1
0
mirror of https://github.com/searxng/searxng.git synced 2024-11-14 00:30:15 +01:00
searxng/searx/engines/seznam.py
Markus Heiser 8205f170ff [mod] pylint all engines without PYLINT_SEARXNG_DISABLE_OPTION
Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>
2024-03-11 14:55:38 +01:00

74 lines
1.9 KiB
Python

# SPDX-License-Identifier: AGPL-3.0-or-later
"""Seznam
"""
from urllib.parse import urlencode
from lxml import html
from searx.network import get
from searx.exceptions import SearxEngineAccessDeniedException
from searx.utils import (
extract_text,
eval_xpath_list,
eval_xpath_getindex,
)
# about
about = {
"website": "https://www.seznam.cz/",
"wikidata_id": "Q3490485",
"official_api_documentation": "https://api.sklik.cz/",
"use_official_api": False,
"require_api_key": False,
"results": "HTML",
"language": "cz",
}
categories = ['general', 'web']
base_url = 'https://search.seznam.cz/'
def request(query, params):
response_index = get(base_url, headers=params['headers'], raise_for_httperror=True)
dom = html.fromstring(response_index.text)
url_params = {
'q': query,
'oq': query,
}
for e in eval_xpath_list(dom, '//input[@type="hidden"]'):
name = e.get('name')
value = e.get('value')
url_params[name] = value
params['url'] = base_url + '?' + urlencode(url_params)
params['cookies'] = response_index.cookies
return params
def response(resp):
if resp.url.path.startswith('/verify'):
raise SearxEngineAccessDeniedException()
results = []
dom = html.fromstring(resp.content.decode())
for result_element in eval_xpath_list(
dom, '//div[@id="searchpage-root"]//div[@class="Layout--left"]/div[@class="f2c528"]'
):
result_data = eval_xpath_getindex(
result_element, './/div[@class="c8774a" or @class="e69e8d a11657"]', 0, default=None
)
if result_data is None:
continue
title_element = eval_xpath_getindex(result_element, './/h3/a', 0)
results.append(
{
'url': title_element.get('href'),
'title': extract_text(title_element),
'content': extract_text(result_data),
}
)
return results