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.

Add languages supported by either all default general engines or 10 engines.

[fix] use english in google engine if no language was set - this prevents guessing the language by the IP of the instance

and refactor method to make it testable without making requests

utils/fetch_languages.py gets languages supported by each engine and
generates engines_languages.json with each engine's supported language.

Engines:
 * Bing images
 * Flickr (noapi)
 * Google
 * Google Images
 * Google News

nid/pref cookies are also removed

TODO check other engines too

[enh] google engine : avoid some "sorry google" by adding another cookie : NID. This cookie is specific by hostname.
This allow to send request to google.* (according to the search language).
Before this commit, request in other languages than english was sent to www.google.com which was redirected to www.google.*
The PREF is still use on the www.google.com domain.

Sometimes there is two requests to google (depending of the source IP) : one to google.com, the second to google.fr (for instance).

Going to https://www.google.com/ncr and saving the PREF cookie for future use prevent this (there is no redirection).

But, recently (or not ?), by doing this the search returns English results even if the Accept-Language is specified.

There is still a way to prevent this : going to preference, set the search language. I don't know if this can be done by searx.

For now, a quick fix is to disable the use of the PREF cookie when the search language is not English (google engine will slower but returns excepted results).

[fix] google engine: depending on the IP of the searx instance, each searx request where making two HTTP requests (see https://support.google.com/websearch/answer/873?hl=en )