zaclys/searxng
1
0
Fork 1
mirror of https://github.com/searxng/searxng synced 2024-01-01 19:24:07 +01:00
Code Issues Projects Releases Packages Wiki Activity

[fix] bing engine: query string only support 2 letter language code

Browse source 
The 'language:xx' query string in the request() function does only support the 2
letter language codes from the "Display languages" list.

Examples of items from the "Display languages" not sopported in the query
string:

  pt-BR, pt-BR, mn-Cyrl-MN, chr-cher, zh-Hans, ha-latn, ca-es-valencia

Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
This commit is contained in:
Markus Heiser 2022-04-10 15:43:41 +02:00
parent 864aeebc2f
commit d97e1a479f
3 changed files with 64 additions and 83 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

124
searx/data/engines_languages.json
View file

@ -4,14 +4,13 @@
"am",
"ar",
"as",
"az-latn",
"az",
"be",
"bg",
"bn",
"bs-latn",
"bs",
"ca",
"ca-es-valencia",
"chr-cher",
"chr",
"cs",
"cy",
"da",
@ -29,7 +28,7 @@
"gd",
"gl",
"gu",
"ha-latn",
"ha",
"he",
"hi",
"hr",
@ -46,7 +45,7 @@
"kn",
"ko",
"kok",
"ku-arab",
"ku",
"ky",
"lb",
"lo",
@ -55,7 +54,7 @@
"mi",
"mk",
"ml",
"mn-Cyrl-MN",
"mn",
"mr",
"ms",
"mt",
@ -65,29 +64,26 @@
"nn",
"nso",
"or",
"pa-arab",
"pa-guru",
"pa",
"pl",
"prs",
"pt-BR",
"pt-PT",
"pt",
"quc",
"quz",
"ro",
"ru",
"rw",
"sd-arab",
"sd",
"si",
"sk",
"sl",
"sq",
"sr-cyrl",
"sr-latn",
"sr",
"sv",
"sw",
"ta",
"te",
"tg-cyrl",
"tg",
"th",
"ti",
"tk",
@ -97,13 +93,12 @@
"ug",
"uk",
"ur",
"uz-latn",
"uz",
"vi",
"wo",
"xh",
"yo",
"zh-Hans",
"zh-Hant",
"zh",
"zu"
],
"bing images": [
@ -111,14 +106,13 @@
"am",
"ar",
"as",
"az-latn",
"az",
"be",
"bg",
"bn",
"bs-latn",
"bs",
"ca",
"ca-es-valencia",
"chr-cher",
"chr",
"cs",
"cy",
"da",
@ -136,7 +130,7 @@
"gd",
"gl",
"gu",
"ha-latn",
"ha",
"he",
"hi",
"hr",
@ -153,7 +147,7 @@
"kn",
"ko",
"kok",
"ku-arab",
"ku",
"ky",
"lb",
"lo",
@ -162,7 +156,7 @@
"mi",
"mk",
"ml",
"mn-Cyrl-MN",
"mn",
"mr",
"ms",
"mt",
@ -172,29 +166,26 @@
"nn",
"nso",
"or",
"pa-arab",
"pa-guru",
"pa",
"pl",
"prs",
"pt-BR",
"pt-PT",
"pt",
"quc",
"quz",
"ro",
"ru",
"rw",
"sd-arab",
"sd",
"si",
"sk",
"sl",
"sq",
"sr-cyrl",
"sr-latn",
"sr",
"sv",
"sw",
"ta",
"te",
"tg-cyrl",
"tg",
"th",
"ti",
"tk",
@ -204,13 +195,12 @@
"ug",
"uk",
"ur",
"uz-latn",
"uz",
"vi",
"wo",
"xh",
"yo",
"zh-Hans",
"zh-Hant",
"zh",
"zu"
],
"bing news": [
@ -218,14 +208,13 @@
"am",
"ar",
"as",
"az-latn",
"az",
"be",
"bg",
"bn",
"bs-latn",
"bs",
"ca",
"ca-es-valencia",
"chr-cher",
"chr",
"cs",
"cy",
"da",
@ -243,7 +232,7 @@
"gd",
"gl",
"gu",
"ha-latn",
"ha",
"he",
"hi",
"hr",
@ -260,7 +249,7 @@
"kn",
"ko",
"kok",
"ku-arab",
"ku",
"ky",
"lb",
"lo",
@ -269,7 +258,7 @@
"mi",
"mk",
"ml",
"mn-Cyrl-MN",
"mn",
"mr",
"ms",
"mt",
@ -279,29 +268,26 @@
"nn",
"nso",
"or",
"pa-arab",
"pa-guru",
"pa",
"pl",
"prs",
"pt-BR",
"pt-PT",
"pt",
"quc",
"quz",
"ro",
"ru",
"rw",
"sd-arab",
"sd",
"si",
"sk",
"sl",
"sq",
"sr-cyrl",
"sr-latn",
"sr",
"sv",
"sw",
"ta",
"te",
"tg-cyrl",
"tg",
"th",
"ti",
"tk",
@ -311,13 +297,12 @@
"ug",
"uk",
"ur",
"uz-latn",
"uz",
"vi",
"wo",
"xh",
"yo",
"zh-Hans",
"zh-Hant",
"zh",
"zu"
],
"bing videos": [
@ -325,14 +310,13 @@
"am",
"ar",
"as",
"az-latn",
"az",
"be",
"bg",
"bn",
"bs-latn",
"bs",
"ca",
"ca-es-valencia",
"chr-cher",
"chr",
"cs",
"cy",
"da",
@ -350,7 +334,7 @@
"gd",
"gl",
"gu",
"ha-latn",
"ha",
"he",
"hi",
"hr",
@ -367,7 +351,7 @@
"kn",
"ko",
"kok",
"ku-arab",
"ku",
"ky",
"lb",
"lo",
@ -376,7 +360,7 @@
"mi",
"mk",
"ml",
"mn-Cyrl-MN",
"mn",
"mr",
"ms",
"mt",
@ -386,29 +370,26 @@
"nn",
"nso",
"or",
"pa-arab",
"pa-guru",
"pa",
"pl",
"prs",
"pt-BR",
"pt-PT",
"pt",
"quc",
"quz",
"ro",
"ru",
"rw",
"sd-arab",
"sd",
"si",
"sk",
"sl",
"sq",
"sr-cyrl",
"sr-latn",
"sr",
"sv",
"sw",
"ta",
"te",
"tg-cyrl",
"tg",
"th",
"ti",
"tk",
@ -418,13 +399,12 @@
"ug",
"uk",
"ur",
"uz-latn",
"uz",
"vi",
"wo",
"xh",
"yo",
"zh-Hans",
"zh-Hant",
"zh",
"zu"
],
"dailymotion": [

21
searx/engines/bing.py
View file

@ -131,21 +131,24 @@ def _fetch_supported_languages(resp):
lang_tags = set()
dom = html.fromstring(resp.text)
lang_links = eval_xpath(dom, '//div[@id="language-section"]//li')
for _li in lang_links:
# Selector to get items from "Display language"
ui_lang_links = eval_xpath(dom, '//div[@id="language-section"]//li')
for _li in ui_lang_links:
href = eval_xpath(_li, './/@href')[0]
(_scheme, _netloc, _path, _params, query, _fragment) = urlparse(href)
query = parse_qs(query, keep_blank_values=True)
# fmt: off
setlang = query.get('setlang', [None, ])[0]
# example: 'mn-Cyrl-MN' --> '['mn', 'Cyrl-MN']
lang, nation = (setlang.split('-', maxsplit=1) + [None,])[:2] # fmt: skip
# fmt: on
# The 'language:xx' query string in the request function (above) does
# only support the 2 letter language codes from the "Display languages"
# list. Examples of items from the "Display languages" not sopported in
# the query string:
# 'mn-Cyrl-MN', 'chr-cher', 'zh-Hans', ha-latn, 'ca-es-valencia'
tag = lang + '-' + nation if nation else lang
lang_tags.add(tag)
setlang = query.get('setlang', [None, ])[0]
lang = setlang.split('-')[0]
lang_tags.add(lang)
return list(lang_tags)

2
searx/languages.py
View file

@ -49,8 +49,6 @@ language_codes = (
('nl-BE', 'Nederlands', 'België', 'Dutch', '\U0001f1e7\U0001f1ea'),
('nl-NL', 'Nederlands', 'Nederland', 'Dutch', '\U0001f1f3\U0001f1f1'),
('pl-PL', 'Polski', 'Polska', 'Polish', '\U0001f1f5\U0001f1f1'),
('pt', 'Português', '', 'Portuguese', '\U0001f310'),
('pt-BR', 'Português', 'Brasil', 'Portuguese', '\U0001f1e7\U0001f1f7'),
('pt-PT', 'Português', 'Portugal', 'Portuguese', '\U0001f1f5\U0001f1f9'),
('ro-RO', 'Română', 'România', 'Romanian', '\U0001f1f7\U0001f1f4'),
('ru-RU', 'Русский', 'Россия', 'Russian', '\U0001f1f7\U0001f1fa'),