mirror of
https://github.com/searxng/searxng.git
synced 2024-11-30 14:41:00 +00:00
153 lines
4.6 KiB
Python
153 lines
4.6 KiB
Python
"""
|
|
Yahoo (Web)
|
|
|
|
@website https://search.yahoo.com/web
|
|
@provide-api yes (https://developer.yahoo.com/boss/search/),
|
|
$0.80/1000 queries
|
|
|
|
@using-api no (because pricing)
|
|
@results HTML (using search portal)
|
|
@stable no (HTML can change)
|
|
@parse url, title, content, suggestion
|
|
"""
|
|
|
|
from lxml import html
|
|
from searx.engines.xpath import extract_text, extract_url
|
|
from searx.url_utils import unquote, urlencode
|
|
|
|
# engine dependent config
|
|
categories = ['general']
|
|
paging = True
|
|
language_support = True
|
|
time_range_support = True
|
|
|
|
# search-url
|
|
base_url = 'https://search.yahoo.com/'
|
|
search_url = 'search?{query}&b={offset}&fl=1&vl=lang_{lang}'
|
|
search_url_with_time = 'search?{query}&b={offset}&fl=1&vl=lang_{lang}&age={age}&btf={btf}&fr2=time'
|
|
|
|
supported_languages_url = 'https://search.yahoo.com/web/advanced'
|
|
|
|
# specific xpath variables
|
|
results_xpath = "//div[contains(concat(' ', normalize-space(@class), ' '), ' Sr ')]"
|
|
url_xpath = './/h3/a/@href'
|
|
title_xpath = './/h3/a'
|
|
content_xpath = './/div[@class="compText aAbs"]'
|
|
suggestion_xpath = "//div[contains(concat(' ', normalize-space(@class), ' '), ' AlsoTry ')]//a"
|
|
|
|
time_range_dict = {'day': ['1d', 'd'],
|
|
'week': ['1w', 'w'],
|
|
'month': ['1m', 'm']}
|
|
|
|
|
|
# remove yahoo-specific tracking-url
|
|
def parse_url(url_string):
|
|
endings = ['/RS', '/RK']
|
|
endpositions = []
|
|
start = url_string.find('http', url_string.find('/RU=') + 1)
|
|
|
|
for ending in endings:
|
|
endpos = url_string.rfind(ending)
|
|
if endpos > -1:
|
|
endpositions.append(endpos)
|
|
|
|
if start == 0 or len(endpositions) == 0:
|
|
return url_string
|
|
else:
|
|
end = min(endpositions)
|
|
return unquote(url_string[start:end])
|
|
|
|
|
|
def _get_url(query, offset, language, time_range):
|
|
if time_range in time_range_dict:
|
|
return base_url + search_url_with_time.format(offset=offset,
|
|
query=urlencode({'p': query}),
|
|
lang=language,
|
|
age=time_range_dict[time_range][0],
|
|
btf=time_range_dict[time_range][1])
|
|
return base_url + search_url.format(offset=offset,
|
|
query=urlencode({'p': query}),
|
|
lang=language)
|
|
|
|
|
|
def _get_language(params):
|
|
if params['language'] == 'all':
|
|
return 'en'
|
|
elif params['language'][:2] == 'zh':
|
|
if params['language'] == 'zh' or params['language'] == 'zh-CH':
|
|
return 'szh'
|
|
else:
|
|
return 'tzh'
|
|
else:
|
|
return params['language'].split('-')[0]
|
|
|
|
|
|
# do search-request
|
|
def request(query, params):
|
|
if params['time_range'] and params['time_range'] not in time_range_dict:
|
|
return params
|
|
|
|
offset = (params['pageno'] - 1) * 10 + 1
|
|
language = _get_language(params)
|
|
|
|
params['url'] = _get_url(query, offset, language, params['time_range'])
|
|
|
|
# TODO required?
|
|
params['cookies']['sB'] = 'fl=1&vl=lang_{lang}&sh=1&rw=new&v=1'\
|
|
.format(lang=language)
|
|
|
|
return params
|
|
|
|
|
|
# get response from search-request
|
|
def response(resp):
|
|
results = []
|
|
|
|
dom = html.fromstring(resp.text)
|
|
|
|
try:
|
|
results_num = int(dom.xpath('//div[@class="compPagination"]/span[last()]/text()')[0]
|
|
.split()[0].replace(',', ''))
|
|
results.append({'number_of_results': results_num})
|
|
except:
|
|
pass
|
|
|
|
# parse results
|
|
for result in dom.xpath(results_xpath):
|
|
try:
|
|
url = parse_url(extract_url(result.xpath(url_xpath), search_url))
|
|
title = extract_text(result.xpath(title_xpath)[0])
|
|
except:
|
|
continue
|
|
|
|
content = extract_text(result.xpath(content_xpath)[0])
|
|
|
|
# append result
|
|
results.append({'url': url,
|
|
'title': title,
|
|
'content': content})
|
|
|
|
# if no suggestion found, return results
|
|
suggestions = dom.xpath(suggestion_xpath)
|
|
if not suggestions:
|
|
return results
|
|
|
|
# parse suggestion
|
|
for suggestion in suggestions:
|
|
# append suggestion
|
|
results.append({'suggestion': extract_text(suggestion)})
|
|
|
|
# return results
|
|
return results
|
|
|
|
|
|
# get supported languages from their site
|
|
def _fetch_supported_languages(resp):
|
|
supported_languages = []
|
|
dom = html.fromstring(resp.text)
|
|
options = dom.xpath('//div[@id="yschlang"]/span/label/input')
|
|
for option in options:
|
|
code = option.xpath('./@value')[0][5:].replace('_', '-')
|
|
supported_languages.append(code)
|
|
|
|
return supported_languages
|