def coverage(font, threshold=10):
cmap = set(chr(c) for c in font.getBestCmap())
languages = set()
scripts = set()
partial = {}
for locale in Locale.getAvailableLocales():
data = LocaleData(locale)
examplar = set("".join(data.getExemplarSet()))
if not cmap.isdisjoint(examplar):
locale = Locale(locale)
locale.addLikelySubtags()
diff = examplar - cmap
if not diff:
scripts.add(locale.getDisplayScript())
languages.add(locale.getDisplayLanguage())
elif len(diff) <= threshold:
partial[locale.getDisplayLanguage()] = diff
return scripts, languages, partial
try:
from icu import Locale, Collator
HAVE_ICU = True
except ImportError:
try:
from PyICU import Locale, Collator
HAVE_ICU = True
except ImportError as err:
# No logger, save the warning message for later.
_icu_err = (
"ICU not loaded because %s. Localization will be impaired. "
"Use your package manager to install PyICU" % str(err))
ICU_LOCALES = None
if HAVE_ICU:
ICU_LOCALES = Locale.getAvailableLocales()
# Map of languages for converting to Microsoft locales and naming
# locales for display to the user. It's important to add to this list
# when a new translation is added. Note the dummy _(): That's just to
# get xgettext to include the string in gramps.pot; actual translation
# is done in _get_language_string() below.
# (The gramps officially-supported language list is ALL_LINGUAS in setup.py)
_ = lambda x: x
_LOCALE_NAMES = {
'ar': ('Arabic_Saudi Arabia', '1256', _("Arabic")),
'bg': ('Bulgrian_Bulgaria', '1251', _("Bulgarian")),
'br': (None, None, _("Breton")), #Windows has no translation for Breton
'ca': ('Catalan_Spain', '1252', _("Catalan")),
'cs': ('Czech_Czech Republic', '1250', _("Czech")),
'da': ('Danish_Denmark', '1252', _("Danish")),
# -*- coding: utf-8 -*-
"""
"""
import os
import json
from icu import Locale
BASE_PATH = os.path.dirname(os.path.abspath(__file__))
locales = []
for locale in Locale.getAvailableLocales().values():
locales.append({'locale': locale.getName(),
'name': locale.getDisplayName(locale)})
json.dump(locales, open(os.path.join(BASE_PATH, 'locales.json'), 'w'))
# LOG.setLevel(logging.DEBUG)
try:
from icu import Locale, Collator
HAVE_ICU = True
except ImportError:
try:
from PyICU import Locale, Collator
HAVE_ICU = True
except ImportError as err:
# No logger, save the warning message for later.
_icu_err = ("ICU not loaded because %s. Localization will be impaired. "
"Use your package manager to install PyICU" % str(err))
ICU_LOCALES = None
if HAVE_ICU:
ICU_LOCALES = Locale.getAvailableLocales()
# Map of languages for converting to Microsoft locales and naming
# locales for display to the user. It's important to add to this list
# when a new translation is added. Note the dummy _(): That's just to
# get xgettext to include the string in wearnow.pot; actual translation
# is done in _get_language_string() below.
# (The wearnow officially-supported language list is ALL_LINGUAS in setup.py)
_ = lambda x: x
_LOCALE_NAMES = {
'ar': ('Arabic_Saudi Arabia', '1256', _("Arabic")),
'bg': ('Bulgrian_Bulgaria', '1251', _("Bulgarian")),
'br': (None, None, _("Breton")), #Windows has no translation for Breton
'ca': ('Catalan_Spain', '1252', _("Catalan")),
'cs': ('Czech_Czech Republic', '1250', _("Czech")),
'da': ('Danish_Denmark', '1252', _("Danish")),
fonts = get_sys_fonts()
font_names = np.array(list(fonts.keys()))
print("Found %d system fonts" % len(fonts), file=sys.stderr)
# Make a huge sparse binary matrix that gives the availability of glyphs for each char in each font
glyph_avail = lil_matrix(
(max(b['stop'] for b in unicode_blocks.values()), len(font_names)),
dtype=np.uint8)
for i, (name, font) in tqdm(enumerate(fonts.items()),
total=len(fonts),
desc="Checking glyph availability"):
chars = get_unicode_tables_by_font(font)
glyph_avail[chars, i] = 1
# Package all locales
locales = [(k, v) for k, v in Locale.getAvailableLocales().items()
if k in [
'en', 'te', 'th', 'vi', 'ar', 'he', 'km', 'ta', 'gu', 'bn',
'ml', 'el', 'ru', 'ko', 'zh', 'ja'
]]
for code, locale in tqdm(locales, desc="Packaging locales"):
chars = get_locale_chars(code, unicode=True)
name = locale.getDisplayName().encode('ascii',
'ignore').decode('ascii')
char_codes = np.sort(list(
chain(*chars.values()))) # Unicode code of each char
row_by_code = dict(zip(char_codes, range(len(
char_codes)))) # Where each code appears in the resulting matrix
# Convert the char dict to use matrix indices
