Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.11/site-packages/babel/core.py: 37%

1152def negotiate_locale(preferred: Iterable[str], available: Iterable[str], sep: str = '_', aliases: Mapping[str, str] = LOCALE_ALIASES) -> str | None:

1153 """Find the best match between available and requested locale strings.

1154

1155 >>> negotiate_locale(['de_DE', 'en_US'], ['de_DE', 'de_AT'])

1156 'de_DE'

1157 >>> negotiate_locale(['de_DE', 'en_US'], ['en', 'de'])

1158 'de'

1159

1160 Case is ignored by the algorithm, the result uses the case of the preferred

1161 locale identifier:

1162

1163 >>> negotiate_locale(['de_DE', 'en_US'], ['de_de', 'de_at'])

1164 'de_DE'

1165

1166 >>> negotiate_locale(['de_DE', 'en_US'], ['de_de', 'de_at'])

1167 'de_DE'

1168

1169 By default, some web browsers unfortunately do not include the territory

1170 in the locale identifier for many locales, and some don't even allow the

1171 user to easily add the territory. So while you may prefer using qualified

1172 locale identifiers in your web-application, they would not normally match

1173 the language-only locale sent by such browsers. To workaround that, this

1174 function uses a default mapping of commonly used language-only locale

1175 identifiers to identifiers including the territory:

1176

1177 >>> negotiate_locale(['ja', 'en_US'], ['ja_JP', 'en_US'])

1178 'ja_JP'

1179

1180 Some browsers even use an incorrect or outdated language code, such as "no"

1181 for Norwegian, where the correct locale identifier would actually be "nb_NO"

1182 (Bokmål) or "nn_NO" (Nynorsk). The aliases are intended to take care of

1183 such cases, too:

1184

1185 >>> negotiate_locale(['no', 'sv'], ['nb_NO', 'sv_SE'])

1186 'nb_NO'

1187

1188 You can override this default mapping by passing a different `aliases`

1189 dictionary to this function, or you can bypass the behavior althogher by

1190 setting the `aliases` parameter to `None`.

1191

1192 :param preferred: the list of locale strings preferred by the user

1193 :param available: the list of locale strings available

1194 :param sep: character that separates the different parts of the locale

1195 strings

1196 :param aliases: a dictionary of aliases for locale identifiers

1197 """

1198 available = [a.lower() for a in available if a]

1199 for locale in preferred:

1200 ll = locale.lower()

1201 if ll in available:

1202 return locale

1203 if aliases:

1204 alias = aliases.get(ll)

1205 if alias:

1206 alias = alias.replace('_', sep)

1207 if alias.lower() in available:

1208 return alias

1209 parts = locale.split(sep)

1210 if len(parts) > 1 and parts[0].lower() in available:

1211 return parts[0]

1212 return None

1213

1214

1215def parse_locale(

1216 identifier: str,

1217 sep: str = '_',

1219 """Parse a locale identifier into a tuple of the form ``(language,

1220 territory, script, variant, modifier)``.

1221

1222 >>> parse_locale('zh_CN')

1223 ('zh', 'CN', None, None)

1224 >>> parse_locale('zh_Hans_CN')

1225 ('zh', 'CN', 'Hans', None)

1226 >>> parse_locale('ca_es_valencia')

1227 ('ca', 'ES', None, 'VALENCIA')

1228 >>> parse_locale('en_150')

1229 ('en', '150', None, None)

1230 >>> parse_locale('en_us_posix')

1231 ('en', 'US', None, 'POSIX')

1232 >>> parse_locale('it_IT@euro')

1233 ('it', 'IT', None, None, 'euro')

1234 >>> parse_locale('it_IT@custom')

1235 ('it', 'IT', None, None, 'custom')

1236 >>> parse_locale('it_IT@')

1237 ('it', 'IT', None, None)

1238

1239 The default component separator is "_", but a different separator can be

1240 specified using the `sep` parameter.

1241

1242 The optional modifier is always separated with "@" and at the end:

1243

1244 >>> parse_locale('zh-CN', sep='-')

1245 ('zh', 'CN', None, None)

1246 >>> parse_locale('zh-CN@custom', sep='-')

1247 ('zh', 'CN', None, None, 'custom')

1248

1249 If the identifier cannot be parsed into a locale, a `ValueError` exception

1250 is raised:

1251

1252 >>> parse_locale('not_a_LOCALE_String')

1253 Traceback (most recent call last):

1254 ...

1255 ValueError: 'not_a_LOCALE_String' is not a valid locale identifier

1256

1257 Encoding information is removed from the identifier, while modifiers are

1258 kept:

1259

1260 >>> parse_locale('en_US.UTF-8')

1261 ('en', 'US', None, None)

1262 >>> parse_locale('de_DE.iso885915@euro')

1263 ('de', 'DE', None, None, 'euro')

1264

1265 See :rfc:`4646` for more information.

1266

1267 :param identifier: the locale identifier string

1268 :param sep: character that separates the different components of the locale

1269 identifier

1270 :raise `ValueError`: if the string does not appear to be a valid locale

1271 identifier

1272 """

1273 if not identifier:

1274 raise ValueError("empty locale identifier")

1275 identifier, _, modifier = identifier.partition('@')

1276 if '.' in identifier:

1277 # this is probably the charset/encoding, which we don't care about

1278 identifier = identifier.split('.', 1)[0]

1279

1280 parts = identifier.split(sep)

1281 lang = parts.pop(0).lower()

1282 if not lang.isalpha():

1283 raise ValueError(f"expected only letters, got {lang!r}")

1284

1285 script = territory = variant = None

1286 if parts and len(parts[0]) == 4 and parts[0].isalpha():

1287 script = parts.pop(0).title()

1288

1289 if parts:

1290 if len(parts[0]) == 2 and parts[0].isalpha():

1291 territory = parts.pop(0).upper()

1292 elif len(parts[0]) == 3 and parts[0].isdigit():

1293 territory = parts.pop(0)

1294

1295 if parts and (

1296 len(parts[0]) == 4 and parts[0][0].isdigit() or

1297 len(parts[0]) >= 5 and parts[0][0].isalpha()

1298 ):

1299 variant = parts.pop().upper()

1300

1301 if parts:

1302 raise ValueError(f"{identifier!r} is not a valid locale identifier")

1303

1304 # TODO(3.0): always return a 5-tuple

1305 if modifier:

1306 return lang, territory, script, variant, modifier

1307 else:

1308 return lang, territory, script, variant

1309

1310

1311def get_locale_identifier(

1312 tup: tuple[str]

1313 | tuple[str, str | None]

1314 | tuple[str, str | None, str | None]

1315 | tuple[str, str | None, str | None, str | None]

1317 sep: str = "_",

1318) -> str:

1319 """The reverse of :func:`parse_locale`. It creates a locale identifier out

1320 of a ``(language, territory, script, variant, modifier)`` tuple. Items can be set to

1321 ``None`` and trailing ``None``\\s can also be left out of the tuple.

1322

1323 >>> get_locale_identifier(('de', 'DE', None, '1999', 'custom'))

1324 'de_DE_1999@custom'

1325 >>> get_locale_identifier(('fi', None, None, None, 'custom'))

1326 'fi@custom'

1327

1328

1329 .. versionadded:: 1.0

1330

1331 :param tup: the tuple as returned by :func:`parse_locale`.

1332 :param sep: the separator for the identifier.

1333 """

1334 tup = tuple(tup[:5]) # type: ignore # length should be no more than 5

1335 lang, territory, script, variant, modifier = tup + (None,) * (5 - len(tup))

1336 ret = sep.join(filter(None, (lang, script, territory, variant)))

1337 return f'{ret}@{modifier}' if modifier else ret