Merge pull request #1252 from MarcAbonce/search-languages

[mod] Refactor engine's search language handling

searx/engines/__init__.py

@@ -20,13 +20,14 @@ import sys
 import threading
 from os.path import realpath, dirname
 from io import open
+from babel.localedata import locale_identifiers
 from flask_babel import gettext
 from operator import itemgetter
 from json import loads
 from requests import get
 from searx import settings
 from searx import logger
-from searx.utils import load_module
+from searx.utils import load_module, match_language
 
 
 logger = logger.getChild('engines')
@@ -38,6 +39,8 @@ engines = {}
 categories = {'general': []}
 
 languages = loads(open(engine_dir + '/../data/engines_languages.json', 'r', encoding='utf-8').read())
+babel_langs = [lang_parts[0] + '-' + lang_parts[-1] if len(lang_parts) > 1 else lang_parts[0]
+               for lang_parts in (lang_code.split('_') for lang_code in locale_identifiers())]
 
 engine_shortcuts = {}
 engine_default_args = {'paging': False,
@@ -97,6 +100,22 @@ def load_engine(engine_data):
     if engine_data['name'] in languages:
         setattr(engine, 'supported_languages', languages[engine_data['name']])
 
+    # find custom aliases for non standard language codes
+    if hasattr(engine, 'supported_languages'):
+        if hasattr(engine, 'language_aliases'):
+            language_aliases = getattr(engine, 'language_aliases')
+        else:
+            language_aliases = {}
+
+        for engine_lang in getattr(engine, 'supported_languages'):
+            iso_lang = match_language(engine_lang, babel_langs, fallback=None)
+            if iso_lang and iso_lang != engine_lang and not engine_lang.startswith(iso_lang) and \
+               iso_lang not in getattr(engine, 'supported_languages'):
+                language_aliases[iso_lang] = engine_lang
+
+        if language_aliases:
+            setattr(engine, 'language_aliases', language_aliases)
+
     # assign language fetching method if auxiliary method exists
     if hasattr(engine, '_fetch_supported_languages'):
         setattr(engine, 'fetch_supported_languages',

searx/engines/archlinux.py

@@ -99,13 +99,13 @@ supported_languages = dict(lang_urls, **main_langs)
 
 # do search-request
 def request(query, params):
-    # translate the locale (e.g. 'en_US') to language code ('en')
+    # translate the locale (e.g. 'en-US') to language code ('en')
     language = locale_to_lang_code(params['language'])
 
     # if our language is hosted on the main site, we need to add its name
     # to the query in order to narrow the results to that language
     if language in main_langs:
-        query += '(' + main_langs[language] + ')'
+        query += b' (' + main_langs[language] + b')'
 
     # prepare the request parameters
     query = urlencode({'search': query})

searx/engines/bing.py

@@ -16,12 +16,14 @@
 from lxml import html
 from searx.engines.xpath import extract_text
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general']
 paging = True
 language_support = True
 supported_languages_url = 'https://www.bing.com/account/general'
+language_aliases = {'zh-CN': 'zh-CHS', 'zh-TW': 'zh-CHT', 'zh-HK': 'zh-CHT'}
 
 # search-url
 base_url = 'https://www.bing.com/'
@@ -32,9 +34,9 @@ search_string = 'search?{query}&first={offset}'
 def request(query, params):
     offset = (params['pageno'] - 1) * 10 + 1
 
-    lang = params['language'].split('-')[0].upper()
+    lang = match_language(params['language'], supported_languages, language_aliases)
 
-    query = u'language:{} {}'.format(lang, query.decode('utf-8')).encode('utf-8')
+    query = u'language:{} {}'.format(lang.split('-')[0].upper(), query.decode('utf-8')).encode('utf-8')
 
     search_path = search_string.format(
         query=urlencode({'q': query}),

searx/engines/bing_images.py

@@ -19,6 +19,7 @@ from lxml import html
 from json import loads
 import re
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['images']
@@ -46,26 +47,6 @@ safesearch_types = {2: 'STRICT',
 _quote_keys_regex = re.compile('({|,)([a-z][a-z0-9]*):(")', re.I | re.U)
 
 
-# get supported region code
-def get_region_code(lang, lang_list=None):
-    region = None
-    if lang in (lang_list or supported_languages):
-        region = lang
-    elif lang.startswith('no'):
-        region = 'nb-NO'
-    else:
-        # try to get a supported country code with language
-        lang = lang.split('-')[0]
-        for lc in (lang_list or supported_languages):
-            if lang == lc.split('-')[0]:
-                region = lc
-                break
-    if region:
-        return region.lower()
-    else:
-        return 'en-us'
-
-
 # do search-request
 def request(query, params):
     offset = (params['pageno'] - 1) * 10 + 1
@@ -74,7 +55,7 @@ def request(query, params):
         query=urlencode({'q': query}),
         offset=offset)
 
-    language = get_region_code(params['language'])
+    language = match_language(params['language'], supported_languages).lower()
 
     params['cookies']['SRCHHPGUSR'] = \
         'ADLT=' + safesearch_types.get(params['safesearch'], 'DEMOTE')

searx/engines/bing_news.py

@@ -14,8 +14,8 @@
 from datetime import datetime
 from dateutil import parser
 from lxml import etree
-from searx.utils import list_get
-from searx.engines.bing import _fetch_supported_languages, supported_languages_url
+from searx.utils import list_get, match_language
+from searx.engines.bing import _fetch_supported_languages, supported_languages_url, language_aliases
 from searx.url_utils import urlencode, urlparse, parse_qsl
 
 # engine dependent config
@@ -71,7 +71,7 @@ def request(query, params):
 
     offset = (params['pageno'] - 1) * 10 + 1
 
-    language = params['language']
+    language = match_language(params['language'], supported_languages, language_aliases)
 
     params['url'] = _get_url(query, language, offset, params['time_range'])
 

searx/engines/bing_videos.py

@@ -12,9 +12,10 @@
 
 from json import loads
 from lxml import html
-from searx.engines.bing_images import _fetch_supported_languages, supported_languages_url, get_region_code
+from searx.engines.bing_images import _fetch_supported_languages, supported_languages_url
 from searx.engines.xpath import extract_text
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 
 categories = ['videos']
@@ -47,8 +48,8 @@ def request(query, params):
         'ADLT=' + safesearch_types.get(params['safesearch'], 'DEMOTE')
 
     # language cookie
-    region = get_region_code(params['language'], lang_list=supported_languages)
-    params['cookies']['_EDGE_S'] = 'mkt=' + region + '&F=1'
+    language = match_language(params['language'], supported_languages).lower()
+    params['cookies']['_EDGE_S'] = 'mkt=' + language + '&F=1'
 
     # query and paging
     params['url'] = search_url.format(query=urlencode({'q': query}),

searx/engines/dailymotion.py

@@ -15,6 +15,7 @@
 from json import loads
 from datetime import datetime
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['videos']
@@ -32,7 +33,7 @@ supported_languages_url = 'https://api.dailymotion.com/languages'
 
 # do search-request
 def request(query, params):
-    locale = params['language']
+    locale = match_language(params['language'], supported_languages)
 
     params['url'] = search_url.format(
         query=urlencode({'search': query, 'localization': locale}),

searx/engines/duckduckgo.py

@@ -18,14 +18,25 @@ from json import loads
 from searx.engines.xpath import extract_text
 from searx.poolrequests import get
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general']
 paging = True
 language_support = True
-supported_languages_url = 'https://duckduckgo.com/d2030.js'
+supported_languages_url = 'https://duckduckgo.com/util/u172.js'
 time_range_support = True
 
+language_aliases = {
+    'ar-SA': 'ar-XA',
+    'es-419': 'es-XL',
+    'ja': 'jp-JP',
+    'ko': 'kr-KR',
+    'sl-SI': 'sl-SL',
+    'zh-TW': 'tzh-TW',
+    'zh-HK': 'tzh-HK'
+}
+
 # search-url
 url = 'https://duckduckgo.com/html?{query}&s={offset}&dc={dc_param}'
 time_range_url = '&df={range}'
@@ -42,34 +53,12 @@ content_xpath = './/a[@class="result__snippet"]'
 
 
 # match query's language to a region code that duckduckgo will accept
-def get_region_code(lang, lang_list=None):
-    # custom fixes for languages
-    if lang[:2] == 'ja':
-        region_code = 'jp-jp'
-    elif lang[:2] == 'sl':
-        region_code = 'sl-sl'
-    elif lang == 'zh-TW':
-        region_code = 'tw-tzh'
-    elif lang == 'zh-HK':
-        region_code = 'hk-tzh'
-    elif lang[-2:] == 'SA':
-        region_code = 'xa-' + lang.split('-')[0]
-    elif lang[-2:] == 'GB':
-        region_code = 'uk-' + lang.split('-')[0]
-    else:
-        region_code = lang.split('-')
-        if len(region_code) == 2:
-            # country code goes first
-            region_code = region_code[1].lower() + '-' + region_code[0].lower()
-        else:
-            # tries to get a country code from language
-            region_code = region_code[0].lower()
-            for lc in (lang_list or supported_languages):
-                lc = lc.split('-')
-                if region_code == lc[0]:
-                    region_code = lc[1].lower() + '-' + lc[0].lower()
-                    break
-    return region_code
+def get_region_code(lang, lang_list=[]):
+    lang_code = match_language(lang, lang_list, language_aliases, 'wt-WT')
+    lang_parts = lang_code.split('-')
+
+    # country code goes first
+    return lang_parts[1].lower() + '-' + lang_parts[0].lower()
 
 
 # do search-request
@@ -79,7 +68,7 @@ def request(query, params):
 
     offset = (params['pageno'] - 1) * 30
 
-    region_code = get_region_code(params['language'])
+    region_code = get_region_code(params['language'], supported_languages)
     params['url'] = url.format(
         query=urlencode({'q': query, 'kl': region_code}), offset=offset, dc_param=offset)
 

searx/engines/duckduckgo_definitions.py

@@ -2,9 +2,9 @@ import json
 from lxml import html
 from re import compile
 from searx.engines.xpath import extract_text
-from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url
+from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url, language_aliases
 from searx.url_utils import urlencode
-from searx.utils import html_to_text
+from searx.utils import html_to_text, match_language
 
 url = 'https://api.duckduckgo.com/'\
     + '?{query}&format=json&pretty=0&no_redirect=1&d=1'
@@ -24,7 +24,8 @@ def result_to_text(url, text, htmlResult):
 
 def request(query, params):
     params['url'] = url.format(query=urlencode({'q': query}))
-    params['headers']['Accept-Language'] = params['language'].split('-')[0]
+    language = match_language(params['language'], supported_languages, language_aliases)
+    params['headers']['Accept-Language'] = language.split('-')[0]
     return params
 
 

searx/engines/duckduckgo_images.py

@@ -15,7 +15,10 @@
 
 from json import loads
 from searx.engines.xpath import extract_text
-from searx.engines.duckduckgo import _fetch_supported_languages, supported_languages_url, get_region_code
+from searx.engines.duckduckgo import (
+    _fetch_supported_languages, supported_languages_url,
+    get_region_code, language_aliases
+)
 from searx.poolrequests import get
 from searx.url_utils import urlencode
 

searx/engines/google.py

@@ -14,6 +14,7 @@ from lxml import html, etree
 from searx.engines.xpath import extract_text, extract_url
 from searx import logger
 from searx.url_utils import urlencode, urlparse, parse_qsl
+from searx.utils import match_language
 
 logger = logger.getChild('google engine')
 
@@ -72,7 +73,7 @@ country_to_hostname = {
     'RO': 'www.google.ro',  # Romania
     'RU': 'www.google.ru',  # Russia
     'SK': 'www.google.sk',  # Slovakia
-    'SL': 'www.google.si',  # Slovenia (SL -> si)
+    'SI': 'www.google.si',  # Slovenia
     'SE': 'www.google.se',  # Sweden
     'TH': 'www.google.co.th',  # Thailand
     'TR': 'www.google.com.tr',  # Turkey
@@ -165,22 +166,20 @@ def extract_text_from_dom(result, xpath):
 def request(query, params):
     offset = (params['pageno'] - 1) * 10
 
+    language = match_language(params['language'], supported_languages)
+    language_array = language.split('-')
+    if params['language'].find('-') > 0:
+        country = params['language'].split('-')[1]
+    elif len(language_array) == 2:
+        country = language_array[1]
+    else:
+        country = 'US'
+
     # temporary fix until a way of supporting en-US is found
-    if params['language'] == 'en-US':
-        params['language'] = 'en-GB'
+    if language == 'en-US':
+        country = 'GB'
 
-    if params['language'][:2] == 'jv':
-        language = 'jw'
-        country = 'ID'
-        url_lang = 'lang_jw'
-    else:
-        language_array = params['language'].lower().split('-')
-        if len(language_array) == 2:
-            country = language_array[1]
-        else:
-            country = 'US'
-        language = language_array[0] + ',' + language_array[0] + '-' + country
-        url_lang = 'lang_' + language_array[0]
+    url_lang = 'lang_' + language
 
     if use_locale_domain:
         google_hostname = country_to_hostname.get(country.upper(), default_hostname)
@@ -196,7 +195,7 @@ def request(query, params):
     if params['time_range'] in time_range_dict:
         params['url'] += time_range_search.format(range=time_range_dict[params['time_range']])
 
-    params['headers']['Accept-Language'] = language
+    params['headers']['Accept-Language'] = language + ',' + language + '-' + country
     params['headers']['Accept'] = 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8'
 
     params['google_hostname'] = google_hostname

searx/engines/google_news.py

@@ -13,6 +13,7 @@
 from lxml import html
 from searx.engines.google import _fetch_supported_languages, supported_languages_url
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # search-url
 categories = ['news']
@@ -50,8 +51,9 @@ def request(query, params):
     params['url'] = search_url.format(query=urlencode({'q': query}),
                                       search_options=urlencode(search_options))
 
-    language_array = params['language'].lower().split('-')
-    params['url'] += '&lr=lang_' + language_array[0]
+    language = match_language(params['language'], supported_languages).split('-')[0]
+    if language:
+        params['url'] += '&lr=lang_' + language
 
     return params
 

searx/engines/qwant.py

@@ -14,6 +14,7 @@ from datetime import datetime
 from json import loads
 from searx.utils import html_to_text
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = None
@@ -45,16 +46,8 @@ def request(query, params):
                                    offset=offset)
 
     # add language tag
-    if params['language'] == 'no' or params['language'].startswith('no-'):
-        params['language'] = params['language'].replace('no', 'nb', 1)
-    if params['language'].find('-') < 0:
-        # tries to get a country code from language
-        for lang in supported_languages:
-            lc = lang.split('-')
-            if params['language'] == lc[0]:
-                params['language'] = lang
-                break
-    params['url'] += '&locale=' + params['language'].replace('-', '_').lower()
+    language = match_language(params['language'], supported_languages)
+    params['url'] += '&locale=' + language.replace('-', '_').lower()
 
     return params
 

searx/engines/swisscows.py

@@ -14,6 +14,7 @@ from json import loads
 import re
 from lxml.html import fromstring
 from searx.url_utils import unquote, urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general', 'images']
@@ -35,11 +36,8 @@ regex_img_url_remove_start = re.compile(b'^https?://i\.swisscows\.ch/\?link=')
 
 # do search-request
 def request(query, params):
-    if params['language'].split('-')[0] == 'no':
-        region = 'nb-NO'
-    else:
-        region = params['language']
-        ui_language = params['language'].split('-')[0]
+    region = match_language(params['language'], supported_languages)
+    ui_language = region.split('-')[0]
 
     search_path = search_string.format(
         query=urlencode({'query': query, 'uiLanguage': ui_language, 'region': region}),

searx/engines/wikidata.py

@@ -16,6 +16,7 @@ from searx.poolrequests import get
 from searx.engines.xpath import extract_text
 from searx.engines.wikipedia import _fetch_supported_languages, supported_languages_url
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 from json import loads
 from lxml.html import fromstring
@@ -56,7 +57,7 @@ calendar_name_xpath = './/sup[contains(@class,"wb-calendar-name")]'
 
 
 def request(query, params):
-    language = params['language'].split('-')[0]
+    language = match_language(params['language'], supported_languages).split('-')[0]
 
     params['url'] = url_search.format(
         query=urlencode({'label': query, 'language': language}))
@@ -68,7 +69,7 @@ def response(resp):
     html = fromstring(resp.text)
     wikidata_ids = html.xpath(wikidata_ids_xpath)
 
-    language = resp.search_params['language'].split('-')[0]
+    language = match_language(resp.search_params['language'], supported_languages).split('-')[0]
 
     # TODO: make requests asynchronous to avoid timeout when result_count > 1
     for wikidata_id in wikidata_ids[:result_count]:

searx/engines/wikipedia.py

@@ -13,6 +13,7 @@
 from json import loads
 from lxml.html import fromstring
 from searx.url_utils import quote, urlencode
+from searx.utils import match_language
 
 # search-url
 base_url = u'https://{language}.wikipedia.org/'
@@ -30,13 +31,7 @@ supported_languages_url = 'https://meta.wikimedia.org/wiki/List_of_Wikipedias'
 
 # set language in base_url
 def url_lang(lang):
-    lang = lang.split('-')[0]
-    if lang not in supported_languages:
-        language = 'en'
-    else:
-        language = lang
-
-    return language
+    return match_language(lang, supported_languages).split('-')[0]
 
 
 # do search-request

searx/engines/yahoo.py

@@ -14,6 +14,7 @@
 from lxml import html
 from searx.engines.xpath import extract_text, extract_url
 from searx.url_utils import unquote, urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['general']
@@ -39,6 +40,8 @@ time_range_dict = {'day': ['1d', 'd'],
                    'week': ['1w', 'w'],
                    'month': ['1m', 'm']}
 
+language_aliases = {'zh-CN': 'zh-CHS', 'zh-TW': 'zh-CHT', 'zh-HK': 'zh-CHT'}
+
 
 # remove yahoo-specific tracking-url
 def parse_url(url_string):
@@ -70,23 +73,16 @@ def _get_url(query, offset, language, time_range):
                                         lang=language)
 
 
-def _get_language(params):
-    if params['language'][:2] == 'zh':
-        if params['language'] == 'zh' or params['language'] == 'zh-CH':
-            return 'szh'
-        else:
-            return 'tzh'
-    else:
-        return params['language'].split('-')[0]
-
-
 # do search-request
 def request(query, params):
     if params['time_range'] and params['time_range'] not in time_range_dict:
         return params
 
     offset = (params['pageno'] - 1) * 10 + 1
-    language = _get_language(params)
+    language = match_language(params['language'], supported_languages, language_aliases)
+    if language not in language_aliases.values():
+        language = language.split('-')[0]
+    language = language.replace('-', '_').lower()
 
     params['url'] = _get_url(query, offset, language, params['time_range'])
 
@@ -145,7 +141,11 @@ def _fetch_supported_languages(resp):
     dom = html.fromstring(resp.text)
     options = dom.xpath('//div[@id="yschlang"]/span/label/input')
     for option in options:
-        code = option.xpath('./@value')[0][5:].replace('_', '-')
+        code_parts = option.xpath('./@value')[0][5:].split('_')
+        if len(code_parts) == 2:
+            code = code_parts[0] + '-' + code_parts[1].upper()
+        else:
+            code = code_parts[0]
         supported_languages.append(code)
 
     return supported_languages

searx/engines/yahoo_news.py

@@ -13,9 +13,12 @@ import re
 from datetime import datetime, timedelta
 from lxml import html
 from searx.engines.xpath import extract_text, extract_url
-from searx.engines.yahoo import parse_url, _fetch_supported_languages, supported_languages_url
+from searx.engines.yahoo import (
+    parse_url, _fetch_supported_languages, supported_languages_url, language_aliases
+)
 from dateutil import parser
 from searx.url_utils import urlencode
+from searx.utils import match_language
 
 # engine dependent config
 categories = ['news']
@@ -38,7 +41,7 @@ suggestion_xpath = '//div[contains(@class,"VerALSOTRY")]//a'
 def request(query, params):
     offset = (params['pageno'] - 1) * 10 + 1
 
-    language = params['language'].split('-')[0]
+    language = match_language(params['language'], supported_languages, language_aliases).split('-')[0]
 
     params['url'] = search_url.format(offset=offset,
                                       query=urlencode({'p': query}),

searx/languages.py

@@ -5,11 +5,7 @@
 language_codes = (
     (u"ar-SA", u"العربية", u"", u"Arabic"),
     (u"bg-BG", u"Български", u"", u"Bulgarian"),
-    (u"ca", u"Català", u"", u"Catalan"),
-    (u"ca-AD", u"Català", u"Andorra", u"Catalan"),
-    (u"ca-CT", u"Català", u"", u"Catalan"),
-    (u"ca-ES", u"Català", u"Espanya", u"Catalan"),
-    (u"ca-FR", u"Català", u"França", u"Catalan"),
+    (u"ca-ES", u"Català", u"", u"Catalan"),
     (u"cs-CZ", u"Čeština", u"", u"Czech"),
     (u"da-DK", u"Dansk", u"", u"Danish"),
     (u"de", u"Deutsch", u"", u"German"),
@@ -21,55 +17,51 @@ language_codes = (
     (u"en-AU", u"English", u"Australia", u"English"),
     (u"en-CA", u"English", u"Canada", u"English"),
     (u"en-GB", u"English", u"United Kingdom", u"English"),
-    (u"en-ID", u"English", u"Indonesia", u"English"),
-    (u"en-IE", u"English", u"Ireland", u"English"),
     (u"en-IN", u"English", u"India", u"English"),
     (u"en-MY", u"English", u"Malaysia", u"English"),
-    (u"en-NZ", u"English", u"New Zealand", u"English"),
-    (u"en-PH", u"English", u"Philippines", u"English"),
-    (u"en-SG", u"English", u"Singapore", u"English"),
     (u"en-US", u"English", u"United States", u"English"),
-    (u"en-ZA", u"English", u"South Africa", u"English"),
     (u"es", u"Español", u"", u"Spanish"),
-    (u"es-AD", u"Español", u"Andorra", u"Spanish"),
     (u"es-AR", u"Español", u"Argentina", u"Spanish"),
-    (u"es-CL", u"Español", u"Chile", u"Spanish"),
-    (u"es-CO", u"Español", u"Colombia", u"Spanish"),
     (u"es-ES", u"Español", u"España", u"Spanish"),
     (u"es-MX", u"Español", u"México", u"Spanish"),
-    (u"es-PE", u"Español", u"Perú", u"Spanish"),
-    (u"es-US", u"Español", u"Estados Unidos", u"Spanish"),
     (u"et-EE", u"Eesti", u"", u"Estonian"),
+    (u"fa-IR", u"فارسی", u"", u"Persian"),
     (u"fi-FI", u"Suomi", u"", u"Finnish"),
     (u"fr", u"Français", u"", u"French"),
-    (u"fr-AD", u"Français", u"Andorre", u"French"),
     (u"fr-BE", u"Français", u"Belgique", u"French"),
     (u"fr-CA", u"Français", u"Canada", u"French"),
     (u"fr-CH", u"Français", u"Suisse", u"French"),
     (u"fr-FR", u"Français", u"France", u"French"),
     (u"he-IL", u"עברית", u"", u"Hebrew"),
+    (u"hr-HR", u"Hrvatski", u"", u"Croatian"),
     (u"hu-HU", u"Magyar", u"", u"Hungarian"),
-    (u"it", u"Italiano", u"", u"Italian"),
-    (u"it-CH", u"Italiano", u"Svizzera", u"Italian"),
-    (u"it-IT", u"Italiano", u"Italia", u"Italian"),
+    (u"id-ID", u"Indonesia", u"", u"Indonesian"),
+    (u"is-IS", u"Íslenska", u"", u"Icelandic"),
+    (u"it-IT", u"Italiano", u"", u"Italian"),
     (u"ja-JP", u"日本語", u"", u"Japanese"),
     (u"ko-KR", u"한국어", u"", u"Korean"),
+    (u"lt-LT", u"Lietuvių", u"", u"Lithuanian"),
+    (u"lv-LV", u"Latviešu", u"", u"Latvian"),
+    (u"ms-MY", u"Bahasa Melayu", u"", u"Malay"),
+    (u"nb-NO", u"Norsk Bokmål", u"", u"Norwegian Bokmål"),
     (u"nl", u"Nederlands", u"", u"Dutch"),
     (u"nl-BE", u"Nederlands", u"België", u"Dutch"),
     (u"nl-NL", u"Nederlands", u"Nederland", u"Dutch"),
-    (u"no-NO", u"Norsk", u"", u"Norwegian"),
     (u"pl-PL", u"Polski", u"", u"Polish"),
     (u"pt", u"Português", u"", u"Portuguese"),
-    (u"pt-AD", u"Português", u"Andorra", u"Portuguese"),
     (u"pt-BR", u"Português", u"Brasil", u"Portuguese"),
     (u"pt-PT", u"Português", u"Portugal", u"Portuguese"),
     (u"ro-RO", u"Română", u"", u"Romanian"),
     (u"ru-RU", u"Русский", u"", u"Russian"),
+    (u"sk-SK", u"Slovenčina", u"", u"Slovak"),
+    (u"sl-SI", u"Slovenščina", u"", u"Slovenian"),
+    (u"sr-RS", u"Српски", u"", u"Serbian"),
     (u"sv-SE", u"Svenska", u"", u"Swedish"),
     (u"th-TH", u"ไทย", u"", u"Thai"),
     (u"tr-TR", u"Türkçe", u"", u"Turkish"),
+    (u"uk-UA", u"Українська", u"", u"Ukrainian"),
+    (u"vi-VN", u"Tiếng Việt", u"", u"Vietnamese"),
     (u"zh", u"中文", u"", u"Chinese"),
     (u"zh-CN", u"中文", u"中国", u"Chinese"),
-    (u"zh-HK", u"中文", u"香港", u"Chinese"),
-    (u"zh-TW", u"中文", u"台湾", u"Chinese")
+    (u"zh-TW", u"中文", u"台灣", u"Chinese")
 )

searx/preferences.py

@@ -115,10 +115,6 @@ class SearchLanguageSetting(EnumStringSetting):
                 pass
             elif lang in self.choices:
                 data = lang
-            elif data == 'nb-NO':
-                data = 'no-NO'
-            elif data == 'ar-XA':
-                data = 'ar-SA'
             else:
                 data = self.value
         self.value = data

searx/query.py

@@ -96,9 +96,13 @@ class RawTextQuery(object):
                                 break
 
                 # user may set a valid, yet not selectable language
-                if not self.languages and VALID_LANGUAGE_CODE.match(lang):
-                    self.languages.append(lang)
-                    parse_next = True
+                if VALID_LANGUAGE_CODE.match(lang):
+                    lang_parts = lang.split('-')
+                    if len(lang_parts) > 1:
+                        lang = lang_parts[0].lower() + '-' + lang_parts[1].upper()
+                    if lang not in self.languages:
+                        self.languages.append(lang)
+                        parse_next = True
 
             # this force a engine or category
             if query_part[0] == '!' or query_part[0] == '?':

searx/templates/oscar/preferences.html

@@ -187,7 +187,7 @@
                                     </td>
                                     <th>{{ search_engine.name }}</th>
 				    <td class="name">{{ shortcuts[search_engine.name] }}</td>
-					<td>{{ support_toggle(current_language in search_engine.supported_languages or current_language.split('-')[0] in search_engine.supported_languages) }}</td>
+					<td>{{ support_toggle(stats[search_engine.name].supports_selected_language) }}</td>
 					<td>{{ support_toggle(search_engine.safesearch==True) }}</td>
 					<td>{{ support_toggle(search_engine.time_range_support==True) }}</td>
 					<td class="{{ 'danger' if stats[search_engine.name]['warn_time'] else '' }}">{{ 'N/A' if stats[search_engine.name].time==None else stats[search_engine.name].time }}</td>
@@ -197,7 +197,7 @@
 					<td class="{{ 'danger' if stats[search_engine.name]['warn_time'] else '' }}">{{ 'N/A' if stats[search_engine.name].time==None else stats[search_engine.name].time }}</td>
 					<td>{{ support_toggle(search_engine.time_range_support==True) }}</td>
 					<td>{{ support_toggle(search_engine.safesearch==True) }}</td>
-					<td>{{ support_toggle(current_language in search_engine.supported_languages or current_language.split('-')[0] in search_engine.supported_languages) }}</td>
+					<td>{{ support_toggle(stats[search_engine.name].supports_selected_language) }}</td>
 					<td>{{ shortcuts[search_engine.name] }}</td>
                                     <th>{{ search_engine.name }}</th>
                                     <td class="onoff-checkbox">

searx/utils.py

@@ -4,6 +4,7 @@ import hmac
 import os
 import re
 
+from babel.core import get_global
 from babel.dates import format_date
 from codecs import getincrementalencoder
 from imp import load_source
@@ -12,6 +13,7 @@ from os.path import splitext, join
 from random import choice
 import sys
 
+from searx import settings
 from searx.version import VERSION_STRING
 from searx.languages import language_codes
 from searx import settings
@@ -322,6 +324,65 @@ def is_valid_lang(lang):
         return False
 
 
+# auxiliary function to match lang_code in lang_list
+def _match_language(lang_code, lang_list=[], custom_aliases={}):
+    # replace language code with a custom alias if necessary
+    if lang_code in custom_aliases:
+        lang_code = custom_aliases[lang_code]
+
+    if lang_code in lang_list:
+        return lang_code
+
+    # try to get the most likely country for this language
+    subtags = get_global('likely_subtags').get(lang_code)
+    if subtags:
+        subtag_parts = subtags.split('_')
+        new_code = subtag_parts[0] + '-' + subtag_parts[-1]
+        if new_code in custom_aliases:
+            new_code = custom_aliases[new_code]
+        if new_code in lang_list:
+            return new_code
+
+    # try to get the any supported country for this language
+    for lc in lang_list:
+        if lang_code == lc.split('-')[0]:
+            return lc
+
+    return None
+
+
+# get the language code from lang_list that best matches locale_code
+def match_language(locale_code, lang_list=[], custom_aliases={}, fallback='en-US'):
+    # try to get language from given locale_code
+    language = _match_language(locale_code, lang_list, custom_aliases)
+    if language:
+        return language
+
+    locale_parts = locale_code.split('-')
+    lang_code = locale_parts[0]
+
+    # try to get language using an equivalent country code
+    if len(locale_parts) > 1:
+        country_alias = get_global('territory_aliases').get(locale_parts[-1])
+        if country_alias:
+            language = _match_language(lang_code + '-' + country_alias[0], lang_list, custom_aliases)
+            if language:
+                return language
+
+    # try to get language using an equivalent language code
+    alias = get_global('language_aliases').get(lang_code)
+    if alias:
+        language = _match_language(alias, lang_list, custom_aliases)
+        if language:
+            return language
+
+    if lang_code != locale_code:
+        # try to get language from given language without giving the country
+        language = _match_language(lang_code, lang_list, custom_aliases)
+
+    return language or fallback
+
+
 def load_module(filename, module_dir):
     modname = splitext(filename)[0]
     if modname in sys.modules:

searx/webapp.py

@@ -58,16 +58,16 @@ from searx.engines import (
 from searx.utils import (
     UnicodeWriter, highlight_content, html_to_text, get_resources_directory,
     get_static_files, get_result_templates, get_themes, gen_useragent,
-    dict_subset, prettify_url
+    dict_subset, prettify_url, match_language
 )
 from searx.version import VERSION_STRING
-from searx.languages import language_codes
+from searx.languages import language_codes as languages
 from searx.search import SearchWithPlugins, get_search_query_from_webapp
 from searx.query import RawTextQuery
 from searx.autocomplete import searx_bang, backends as autocomplete_backends
 from searx.plugins import plugins
 from searx.plugins.oa_doi_rewrite import get_doi_resolver
-from searx.preferences import Preferences, ValidationException
+from searx.preferences import Preferences, ValidationException, LANGUAGE_CODES
 from searx.answerers import answerers
 from searx.url_utils import urlencode, urlparse, urljoin
 from searx.utils import new_hmac
@@ -133,7 +133,7 @@ if not searx_debug \
 babel = Babel(app)
 
 rtl_locales = ['ar', 'arc', 'bcc', 'bqi', 'ckb', 'dv', 'fa', 'glk', 'he',
-               'ku', 'mzn', 'pnb'', ''ps', 'sd', 'ug', 'ur', 'yi']
+               'ku', 'mzn', 'pnb', 'ps', 'sd', 'ug', 'ur', 'yi']
 
 # used when translating category names
 _category_names = (gettext('files'),
@@ -352,9 +352,11 @@ def render(template_name, override_theme=None, **kwargs):
 
     kwargs['safesearch'] = str(request.preferences.get_value('safesearch'))
 
-    kwargs['language_codes'] = language_codes
+    kwargs['language_codes'] = languages
     if 'current_language' not in kwargs:
-        kwargs['current_language'] = request.preferences.get_value('language')
+        kwargs['current_language'] = match_language(request.preferences.get_value('language'),
+                                                    LANGUAGE_CODES,
+                                                    fallback=settings['search']['language'])
 
     # override url_for function in templates
     kwargs['url_for'] = url_for_theme
@@ -590,7 +592,9 @@ def index():
         infoboxes=result_container.infoboxes,
         paging=result_container.paging,
         unresponsive_engines=result_container.unresponsive_engines,
-        current_language=search_query.lang,
+        current_language=match_language(search_query.lang,
+                                        LANGUAGE_CODES,
+                                        fallback=settings['search']['language']),
         base_url=get_base_url(),
         theme=get_current_theme_name(),
         favicons=global_favicons[themes.index(get_current_theme_name())]
@@ -687,6 +691,10 @@ def preferences():
                              'warn_time': False}
             if e.timeout > settings['outgoing']['request_timeout']:
                 stats[e.name]['warn_timeout'] = True
+            if match_language(request.preferences.get_value('language'),
+                              getattr(e, 'supported_languages', []),
+                              getattr(e, 'language_aliases', {}), None):
+                stats[e.name]['supports_selected_language'] = True
 
     # get first element [0], the engine time,
     # and then the second element [1] : the time (the first one is the label)

tests/unit/engines/test_archlinux.py

@@ -19,12 +19,17 @@ class TestArchLinuxEngine(SearxTestCase):
         query = 'test_query'
         dic = defaultdict(dict)
         dic['pageno'] = 1
-        dic['language'] = 'en_US'
+        dic['language'] = 'en-US'
         params = archlinux.request(query, dic)
         self.assertTrue('url' in params)
         self.assertTrue(query in params['url'])
         self.assertTrue('wiki.archlinux.org' in params['url'])
 
+        for lang, name in archlinux.main_langs:
+            dic['language'] = lang
+            params = archlinux.request(query, dic)
+            self.assertTrue(name in params['url'])
+
         for lang, domain in domains.items():
             dic['language'] = lang
             params = archlinux.request(query, dic)

tests/unit/engines/test_bing.py

@@ -7,6 +7,7 @@ from searx.testing import SearxTestCase
 class TestBingEngine(SearxTestCase):
 
     def test_request(self):
+        bing.supported_languages = ['en', 'fr', 'zh-CHS', 'zh-CHT', 'pt-PT', 'pt-BR']
         query = u'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 0

tests/unit/engines/test_bing_images.py

@@ -9,7 +9,6 @@ class TestBingImagesEngine(SearxTestCase):
 
     def test_request(self):
         bing_images.supported_languages = ['fr-FR', 'en-US']
-
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1

tests/unit/engines/test_bing_news.py

@@ -8,10 +8,11 @@ import lxml
 class TestBingNewsEngine(SearxTestCase):
 
     def test_request(self):
+        bing_news.supported_languages = ['en', 'fr']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1
-        dicto['language'] = 'fr_FR'
+        dicto['language'] = 'fr-FR'
         dicto['time_range'] = ''
         params = bing_news.request(query, dicto)
         self.assertIn('url', params)

tests/unit/engines/test_bing_videos.py

@@ -9,7 +9,6 @@ class TestBingVideosEngine(SearxTestCase):
 
     def test_request(self):
         bing_videos.supported_languages = ['fr-FR', 'en-US']
-
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1

tests/unit/engines/test_dailymotion.py

@@ -8,10 +8,11 @@ from searx.testing import SearxTestCase
 class TestDailymotionEngine(SearxTestCase):
 
     def test_request(self):
+        dailymotion.supported_languages = ['en', 'fr']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 0
-        dicto['language'] = 'fr_FR'
+        dicto['language'] = 'fr-FR'
         params = dailymotion.request(query, dicto)
         self.assertTrue('url' in params)
         self.assertTrue(query in params['url'])

tests/unit/engines/test_duckduckgo.py

@@ -1,18 +1,21 @@
 # -*- coding: utf-8 -*-
 from collections import defaultdict
 import mock
-from searx.engines import duckduckgo
+from searx.engines import load_engine, duckduckgo
 from searx.testing import SearxTestCase
 
 
 class TestDuckduckgoEngine(SearxTestCase):
 
     def test_request(self):
+        duckduckgo = load_engine({'engine': 'duckduckgo', 'name': 'duckduckgo'})
+
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1
-        dicto['language'] = 'de-CH'
         dicto['time_range'] = ''
+
+        dicto['language'] = 'de-CH'
         params = duckduckgo.request(query, dicto)
         self.assertIn('url', params)
         self.assertIn(query, params['url'])
@@ -20,16 +23,19 @@ class TestDuckduckgoEngine(SearxTestCase):
         self.assertIn('ch-de', params['url'])
         self.assertIn('s=0', params['url'])
 
-        # when ddg uses non standard code
+        # when ddg uses non standard codes
+        dicto['language'] = 'zh-HK'
+        params = duckduckgo.request(query, dicto)
+        self.assertIn('hk-tzh', params['url'])
+
         dicto['language'] = 'en-GB'
         params = duckduckgo.request(query, dicto)
         self.assertIn('uk-en', params['url'])
 
         # no country given
-        duckduckgo.supported_languages = ['de-CH', 'en-US']
-        dicto['language'] = 'de'
+        dicto['language'] = 'en'
         params = duckduckgo.request(query, dicto)
-        self.assertIn('ch-de', params['url'])
+        self.assertIn('us-en', params['url'])
 
     def test_no_url_in_request_year_time_range(self):
         dicto = defaultdict(dict)

tests/unit/engines/test_duckduckgo_definitions.py

@@ -18,6 +18,7 @@ class TestDDGDefinitionsEngine(SearxTestCase):
         self.assertEqual(result, 'Text in link')
 
     def test_request(self):
+        duckduckgo_definitions.supported_languages = ['en-US', 'es-ES']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1

tests/unit/engines/test_duckduckgo_images.py

@@ -9,7 +9,6 @@ class TestDuckduckgoImagesEngine(SearxTestCase):
 
     def test_request(self):
         duckduckgo_images.supported_languages = ['de-CH', 'en-US']
-
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['is_test'] = True

tests/unit/engines/test_google.py

@@ -15,6 +15,8 @@ class TestGoogleEngine(SearxTestCase):
         return response
 
     def test_request(self):
+        google.supported_languages = ['en', 'fr', 'zh-CN']
+
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1
@@ -31,6 +33,11 @@ class TestGoogleEngine(SearxTestCase):
         self.assertIn('google.co', params['url'])
         self.assertIn('en', params['headers']['Accept-Language'])
 
+        dicto['language'] = 'zh'
+        params = google.request(query, dicto)
+        self.assertIn('google.com', params['url'])
+        self.assertIn('zh-CN', params['headers']['Accept-Language'])
+
     def test_response(self):
         self.assertRaises(AttributeError, google.response, None)
         self.assertRaises(AttributeError, google.response, [])

tests/unit/engines/test_google_news.py

@@ -9,6 +9,7 @@ from searx.testing import SearxTestCase
 class TestGoogleNewsEngine(SearxTestCase):
 
     def test_request(self):
+        google_news.supported_languages = ['en-US', 'fr-FR']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1

tests/unit/engines/test_qwant.py

@@ -7,6 +7,7 @@ from searx.testing import SearxTestCase
 class TestQwantEngine(SearxTestCase):
 
     def test_request(self):
+        qwant.supported_languages = ['en-US', 'fr-CA', 'fr-FR']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 0
@@ -26,7 +27,6 @@ class TestQwantEngine(SearxTestCase):
         self.assertIn('en_us', params['url'])
         self.assertIn('news', params['url'])
 
-        qwant.supported_languages = ['en', 'fr-FR', 'fr-CA']
         dicto['language'] = 'fr'
         params = qwant.request(query, dicto)
         self.assertIn('fr_fr', params['url'])

tests/unit/engines/test_swisscows.py

@@ -7,6 +7,7 @@ from searx.testing import SearxTestCase
 class TestSwisscowsEngine(SearxTestCase):
 
     def test_request(self):
+        swisscows.supported_languages = ['de-AT', 'de-DE']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1

tests/unit/engines/test_wikidata.py

@@ -9,6 +9,7 @@ from searx.testing import SearxTestCase
 class TestWikidataEngine(SearxTestCase):
 
     def test_request(self):
+        wikidata.supported_languages = ['en', 'es']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['language'] = 'en-US'

tests/unit/engines/test_yahoo.py

@@ -25,11 +25,12 @@ class TestYahooEngine(SearxTestCase):
         self.assertEqual('https://this.is.the.url/', url)
 
     def test_request(self):
+        yahoo.supported_languages = ['en', 'fr', 'zh-CHT', 'zh-CHS']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1
         dicto['time_range'] = ''
-        dicto['language'] = 'fr_FR'
+        dicto['language'] = 'fr-FR'
         params = yahoo.request(query, dicto)
         self.assertIn('url', params)
         self.assertIn(query, params['url'])
@@ -39,6 +40,16 @@ class TestYahooEngine(SearxTestCase):
         self.assertIn('sB', params['cookies'])
         self.assertIn('fr', params['cookies']['sB'])
 
+        dicto['language'] = 'zh'
+        params = yahoo.request(query, dicto)
+        self.assertIn('zh_chs', params['url'])
+        self.assertIn('zh_chs', params['cookies']['sB'])
+
+        dicto['language'] = 'zh-TW'
+        params = yahoo.request(query, dicto)
+        self.assertIn('zh_cht', params['url'])
+        self.assertIn('zh_cht', params['cookies']['sB'])
+
     def test_no_url_in_request_year_time_range(self):
         dicto = defaultdict(dict)
         query = 'test_query'
@@ -168,5 +179,5 @@ class TestYahooEngine(SearxTestCase):
         self.assertEqual(type(languages), list)
         self.assertEqual(len(languages), 3)
         self.assertIn('ar', languages)
-        self.assertIn('zh-chs', languages)
-        self.assertIn('zh-cht', languages)
+        self.assertIn('zh-CHS', languages)
+        self.assertIn('zh-CHT', languages)

tests/unit/engines/test_yahoo_news.py

@@ -9,10 +9,11 @@ from searx.testing import SearxTestCase
 class TestYahooNewsEngine(SearxTestCase):
 
     def test_request(self):
+        yahoo_news.supported_languages = ['en', 'fr']
         query = 'test_query'
         dicto = defaultdict(dict)
         dicto['pageno'] = 1
-        dicto['language'] = 'fr_FR'
+        dicto['language'] = 'fr-FR'
         params = yahoo_news.request(query, dicto)
         self.assertIn('url', params)
         self.assertIn(query, params['url'])

tests/unit/test_utils.py

@@ -65,6 +65,31 @@ class TestUtils(SearxTestCase):
         for test_url, expected in data:
             self.assertEqual(utils.prettify_url(test_url, max_length=32), expected)
 
+    def test_match_language(self):
+        self.assertEqual(utils.match_language('es', ['es']), 'es')
+        self.assertEqual(utils.match_language('es', [], fallback='fallback'), 'fallback')
+        self.assertEqual(utils.match_language('ja', ['jp'], {'ja': 'jp'}), 'jp')
+
+        aliases = {'en-GB': 'en-UK', 'he': 'iw'}
+
+        # guess country
+        self.assertEqual(utils.match_language('de-DE', ['de']), 'de')
+        self.assertEqual(utils.match_language('de', ['de-DE']), 'de-DE')
+        self.assertEqual(utils.match_language('es-CO', ['es-AR', 'es-ES', 'es-MX']), 'es-ES')
+        self.assertEqual(utils.match_language('es-CO', ['es-MX']), 'es-MX')
+        self.assertEqual(utils.match_language('en-UK', ['en-AU', 'en-GB', 'en-US']), 'en-GB')
+        self.assertEqual(utils.match_language('en-GB', ['en-AU', 'en-UK', 'en-US'], aliases), 'en-UK')
+
+        # language aliases
+        self.assertEqual(utils.match_language('iw', ['he']), 'he')
+        self.assertEqual(utils.match_language('he', ['iw'], aliases), 'iw')
+        self.assertEqual(utils.match_language('iw-IL', ['he']), 'he')
+        self.assertEqual(utils.match_language('he-IL', ['iw'], aliases), 'iw')
+        self.assertEqual(utils.match_language('iw', ['he-IL']), 'he-IL')
+        self.assertEqual(utils.match_language('he', ['iw-IL'], aliases), 'iw-IL')
+        self.assertEqual(utils.match_language('iw-IL', ['he-IL']), 'he-IL')
+        self.assertEqual(utils.match_language('he-IL', ['iw-IL'], aliases), 'iw-IL')
+
 
 class TestHTMLTextExtractor(SearxTestCase):
 

utils/fetch_languages.py

@@ -2,83 +2,27 @@
 
 # This script generates languages.py from intersecting each engine's supported languages.
 #
-# The country names are obtained from http://api.geonames.org which requires registering as a user.
-#
 # Output files (engines_languages.json and languages.py)
 # are written in current directory to avoid overwriting in case something goes wrong.
 
-from requests import get
-from lxml.html import fromstring
-from json import loads, dump
+from json import dump
 import io
 from sys import path
+from babel import Locale, UnknownLocaleError
+from babel.languages import get_global
+
 path.append('../searx')  # noqa
 from searx import settings
-from searx.url_utils import urlencode
 from searx.engines import initialize_engines, engines
 
-# Geonames API for country names.
-geonames_user = ''  # ADD USER NAME HERE
-country_names_url = 'http://api.geonames.org/countryInfoJSON?{parameters}'
-
 # Output files.
 engines_languages_file = 'engines_languages.json'
 languages_file = 'languages.py'
 
-engines_languages = {}
-
-
-# To filter out invalid codes and dialects.
-def valid_code(lang_code):
-    # filter invalid codes
-    # sl-SL is technically not invalid, but still a mistake
-    invalid_codes = ['sl-SL', 'wt-WT', 'jw']
-    invalid_countries = ['UK', 'XA', 'XL']
-    if lang_code[:2] == 'xx'\
-       or lang_code in invalid_codes\
-       or lang_code[-2:] in invalid_countries\
-       or is_dialect(lang_code):
-        return False
-
-    return True
-
-
-# Language codes with any additional tags other than language and country.
-def is_dialect(lang_code):
-    lang_code = lang_code.split('-')
-    if len(lang_code) > 2 or len(lang_code[0]) > 3:
-        return True
-    if len(lang_code) == 2 and len(lang_code[1]) > 2:
-        return True
-
-    return False
-
-
-# Get country name in specified language.
-def get_country_name(locale):
-    if geonames_user is '':
-        return ''
-
-    locale = locale.split('-')
-    if len(locale) != 2:
-        return ''
-
-    url = country_names_url.format(parameters=urlencode({'lang': locale[0],
-                                                         'country': locale[1],
-                                                         'username': geonames_user}))
-    response = get(url)
-    json = loads(response.text)
-    content = json.get('geonames', None)
-    if content is None or len(content) != 1:
-        print("No country name found for " + locale[0] + "-" + locale[1])
-        return ''
-
-    return content[0].get('countryName', '')
-
 
 # Fetchs supported languages for each engine and writes json file with those.
 def fetch_supported_languages():
-    initialize_engines(settings['engines'])
+    engines_languages = {}
     for engine_name in engines:
         if hasattr(engines[engine_name], 'fetch_supported_languages'):
             try:
@@ -90,81 +34,135 @@ def fetch_supported_languages():
     with io.open(engines_languages_file, "w", encoding="utf-8") as f:
         dump(engines_languages, f, ensure_ascii=False)
 
+    return engines_languages
+
+
+# Get babel Locale object from lang_code if possible.
+def get_locale(lang_code):
+    try:
+        locale = Locale.parse(lang_code, sep='-')
+        return locale
+    except (UnknownLocaleError, ValueError):
+        return None
+
+
+# Append engine_name to list of engines that support locale.
+def add_engine_counter(lang_code, engine_name, languages):
+    if lang_code in languages:
+        if 'counter' not in languages[lang_code]:
+            languages[lang_code]['counter'] = [engine_name]
+        elif engine_name not in languages[lang_code]['counter']:
+            languages[lang_code]['counter'].append(engine_name)
 
-# Join all language lists.
-# Iterate all languages supported by each engine.
-def join_language_lists():
-    global languages
-    # include wikipedia first for more accurate language names
-    languages = {code: lang for code, lang
-                 in engines_languages['wikipedia'].items()
-                 if valid_code(code)}
 
+# Join all language lists.
+# TODO: Add language names from engine's language list if name not known by babel.
+def join_language_lists(engines_languages):
+    language_list = {}
     for engine_name in engines_languages:
-        for locale in engines_languages[engine_name]:
-            if valid_code(locale):
-                # if language is not on list or if it has no name yet
-                if locale not in languages or not languages[locale].get('name'):
-                    if isinstance(engines_languages[engine_name], dict):
-                        languages[locale] = engines_languages[engine_name][locale]
-                    else:
-                        languages[locale] = {}
-
-            # add to counter of engines that support given language
-            lang = locale.split('-')[0]
-            if lang in languages:
-                if 'counter' not in languages[lang]:
-                    languages[lang]['counter'] = [engine_name]
-                elif engine_name not in languages[lang]['counter']:
-                    languages[lang]['counter'].append(engine_name)
-
-    # filter list to include only languages supported by most engines
-    min_supported_engines = int(0.70 * len(engines_languages))
-    languages = {code: lang for code, lang
-                 in languages.items()
-                 if len(lang.get('counter', [])) >= min_supported_engines or
-                 len(languages.get(code.split('-')[0], {}).get('counter', [])) >= min_supported_engines}
-
-    # get locales that have no name or country yet
-    for locale in languages.keys():
-        # try to get language names
-        if not languages[locale].get('name'):
-            name = languages.get(locale.split('-')[0], {}).get('name', None)
-            if name:
-                languages[locale]['name'] = name
-            else:
-                # filter out locales with no name
-                del languages[locale]
-                continue
-
-        # try to get language name in english
-        if not languages[locale].get('english_name'):
-            languages[locale]['english_name'] = languages.get(locale.split('-')[0], {}).get('english_name', '')
-
-        # try to get country name
-        if locale.find('-') > 0 and not languages[locale].get('country'):
-            languages[locale]['country'] = get_country_name(locale) or ''
-
-
-# Remove countryless language if language is featured in only one country.
-def filter_single_country_languages():
-    prev_lang = None
-    prev_code = None
-    for code in sorted(languages):
-        lang = code.split('-')[0]
-        if lang == prev_lang:
+        for lang_code in engines_languages[engine_name]:
+
+            # apply custom fixes if necessary
+            if lang_code in getattr(engines[engine_name], 'language_aliases', {}).values():
+                lang_code = next(lc for lc, alias in engines[engine_name].language_aliases.items()
+                                 if lang_code == alias)
+
+            locale = get_locale(lang_code)
+
+            # ensure that lang_code uses standard language and country codes
+            if locale and locale.territory:
+                lang_code = locale.language + '-' + locale.territory
+
+            # add locale if it's not in list
+            if lang_code not in language_list:
+                if locale:
+                    language_list[lang_code] = {'name': locale.get_language_name().title(),
+                                                'english_name': locale.english_name,
+                                                'country': locale.get_territory_name() or ''}
+
+                    # also add language without country
+                    if locale.language not in language_list:
+                        language_list[locale.language] = {'name': locale.get_language_name().title(),
+                                                          'english_name': locale.english_name}
+                else:
+                    language_list[lang_code] = {}
+
+            # count engine for both language_country combination and language alone
+            add_engine_counter(lang_code, engine_name, language_list)
+            add_engine_counter(lang_code.split('-')[0], engine_name, language_list)
+
+    return language_list
+
+
+# Filter language list so it only includes the most supported languages and countries.
+def filter_language_list(all_languages):
+    min_supported_engines = 10
+    main_engines = [engine_name for engine_name in engines.keys()
+                    if 'general' in engines[engine_name].categories and
+                       engines[engine_name].supported_languages and
+                       not engines[engine_name].disabled]
+
+    # filter list to include only languages supported by most engines or all default general engines
+    filtered_languages = {code: lang for code, lang
+                          in all_languages.items()
+                          if (len(lang.get('counter', [])) >= min_supported_engines or
+                              all(main_engine in lang.get('counter', [])
+                                  for main_engine in main_engines))}
+
+    return filtered_languages
+
+
+# Add country codes to languages without one and filter out language codes.
+def assign_country_codes(filtered_languages, all_languages):
+    sorted_languages = sorted(all_languages,
+                              key=lambda lang: len(all_languages[lang].get('counter', [])),
+                              reverse=True)
+    previous_lang = None
+    previous_code = None
+    countries = 0
+    for current_code in sorted(filtered_languages):
+        current_lang = current_code.split('-')[0]
+
+        # count country codes per language
+        if current_lang == previous_lang:
             countries += 1
+
         else:
-            if prev_lang is not None and countries == 1:
-                del languages[prev_lang]
-                languages[prev_code]['country'] = ''
+            if previous_lang is not None:
+                # if language has no single country code
+                if countries == 0:
+                    # try to get country code with most supported engines
+                    for l in sorted_languages:
+                        l_parts = l.split('-')
+                        if len(l_parts) == 2 and l_parts[0] == previous_lang:
+                            filtered_languages[l] = all_languages[l]
+                            filtered_languages[l]['country'] = ''
+                            countries = 1
+                            break
+
+                    if countries == 0:
+                        # get most likely country code from babel
+                        subtags = get_global('likely_subtags').get(previous_lang)
+                        if subtags:
+                            subtag_parts = subtags.split('_')
+                            new_code = subtag_parts[0] + '-' + subtag_parts[-1]
+                            filtered_languages[new_code] = all_languages[previous_lang]
+                            countries = 1
+
+                if countries == 1:
+                    # remove countryless version of language if there's only one country
+                    del filtered_languages[previous_lang]
+                    if previous_code in filtered_languages:
+                        filtered_languages[previous_code]['country'] = ''
+
             countries = 0
-            prev_lang = lang
-        prev_code = code
+            previous_lang = current_lang
+
+        previous_code = current_code
 
 
 # Write languages.py.
-def write_languages_file():
+def write_languages_file(languages):
     new_file = open(languages_file, 'wb')
     file_content = '# -*- coding: utf-8 -*-\n'\
                    + '# list of language codes\n'\
@@ -183,7 +181,9 @@ def write_languages_file():
 
 
 if __name__ == "__main__":
-    fetch_supported_languages()
-    join_language_lists()
-    filter_single_country_languages()
-    write_languages_file()
+    initialize_engines(settings['engines'])
+    engines_languages = fetch_supported_languages()
+    all_languages = join_language_lists(engines_languages)
+    filtered_languages = filter_language_list(all_languages)
+    assign_country_codes(filtered_languages, all_languages)
+    write_languages_file(filtered_languages)