.path) |
22
|
|
|
|
23
|
|
|
@property |
24
|
|
|
def location(self): |
25
|
|
|
return str(self.path.relative_to(os.getcwd())) |
26
|
|
|
|
27
|
|
|
def get(self, name): |
28
|
|
|
path = self.path.joinpath(self.default_machine_id) if self.default_machine_id else self.path |
29
|
|
|
if not path.exists(): |
30
|
|
|
path.mkdir(parents=True) |
31
|
|
|
return path.joinpath(name) |
32
|
|
|
|
33
|
|
|
@property |
34
|
|
|
def _next_num(self): |
35
|
|
|
files = self.query("[0-9][0-9][0-9][0-9]_*") |
36
|
|
|
files.sort(reverse=True) |
37
|
|
|
if not files: |
38
|
|
|
return "0001" |
39
|
|
|
for f in files: |
40
|
|
|
try: |
41
|
|
|
return "%04i" % (int(str(f.name).split('_')[0]) + 1) |
42
|
|
|
except ValueError: |
43
|
|
|
raise |
44
|
|
|
|
45
|
|
|
def save(self, output_json, save): |
46
|
|
|
output_file = self.get("%s_%s.json" % (self._next_num, save)) |
47
|
|
|
assert not output_file.exists() |
48
|
|
|
with output_file.open('wb') as fh: |
49
|
|
|
fh.write(safe_dumps(output_json, ensure_ascii=True, indent=4).encode()) |
50
|
|
|
self.logger.info("Saved benchmark data in: %s" % output_file) |
51
|
|
|
|
52
|
|
|
def query(self, *globs_or_files): |
53
|
|
|
files = [] |
54
|
|
|
globs = [] |
55
|
|
|
if not globs_or_files: |
56
|
|
|
globs_or_files = "*", |
57
|
|
|
|
58
|
|
|
for globish in globs_or_files: |
59
|
|
|
candidate = Path(globish) |
60
|
|
|
try: |
61
|
|
|
is_file = candidate.is_file() |
62
|
|
|
except OSError: |
63
|
|
|
is_file = False |
64
|
|
|
if is_file: |
65
|
|
|
files.append(candidate) |
66
|
|
|
continue |
67
|
|
|
|
68
|
|
|
parts = candidate.parts |
69
|
|
|
if len(parts) > 2: |
70
|
|
|
raise ValueError("{0!r} isn't an existing file or acceptable glob. " |
71
|
|
|
"Expected 'platform-glob/filename-glob' or 'filename-glob'.".format(globish)) |
72
|
|
|
elif len(parts) == 2: |
73
|
|
|
platform_glob, filename_glob = parts |
74
|
|
|
else: |
75
|
|
|
platform_glob = self.default_machine_id or "*" |
76
|
|
|
filename_glob, = parts or [''] |
77
|
|
|
|
78
|
|
|
filename_glob = filename_glob.rstrip("*") + "*.json" |
79
|
|
|
globs.append((platform_glob, filename_glob)) |
80
|
|
|
|
81
|
|
|
files.extend(( |
82
|
|
|
file |
83
|
|
|
for platform_glob, filename_glob in globs |
84
|
|
|
for path in self.path.glob(platform_glob) |
85
|
|
|
for file in path.glob(filename_glob) |
86
|
|
|
)) |
87
|
|
|
files.extend(( |
88
|
|
|
file |
89
|
|
|
for _, filename_glob in globs |
90
|
|
|
for file in self.path.glob(filename_glob) |
91
|
|
|
)) |
92
|
|
|
return sorted(files, key=lambda file: (file.name, file.parent)) |
93
|
|
|
|
94
|
|
|
def load(self, *globs_or_files): |
95
|
|
|
if not globs_or_files: |
96
|
|
|
globs_or_files = '[0-9][0-9][0-9][0-9]_', |
97
|
|
|
|
98
|
|
|
for file in self.query(*globs_or_files): |
99
|
|
|
if file in self._cache: |
100
|
|
|
data = self._cache[file] |
101
|
|
|
else: |
102
|
|
|
with file.open("rU") as fh: |
103
|
|
|
try: |
104
|
|
|
data = json.load(fh) |
105
|
|
|
except Exception as exc: |
106
|
|
|
self.logger.warn("BENCHMARK-C5", |
107
|
|
|
"Failed to load {0}: {1}".format(file, exc), fslocation=self.location) |
108
|
|
|
continue |
109
|
|
|
self._cache[file] = data |
110
|
|
|
try: |
111
|
|
|
relpath = file.relative_to(self.path) |
112
|
|
|
except ValueError: |
113
|
|
|
relpath = file |
114
|
|
|
yield relpath, data |
115
|
|
|
|
116
|
|
|
def load_benchmarks(self, *globs_or_files): |
117
|
|
|
sources = [ |
118
|
|
|
(short_filename(path), path, data) |
119
|
|
|
for path, data in self.load(*globs_or_files) |
120
|
|
|
] |
121
|
|
|
common = len(commonpath([src for src, _, _ in sources])) if sources else 0 |
122
|
|
|
for source, path, data in sources: |
123
|
|
|
source = source[common:].lstrip(r'\/') |
124
|
|
|
|
125
|
|
|
for bench in data["benchmarks"]: |
126
|
|
|
bench.update(bench.pop("stats")) |
127
|
|
|
bench['path'] = str(path) |
128
|
|
|
bench['source'] = source |
129
|
|
|
yield bench |
130
|
|
|
|