[fix] Google API changed

It seems that Google is rolling out a modified WEB API [1][2].

In the past there was only the UI language in the `hl` argument but nowadays it
seems a combination of the UI language and the "search region" is mixed in this
argument and the `gl` argument has been removed.  I'm very surprised that google
is starting to mix the parameters of the UI with the parameters of the search
index.

This patch modifies the get_google_info(..) function.  Beside Google-WEB this
function is also used by other Google services, here are some examples to test
region & language of ..

- Google-WEB:    `!go dragon boat :en-CA`
- Google-News:   `!gon dragon boat :en-CA`
- Google-Videos: `!gov bmw :en-CA`
- Goolge-Images  `!goi bmw :en-CA`

- [1] https://github.com/searxng/searxng/issues/2515#issuecomment-1606294635
- [2] https://github.com/searxng/searxng/issues/2515#issuecomment-1607150817

Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
This commit is contained in:
Markus Heiser 2023-06-26 13:33:12 +02:00 committed by Markus Heiser
parent e8706fb738
commit da7c30291d

View file

@ -172,7 +172,8 @@ def get_google_info(params, eng_traits):
# The Interface Language:
# https://developers.google.com/custom-search/docs/xml_results_appendices#interfaceLanguages
ret_val['params']['hl'] = lang_code
# https://github.com/searxng/searxng/issues/2515#issuecomment-1607150817
ret_val['params']['hl'] = f'{lang_code}-{country}'
# lr parameter:
# The lr (language restrict) parameter restricts search results to
@ -212,7 +213,8 @@ def get_google_info(params, eng_traits):
# countries other than the United States.
# https://developers.google.com/custom-search/docs/xml_results#glsp
ret_val['params']['gl'] = country
# https://github.com/searxng/searxng/issues/2515#issuecomment-1606294635
# ret_val['params']['gl'] = country
# ie parameter:
# The ie parameter sets the character encoding scheme that should be used