mirror of
https://github.com/searxng/searxng.git
synced 2024-11-22 20:17:45 +01:00
c3daa08537
Xpath engine and results template changed to account for the fact that archive.org doesn't cache .onions, though some onion engines migth have their own cache. Disabled by default. Can be enabled by setting the SOCKS proxies to wherever Tor is listening and setting using_tor_proxy as True. Requires Tor and updating packages. To avoid manually adding the timeout on each engine, you can set extra_proxy_timeout to account for Tor's (or whatever proxy used) extra time.
83 lines
2.3 KiB
Python
83 lines
2.3 KiB
Python
"""
|
|
Ahmia (Onions)
|
|
|
|
@website http://msydqstlz2kzerdg.onion
|
|
@provides-api no
|
|
|
|
@using-api no
|
|
@results HTML
|
|
@stable no
|
|
@parse url, title, content
|
|
"""
|
|
|
|
from urllib.parse import urlencode, urlparse, parse_qs
|
|
from lxml.html import fromstring
|
|
from searx.engines.xpath import extract_url, extract_text
|
|
|
|
# engine config
|
|
categories = ['onions']
|
|
paging = True
|
|
page_size = 10
|
|
|
|
# search url
|
|
search_url = 'http://msydqstlz2kzerdg.onion/search/?{query}'
|
|
time_range_support = True
|
|
time_range_dict = {'day': 1,
|
|
'week': 7,
|
|
'month': 30}
|
|
|
|
# xpaths
|
|
results_xpath = '//li[@class="result"]'
|
|
url_xpath = './h4/a/@href'
|
|
title_xpath = './h4/a[1]'
|
|
content_xpath = './/p[1]'
|
|
correction_xpath = '//*[@id="didYouMean"]//a'
|
|
number_of_results_xpath = '//*[@id="totalResults"]'
|
|
|
|
|
|
def request(query, params):
|
|
params['url'] = search_url.format(query=urlencode({'q': query}))
|
|
|
|
if params['time_range'] in time_range_dict:
|
|
params['url'] += '&' + urlencode({'d': time_range_dict[params['time_range']]})
|
|
|
|
return params
|
|
|
|
|
|
def response(resp):
|
|
results = []
|
|
dom = fromstring(resp.text)
|
|
|
|
# trim results so there's not way too many at once
|
|
first_result_index = page_size * (resp.search_params.get('pageno', 1) - 1)
|
|
all_results = dom.xpath(results_xpath)
|
|
trimmed_results = all_results[first_result_index:first_result_index + page_size]
|
|
|
|
# get results
|
|
for result in trimmed_results:
|
|
# remove ahmia url and extract the actual url for the result
|
|
raw_url = extract_url(result.xpath(url_xpath), search_url)
|
|
cleaned_url = parse_qs(urlparse(raw_url).query).get('redirect_url', [''])[0]
|
|
|
|
title = extract_text(result.xpath(title_xpath))
|
|
content = extract_text(result.xpath(content_xpath))
|
|
|
|
results.append({'url': cleaned_url,
|
|
'title': title,
|
|
'content': content,
|
|
'is_onion': True})
|
|
|
|
# get spelling corrections
|
|
for correction in dom.xpath(correction_xpath):
|
|
results.append({'correction': extract_text(correction)})
|
|
|
|
# get number of results
|
|
number_of_results = dom.xpath(number_of_results_xpath)
|
|
if number_of_results:
|
|
try:
|
|
results.append({'number_of_results': int(extract_text(number_of_results))})
|
|
except:
|
|
pass
|
|
|
|
return results
|