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
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
|
#!/usr/bin/env python3
import argparse
import logging
import multiprocessing
import os
import signal
import sys
from itertools import chain
from multiprocessing import Pool
from multiprocessing.util import Finalize
from pathlib import Path
from urllib.parse import urlparse
from fontTools.subset import Options, Subsetter, load_font
from selenium import webdriver
logging.basicConfig(format='[%(relativeCreated)d] %(message)s')
logger = logging.getLogger('websubset')
logger.setLevel(logging.INFO)
EXTRACT_SCRIPT = r'''
let whitelist = new Set(arguments[0]);
let walker = document.createTreeWalker(document.body, NodeFilter.SHOW_TEXT);
let node, dict = {};
while (node = walker.nextNode()) {
let cs = getComputedStyle(node.parentNode);
let css = k => cs.getPropertyValue(k);
if (css('display') == 'none')
continue;
let k = css('font-family').replace(/"/g, '').replace(/,.*/, '') + ';' +
css('font-weight') + ';' +
css('font-style');
if (!whitelist.has(k))
continue;
if (!(k in dict))
dict[k] = '';
dict[k] += node.nodeValue;
}
return dict;
'''
def gen_font_face(font):
if 'fontfile' not in font:
return ''
return ''.join([
'@font-face{',
'font-family:"', font['family'], '";',
'font-weight:', font['weight'], ';',
'font-style:', font['style'], ';',
'src: url("', font["fontfile"], '");',
'}'])
DRIVER = None
def stop_driver():
global DRIVER
if DRIVER:
DRIVER.quit()
DRIVER = None
def hook_sig(signum):
orig_handler = signal.getsignal(signum)
if orig_handler is None:
raise Exception('{signum} handler is None')
def term_handler(*_):
stop_driver()
signal.signal(signum, orig_handler)
os.kill(os.getpid(), signum)
signal.signal(signum, term_handler)
def start_wworker(driver_name):
# clamp selenium wasteful sleeps to 0.1s
import time
from time import sleep
time.sleep = lambda secs: sleep(min(secs, 0.1))
hook_sig(signal.SIGTERM)
global DRIVER
if driver_name == 'chrome':
chrome_options = webdriver.chrome.options.Options()
chrome_options.headless = True
chrome_options.experimental_options["prefs"] = {
"profile.default_content_setting_values.images": 2
}
DRIVER = webdriver.Chrome(options=chrome_options, desired_capabilities={'detach': True})
elif driver_name == 'firefox':
firefox_profile = webdriver.FirefoxProfile()
firefox_profile.set_preference('permissions.default.image', 2)
firefox_options = webdriver.firefox.options.Options()
firefox_options.headless = True
DRIVER = webdriver.Firefox(firefox_profile=firefox_profile, options=firefox_options)
else:
raise Exception('unknown driver name')
Finalize(DRIVER, stop_driver, exitpriority=16)
def is_uri(path):
parsed = urlparse(path)
return parsed.scheme and parsed.netloc
def make_uri(path):
if is_uri(path):
return path
else:
return Path(path).resolve().as_uri()
def extract(path, fonts, screenshots):
logger.info('fetching %s', path)
DRIVER.get(make_uri(path))
if screenshots:
logger.info('replacing fonts for %s', path)
height = DRIVER.execute_script(''.join([
"let style = document.createElement('style'); style.innerHTML = '",
''.join(gen_font_face(font) for font in fonts),
"'; document.body.appendChild(style); return document.documentElement.scrollHeight"]))
logger.info('taking pre-screenshot for %s', path)
DRIVER.set_window_size(1920, height)
screenshot = DRIVER.get_screenshot_as_png()
else:
screenshot = None
logger.info('extracting text from %s', path)
whitelist = [';'.join((f['family'], f['weight'], f['style'])) for f in fonts]
return (path, DRIVER.execute_script(EXTRACT_SCRIPT, whitelist), screenshot)
def get_fontdesc(fonts, fontspec):
font_match = dict(zip(('family', 'weight', 'style'), fontspec.split(';')))
for font in fonts:
if font_match.items() <= font.items():
return font
return None
def subset(fontdesc, text, fts_opts):
fontfile = fontdesc['fontfile']
logger.info('subsetting %s', fontfile)
font = load_font(fontfile, fts_opts, dontLoadGlyphNames=True)
subsetter = Subsetter(options=fts_opts)
subsetter.populate(text=text)
subsetter.subset(font)
ret = []
outfile = fontfile[:fontfile.rindex('.')] + '.subset.woff2'
ret.append((font, 'woff2', outfile))
return ret
def write_subset(font, flavor, outfile):
logger.info('writing %s', outfile)
font.flavor = flavor
font.save(outfile)
def verify(path, screenshot_begin_png):
logger.info('refetching %s', path)
DRIVER.get(make_uri(path))
from io import BytesIO
from PIL import Image, ImageChops
screenshot_begin = Image.open(BytesIO(screenshot_begin_png), formats=('PNG',)).convert('RGB')
logger.info('taking post-screenshot for %s', path)
DRIVER.set_window_size(*screenshot_begin.size)
screenshot_end_png = DRIVER.get_screenshot_as_png()
screenshot_end = Image.open(BytesIO(screenshot_end_png), formats=('PNG',)).convert('RGB')
logger.info('checking %s screenshots for %s', 'x'.join(map(str, screenshot_end.size)), path)
if ImageChops.difference(screenshot_begin, screenshot_end).getbbox():
raise Exception(f'screenshots do not match for {path}')
def main(argv):
parser = argparse.ArgumentParser(description='Web Font Subsetter', epilog='see pyftsubset --help for additional options')
parser.add_argument('--driver', help='selenium driver name (chrome or firefox)', default='chrome')
parser.add_argument('--no-screenshots', help='skip screenshot validation', action='store_true')
parser.add_argument('--font', help='add font (fontfile:family:weight:style)', action='append')
parser.add_argument('file', help='html files', nargs='+')
args, leftover = parser.parse_known_intermixed_args(argv)
options = Options()
files = args.file + options.parse_opts(leftover)
if any([file[0] == '-' for file in files]):
parser.print_usage()
raise Exception('bad arguments')
if options.with_zopfli:
from fontTools.ttLib import sfnt
sfnt.USE_ZOPFLI = True
ncpus = len(os.sched_getaffinity(0))
fonts = [dict(zip(['fontfile', 'family', 'weight', 'style'], font.split(':'))) for font in args.font]
nwworkers = min(len(files), ncpus)
nfworkers = min(len(fonts), ncpus)
logger.info('using %d web workers, %d font workers', nwworkers, nfworkers)
with Pool(nfworkers) as fpool, \
Pool(nwworkers, start_wworker, (args.driver,)) as wpool:
all_font_texts = {}
screenshots = []
extract_args = ((file, fonts, not args.no_screenshots) for file in args.file)
extracted = wpool.starmap(extract, extract_args)
for path, font_texts, screenshot in extracted:
if not args.no_screenshots:
screenshots.append((path, screenshot))
for fontspec, text in font_texts.items():
if fontspec in all_font_texts:
all_font_texts[fontspec] |= set(text)
else:
all_font_texts[fontspec] = set(text)
if args.no_screenshots:
logger.info('shutting down web workers early')
wpool.close()
subset_args = (
(get_fontdesc(fonts, fontspec), ''.join(text), options)
for fontspec, text in all_font_texts.items())
subsetted = fpool.starmap(subset, subset_args)
fpool.starmap(write_subset, chain(*subsetted))
if not args.no_screenshots:
wpool.starmap(verify, screenshots, 1)
if multiprocessing.active_children():
logger.info('waiting for workers')
for proc in multiprocessing.active_children():
proc.join()
logger.info('exiting successfully')
if __name__ == '__main__':
main(sys.argv[1:])
|
