Ponysearch/searx/search/processors/online.py

211 lines
7.5 KiB
Python
Raw Normal View History

# SPDX-License-Identifier: AGPL-3.0-or-later
# lint: pylint
"""Processores for engine-type: ``online``
"""
from time import time
2021-03-18 19:59:01 +01:00
import asyncio
import httpx
import searx.network
from searx import logger
from searx.utils import gen_useragent
from searx.exceptions import (
SearxEngineAccessDeniedException,
SearxEngineCaptchaException,
SearxEngineTooManyRequestsException,
)
2021-04-14 17:23:15 +02:00
from searx.metrics.error_recorder import count_error
from .abstract import EngineProcessor
logger = logger.getChild('searx.search.processor.online')
2020-12-17 16:49:48 +01:00
def default_request_params():
"""Default request parameters for ``online`` engines."""
2020-12-17 16:49:48 +01:00
return {
'method': 'GET',
'headers': {},
'data': {},
'url': '',
'cookies': {},
'verify': True,
'auth': None
}
class OnlineProcessor(EngineProcessor):
"""Processor class for ``online`` engines."""
engine_type = 'online'
def get_params(self, search_query, engine_category):
params = super().get_params(search_query, engine_category)
if params is None:
return None
# add default params
2020-12-17 16:49:48 +01:00
params.update(default_request_params())
# add an user agent
params['headers']['User-Agent'] = gen_useragent()
return params
def _send_http_request(self, params):
# create dictionary which contain all
# informations about the request
request_args = dict(
headers=params['headers'],
cookies=params['cookies'],
verify=params['verify'],
auth=params['auth']
)
# max_redirects
max_redirects = params.get('max_redirects')
if max_redirects:
request_args['max_redirects'] = max_redirects
# allow_redirects
if 'allow_redirects' in params:
request_args['allow_redirects'] = params['allow_redirects']
# soft_max_redirects
soft_max_redirects = params.get('soft_max_redirects', max_redirects or 0)
# raise_for_status
request_args['raise_for_httperror'] = params.get('raise_for_httperror', True)
# specific type of request (GET or POST)
if params['method'] == 'GET':
req = searx.network.get
else:
req = searx.network.post
request_args['data'] = params['data']
# send the request
response = req(params['url'], **request_args)
# check soft limit of the redirect count
if len(response.history) > soft_max_redirects:
# unexpected redirect : record an error
# but the engine might still return valid results.
status_code = str(response.status_code or '')
2021-03-18 19:59:01 +01:00
reason = response.reason_phrase or ''
hostname = response.url.host
2021-04-14 17:23:15 +02:00
count_error(self.engine_name,
'{} redirects, maximum: {}'.format(len(response.history), soft_max_redirects),
(status_code, reason, hostname),
secondary=True)
return response
def _search_basic(self, query, params):
# update request parameters dependent on
# search-engine (contained in engines folder)
self.engine.request(query, params)
# ignoring empty urls
if params['url'] is None:
return None
if not params['url']:
return None
# send request
response = self._send_http_request(params)
# parse the response
response.search_params = params
return self.engine.response(response)
def search(self, query, params, result_container, start_time, timeout_limit):
# set timeout for all HTTP requests
searx.network.set_timeout_for_thread(timeout_limit, start_time=start_time)
# reset the HTTP total time
searx.network.reset_time_for_thread()
# set the network
searx.network.set_context_network_name(self.engine_name)
try:
# send requests and parse the results
search_results = self._search_basic(query, params)
self.extend_container(result_container, start_time, search_results)
except (httpx.TimeoutException, asyncio.TimeoutError) as e:
# requests timeout (connect or read)
self.handle_exception(result_container, e, suspend=True)
logger.error("engine {0} : HTTP requests timeout"
"(search duration : {1} s, timeout: {2} s) : {3}"
.format(self.engine_name, time() - start_time,
timeout_limit,
e.__class__.__name__))
except (httpx.HTTPError, httpx.StreamError) as e:
# other requests exception
self.handle_exception(result_container, e, suspend=True)
logger.exception("engine {0} : requests exception"
"(search duration : {1} s, timeout: {2} s) : {3}"
.format(self.engine_name, time() - start_time,
timeout_limit,
e))
except SearxEngineCaptchaException as e:
self.handle_exception(result_container, e, suspend=True)
logger.exception('engine {0} : CAPTCHA'.format(self.engine_name))
except SearxEngineTooManyRequestsException as e:
self.handle_exception(result_container, e, suspend=True)
logger.exception('engine {0} : Too many requests'.format(self.engine_name))
except SearxEngineAccessDeniedException as e:
self.handle_exception(result_container, e, suspend=True)
logger.exception('engine {0} : Searx is blocked'.format(self.engine_name))
except Exception as e: # pylint: disable=broad-except
self.handle_exception(result_container, e)
logger.exception('engine {0} : exception : {1}'.format(self.engine_name, e))
2020-12-24 09:28:16 +01:00
def get_default_tests(self):
tests = {}
tests['simple'] = {
'matrix': {'query': ('life', 'computer')},
2020-12-24 09:28:16 +01:00
'result_container': ['not_empty'],
}
if getattr(self.engine, 'paging', False):
tests['paging'] = {
'matrix': {'query': 'time',
'pageno': (1, 2, 3)},
'result_container': ['not_empty'],
'test': ['unique_results']
}
if 'general' in self.engine.categories:
# avoid documentation about HTML tags (<time> and <input type="time">)
tests['paging']['matrix']['query'] = 'news'
2020-12-24 09:28:16 +01:00
if getattr(self.engine, 'time_range', False):
tests['time_range'] = {
'matrix': {'query': 'news',
2020-12-24 09:28:16 +01:00
'time_range': (None, 'day')},
'result_container': ['not_empty'],
'test': ['unique_results']
}
if getattr(self.engine, 'supported_languages', []):
2020-12-24 09:28:16 +01:00
tests['lang_fr'] = {
'matrix': {'query': 'paris', 'lang': 'fr'},
'result_container': ['not_empty', ('has_language', 'fr')],
2020-12-24 09:28:16 +01:00
}
tests['lang_en'] = {
'matrix': {'query': 'paris', 'lang': 'en'},
'result_container': ['not_empty', ('has_language', 'en')],
2020-12-24 09:28:16 +01:00
}
if getattr(self.engine, 'safesearch', False):
tests['safesearch'] = {
'matrix': {'query': 'porn',
'safesearch': (0, 2)},
'test': ['unique_results']
}
return tests