This commit is contained in:
Joseph Cheung 2023-02-24 09:15:34 +08:00
parent 203079aecf
commit 8914f86883
8 changed files with 607 additions and 4525 deletions

View file

@ -5,20 +5,20 @@
"""
# pylint: disable=use-dict-literal
import json
from json import loads
from urllib.parse import urlencode
import lxml
from lxml import etree
from httpx import HTTPError
from searx import settings
from searx.engines import (
engines,
google,
)
from searx.data import ENGINES_LANGUAGES
from searx.network import get as http_get
from searx.exceptions import SearxEngineResponseException
# a fetch_supported_languages() for XPath engines isn't available right now
# _brave = ENGINES_LANGUAGES['brave'].keys()
def get(*args, **kwargs):
if 'timeout' not in kwargs:
@ -55,58 +55,34 @@ def dbpedia(query, _lang):
results = []
if response.ok:
dom = lxml.etree.fromstring(response.content)
dom = etree.fromstring(response.content)
results = dom.xpath('//Result/Label//text()')
return results
def duckduckgo(query, sxng_locale):
"""Autocomplete from DuckDuckGo. Supports DuckDuckGo's languages"""
def duckduckgo(query, _lang):
# duckduckgo autocompleter
url = 'https://ac.duckduckgo.com/ac/?{0}&type=list'
traits = engines['duckduckgo'].traits
args = {
'q': query,
'kl': traits.get_region(sxng_locale, traits.all_locale),
}
url = 'https://duckduckgo.com/ac/?type=list&' + urlencode(args)
resp = get(url)
ret_val = []
if resp.ok:
j = resp.json()
if len(j) > 1:
ret_val = j[1]
return ret_val
resp = loads(get(url.format(urlencode(dict(q=query)))).text)
if len(resp) > 1:
return resp[1]
return []
def google_complete(query, sxng_locale):
"""Autocomplete from Google. Supports Google's languages and subdomains
(:py:obj:`searx.engines.google.get_google_info`) by using the async REST
API::
def google(query, lang):
# google autocompleter
autocomplete_url = 'https://suggestqueries.google.com/complete/search?client=toolbar&'
https://{subdomain}/complete/search?{args}
response = get(autocomplete_url + urlencode(dict(hl=lang, q=query)))
"""
google_info = google.get_google_info({'searxng_locale': sxng_locale}, engines['google'].traits)
url = 'https://{subdomain}/complete/search?{args}'
args = urlencode(
{
'q': query,
'client': 'gws-wiz',
'hl': google_info['params']['hl'],
}
)
results = []
resp = get(url.format(subdomain=google_info['subdomain'], args=args))
if resp.ok:
json_txt = resp.text[resp.text.find('[') : resp.text.find(']', -3) + 1]
data = json.loads(json_txt)
for item in data[0]:
results.append(lxml.html.fromstring(item[0]).text_content())
if response.ok:
dom = etree.fromstring(response.text)
results = dom.xpath('//suggestion/@data')
return results
@ -133,9 +109,9 @@ def seznam(query, _lang):
]
def startpage(query, sxng_locale):
"""Autocomplete from Startpage. Supports Startpage's languages"""
lui = engines['startpage'].traits.get_language(sxng_locale, 'english')
def startpage(query, lang):
# startpage autocompleter
lui = ENGINES_LANGUAGES['startpage'].get(lang, 'english')
url = 'https://startpage.com/suggestions?{query}'
resp = get(url.format(query=urlencode({'q': query, 'segment': 'startpage.udog', 'lui': lui})))
data = resp.json()
@ -146,20 +122,20 @@ def swisscows(query, _lang):
# swisscows autocompleter
url = 'https://swisscows.ch/api/suggest?{query}&itemsCount=5'
resp = json.loads(get(url.format(query=urlencode({'query': query}))).text)
resp = loads(get(url.format(query=urlencode({'query': query}))).text)
return resp
def qwant(query, sxng_locale):
"""Autocomplete from Qwant. Supports Qwant's regions."""
def qwant(query, lang):
# qwant autocompleter (additional parameter : lang=en_en&count=xxx )
url = 'https://api.qwant.com/api/suggest?{query}'
resp = get(url.format(query=urlencode({'q': query, 'lang': lang})))
results = []
locale = engines['startpage'].traits.get_region(sxng_locale, 'en_US')
url = 'https://api.qwant.com/v3/suggest?{query}'
resp = get(url.format(query=urlencode({'q': query, 'locale': locale, 'version': '2'})))
if resp.ok:
data = resp.json()
data = loads(resp.text)
if data['status'] == 'success':
for item in data['data']['items']:
results.append(item['value'])
@ -167,38 +143,21 @@ def qwant(query, sxng_locale):
return results
def wikipedia(query, sxng_locale):
"""Autocomplete from Wikipedia. Supports Wikipedia's languages (aka netloc)."""
results = []
eng_traits = engines['wikipedia'].traits
wiki_lang = eng_traits.get_language(sxng_locale, 'en')
wiki_netloc = eng_traits.custom['wiki_netloc'].get(wiki_lang, 'en.wikipedia.org')
def wikipedia(query, lang):
# wikipedia autocompleter
url = 'https://' + lang + '.wikipedia.org/w/api.php?action=opensearch&{0}&limit=10&namespace=0&format=json'
url = 'https://{wiki_netloc}/w/api.php?{args}'
args = urlencode(
{
'action': 'opensearch',
'format': 'json',
'formatversion': '2',
'search': query,
'namespace': '0',
'limit': '10',
}
)
resp = get(url.format(args=args, wiki_netloc=wiki_netloc))
if resp.ok:
data = resp.json()
if len(data) > 1:
results = data[1]
return results
resp = loads(get(url.format(urlencode(dict(search=query)))).text)
if len(resp) > 1:
return resp[1]
return []
def yandex(query, _lang):
# yandex autocompleter
url = "https://suggest.yandex.com/suggest-ff.cgi?{0}"
resp = json.loads(get(url.format(urlencode(dict(part=query)))).text)
resp = loads(get(url.format(urlencode(dict(part=query)))).text)
if len(resp) > 1:
return resp[1]
return []
@ -207,7 +166,7 @@ def yandex(query, _lang):
backends = {
'dbpedia': dbpedia,
'duckduckgo': duckduckgo,
'google': google_complete,
'google': google,
'seznam': seznam,
'startpage': startpage,
'swisscows': swisscows,
@ -218,11 +177,12 @@ backends = {
}
def search_autocomplete(backend_name, query, sxng_locale):
def search_autocomplete(backend_name, query, lang):
backend = backends.get(backend_name)
if backend is None:
return []
try:
return backend(query, sxng_locale)
return backend(query, lang)
except (HTTPError, SearxEngineResponseException):
return []

169
searx/baidu.py Normal file
View file

@ -0,0 +1,169 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
"""Bing (Web)
- https://github.com/searx/searx/issues/2019#issuecomment-648227442
"""
import re
from urllib.parse import urlencode
from lxml import html
from searx.utils import eval_xpath, extract_text, eval_xpath_list, eval_xpath_getindex
from searx.network import raise_for_httperror, multi_requests, get, Request
from searx.exceptions import SearxEngineCaptchaException
about = {
"website": 'https://www.baidu.com',
"wikidata_id": 'Q14772',
"official_api_documentation": 'https://apis.baidu.com/',
"use_official_api": False,
"require_api_key": False,
"results": 'HTML',
"language": 'zn',
}
# engine dependent config
categories = ['general', 'web']
paging = False
time_range_support = False
safesearch = False
base_url = 'https://www.baidu.com/'
search_string = 's?{query}'
skip_tpls = ('img_normal', 'short_video', 'yl_music_song', 'dict3', 'recommend_list')
desc_xpath_per_tpl = {
'se_com_default': './/span[contains(@class, "content-right_8Zs40")]',
'kaifa_pc_open_source_software': './/p[contains(@class, "c-color-text")]',
'bk_polysemy': './/div/@aria-label',
'se_st_single_video_zhanzhang': './/span[contains(@class, "c-span-last")]//p[2]',
}
def get_initial_parameters(params):
resp_index = get(base_url, headers=params['headers'], raise_for_httperror=True)
dom = html.fromstring(resp_index.text)
query_params = {}
for ielement in eval_xpath_list(dom, '//form[@id="form"]//input[@name]'):
name = ielement.attrib.get('name')
value = ielement.attrib.get('value')
query_params[name] = value
return query_params, resp_index.cookies
def request(query, params):
params['headers'].update(
{
'Accept-Language': 'en-US,en;q=0.5',
'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',
'Sec-Fetch-Dest': 'document',
'Sec-Fetch-Mode': 'navigate',
'Sec-Fetch-Site': 'same-origin',
'Sec-Fetch-User': '?1',
'Sec-GPC': '1',
'Upgrade-Insecure-Requests': '1',
'TE': 'trailers',
}
)
query_params, cookies = get_initial_parameters(params)
query_params['wd'] = query
params['url'] = base_url + search_string.format(query=urlencode(query_params))
params['cookies'] = cookies
params['raise_for_httperror'] = False
return params
def response(resp):
results = []
if resp.url.host == 'wappass.baidu.com' or resp.url.path.startswith('/static/captcha'):
raise SearxEngineCaptchaException()
raise_for_httperror(resp)
dom = html.fromstring(resp.text)
# follow redirect but don't use the result page to reduce the CAPTCHA issue
redirect_element = eval_xpath_getindex(dom, '//noscript/meta[@http-equiv="refresh"]/@content', 0, default=None)
if redirect_element and redirect_element.startswith('0; url='):
get(
base_url + redirect_element[8:],
headers=resp.search_params['headers'],
cookies=resp.search_params['cookies'],
)
for result in eval_xpath_list(dom, '//div[contains(@id,"content_left")]/div[contains(@class, "c-container")]'):
tpl = result.attrib.get('tpl')
if tpl in skip_tpls:
continue
if tpl == 'kaifa_pc_blog_weak':
# skip the result to kaifa.baidu.com (search engine for IT)
# but includes results from kaifa
for r2 in eval_xpath_list(result, './/div[contains(@class, "c-gap-bottom-small")]'):
title = extract_text(eval_xpath(r2, './/div[@class="c-row"]//a'))
url = extract_text(eval_xpath(r2, './/div[@class="c-row"]//a/@href'))
content = extract_text(eval_xpath(r2, '//span[@class="c-line-clamp2"]'))
results.append(
{
'url': url,
'title': title,
'content': content,
}
)
continue
# normal results
title = extract_text(eval_xpath(result, './/h3/a'))
url = extract_text(eval_xpath(result, './/h3/a/@href'))
if not title or not url:
continue
content = None
if tpl in desc_xpath_per_tpl:
# try the XPath for the Baidu template
content = extract_text(eval_xpath(result, desc_xpath_per_tpl[tpl]))
if not content:
# no content was found: try all the XPath from the Baidu templates
for xp in desc_xpath_per_tpl.values():
content = extract_text(eval_xpath(result, xp))
if content:
break
results.append(
{
'url': url,
'title': title,
'content': content,
}
)
# resolve the Baidu redirections
# note: Baidu does not support HTTP/2
request_list = [
Request.get(
u['url'].replace('http://www.baidu.com/link?url=', 'https://www.baidu.com/link?url='),
allow_redirects=False,
headers=resp.search_params['headers'],
)
for u in results
]
response_list = multi_requests(request_list)
for i, redirect_response in enumerate(response_list):
if not isinstance(redirect_response, Exception):
results[i]['url'] = redirect_response.headers['location']
return results
def debug_write_content_to_file(text):
RE_STYLE_ELEMENT = re.compile(r'<style[^>]*>[^<]+</style>')
RE_SCRIPT_ELEMENT = re.compile(r'<script[^>]*>[^<]+</script>')
RE_COMMENT_ELEMENT = re.compile(r'\<\!\-\-[^-]+\-\-\>')
with open('baidu.html', 'wt', encoding='utf-8') as f:
text = RE_STYLE_ELEMENT.sub("", text)
text = RE_SCRIPT_ELEMENT.sub("", text)
text = RE_COMMENT_ELEMENT.sub("", text)
text = "\n".join([ll.rstrip() for ll in text.splitlines() if ll.strip()])
f.write(text)

File diff suppressed because it is too large Load diff

View file

@ -1,39 +1,34 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
"""This is the implementation of the Google WEB engine. Some of this
implementations (manly the :py:obj:`get_google_info`) are shared by other
engines:
"""This is the implementation of the google WEB engine. Some of this
implementations are shared by other engines:
- :ref:`google images engine`
- :ref:`google news engine`
- :ref:`google videos engine`
- :ref:`google scholar engine`
- :ref:`google autocomplete`
The google WEB engine itself has a special setup option:
.. code:: yaml
- name: google
...
use_mobile_ui: false
``use_mobile_ui``: (default: ``false``)
Enables to use *mobile endpoint* to bypass the google blocking (see
:issue:`159`). On the mobile UI of Google Search, the button :guilabel:`More
results` is not affected by Google rate limiting and we can still do requests
while actively blocked by the original Google search. By activate
``use_mobile_ui`` this behavior is simulated by adding the parameter
``async=use_ac:true,_fmt:pc`` to the :py:func:`request`.
"""
from typing import TYPE_CHECKING
import re
from urllib.parse import urlencode
from lxml import html
import babel
import babel.core
import babel.languages
from searx.utils import extract_text, eval_xpath, eval_xpath_list, eval_xpath_getindex
from searx.locales import language_tag, region_tag, get_offical_locales
from searx import network
from searx.utils import match_language, extract_text, eval_xpath, eval_xpath_list, eval_xpath_getindex
from searx.exceptions import SearxEngineCaptchaException
from searx.enginelib.traits import EngineTraits
if TYPE_CHECKING:
import logging
logger: logging.Logger
traits: EngineTraits
# about
about = {
@ -50,6 +45,64 @@ categories = ['general', 'web']
paging = True
time_range_support = True
safesearch = True
send_accept_language_header = True
use_mobile_ui = False
supported_languages_url = 'https://www.google.com/preferences?#languages'
# based on https://en.wikipedia.org/wiki/List_of_Google_domains and tests
google_domains = {
'BG': 'google.bg', # Bulgaria
'CZ': 'google.cz', # Czech Republic
'DE': 'google.de', # Germany
'DK': 'google.dk', # Denmark
'AT': 'google.at', # Austria
'CH': 'google.ch', # Switzerland
'GR': 'google.gr', # Greece
'AU': 'google.com.au', # Australia
'CA': 'google.ca', # Canada
'GB': 'google.co.uk', # United Kingdom
'ID': 'google.co.id', # Indonesia
'IE': 'google.ie', # Ireland
'IN': 'google.co.in', # India
'MY': 'google.com.my', # Malaysia
'NZ': 'google.co.nz', # New Zealand
'PH': 'google.com.ph', # Philippines
'SG': 'google.com.sg', # Singapore
'US': 'google.com', # United States (google.us) redirects to .com
'ZA': 'google.co.za', # South Africa
'AR': 'google.com.ar', # Argentina
'CL': 'google.cl', # Chile
'ES': 'google.es', # Spain
'MX': 'google.com.mx', # Mexico
'EE': 'google.ee', # Estonia
'FI': 'google.fi', # Finland
'BE': 'google.be', # Belgium
'FR': 'google.fr', # France
'IL': 'google.co.il', # Israel
'HR': 'google.hr', # Croatia
'HU': 'google.hu', # Hungary
'IT': 'google.it', # Italy
'JP': 'google.co.jp', # Japan
'KR': 'google.co.kr', # South Korea
'LT': 'google.lt', # Lithuania
'LV': 'google.lv', # Latvia
'NO': 'google.no', # Norway
'NL': 'google.nl', # Netherlands
'PL': 'google.pl', # Poland
'BR': 'google.com.br', # Brazil
'PT': 'google.pt', # Portugal
'RO': 'google.ro', # Romania
'RU': 'google.ru', # Russia
'SK': 'google.sk', # Slovakia
'SI': 'google.si', # Slovenia
'SE': 'google.se', # Sweden
'TH': 'google.co.th', # Thailand
'TR': 'google.com.tr', # Turkey
'UA': 'google.com.ua', # Ukraine
'CN': 'google.com.hk', # There is no google.cn, we use .com.hk for zh-CN
'HK': 'google.com.hk', # Hong Kong
'TW': 'google.com.tw', # Taiwan
}
time_range_dict = {'day': 'd', 'week': 'w', 'month': 'm', 'year': 'y'}
@ -62,50 +115,47 @@ filter_mapping = {0: 'off', 1: 'medium', 2: 'high'}
results_xpath = './/div[@data-sokoban-container]'
title_xpath = './/a/h3[1]'
href_xpath = './/a[h3]/@href'
content_xpath = './/div[@data-content-feature]'
content_xpath = './/div[@data-content-feature=1]'
# google *sections* are no usual *results*, we ignore them
g_section_with_header = './g-section-with-header'
# Suggestions are links placed in a *card-section*, we extract only the text
# from the links not the links itself.
suggestion_xpath = '//div[contains(@class, "EIaa9b")]//a'
# UI_ASYNC = 'use_ac:true,_fmt:html' # returns a HTTP 500 when user search for
# # celebrities like '!google natasha allegri'
# # or '!google chris evans'
UI_ASYNC = 'use_ac:true,_fmt:prog'
"""Format of the response from UI's async request."""
def get_google_info(params, eng_traits):
"""Composing various (language) properties for the google engines (:ref:`google
API`).
def get_lang_info(params, lang_list, custom_aliases, supported_any_language):
"""Composing various language properties for the google engines.
This function is called by the various google engines (:ref:`google web
engine`, :ref:`google images engine`, :ref:`google news engine` and
:ref:`google videos engine`).
:param dict param: Request parameters of the engine. At least
a ``searxng_locale`` key should be in the dictionary.
:param dict param: request parameters of the engine
:param eng_traits: Engine's traits fetched from google preferences
(:py:obj:`searx.enginelib.traits.EngineTraits`)
:param list lang_list: list of supported languages of the engine
:py:obj:`ENGINES_LANGUAGES[engine-name] <searx.data.ENGINES_LANGUAGES>`
:param dict lang_list: custom aliases for non standard language codes
(used when calling :py:func:`searx.utils.match_language`)
:param bool supported_any_language: When a language is not specified, the
language interpretation is left up to Google to decide how the search
results should be delivered. This argument is ``True`` for the google
engine and ``False`` for the other engines (google-images, -news,
-scholar, -videos).
:rtype: dict
:returns:
Py-Dictionary with the key/value pairs:
language:
The language code that is used by google (e.g. ``lang_en`` or
``lang_zh-TW``)
Return value from :py:func:`searx.utils.match_language`
country:
The country code that is used by google (e.g. ``US`` or ``TW``)
locale:
A instance of :py:obj:`babel.core.Locale` build from the
``searxng_locale`` value.
The country code (e.g. US, AT, CA, FR, DE ..)
subdomain:
Google subdomain :py:obj:`google_domains` that fits to the country
@ -115,67 +165,52 @@ def get_google_info(params, eng_traits):
Py-Dictionary with additional request arguments (can be passed to
:py:func:`urllib.parse.urlencode`).
- ``hl`` parameter: specifies the interface language of user interface.
- ``lr`` parameter: restricts search results to documents written in
a particular language.
- ``cr`` parameter: restricts search results to documents
originating in a particular country.
- ``ie`` parameter: sets the character encoding scheme that should
be used to interpret the query string ('utf8').
- ``oe`` parameter: sets the character encoding scheme that should
be used to decode the XML result ('utf8').
headers:
Py-Dictionary with additional HTTP headers (can be passed to
request's headers)
- ``Accept: '*/*``
"""
ret_val = {
'language': None,
'country': None,
'subdomain': None,
'params': {},
'headers': {},
'cookies': {},
'locale': None,
}
sxng_locale = params.get('searxng_locale', 'all')
try:
locale = babel.Locale.parse(sxng_locale, sep='-')
except babel.core.UnknownLocaleError:
locale = None
# language ...
eng_lang = eng_traits.get_language(sxng_locale, 'lang_en')
lang_code = eng_lang.split('_')[-1] # lang_zh-TW --> zh-TW / lang_en --> en
country = eng_traits.get_region(sxng_locale, eng_traits.all_locale)
_lang = params['language']
_any_language = _lang.lower() == 'all'
if _any_language:
_lang = 'en-US'
language = match_language(_lang, lang_list, custom_aliases)
ret_val['language'] = language
# Test zh_hans & zh_hant --> in the topmost links in the result list of list
# TW and HK you should a find wiktionary.org zh_hant link. In the result
# list of zh-CN should not be no hant link instead you should find
# zh.m.wikipedia.org/zh somewhere in the top.
# country ...
# '!go 日 :zh-TW' --> https://zh.m.wiktionary.org/zh-hant/%E6%97%A5
# '!go 日 :zh-CN' --> https://zh.m.wikipedia.org/zh/%E6%97%A5
ret_val['language'] = eng_lang
_l = _lang.split('-')
if len(_l) == 2:
country = _l[1]
else:
country = _l[0].upper()
if country == 'EN':
country = 'US'
ret_val['country'] = country
ret_val['locale'] = locale
ret_val['subdomain'] = eng_traits.custom['supported_domains'].get(country.upper(), 'www.google.com')
# subdomain ...
ret_val['subdomain'] = 'www.' + google_domains.get(country.upper(), 'google.com')
# params & headers
lang_country = '%s-%s' % (language, country) # (en-US, en-EN, de-DE, de-AU, fr-FR ..)
# hl parameter:
# The hl parameter specifies the interface language (host language) of
# your user interface. To improve the performance and the quality of your
# search results, you are strongly encouraged to set this parameter
# explicitly.
# https://developers.google.com/custom-search/docs/xml_results#hlsp
# The Interface Language:
# https://developers.google.com/custom-search/docs/xml_results#hlsp The
# Interface Language:
# https://developers.google.com/custom-search/docs/xml_results_appendices#interfaceLanguages
ret_val['params']['hl'] = lang_code
ret_val['params']['hl'] = lang_list.get(lang_country, language)
# lr parameter:
# The lr (language restrict) parameter restricts search results to
@ -183,72 +218,22 @@ def get_google_info(params, eng_traits):
# https://developers.google.com/custom-search/docs/xml_results#lrsp
# Language Collection Values:
# https://developers.google.com/custom-search/docs/xml_results_appendices#languageCollections
#
# To select 'all' languages an empty 'lr' value is used.
#
# Different to other google services, Google Schloar supports to select more
# than one language. The languages are seperated by a pipe '|' (logical OR).
# By example: &lr=lang_zh-TW%7Clang_de selects articles written in
# traditional chinese OR german language.
ret_val['params']['lr'] = eng_lang
if sxng_locale == 'all':
ret_val['params']['lr'] = ''
if _any_language and supported_any_language:
# cr parameter:
# The cr parameter restricts search results to documents originating in a
# particular country.
# https://developers.google.com/custom-search/docs/xml_results#crsp
# interpretation is left up to Google (based on whoogle)
#
# - add parameter ``source=lnt``
# - don't use parameter ``lr``
# - don't add a ``Accept-Language`` HTTP header.
ret_val['params']['cr'] = 'country' + country
if sxng_locale == 'all':
ret_val['params']['cr'] = ''
ret_val['params']['source'] = 'lnt'
# gl parameter: (mandatory by Geeogle News)
# The gl parameter value is a two-letter country code. For WebSearch
# results, the gl parameter boosts search results whose country of origin
# matches the parameter value. See the Country Codes section for a list of
# valid values.
# Specifying a gl parameter value in WebSearch requests should improve the
# relevance of results. This is particularly true for international
# customers and, even more specifically, for customers in English-speaking
# countries other than the United States.
# https://developers.google.com/custom-search/docs/xml_results#glsp
else:
ret_val['params']['gl'] = country
# ie parameter:
# The ie parameter sets the character encoding scheme that should be used
# to interpret the query string. The default ie value is latin1.
# https://developers.google.com/custom-search/docs/xml_results#iesp
ret_val['params']['ie'] = 'utf8'
# oe parameter:
# The oe parameter sets the character encoding scheme that should be used
# to decode the XML result. The default oe value is latin1.
# https://developers.google.com/custom-search/docs/xml_results#oesp
ret_val['params']['oe'] = 'utf8'
# num parameter:
# The num parameter identifies the number of search results to return.
# The default num value is 10, and the maximum value is 20. If you request
# more than 20 results, only 20 results will be returned.
# https://developers.google.com/custom-search/docs/xml_results#numsp
# HINT: seems to have no effect (tested in google WEB & Images)
# ret_val['params']['num'] = 20
# HTTP headers
ret_val['headers']['Accept'] = '*/*'
# Cookies
# - https://github.com/searxng/searxng/pull/1679#issuecomment-1235432746
# - https://github.com/searxng/searxng/issues/1555
ret_val['cookies']['CONSENT'] = "YES+"
# restricts search results to documents written in a particular
# language.
ret_val['params']['lr'] = "lang_" + lang_list.get(lang_country, language)
return ret_val
@ -260,34 +245,33 @@ def detect_google_sorry(resp):
def request(query, params):
"""Google search request"""
# pylint: disable=line-too-long
offset = (params['pageno'] - 1) * 10
google_info = get_google_info(params, traits)
lang_info = get_lang_info(params, supported_languages, language_aliases, True)
additional_parameters = {}
if use_mobile_ui:
additional_parameters = {
'asearch': 'arc',
'async': 'use_ac:true,_fmt:prog',
}
# https://www.google.de/search?q=corona&hl=de&lr=lang_de&start=0&tbs=qdr%3Ad&safe=medium
query_url = (
'https://'
+ google_info['subdomain']
+ lang_info['subdomain']
+ '/search'
+ "?"
+ urlencode(
{
'q': query,
**google_info['params'],
'filter': '0',
**lang_info['params'],
'ie': "utf8",
'oe': "utf8",
'start': offset,
# 'vet': '12ahUKEwik3ZbIzfn7AhXMX_EDHbUDBh0QxK8CegQIARAC..i',
# 'ved': '2ahUKEwik3ZbIzfn7AhXMX_EDHbUDBh0Q_skCegQIARAG',
# 'cs' : 1,
# 'sa': 'N',
# 'yv': 3,
# 'prmd': 'vin',
# 'ei': 'GASaY6TxOcy_xc8PtYeY6AE',
# 'sa': 'N',
# 'sstk': 'AcOHfVkD7sWCSAheZi-0tx_09XDO55gTWY0JNq3_V26cNN-c8lfD45aZYPI8s_Bqp8s57AHz5pxchDtAGCA_cikAWSjy9kw3kgg'
# formally known as use_mobile_ui
'asearch': 'arc',
'async': UI_ASYNC,
'filter': '0',
**additional_parameters,
}
)
)
@ -298,45 +282,25 @@ def request(query, params):
query_url += '&' + urlencode({'safe': filter_mapping[params['safesearch']]})
params['url'] = query_url
params['cookies'] = google_info['cookies']
params['headers'].update(google_info['headers'])
params['cookies']['CONSENT'] = "YES+"
params['headers'].update(lang_info['headers'])
if use_mobile_ui:
params['headers']['Accept'] = '*/*'
else:
params['headers']['Accept'] = 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8'
return params
# (function(){var s='data:image/jpeg;base64,/9j/4AAQ ...
# ... DX/Ff5XSpSgdU32xSlKDJ//9k\x3d';var ii=['dimg_21'];_setImagesSrc(ii,s);})();
RE_DATA_IMAGE = re.compile(r"'(data:image[^']*)'[^']*ii=\['([^']*)'\];_setImagesSrc")
def _parse_data_images(dom):
data_image_map = {}
for _script in eval_xpath_list(dom, "//script[@nonce]"):
script = _script.text
if not script:
continue
script = RE_DATA_IMAGE.search(script)
if not script:
continue
data_image_map[script.group(2)] = script.group(1).replace(r'\x3d', '=')
logger.debug('data:image objects --> %s', list(data_image_map.keys()))
return data_image_map
def response(resp):
"""Get response from google's search request"""
# pylint: disable=too-many-branches, too-many-statements
detect_google_sorry(resp)
results = []
# convert the text to dom
dom = html.fromstring(resp.text)
data_image_map = []
if '_fmt:html' in UI_ASYNC:
# in this format images are embedded by a bse64 encoded 'data:image'
data_image_map = _parse_data_images(dom)
# results --> answer
answer_list = eval_xpath(dom, '//div[contains(@class, "LGOjhe")]')
if answer_list:
@ -345,9 +309,20 @@ def response(resp):
else:
logger.debug("did not find 'answer'")
# results --> number_of_results
if not use_mobile_ui:
try:
_txt = eval_xpath_getindex(dom, '//div[@id="result-stats"]//text()', 0)
_digit = ''.join([n for n in _txt if n.isdigit()])
number_of_results = int(_digit)
results.append({'number_of_results': number_of_results})
except Exception as e: # pylint: disable=broad-except
logger.debug("did not 'number_of_results'")
logger.error(e, exc_info=True)
# parse results
for result in eval_xpath_list(dom, results_xpath): # pylint: disable=too-many-nested-blocks
for result in eval_xpath_list(dom, results_xpath):
# google *sections*
if extract_text(eval_xpath(result, g_section_with_header)):
@ -364,31 +339,13 @@ def response(resp):
url = eval_xpath_getindex(result, href_xpath, 0, None)
if url is None:
continue
content = []
img_list = []
for content_feature in eval_xpath(result, content_xpath):
val = content_feature.attrib['data-content-feature']
if val in ['1', '2']:
txt = extract_text(content_feature, allow_none=True)
if txt:
content.append(txt)
elif '0' in val:
img = content_feature.xpath('.//img/@src')
if img:
img = img[0]
if img.startswith('data:image'):
img_id = content_feature.xpath('.//img/@id')
if img_id:
img = data_image_map.get(img_id[0])
img_list.append(img)
if not content:
content = extract_text(eval_xpath_getindex(result, content_xpath, 0, default=None), allow_none=True)
if content is None:
logger.debug('ignoring item from the result_xpath list: missing content of title "%s"', title)
continue
content = ' / '.join(content)
img_src = img_list[0] if img_list else None
results.append({'url': url, 'title': title, 'content': content, 'img_src': img_src})
logger.debug('add link to results: %s', title)
results.append({'url': url, 'title': title, 'content': content})
except Exception as e: # pylint: disable=broad-except
logger.error(e, exc_info=True)
@ -404,107 +361,15 @@ def response(resp):
# get supported languages from their site
skip_countries = [
# official language of google-country not in google-languages
'AL', # Albanien (sq)
'AZ', # Aserbaidschan (az)
'BD', # Bangladesch (bn)
'BN', # Brunei Darussalam (ms)
'BT', # Bhutan (dz)
'ET', # Äthiopien (am)
'GE', # Georgien (ka, os)
'GL', # Grönland (kl)
'KH', # Kambodscha (km)
'LA', # Laos (lo)
'LK', # Sri Lanka (si, ta)
'ME', # Montenegro (sr)
'MK', # Nordmazedonien (mk, sq)
'MM', # Myanmar (my)
'MN', # Mongolei (mn)
'MV', # Malediven (dv) // dv_MV is unknown by babel
'MY', # Malaysia (ms)
'NP', # Nepal (ne)
'TJ', # Tadschikistan (tg)
'TM', # Turkmenistan (tk)
'UZ', # Usbekistan (uz)
]
def fetch_traits(engine_traits: EngineTraits, add_domains: bool = True):
"""Fetch languages from Google."""
# pylint: disable=import-outside-toplevel, too-many-branches
engine_traits.custom['supported_domains'] = {}
resp = network.get('https://www.google.com/preferences')
if not resp.ok:
raise RuntimeError("Response from Google's preferences is not OK.")
def _fetch_supported_languages(resp):
ret_val = {}
dom = html.fromstring(resp.text)
# supported language codes
radio_buttons = eval_xpath_list(dom, '//*[@id="langSec"]//input[@name="lr"]')
lang_map = {'no': 'nb'}
for x in eval_xpath_list(dom, '//*[@id="langSec"]//input[@name="lr"]'):
for x in radio_buttons:
name = x.get("data-name")
code = x.get("value").split('_')[-1]
ret_val[code] = {"name": name}
eng_lang = x.get("value").split('_')[-1]
try:
locale = babel.Locale.parse(lang_map.get(eng_lang, eng_lang), sep='-')
except babel.UnknownLocaleError:
print("ERROR: %s -> %s is unknown by babel" % (x.get("data-name"), eng_lang))
continue
sxng_lang = language_tag(locale)
conflict = engine_traits.languages.get(sxng_lang)
if conflict:
if conflict != eng_lang:
print("CONFLICT: babel %s --> %s, %s" % (sxng_lang, conflict, eng_lang))
continue
engine_traits.languages[sxng_lang] = 'lang_' + eng_lang
# alias languages
engine_traits.languages['zh'] = 'lang_zh-CN'
# supported region codes
for x in eval_xpath_list(dom, '//*[@name="region"]/..//input[@name="region"]'):
eng_country = x.get("value")
if eng_country in skip_countries:
continue
if eng_country == 'ZZ':
engine_traits.all_locale = 'ZZ'
continue
sxng_locales = get_offical_locales(eng_country, engine_traits.languages.keys(), regional=True)
if not sxng_locales:
print("ERROR: can't map from google country %s (%s) to a babel region." % (x.get('data-name'), eng_country))
continue
for sxng_locale in sxng_locales:
engine_traits.regions[region_tag(sxng_locale)] = eng_country
# alias regions
engine_traits.regions['zh-CN'] = 'HK'
# supported domains
if add_domains:
resp = network.get('https://www.google.com/supported_domains')
if not resp.ok:
raise RuntimeError("Response from https://www.google.com/supported_domains is not OK.")
for domain in resp.text.split():
domain = domain.strip()
if not domain or domain in [
'.google.com',
]:
continue
region = domain.split('.')[-1].upper()
engine_traits.custom['supported_domains'][region] = 'www' + domain
if region == 'HK':
# There is no google.cn, we use .com.hk for zh-CN
engine_traits.custom['supported_domains']['CN'] = 'www' + domain
return ret_val

View file

@ -1,38 +1,31 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
"""This is the implementation of the Google Images engine using the internal
Google API used by the Google Go Android app.
"""This is the implementation of the google images engine using the google
internal API used the Google Go Android app.
This internal API offer results in
- JSON (``_fmt:json``)
- Protobuf_ (``_fmt:pb``)
- Protobuf_ compressed? (``_fmt:pc``)
- HTML (``_fmt:html``)
- Protobuf_ encoded in JSON (``_fmt:jspb``).
- JSON (_fmt:json)
- Protobuf (_fmt:pb)
- Protobuf compressed? (_fmt:pc)
- HTML (_fmt:html)
- Protobuf encoded in JSON (_fmt:jspb).
.. _Protobuf: https://en.wikipedia.org/wiki/Protocol_Buffers
"""
from typing import TYPE_CHECKING
from urllib.parse import urlencode
from json import loads
from searx.engines.google import fetch_traits # pylint: disable=unused-import
from searx.engines.google import (
get_google_info,
get_lang_info,
time_range_dict,
detect_google_sorry,
)
if TYPE_CHECKING:
import logging
from searx.enginelib.traits import EngineTraits
logger: logging.Logger
traits: EngineTraits
# pylint: disable=unused-import
from searx.engines.google import supported_languages_url, _fetch_supported_languages
# pylint: enable=unused-import
# about
about = {
@ -47,6 +40,7 @@ about = {
# engine dependent config
categories = ['images', 'web']
paging = True
use_locale_domain = True
time_range_support = True
safesearch = True
send_accept_language_header = True
@ -57,18 +51,20 @@ filter_mapping = {0: 'images', 1: 'active', 2: 'active'}
def request(query, params):
"""Google-Image search request"""
google_info = get_google_info(params, traits)
lang_info = get_lang_info(params, supported_languages, language_aliases, False)
query_url = (
'https://'
+ google_info['subdomain']
+ lang_info['subdomain']
+ '/search'
+ "?"
+ urlencode(
{
'q': query,
'tbm': "isch",
**google_info['params'],
**lang_info['params'],
'ie': "utf8",
'oe': "utf8",
'asearch': 'isch',
'async': '_fmt:json,p:1,ijn:' + str(params['pageno']),
}
@ -81,8 +77,9 @@ def request(query, params):
query_url += '&' + urlencode({'safe': filter_mapping[params['safesearch']]})
params['url'] = query_url
params['cookies'] = google_info['cookies']
params['headers'].update(google_info['headers'])
params['headers'].update(lang_info['headers'])
params['headers']['User-Agent'] = 'NSTN/3.60.474802233.release Dalvik/2.1.0 (Linux; U; Android 12; US) gzip'
params['headers']['Accept'] = '*/*'
return params
@ -114,11 +111,7 @@ def response(resp):
copyright_notice = item["result"].get('iptc', {}).get('copyright_notice')
if copyright_notice:
result_item['source'] += ' | ' + copyright_notice
freshness_date = item["result"].get("freshness_date")
if freshness_date:
result_item['source'] += ' | ' + freshness_date
result_item['source'] += ' / ' + copyright_notice
file_size = item.get('gsa', {}).get('file_size')
if file_size:

View file

@ -1,40 +1,24 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
"""This is the implementation of the Google News engine.
"""This is the implementation of the google news engine. The google news API
ignores some parameters from the common :ref:`google API`:
Google News has a different region handling compared to Google WEB.
- the ``ceid`` argument has to be set (:py:obj:`ceid_list`)
- the hl_ argument has to be set correctly (and different to Google WEB)
- the gl_ argument is mandatory
If one of this argument is not set correctly, the request is redirected to
CONSENT dialog::
https://consent.google.com/m?continue=
The google news API ignores some parameters from the common :ref:`google API`:
- num_ : the number of search results is ignored / there is no paging all
results for a query term are in the first response.
- num_ : the number of search results is ignored
- save_ : is ignored / Google-News results are always *SafeSearch*
.. _hl: https://developers.google.com/custom-search/docs/xml_results#hlsp
.. _gl: https://developers.google.com/custom-search/docs/xml_results#glsp
.. _num: https://developers.google.com/custom-search/docs/xml_results#numsp
.. _save: https://developers.google.com/custom-search/docs/xml_results#safesp
"""
from typing import TYPE_CHECKING
# pylint: disable=invalid-name
import binascii
import re
from urllib.parse import urlencode
from base64 import b64decode
from lxml import html
import babel
from searx import locales
from searx.utils import (
eval_xpath,
eval_xpath_list,
@ -42,19 +26,18 @@ from searx.utils import (
extract_text,
)
from searx.engines.google import fetch_traits as _fetch_traits # pylint: disable=unused-import
# pylint: disable=unused-import
from searx.engines.google import (
get_google_info,
supported_languages_url,
_fetch_supported_languages,
)
# pylint: enable=unused-import
from searx.engines.google import (
get_lang_info,
detect_google_sorry,
)
from searx.enginelib.traits import EngineTraits
if TYPE_CHECKING:
import logging
logger: logging.Logger
traits: EngineTraits
# about
about = {
@ -66,77 +49,70 @@ about = {
"results": 'HTML',
}
# compared to other google engines google-news has a different time range
# support. The time range is included in the search term.
time_range_dict = {
'day': 'when:1d',
'week': 'when:7d',
'month': 'when:1m',
'year': 'when:1y',
}
# engine dependent config
categories = ['news']
paging = False
time_range_support = False
use_locale_domain = True
time_range_support = True
# Google-News results are always *SafeSearch*. Option 'safesearch' is set to
# False here, otherwise checker will report safesearch-errors::
#
# safesearch : results are identitical for safesearch=0 and safesearch=2
safesearch = True
# send_accept_language_header = True
safesearch = False
send_accept_language_header = True
def request(query, params):
"""Google-News search request"""
sxng_locale = params.get('searxng_locale', 'en-US')
ceid = locales.get_engine_locale(sxng_locale, traits.custom['ceid'], default='US:en')
google_info = get_google_info(params, traits)
google_info['subdomain'] = 'news.google.com' # google news has only one domain
lang_info = get_lang_info(params, supported_languages, language_aliases, False)
ceid_region, ceid_lang = ceid.split(':')
ceid_lang, ceid_suffix = (
ceid_lang.split('-')
+ [
None,
]
)[:2]
# google news has only one domain
lang_info['subdomain'] = 'news.google.com'
google_info['params']['hl'] = ceid_lang
ceid = "%s:%s" % (lang_info['country'], lang_info['language'])
if ceid_suffix and ceid_suffix not in ['Hans', 'Hant']:
# google news redirects en to en-US
if lang_info['params']['hl'] == 'en':
lang_info['params']['hl'] = 'en-US'
if ceid_region.lower() == ceid_lang:
google_info['params']['hl'] = ceid_lang + '-' + ceid_region
else:
google_info['params']['hl'] = ceid_lang + '-' + ceid_suffix
elif ceid_region.lower() != ceid_lang:
if ceid_region in ['AT', 'BE', 'CH', 'IL', 'SA', 'IN', 'BD', 'PT']:
google_info['params']['hl'] = ceid_lang
else:
google_info['params']['hl'] = ceid_lang + '-' + ceid_region
google_info['params']['lr'] = 'lang_' + ceid_lang.split('-')[0]
google_info['params']['gl'] = ceid_region
# Very special to google-news compared to other google engines, the time
# range is included in the search term.
if params['time_range']:
query += ' ' + time_range_dict[params['time_range']]
query_url = (
'https://'
+ google_info['subdomain']
+ "/search?"
+ urlencode(
{
'q': query,
**google_info['params'],
}
)
# ceid includes a ':' character which must not be urlencoded
+ lang_info['subdomain']
+ '/search'
+ "?"
+ urlencode({'q': query, **lang_info['params'], 'ie': "utf8", 'oe': "utf8", 'gl': lang_info['country']})
+ ('&ceid=%s' % ceid)
)
) # ceid includes a ':' character which must not be urlencoded
params['url'] = query_url
params['cookies'] = google_info['cookies']
params['headers'].update(google_info['headers'])
params['cookies']['CONSENT'] = "YES+"
params['headers'].update(lang_info['headers'])
params['headers']['Accept'] = 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8'
return params
def response(resp):
"""Get response from google's search request"""
results = []
detect_google_sorry(resp)
# convert the text to dom
@ -176,8 +152,8 @@ def response(resp):
# The pub_date is mostly a string like 'yesertday', not a real
# timezone date or time. Therefore we can't use publishedDate.
pub_date = extract_text(eval_xpath(result, './article//time'))
pub_origin = extract_text(eval_xpath(result, './article//a[@data-n-tid]'))
pub_date = extract_text(eval_xpath(result, './article/div[1]/div[1]/time'))
pub_origin = extract_text(eval_xpath(result, './article/div[1]/div[1]/a'))
content = ' / '.join([x for x in [pub_origin, pub_date] if x])
@ -198,127 +174,3 @@ def response(resp):
# return results
return results
ceid_list = [
'AE:ar',
'AR:es-419',
'AT:de',
'AU:en',
'BD:bn',
'BE:fr',
'BE:nl',
'BG:bg',
'BR:pt-419',
'BW:en',
'CA:en',
'CA:fr',
'CH:de',
'CH:fr',
'CL:es-419',
'CN:zh-Hans',
'CO:es-419',
'CU:es-419',
'CZ:cs',
'DE:de',
'EG:ar',
'ES:es',
'ET:en',
'FR:fr',
'GB:en',
'GH:en',
'GR:el',
'HK:zh-Hant',
'HU:hu',
'ID:en',
'ID:id',
'IE:en',
'IL:en',
'IL:he',
'IN:bn',
'IN:en',
'IN:hi',
'IN:ml',
'IN:mr',
'IN:ta',
'IN:te',
'IT:it',
'JP:ja',
'KE:en',
'KR:ko',
'LB:ar',
'LT:lt',
'LV:en',
'LV:lv',
'MA:fr',
'MX:es-419',
'MY:en',
'NA:en',
'NG:en',
'NL:nl',
'NO:no',
'NZ:en',
'PE:es-419',
'PH:en',
'PK:en',
'PL:pl',
'PT:pt-150',
'RO:ro',
'RS:sr',
'RU:ru',
'SA:ar',
'SE:sv',
'SG:en',
'SI:sl',
'SK:sk',
'SN:fr',
'TH:th',
'TR:tr',
'TW:zh-Hant',
'TZ:en',
'UA:ru',
'UA:uk',
'UG:en',
'US:en',
'US:es-419',
'VE:es-419',
'VN:vi',
'ZA:en',
'ZW:en',
]
"""List of region/language combinations supported by Google News. Values of the
``ceid`` argument of the Google News REST API."""
_skip_values = [
'ET:en', # english (ethiopia)
'ID:en', # english (indonesia)
'LV:en', # english (latvia)
]
_ceid_locale_map = {'NO:no': 'nb-NO'}
def fetch_traits(engine_traits: EngineTraits):
_fetch_traits(engine_traits, add_domains=False)
engine_traits.custom['ceid'] = {}
for ceid in ceid_list:
if ceid in _skip_values:
continue
region, lang = ceid.split(':')
x = lang.split('-')
if len(x) > 1:
if x[1] not in ['Hant', 'Hans']:
lang = x[0]
sxng_locale = _ceid_locale_map.get(ceid, lang + '-' + region)
try:
locale = babel.Locale.parse(sxng_locale, sep='-')
except babel.UnknownLocaleError:
print("ERROR: %s -> %s is unknown by babel" % (ceid, sxng_locale))
continue
engine_traits.custom['ceid'][locales.region_tag(locale)] = ceid

View file

@ -1,18 +1,19 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
"""This is the implementation of the Google Scholar engine.
"""Google (Scholar)
Compared to other Google services the Scholar engine has a simple GET REST-API
and there does not exists `async` API. Even though the API slightly vintage we
can make use of the :ref:`google API` to assemble the arguments of the GET
request.
For detailed description of the *REST-full* API see: `Query Parameter
Definitions`_.
.. _Query Parameter Definitions:
https://developers.google.com/custom-search/docs/xml_results#WebSearch_Query_Parameter_Definitions
"""
from typing import TYPE_CHECKING
from typing import Optional
# pylint: disable=invalid-name
from urllib.parse import urlencode
from datetime import datetime
from typing import Optional
from lxml import html
from searx.utils import (
@ -22,21 +23,19 @@ from searx.utils import (
extract_text,
)
from searx.exceptions import SearxEngineCaptchaException
from searx.engines.google import fetch_traits # pylint: disable=unused-import
from searx.engines.google import (
get_google_info,
get_lang_info,
time_range_dict,
detect_google_sorry,
)
from searx.enginelib.traits import EngineTraits
if TYPE_CHECKING:
import logging
# pylint: disable=unused-import
from searx.engines.google import (
supported_languages_url,
_fetch_supported_languages,
)
logger: logging.Logger
traits: EngineTraits
# pylint: enable=unused-import
# about
about = {
@ -52,62 +51,53 @@ about = {
categories = ['science', 'scientific publications']
paging = True
language_support = True
use_locale_domain = True
time_range_support = True
safesearch = False
send_accept_language_header = True
def time_range_args(params):
"""Returns a dictionary with a time range arguments based on
``params['time_range']``.
Google Scholar supports a detailed search by year. Searching by *last
month* or *last week* (as offered by SearXNG) is uncommon for scientific
publications and is not supported by Google Scholar.
To limit the result list when the users selects a range, all the SearXNG
ranges (*day*, *week*, *month*, *year*) are mapped to *year*. If no range
is set an empty dictionary of arguments is returned. Example; when
user selects a time range (current year minus one in 2022):
.. code:: python
{ 'as_ylo' : 2021 }
def time_range_url(params):
"""Returns a URL query component for a google-Scholar time range based on
``params['time_range']``. Google-Scholar does only support ranges in years.
To have any effect, all the Searx ranges (*day*, *week*, *month*, *year*)
are mapped to *year*. If no range is set, an empty string is returned.
Example::
&as_ylo=2019
"""
ret_val = {}
# as_ylo=2016&as_yhi=2019
ret_val = ''
if params['time_range'] in time_range_dict:
ret_val['as_ylo'] = datetime.now().year - 1
return ret_val
def detect_google_captcha(dom):
"""In case of CAPTCHA Google Scholar open its own *not a Robot* dialog and is
not redirected to ``sorry.google.com``.
"""
if eval_xpath(dom, "//form[@id='gs_captcha_f']"):
raise SearxEngineCaptchaException()
ret_val = urlencode({'as_ylo': datetime.now().year - 1})
return '&' + ret_val
def request(query, params):
"""Google-Scholar search request"""
google_info = get_google_info(params, traits)
offset = (params['pageno'] - 1) * 10
lang_info = get_lang_info(params, supported_languages, language_aliases, False)
# subdomain is: scholar.google.xy
google_info['subdomain'] = google_info['subdomain'].replace("www.", "scholar.")
lang_info['subdomain'] = lang_info['subdomain'].replace("www.", "scholar.")
args = {
'q': query,
**google_info['params'],
'start': (params['pageno'] - 1) * 10,
'as_sdt': '2007', # include patents / to disable set '0,5'
'as_vis': '0', # include citations / to disable set '1'
}
args.update(time_range_args(params))
query_url = (
'https://'
+ lang_info['subdomain']
+ '/scholar'
+ "?"
+ urlencode({'q': query, **lang_info['params'], 'ie': "utf8", 'oe': "utf8", 'start': offset})
)
params['url'] = 'https://' + google_info['subdomain'] + '/scholar?' + urlencode(args)
params['cookies'] = google_info['cookies']
params['headers'].update(google_info['headers'])
query_url += time_range_url(params)
params['url'] = query_url
params['cookies']['CONSENT'] = "YES+"
params['headers'].update(lang_info['headers'])
params['headers']['Accept'] = 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8'
# params['google_subdomain'] = subdomain
return params
@ -148,15 +138,19 @@ def parse_gs_a(text: Optional[str]):
def response(resp): # pylint: disable=too-many-locals
"""Parse response from Google Scholar"""
"""Get response from google's search request"""
results = []
detect_google_sorry(resp)
# which subdomain ?
# subdomain = resp.search_params.get('google_subdomain')
# convert the text to dom
dom = html.fromstring(resp.text)
detect_google_captcha(dom)
# parse results
for result in eval_xpath_list(dom, '//div[@data-rp]'):
for result in eval_xpath_list(dom, '//div[@data-cid]'):
title = extract_text(eval_xpath(result, './/h3[1]//a'))
@ -164,7 +158,7 @@ def response(resp): # pylint: disable=too-many-locals
# this is a [ZITATION] block
continue
pub_type = extract_text(eval_xpath(result, './/span[@class="gs_ctg2"]'))
pub_type = extract_text(eval_xpath(result, './/span[@class="gs_ct1"]'))
if pub_type:
pub_type = pub_type[1:-1].lower()

View file

@ -1,6 +1,6 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
"""This is the implementation of the Google Videos engine.
"""This is the implementation of the google videos engine.
.. admonition:: Content-Security-Policy (CSP)
@ -14,8 +14,9 @@
"""
from typing import TYPE_CHECKING
# pylint: disable=invalid-name
import re
from urllib.parse import urlencode
from lxml import html
@ -26,22 +27,20 @@ from searx.utils import (
extract_text,
)
from searx.engines.google import fetch_traits # pylint: disable=unused-import
from searx.engines.google import (
get_google_info,
get_lang_info,
time_range_dict,
filter_mapping,
g_section_with_header,
title_xpath,
suggestion_xpath,
detect_google_sorry,
)
from searx.enginelib.traits import EngineTraits
if TYPE_CHECKING:
import logging
# pylint: disable=unused-import
from searx.engines.google import supported_languages_url, _fetch_supported_languages
logger: logging.Logger
traits: EngineTraits
# pylint: enable=unused-import
# about
about = {
@ -56,32 +55,70 @@ about = {
# engine dependent config
categories = ['videos', 'web']
paging = True
paging = False
language_support = True
use_locale_domain = True
time_range_support = True
safesearch = True
send_accept_language_header = True
RE_CACHE = {}
def _re(regexpr):
"""returns compiled regular expression"""
RE_CACHE[regexpr] = RE_CACHE.get(regexpr, re.compile(regexpr))
return RE_CACHE[regexpr]
def scrap_out_thumbs_src(dom):
ret_val = {}
thumb_name = 'dimg_'
for script in eval_xpath_list(dom, '//script[contains(., "google.ldi={")]'):
_script = script.text
# "dimg_35":"https://i.ytimg.c....",
_dimurl = _re("s='([^']*)").findall(_script)
for k, v in _re('(' + thumb_name + '[0-9]*)":"(http[^"]*)').findall(_script):
v = v.replace(r'\u003d', '=')
v = v.replace(r'\u0026', '&')
ret_val[k] = v
logger.debug("found %s imgdata for: %s", thumb_name, ret_val.keys())
return ret_val
def scrap_out_thumbs(dom):
"""Scrap out thumbnail data from <script> tags."""
ret_val = {}
thumb_name = 'dimg_'
for script in eval_xpath_list(dom, '//script[contains(., "_setImagesSrc")]'):
_script = script.text
# var s='data:image/jpeg;base64, ...'
_imgdata = _re("s='([^']*)").findall(_script)
if not _imgdata:
continue
# var ii=['dimg_17']
for _vidthumb in _re(r"(%s\d+)" % thumb_name).findall(_script):
# At least the equal sign in the URL needs to be decoded
ret_val[_vidthumb] = _imgdata[0].replace(r"\x3d", "=")
logger.debug("found %s imgdata for: %s", thumb_name, ret_val.keys())
return ret_val
def request(query, params):
"""Google-Video search request"""
google_info = get_google_info(params, traits)
lang_info = get_lang_info(params, supported_languages, language_aliases, False)
query_url = (
'https://'
+ google_info['subdomain']
+ lang_info['subdomain']
+ '/search'
+ "?"
+ urlencode(
{
'q': query,
'tbm': "vid",
'start': 10 * params['pageno'],
**google_info['params'],
'asearch': 'arc',
'async': 'use_ac:true,_fmt:html',
}
)
+ urlencode({'q': query, 'tbm': "vid", **lang_info['params'], 'ie': "utf8", 'oe': "utf8"})
)
if params['time_range'] in time_range_dict:
@ -90,8 +127,9 @@ def request(query, params):
query_url += '&' + urlencode({'safe': filter_mapping[params['safesearch']]})
params['url'] = query_url
params['cookies'] = google_info['cookies']
params['headers'].update(google_info['headers'])
params['cookies']['CONSENT'] = "YES+"
params['headers'].update(lang_info['headers'])
params['headers']['Accept'] = 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8'
return params
@ -103,30 +141,43 @@ def response(resp):
# convert the text to dom
dom = html.fromstring(resp.text)
vidthumb_imgdata = scrap_out_thumbs(dom)
thumbs_src = scrap_out_thumbs_src(dom)
logger.debug(str(thumbs_src))
# parse results
for result in eval_xpath_list(dom, '//div[contains(@class, "g ")]'):
img_src = eval_xpath_getindex(result, './/img/@src', 0, None)
if img_src is None:
# ignore google *sections*
if extract_text(eval_xpath(result, g_section_with_header)):
logger.debug("ignoring <g-section-with-header>")
continue
title = extract_text(eval_xpath_getindex(result, './/a/h3[1]', 0))
url = eval_xpath_getindex(result, './/a/h3[1]/../@href', 0)
# ingnore articles without an image id / e.g. news articles
img_id = eval_xpath_getindex(result, './/g-img/img/@id', 0, default=None)
if img_id is None:
logger.error("no img_id found in item %s (news article?)", len(results) + 1)
continue
img_src = vidthumb_imgdata.get(img_id, None)
if not img_src:
img_src = thumbs_src.get(img_id, "")
title = extract_text(eval_xpath_getindex(result, title_xpath, 0))
url = eval_xpath_getindex(result, './/div[@class="dXiKIc"]//a/@href', 0)
length = extract_text(eval_xpath(result, './/div[contains(@class, "P7xzyf")]/span/span'))
c_node = eval_xpath_getindex(result, './/div[@class="Uroaid"]', 0)
content = extract_text(c_node)
pub_info = extract_text(eval_xpath(result, './/div[@class="P7xzyf"]'))
length = extract_text(eval_xpath(result, './/div[@class="J1mWY"]'))
pub_info = extract_text(eval_xpath(result, './/div[@class="Zg1NU"]'))
results.append(
{
'url': url,
'title': title,
'content': content,
'length': length,
'author': pub_info,
'thumbnail': img_src,
'length': length,
'template': 'videos.html',
}
)