mirror of
https://github.com/searxng/searxng.git
synced 2024-11-30 14:41:00 +00:00
ca93a01844
The language_support variable is set to True by default, and set to False in only 5 engines. Except the documentation and the /config URL, this variable is not used. This commit remove the variable definition in the engines, and set value according to supported_languages length: False when the length is 0, True otherwise. Close #2485
77 lines
2.3 KiB
Python
77 lines
2.3 KiB
Python
# SPDX-License-Identifier: AGPL-3.0-or-later
|
|
"""
|
|
Duden
|
|
"""
|
|
|
|
import re
|
|
from urllib.parse import quote, urljoin
|
|
from lxml import html
|
|
from searx.utils import extract_text, eval_xpath, eval_xpath_list, eval_xpath_getindex
|
|
|
|
# about
|
|
about = {
|
|
"website": 'https://www.duden.de',
|
|
"wikidata_id": 'Q73624591',
|
|
"official_api_documentation": None,
|
|
"use_official_api": False,
|
|
"require_api_key": False,
|
|
"results": 'HTML',
|
|
}
|
|
|
|
categories = ['general']
|
|
paging = True
|
|
|
|
# search-url
|
|
base_url = 'https://www.duden.de/'
|
|
search_url = base_url + 'suchen/dudenonline/{query}?search_api_fulltext=&page={offset}'
|
|
|
|
|
|
def request(query, params):
|
|
'''pre-request callback
|
|
params<dict>:
|
|
method : POST/GET
|
|
headers : {}
|
|
data : {} # if method == POST
|
|
url : ''
|
|
category: 'search category'
|
|
pageno : 1 # number of the requested page
|
|
'''
|
|
|
|
offset = (params['pageno'] - 1)
|
|
if offset == 0:
|
|
search_url_fmt = base_url + 'suchen/dudenonline/{query}'
|
|
params['url'] = search_url_fmt.format(query=quote(query))
|
|
else:
|
|
params['url'] = search_url.format(offset=offset, query=quote(query))
|
|
# after the last page of results, spelling corrections are returned after a HTTP redirect
|
|
# whatever the page number is
|
|
params['soft_max_redirects'] = 1
|
|
return params
|
|
|
|
|
|
def response(resp):
|
|
'''post-response callback
|
|
resp: requests response object
|
|
'''
|
|
results = []
|
|
|
|
dom = html.fromstring(resp.text)
|
|
|
|
number_of_results_element =\
|
|
eval_xpath_getindex(dom, '//a[@class="active" and contains(@href,"/suchen/dudenonline")]/span/text()',
|
|
0, default=None)
|
|
if number_of_results_element is not None:
|
|
number_of_results_string = re.sub('[^0-9]', '', number_of_results_element)
|
|
results.append({'number_of_results': int(number_of_results_string)})
|
|
|
|
for result in eval_xpath_list(dom, '//section[not(contains(@class, "essay"))]'):
|
|
url = eval_xpath_getindex(result, './/h2/a', 0).get('href')
|
|
url = urljoin(base_url, url)
|
|
title = eval_xpath(result, 'string(.//h2/a)').strip()
|
|
content = extract_text(eval_xpath(result, './/p'))
|
|
# append result
|
|
results.append({'url': url,
|
|
'title': title,
|
|
'content': content})
|
|
|
|
return results
|