mirror of
https://github.com/searxng/searxng.git
synced 2024-11-18 08:51:06 +00:00
184 lines
5.2 KiB
Python
184 lines
5.2 KiB
Python
# SPDX-License-Identifier: AGPL-3.0-or-later
|
|
# lint: pylint
|
|
"""DuckDuckGo Lite
|
|
"""
|
|
|
|
from json import loads
|
|
|
|
from lxml.html import fromstring
|
|
|
|
from searx.utils import (
|
|
dict_subset,
|
|
eval_xpath,
|
|
eval_xpath_getindex,
|
|
extract_text,
|
|
match_language,
|
|
)
|
|
from searx.network import get
|
|
|
|
# about
|
|
about = {
|
|
"website": 'https://lite.duckduckgo.com/lite',
|
|
"wikidata_id": 'Q12805',
|
|
"official_api_documentation": 'https://duckduckgo.com/api',
|
|
"use_official_api": False,
|
|
"require_api_key": False,
|
|
"results": 'HTML',
|
|
}
|
|
|
|
# engine dependent config
|
|
categories = ['general', 'web']
|
|
paging = True
|
|
supported_languages_url = 'https://duckduckgo.com/util/u588.js'
|
|
time_range_support = True
|
|
|
|
language_aliases = {
|
|
'ar-SA': 'ar-XA',
|
|
'es-419': 'es-XL',
|
|
'ja': 'jp-JP',
|
|
'ko': 'kr-KR',
|
|
'sl-SI': 'sl-SL',
|
|
'zh-TW': 'tzh-TW',
|
|
'zh-HK': 'tzh-HK',
|
|
}
|
|
|
|
time_range_dict = {'day': 'd', 'week': 'w', 'month': 'm', 'year': 'y'}
|
|
|
|
# search-url
|
|
url = 'https://lite.duckduckgo.com/lite'
|
|
url_ping = 'https://duckduckgo.com/t/sl_l'
|
|
|
|
# match query's language to a region code that duckduckgo will accept
|
|
def get_region_code(lang, lang_list=None):
|
|
if lang == 'all':
|
|
return None
|
|
|
|
lang_code = match_language(lang, lang_list or [], language_aliases, 'wt-WT')
|
|
lang_parts = lang_code.split('-')
|
|
|
|
# country code goes first
|
|
return lang_parts[1].lower() + '-' + lang_parts[0].lower()
|
|
|
|
|
|
def request(query, params):
|
|
|
|
params['url'] = url
|
|
params['method'] = 'POST'
|
|
|
|
params['data']['q'] = query
|
|
|
|
# The API is not documented, so we do some reverse engineering and emulate
|
|
# what https://lite.duckduckgo.com/lite/ does when you press "next Page"
|
|
# link again and again ..
|
|
|
|
params['headers']['Content-Type'] = 'application/x-www-form-urlencoded'
|
|
|
|
# initial page does not have an offset
|
|
if params['pageno'] == 2:
|
|
# second page does have an offset of 30
|
|
offset = (params['pageno'] - 1) * 30
|
|
params['data']['s'] = offset
|
|
params['data']['dc'] = offset + 1
|
|
|
|
elif params['pageno'] > 2:
|
|
# third and following pages do have an offset of 30 + n*50
|
|
offset = 30 + (params['pageno'] - 2) * 50
|
|
params['data']['s'] = offset
|
|
params['data']['dc'] = offset + 1
|
|
|
|
# initial page does not have additional data in the input form
|
|
if params['pageno'] > 1:
|
|
# request the second page (and more pages) needs 'o' and 'api' arguments
|
|
params['data']['o'] = 'json'
|
|
params['data']['api'] = 'd.js'
|
|
|
|
# initial page does not have additional data in the input form
|
|
if params['pageno'] > 2:
|
|
# request the third page (and more pages) some more arguments
|
|
params['data']['nextParams'] = ''
|
|
params['data']['v'] = ''
|
|
params['data']['vqd'] = ''
|
|
|
|
region_code = get_region_code(params['language'], supported_languages)
|
|
if region_code:
|
|
params['data']['kl'] = region_code
|
|
params['cookies']['kl'] = region_code
|
|
|
|
params['data']['df'] = ''
|
|
if params['time_range'] in time_range_dict:
|
|
params['data']['df'] = time_range_dict[params['time_range']]
|
|
params['cookies']['df'] = time_range_dict[params['time_range']]
|
|
|
|
logger.debug("param data: %s", params['data'])
|
|
logger.debug("param cookies: %s", params['cookies'])
|
|
return params
|
|
|
|
|
|
# get response from search-request
|
|
def response(resp):
|
|
|
|
headers_ping = dict_subset(resp.request.headers, ['User-Agent', 'Accept-Encoding', 'Accept', 'Cookie'])
|
|
get(url_ping, headers=headers_ping)
|
|
|
|
if resp.status_code == 303:
|
|
return []
|
|
|
|
results = []
|
|
doc = fromstring(resp.text)
|
|
|
|
result_table = eval_xpath(doc, '//html/body/form/div[@class="filters"]/table')
|
|
if not len(result_table) >= 3:
|
|
# no more results
|
|
return []
|
|
result_table = result_table[2]
|
|
|
|
tr_rows = eval_xpath(result_table, './/tr')
|
|
|
|
# In the last <tr> is the form of the 'previous/next page' links
|
|
tr_rows = tr_rows[:-1]
|
|
|
|
len_tr_rows = len(tr_rows)
|
|
offset = 0
|
|
|
|
while len_tr_rows >= offset + 4:
|
|
|
|
# assemble table rows we need to scrap
|
|
tr_title = tr_rows[offset]
|
|
tr_content = tr_rows[offset + 1]
|
|
offset += 4
|
|
|
|
# ignore sponsored Adds <tr class="result-sponsored">
|
|
if tr_content.get('class') == 'result-sponsored':
|
|
continue
|
|
|
|
a_tag = eval_xpath_getindex(tr_title, './/td//a[@class="result-link"]', 0, None)
|
|
if a_tag is None:
|
|
continue
|
|
|
|
td_content = eval_xpath_getindex(tr_content, './/td[@class="result-snippet"]', 0, None)
|
|
if td_content is None:
|
|
continue
|
|
|
|
results.append(
|
|
{
|
|
'title': a_tag.text_content(),
|
|
'content': extract_text(td_content),
|
|
'url': a_tag.get('href'),
|
|
}
|
|
)
|
|
|
|
return results
|
|
|
|
|
|
# get supported languages from their site
|
|
def _fetch_supported_languages(resp):
|
|
|
|
# response is a js file with regions as an embedded object
|
|
response_page = resp.text
|
|
response_page = response_page[response_page.find('regions:{') + 8 :]
|
|
response_page = response_page[: response_page.find('}') + 1]
|
|
|
|
regions_json = loads(response_page)
|
|
supported_languages = map((lambda x: x[3:] + '-' + x[:2].upper()), regions_json.keys())
|
|
|
|
return list(supported_languages)
|