refactor engine's search language handling

Add match_language function in utils to match any user given
language code with a list of engine's supported languages.

Also add language_aliases dict on each engine to translate
standard language codes into the custom codes used by the engine.

1 files changed, 61 insertions, 0 deletions

diff --git a/searx/utils.py b/searx/utils.py
index 8f095f3b..77c39290 100644
--- a/searx/utils.py
+++ b/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: