mirror of
https://github.com/searxng/searxng.git
synced 2024-11-19 02:40:11 +01:00
[fix] bing engine: fix paging support, show inital page.
Follow up queries for the pages needed to be fixed. - Split search-term in one for initial query and one for following queries. - Set some headers in HTTP requests, bing needs for paging support. - IMO //div[@class="sa_cc"] does no longer match in a bing response. Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
This commit is contained in:
parent
b2177e5916
commit
6b85607274
@ -1,6 +1,8 @@
|
|||||||
# SPDX-License-Identifier: AGPL-3.0-or-later
|
# SPDX-License-Identifier: AGPL-3.0-or-later
|
||||||
# lint: pylint
|
# lint: pylint
|
||||||
"""Bing (Web)
|
"""Bing (Web)
|
||||||
|
|
||||||
|
- https://github.com/searx/searx/issues/2019#issuecomment-648227442
|
||||||
"""
|
"""
|
||||||
|
|
||||||
import re
|
import re
|
||||||
@ -27,14 +29,27 @@ language_aliases = {'zh-CN': 'zh-CHS', 'zh-TW': 'zh-CHT', 'zh-HK': 'zh-CHT'}
|
|||||||
|
|
||||||
# search-url
|
# search-url
|
||||||
base_url = 'https://www.bing.com/'
|
base_url = 'https://www.bing.com/'
|
||||||
search_string = 'search?{query}&first={offset}'
|
|
||||||
|
# initial query: https://www.bing.com/search?q=foo&search=&form=QBLH
|
||||||
|
inital_query = 'search?{query}&search=&form=QBLH'
|
||||||
|
|
||||||
|
# following queries: https://www.bing.com/search?q=foo&search=&first=11&FORM=PERE
|
||||||
|
page_query = 'search?{query}&search=&first={offset}&FORM=PERE'
|
||||||
|
|
||||||
def _get_offset_from_pageno(pageno):
|
def _get_offset_from_pageno(pageno):
|
||||||
return (pageno - 1) * 10 + 1
|
return (pageno - 1) * 10 + 1
|
||||||
|
|
||||||
def request(query, params):
|
def request(query, params):
|
||||||
|
|
||||||
offset = _get_offset_from_pageno(params.get('pageno', 0))
|
offset = _get_offset_from_pageno(params.get('pageno', 1))
|
||||||
|
|
||||||
|
# logger.debug("params['pageno'] --> %s", params.get('pageno'))
|
||||||
|
# logger.debug(" offset --> %s", offset)
|
||||||
|
|
||||||
|
search_string = page_query
|
||||||
|
if offset == 1:
|
||||||
|
search_string = inital_query
|
||||||
|
|
||||||
if params['language'] == 'all':
|
if params['language'] == 'all':
|
||||||
lang = 'EN'
|
lang = 'EN'
|
||||||
else:
|
else:
|
||||||
@ -49,11 +64,19 @@ def request(query, params):
|
|||||||
search_path = search_string.format(
|
search_path = search_string.format(
|
||||||
query = urlencode({'q': query}),
|
query = urlencode({'q': query}),
|
||||||
offset = offset)
|
offset = offset)
|
||||||
|
|
||||||
|
if offset > 1:
|
||||||
|
referer = base_url + inital_query.format(query = urlencode({'q': query}))
|
||||||
|
params['headers']['Referer'] = referer
|
||||||
|
logger.debug("headers.Referer --> %s", referer )
|
||||||
|
|
||||||
params['url'] = base_url + search_path
|
params['url'] = base_url + search_path
|
||||||
|
params['headers']['Accept-Language'] = "en-US,en;q=0.5"
|
||||||
|
params['headers']['Accept'] = (
|
||||||
|
'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8'
|
||||||
|
)
|
||||||
return params
|
return params
|
||||||
|
|
||||||
|
|
||||||
def response(resp):
|
def response(resp):
|
||||||
|
|
||||||
results = []
|
results = []
|
||||||
@ -63,6 +86,9 @@ def response(resp):
|
|||||||
|
|
||||||
for result in eval_xpath(dom, '//div[@class="sa_cc"]'):
|
for result in eval_xpath(dom, '//div[@class="sa_cc"]'):
|
||||||
|
|
||||||
|
# IMO //div[@class="sa_cc"] does no longer match
|
||||||
|
logger.debug('found //div[@class="sa_cc"] --> %s', result)
|
||||||
|
|
||||||
link = eval_xpath(result, './/h3/a')[0]
|
link = eval_xpath(result, './/h3/a')[0]
|
||||||
url = link.attrib.get('href')
|
url = link.attrib.get('href')
|
||||||
title = extract_text(link)
|
title = extract_text(link)
|
||||||
|
Loading…
Reference in New Issue
Block a user