Merge pull request #2225 from dalf/processors

Processors
This commit is contained in:
Alexandre Flament 2020-12-17 11:49:43 +01:00 committed by GitHub
commit 9b27935f71
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
20 changed files with 786 additions and 667 deletions

View file

@ -33,7 +33,7 @@ Engine .. Paging support **P**
------------------------- -------------------- ------------
Shortcut **S** Language support **L**
Timeout **TO** Time range support **TR**
Disabled **D** Offline **O**
Disabled **D** Engine type **ET**
------------- ----------- -------------------- ------------
Safe search **SS**
------------- ----------- ---------------------------------
@ -62,10 +62,10 @@ Show errors **DE**
- SS
- D
- TR
- O
- W
- D
- DE
- ET
- W
- D
- DE
{% for name, mod in engines.items() %}
@ -79,7 +79,7 @@ Show errors **DE**
- {{(mod.safesearch and "y") or ""}}
- {{(mod.disabled and "y") or ""}}
- {{(mod.time_range_support and "y") or ""}}
- {{(mod.offline and "y") or ""}}
- {{mod.engine_type or ""}}
- {{mod.weight or 1 }}
- {{(mod.disabled and "y") or ""}}
- {{(mod.display_error_messages and "y") or ""}}

View file

@ -37,15 +37,16 @@ settings. However, the standard way is the following:
engine file
-----------
======================= =========== ===========================================
======================= =========== ========================================================
argument type information
======================= =========== ===========================================
======================= =========== ========================================================
categories list pages, in which the engine is working
paging boolean support multible pages
language_support boolean support language choosing
time_range_support boolean support search time range
offline boolean engine runs offline
======================= =========== ===========================================
engine_type str ``online`` by default, other possibles values are
``offline``, ``online_dictionnary``, ``online_currency``
======================= =========== ========================================================
.. _engine settings:
@ -111,22 +112,49 @@ passed arguments
These arguments can be used to construct the search query. Furthermore,
parameters with default value can be redefined for special purposes.
If the ``engine_type`` is ``online```:
====================== ============== ========================================================================
argument type default-value, information
====================== ============== ========================================================================
url str ``''``
method str ``'GET'``
headers set ``{}``
data set ``{}``
cookies set ``{}``
verify bool ``True``
headers.User-Agent str a random User-Agent
category str current category, like ``'general'``
safesearch int ``0``, between ``0`` and ``2`` (normal, moderate, strict)
time_range Optional[str] ``None``, can be ``day``, ``week``, ``month``, ``year``
pageno int current pagenumber
language str specific language code like ``'en_US'``, or ``'all'`` if unspecified
====================== ============== ========================================================================
If the ``engine_type`` is ``online_dictionnary```, in addition to the ``online`` arguments:
====================== ============ ========================================================================
argument type default-value, information
====================== ============ ========================================================================
url string ``''``
method string ``'GET'``
headers set ``{}``
data set ``{}``
cookies set ``{}``
verify boolean ``True``
headers.User-Agent string a random User-Agent
category string current category, like ``'general'``
started datetime current date-time
pageno int current pagenumber
language string specific language code like ``'en_US'``, or ``'all'`` if unspecified
from_lang str specific language code like ``'en_US'``
to_lang str specific language code like ``'en_US'``
query str the text query without the languages
====================== ============ ========================================================================
If the ``engine_type`` is ``online_currency```, in addition to the ``online`` arguments:
====================== ============ ========================================================================
argument type default-value, information
====================== ============ ========================================================================
amount float the amount to convert
from str ISO 4217 code
to str ISO 4217 code
from_name str currency name
to_name str currency name
====================== ============ ========================================================================
parsed arguments
----------------
@ -137,12 +165,12 @@ request:
=================== =========== ==========================================================================
argument type information
=================== =========== ==========================================================================
url string requested url
method string HTTP request method
url str requested url
method str HTTP request method
headers set HTTP header information
data set HTTP data information (parsed if ``method != 'GET'``)
data set HTTP data information
cookies set HTTP cookies
verify boolean Performing SSL-Validity check
verify bool Performing SSL-Validity check
max_redirects int maximum redirects, hard limit
soft_max_redirects int maximum redirects, soft limit. Record an error but don't stop the engine
raise_for_httperror bool True by default: raise an exception if the HTTP code of response is >= 300

View file

@ -53,7 +53,7 @@ engine_default_args = {'paging': False,
'suspend_end_time': 0,
'continuous_errors': 0,
'time_range_support': False,
'offline': False,
'engine_type': 'online',
'display_error_messages': True,
'tokens': []}
@ -142,7 +142,9 @@ def load_engine(engine_data):
'errors': 0
}
if not engine.offline:
engine_type = getattr(engine, 'engine_type', 'online')
if engine_type != 'offline':
engine.stats['page_load_time'] = 0
engine.stats['page_load_count'] = 0
@ -209,7 +211,7 @@ def get_engines_stats(preferences):
else:
score = score_per_result = 0.0
if not engine.offline:
if engine.engine_type != 'offline':
load_times = 0
if engine.stats['page_load_count'] != 0:
load_times = engine.stats['page_load_time'] / float(engine.stats['page_load_count']) # noqa
@ -300,7 +302,7 @@ def initialize_engines(engine_list):
def _set_https_support_for_engine(engine):
# check HTTPS support if it is not disabled
if not engine.offline and not hasattr(engine, 'https_support'):
if engine.engine_type != 'offline' and not hasattr(engine, 'https_support'):
params = engine.request('http_test', {
'method': 'GET',
'headers': {},

View file

@ -23,7 +23,7 @@ from threading import Thread
from searx import logger
offline = True
engine_type = 'offline'
paging = True
command = []
delimiter = {}

View file

@ -1,54 +1,16 @@
import json
import re
import unicodedata
from searx.data import CURRENCIES # NOQA
engine_type = 'online_currency'
categories = []
url = 'https://duckduckgo.com/js/spice/currency/1/{0}/{1}'
weight = 100
parser_re = re.compile('.*?(\\d+(?:\\.\\d+)?) ([^.0-9]+) (?:in|to) ([^.0-9]+)', re.I)
https_support = True
def normalize_name(name):
name = name.lower().replace('-', ' ').rstrip('s')
name = re.sub(' +', ' ', name)
return unicodedata.normalize('NFKD', name).lower()
def name_to_iso4217(name):
global CURRENCIES
name = normalize_name(name)
currency = CURRENCIES['names'].get(name, [name])
return currency[0]
def iso4217_to_name(iso4217, language):
global CURRENCIES
return CURRENCIES['iso4217'].get(iso4217, {}).get(language, iso4217)
def request(query, params):
m = parser_re.match(query)
if not m:
# wrong query
return params
amount, from_currency, to_currency = m.groups()
amount = float(amount)
from_currency = name_to_iso4217(from_currency.strip())
to_currency = name_to_iso4217(to_currency.strip())
params['url'] = url.format(from_currency, to_currency)
params['amount'] = amount
params['from'] = from_currency
params['to'] = to_currency
params['from_name'] = iso4217_to_name(from_currency, 'en')
params['to_name'] = iso4217_to_name(to_currency, 'en')
params['url'] = url.format(params['from'], params['to'])
return params

View file

@ -9,36 +9,24 @@
@parse url, title, content
"""
import re
from urllib.parse import urljoin
from lxml import html
from searx.utils import is_valid_lang, eval_xpath
from searx.utils import eval_xpath
engine_type = 'online_dictionnary'
categories = ['general']
url = 'https://dictzone.com/{from_lang}-{to_lang}-dictionary/{query}'
weight = 100
parser_re = re.compile('.*?([a-z]+)-([a-z]+) ([^ ]+)$', re.I)
results_xpath = './/table[@id="r"]/tr'
https_support = True
def request(query, params):
m = parser_re.match(query)
if not m:
return params
from_lang, to_lang, query = m.groups()
from_lang = is_valid_lang(from_lang)
to_lang = is_valid_lang(to_lang)
if not from_lang or not to_lang:
return params
params['url'] = url.format(from_lang=from_lang[2],
to_lang=to_lang[2],
query=query)
params['url'] = url.format(from_lang=params['from_lang'][2],
to_lang=params['to_lang'][2],
query=params['query'])
return params

View file

@ -8,44 +8,26 @@
@stable yes
@parse url, title, content
"""
import re
from searx.utils import is_valid_lang
engine_type = 'online_dictionnary'
categories = ['general']
url = 'https://api.mymemory.translated.net/get?q={query}&langpair={from_lang}|{to_lang}{key}'
web_url = 'https://mymemory.translated.net/en/{from_lang}/{to_lang}/{query}'
weight = 100
https_support = True
parser_re = re.compile('.*?([a-z]+)-([a-z]+) (.{2,})$', re.I)
api_key = ''
def request(query, params):
m = parser_re.match(query)
if not m:
return params
from_lang, to_lang, query = m.groups()
from_lang = is_valid_lang(from_lang)
to_lang = is_valid_lang(to_lang)
if not from_lang or not to_lang:
return params
if api_key:
key_form = '&key=' + api_key
else:
key_form = ''
params['url'] = url.format(from_lang=from_lang[1],
to_lang=to_lang[1],
query=query,
params['url'] = url.format(from_lang=params['from_lang'][1],
to_lang=params['to_lang'][1],
query=params['query'],
key=key_form)
params['query'] = query
params['from_lang'] = from_lang
params['to_lang'] = to_lang
return params

View file

@ -1,541 +0,0 @@
'''
searx is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
searx is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with searx. If not, see < http://www.gnu.org/licenses/ >.
(C) 2013- by Adam Tauber, <asciimoo@gmail.com>
'''
import typing
import gc
import threading
from time import time
from uuid import uuid4
from urllib.parse import urlparse
from _thread import start_new_thread
import requests.exceptions
import searx.poolrequests as requests_lib
from searx.engines import engines, settings
from searx.answerers import ask
from searx.external_bang import get_bang_url
from searx.utils import gen_useragent
from searx.results import ResultContainer
from searx import logger
from searx.plugins import plugins
from searx.exceptions import (SearxEngineAccessDeniedException, SearxEngineCaptchaException,
SearxEngineTooManyRequestsException,)
from searx.metrology.error_recorder import record_exception, record_error
logger = logger.getChild('search')
max_request_timeout = settings.get('outgoing', {}).get('max_request_timeout' or None)
if max_request_timeout is None:
logger.info('max_request_timeout={0}'.format(max_request_timeout))
else:
if isinstance(max_request_timeout, float):
logger.info('max_request_timeout={0} second(s)'.format(max_request_timeout))
else:
logger.critical('outgoing.max_request_timeout if defined has to be float')
import sys
sys.exit(1)
class EngineRef:
__slots__ = 'name', 'category', 'from_bang'
def __init__(self, name: str, category: str, from_bang: bool=False):
self.name = name
self.category = category
self.from_bang = from_bang
def __repr__(self):
return "EngineRef({!r}, {!r}, {!r})".format(self.name, self.category, self.from_bang)
def __eq__(self, other):
return self.name == other.name and self.category == other.category and self.from_bang == other.from_bang
class SearchQuery:
"""container for all the search parameters (query, language, etc...)"""
__slots__ = 'query', 'engineref_list', 'categories', 'lang', 'safesearch', 'pageno', 'time_range',\
'timeout_limit', 'external_bang'
def __init__(self,
query: str,
engineref_list: typing.List[EngineRef],
categories: typing.List[str],
lang: str,
safesearch: int,
pageno: int,
time_range: typing.Optional[str],
timeout_limit: typing.Optional[float]=None,
external_bang: typing.Optional[str]=None):
self.query = query
self.engineref_list = engineref_list
self.categories = categories
self.lang = lang
self.safesearch = safesearch
self.pageno = pageno
self.time_range = time_range
self.timeout_limit = timeout_limit
self.external_bang = external_bang
def __repr__(self):
return "SearchQuery({!r}, {!r}, {!r}, {!r}, {!r}, {!r}, {!r}, {!r}, {!r})".\
format(self.query, self.engineref_list, self.categories, self.lang, self.safesearch,
self.pageno, self.time_range, self.timeout_limit, self.external_bang)
def __eq__(self, other):
return self.query == other.query\
and self.engineref_list == other.engineref_list\
and self.categories == self.categories\
and self.lang == other.lang\
and self.safesearch == other.safesearch\
and self.pageno == other.pageno\
and self.time_range == other.time_range\
and self.timeout_limit == other.timeout_limit\
and self.external_bang == other.external_bang
def send_http_request(engine, request_params):
# create dictionary which contain all
# informations about the request
request_args = dict(
headers=request_params['headers'],
cookies=request_params['cookies'],
verify=request_params['verify'],
auth=request_params['auth']
)
# setting engine based proxies
if hasattr(engine, 'proxies'):
request_args['proxies'] = requests_lib.get_proxies(engine.proxies)
# max_redirects
max_redirects = request_params.get('max_redirects')
if max_redirects:
request_args['max_redirects'] = max_redirects
# soft_max_redirects
soft_max_redirects = request_params.get('soft_max_redirects', max_redirects or 0)
# raise_for_status
request_args['raise_for_httperror'] = request_params.get('raise_for_httperror', False)
# specific type of request (GET or POST)
if request_params['method'] == 'GET':
req = requests_lib.get
else:
req = requests_lib.post
request_args['data'] = request_params['data']
# send the request
response = req(request_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 '')
reason = response.reason or ''
hostname = str(urlparse(response.url or '').netloc)
record_error(engine.name,
'{} redirects, maximum: {}'.format(len(response.history), soft_max_redirects),
(status_code, reason, hostname))
return response
def search_one_http_request(engine, query, request_params):
# update request parameters dependent on
# search-engine (contained in engines folder)
engine.request(query, request_params)
# ignoring empty urls
if request_params['url'] is None:
return None
if not request_params['url']:
return None
# send request
response = send_http_request(engine, request_params)
# parse the response
response.search_params = request_params
return engine.response(response)
def search_one_http_request_safe(engine_name, query, request_params, result_container, start_time, timeout_limit):
# set timeout for all HTTP requests
requests_lib.set_timeout_for_thread(timeout_limit, start_time=start_time)
# reset the HTTP total time
requests_lib.reset_time_for_thread()
#
engine = engines[engine_name]
# suppose everything will be alright
requests_exception = False
suspended_time = None
try:
# send requests and parse the results
search_results = search_one_http_request(engine, query, request_params)
# check if the engine accepted the request
if search_results is not None:
# yes, so add results
result_container.extend(engine_name, search_results)
# update engine time when there is no exception
engine_time = time() - start_time
page_load_time = requests_lib.get_time_for_thread()
result_container.add_timing(engine_name, engine_time, page_load_time)
with threading.RLock():
engine.stats['engine_time'] += engine_time
engine.stats['engine_time_count'] += 1
# update stats with the total HTTP time
engine.stats['page_load_time'] += page_load_time
engine.stats['page_load_count'] += 1
except Exception as e:
record_exception(engine_name, e)
# Timing
engine_time = time() - start_time
page_load_time = requests_lib.get_time_for_thread()
result_container.add_timing(engine_name, engine_time, page_load_time)
# Record the errors
with threading.RLock():
engine.stats['errors'] += 1
if (issubclass(e.__class__, requests.exceptions.Timeout)):
result_container.add_unresponsive_engine(engine_name, 'HTTP timeout')
# requests timeout (connect or read)
logger.error("engine {0} : HTTP requests timeout"
"(search duration : {1} s, timeout: {2} s) : {3}"
.format(engine_name, engine_time, timeout_limit, e.__class__.__name__))
requests_exception = True
elif (issubclass(e.__class__, requests.exceptions.RequestException)):
result_container.add_unresponsive_engine(engine_name, 'HTTP error')
# other requests exception
logger.exception("engine {0} : requests exception"
"(search duration : {1} s, timeout: {2} s) : {3}"
.format(engine_name, engine_time, timeout_limit, e))
requests_exception = True
elif (issubclass(e.__class__, SearxEngineCaptchaException)):
result_container.add_unresponsive_engine(engine_name, 'CAPTCHA required')
logger.exception('engine {0} : CAPTCHA')
suspended_time = e.suspended_time # pylint: disable=no-member
elif (issubclass(e.__class__, SearxEngineTooManyRequestsException)):
result_container.add_unresponsive_engine(engine_name, 'too many requests')
logger.exception('engine {0} : Too many requests')
suspended_time = e.suspended_time # pylint: disable=no-member
elif (issubclass(e.__class__, SearxEngineAccessDeniedException)):
result_container.add_unresponsive_engine(engine_name, 'blocked')
logger.exception('engine {0} : Searx is blocked')
suspended_time = e.suspended_time # pylint: disable=no-member
else:
result_container.add_unresponsive_engine(engine_name, 'unexpected crash')
# others errors
logger.exception('engine {0} : exception : {1}'.format(engine_name, e))
else:
if getattr(threading.current_thread(), '_timeout', False):
record_error(engine_name, 'Timeout')
# suspend the engine if there is an HTTP error
# or suspended_time is defined
with threading.RLock():
if requests_exception or suspended_time:
# update continuous_errors / suspend_end_time
engine.continuous_errors += 1
if suspended_time is None:
suspended_time = min(settings['search']['max_ban_time_on_fail'],
engine.continuous_errors * settings['search']['ban_time_on_fail'])
engine.suspend_end_time = time() + suspended_time
else:
# reset the suspend variables
engine.continuous_errors = 0
engine.suspend_end_time = 0
def record_offline_engine_stats_on_error(engine, result_container, start_time):
engine_time = time() - start_time
result_container.add_timing(engine.name, engine_time, engine_time)
with threading.RLock():
engine.stats['errors'] += 1
def search_one_offline_request(engine, query, request_params):
return engine.search(query, request_params)
def search_one_offline_request_safe(engine_name, query, request_params, result_container, start_time, timeout_limit):
engine = engines[engine_name]
try:
search_results = search_one_offline_request(engine, query, request_params)
if search_results:
result_container.extend(engine_name, search_results)
engine_time = time() - start_time
result_container.add_timing(engine_name, engine_time, engine_time)
with threading.RLock():
engine.stats['engine_time'] += engine_time
engine.stats['engine_time_count'] += 1
except ValueError as e:
record_exception(engine_name, e)
record_offline_engine_stats_on_error(engine, result_container, start_time)
logger.exception('engine {0} : invalid input : {1}'.format(engine_name, e))
except Exception as e:
record_exception(engine_name, e)
record_offline_engine_stats_on_error(engine, result_container, start_time)
result_container.add_unresponsive_engine(engine_name, 'unexpected crash', str(e))
logger.exception('engine {0} : exception : {1}'.format(engine_name, e))
else:
if getattr(threading.current_thread(), '_timeout', False):
record_error(engine_name, 'Timeout')
def search_one_request_safe(engine_name, query, request_params, result_container, start_time, timeout_limit):
if engines[engine_name].offline:
return search_one_offline_request_safe(engine_name, query, request_params, result_container, start_time, timeout_limit) # noqa
return search_one_http_request_safe(engine_name, query, request_params, result_container, start_time, timeout_limit)
def search_multiple_requests(requests, result_container, start_time, timeout_limit):
search_id = uuid4().__str__()
for engine_name, query, request_params in requests:
th = threading.Thread(
target=search_one_request_safe,
args=(engine_name, query, request_params, result_container, start_time, timeout_limit),
name=search_id,
)
th._timeout = False
th._engine_name = engine_name
th.start()
for th in threading.enumerate():
if th.name == search_id:
remaining_time = max(0.0, timeout_limit - (time() - start_time))
th.join(remaining_time)
if th.is_alive():
th._timeout = True
result_container.add_unresponsive_engine(th._engine_name, 'timeout')
logger.warning('engine timeout: {0}'.format(th._engine_name))
# get default reqest parameter
def default_request_params():
return {
'method': 'GET',
'headers': {},
'data': {},
'url': '',
'cookies': {},
'verify': True,
'auth': None,
'raise_for_httperror': True
}
class Search:
"""Search information container"""
__slots__ = "search_query", "result_container", "start_time", "actual_timeout"
def __init__(self, search_query):
# init vars
super().__init__()
self.search_query = search_query
self.result_container = ResultContainer()
self.start_time = None
self.actual_timeout = None
def search_external_bang(self):
"""
Check if there is a external bang.
If yes, update self.result_container and return True
"""
if self.search_query.external_bang:
self.result_container.redirect_url = get_bang_url(self.search_query)
# This means there was a valid bang and the
# rest of the search does not need to be continued
if isinstance(self.result_container.redirect_url, str):
return True
return False
def search_answerers(self):
"""
Check if an answer return a result.
If yes, update self.result_container and return True
"""
answerers_results = ask(self.search_query)
if answerers_results:
for results in answerers_results:
self.result_container.extend('answer', results)
return True
return False
def _is_accepted(self, engine_name, engine):
# skip suspended engines
if engine.suspend_end_time >= time():
logger.debug('Engine currently suspended: %s', engine_name)
return False
# if paging is not supported, skip
if self.search_query.pageno > 1 and not engine.paging:
return False
# if time_range is not supported, skip
if self.search_query.time_range and not engine.time_range_support:
return False
return True
def _get_params(self, engineref, user_agent):
if engineref.name not in engines:
return None, None
engine = engines[engineref.name]
if not self._is_accepted(engineref.name, engine):
return None, None
# set default request parameters
request_params = {}
if not engine.offline:
request_params = default_request_params()
request_params['headers']['User-Agent'] = user_agent
if hasattr(engine, 'language') and engine.language:
request_params['language'] = engine.language
else:
request_params['language'] = self.search_query.lang
request_params['safesearch'] = self.search_query.safesearch
request_params['time_range'] = self.search_query.time_range
request_params['category'] = engineref.category
request_params['pageno'] = self.search_query.pageno
with threading.RLock():
engine.stats['sent_search_count'] += 1
return request_params, engine.timeout
# do search-request
def _get_requests(self):
# init vars
requests = []
# set default useragent
# user_agent = request.headers.get('User-Agent', '')
user_agent = gen_useragent()
# max of all selected engine timeout
default_timeout = 0
# start search-reqest for all selected engines
for engineref in self.search_query.engineref_list:
# set default request parameters
request_params, engine_timeout = self._get_params(engineref, user_agent)
if request_params is None:
continue
# append request to list
requests.append((engineref.name, self.search_query.query, request_params))
# update default_timeout
default_timeout = max(default_timeout, engine_timeout)
# adjust timeout
actual_timeout = default_timeout
query_timeout = self.search_query.timeout_limit
if max_request_timeout is None and query_timeout is None:
# No max, no user query: default_timeout
pass
elif max_request_timeout is None and query_timeout is not None:
# No max, but user query: From user query except if above default
actual_timeout = min(default_timeout, query_timeout)
elif max_request_timeout is not None and query_timeout is None:
# Max, no user query: Default except if above max
actual_timeout = min(default_timeout, max_request_timeout)
elif max_request_timeout is not None and query_timeout is not None:
# Max & user query: From user query except if above max
actual_timeout = min(query_timeout, max_request_timeout)
logger.debug("actual_timeout={0} (default_timeout={1}, ?timeout_limit={2}, max_request_timeout={3})"
.format(actual_timeout, default_timeout, query_timeout, max_request_timeout))
return requests, actual_timeout
def search_standard(self):
"""
Update self.result_container, self.actual_timeout
"""
requests, self.actual_timeout = self._get_requests()
# send all search-request
if requests:
search_multiple_requests(requests, self.result_container, self.start_time, self.actual_timeout)
start_new_thread(gc.collect, tuple())
# return results, suggestions, answers and infoboxes
return True
# do search-request
def search(self):
self.start_time = time()
if not self.search_external_bang():
if not self.search_answerers():
self.search_standard()
return self.result_container
class SearchWithPlugins(Search):
"""Similar to the Search class but call the plugins."""
__slots__ = 'ordered_plugin_list', 'request'
def __init__(self, search_query, ordered_plugin_list, request):
super().__init__(search_query)
self.ordered_plugin_list = ordered_plugin_list
self.request = request
def search(self):
if plugins.call(self.ordered_plugin_list, 'pre_search', self.request, self):
super().search()
plugins.call(self.ordered_plugin_list, 'post_search', self.request, self)
results = self.result_container.get_ordered_results()
for result in results:
plugins.call(self.ordered_plugin_list, 'on_result', self.request, self, result)
return self.result_container

265
searx/search/__init__.py Normal file
View file

@ -0,0 +1,265 @@
'''
searx is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
searx is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with searx. If not, see < http://www.gnu.org/licenses/ >.
(C) 2013- by Adam Tauber, <asciimoo@gmail.com>
'''
import typing
import gc
import threading
from time import time
from uuid import uuid4
from _thread import start_new_thread
from searx import settings
from searx.answerers import ask
from searx.external_bang import get_bang_url
from searx.results import ResultContainer
from searx import logger
from searx.plugins import plugins
from searx.search.processors import processors, initialize as initialize_processors
logger = logger.getChild('search')
max_request_timeout = settings.get('outgoing', {}).get('max_request_timeout' or None)
if max_request_timeout is None:
logger.info('max_request_timeout={0}'.format(max_request_timeout))
else:
if isinstance(max_request_timeout, float):
logger.info('max_request_timeout={0} second(s)'.format(max_request_timeout))
else:
logger.critical('outgoing.max_request_timeout if defined has to be float')
import sys
sys.exit(1)
def initialize(settings_engines=None):
settings_engines = settings_engines or settings['engines']
initialize_processors(settings_engines)
class EngineRef:
__slots__ = 'name', 'category', 'from_bang'
def __init__(self, name: str, category: str, from_bang: bool=False):
self.name = name
self.category = category
self.from_bang = from_bang
def __repr__(self):
return "EngineRef({!r}, {!r}, {!r})".format(self.name, self.category, self.from_bang)
def __eq__(self, other):
return self.name == other.name and self.category == other.category and self.from_bang == other.from_bang
class SearchQuery:
"""container for all the search parameters (query, language, etc...)"""
__slots__ = 'query', 'engineref_list', 'categories', 'lang', 'safesearch', 'pageno', 'time_range',\
'timeout_limit', 'external_bang'
def __init__(self,
query: str,
engineref_list: typing.List[EngineRef],
categories: typing.List[str],
lang: str,
safesearch: int,
pageno: int,
time_range: typing.Optional[str],
timeout_limit: typing.Optional[float]=None,
external_bang: typing.Optional[str]=None):
self.query = query
self.engineref_list = engineref_list
self.categories = categories
self.lang = lang
self.safesearch = safesearch
self.pageno = pageno
self.time_range = time_range
self.timeout_limit = timeout_limit
self.external_bang = external_bang
def __repr__(self):
return "SearchQuery({!r}, {!r}, {!r}, {!r}, {!r}, {!r}, {!r}, {!r}, {!r})".\
format(self.query, self.engineref_list, self.categories, self.lang, self.safesearch,
self.pageno, self.time_range, self.timeout_limit, self.external_bang)
def __eq__(self, other):
return self.query == other.query\
and self.engineref_list == other.engineref_list\
and self.categories == self.categories\
and self.lang == other.lang\
and self.safesearch == other.safesearch\
and self.pageno == other.pageno\
and self.time_range == other.time_range\
and self.timeout_limit == other.timeout_limit\
and self.external_bang == other.external_bang
class Search:
"""Search information container"""
__slots__ = "search_query", "result_container", "start_time", "actual_timeout"
def __init__(self, search_query):
# init vars
super().__init__()
self.search_query = search_query
self.result_container = ResultContainer()
self.start_time = None
self.actual_timeout = None
def search_external_bang(self):
"""
Check if there is a external bang.
If yes, update self.result_container and return True
"""
if self.search_query.external_bang:
self.result_container.redirect_url = get_bang_url(self.search_query)
# This means there was a valid bang and the
# rest of the search does not need to be continued
if isinstance(self.result_container.redirect_url, str):
return True
return False
def search_answerers(self):
"""
Check if an answer return a result.
If yes, update self.result_container and return True
"""
answerers_results = ask(self.search_query)
if answerers_results:
for results in answerers_results:
self.result_container.extend('answer', results)
return True
return False
# do search-request
def _get_requests(self):
# init vars
requests = []
# max of all selected engine timeout
default_timeout = 0
# start search-reqest for all selected engines
for engineref in self.search_query.engineref_list:
processor = processors[engineref.name]
# set default request parameters
request_params = processor.get_params(self.search_query, engineref.category)
if request_params is None:
continue
# append request to list
requests.append((engineref.name, self.search_query.query, request_params))
# update default_timeout
default_timeout = max(default_timeout, processor.engine.timeout)
# adjust timeout
actual_timeout = default_timeout
query_timeout = self.search_query.timeout_limit
if max_request_timeout is None and query_timeout is None:
# No max, no user query: default_timeout
pass
elif max_request_timeout is None and query_timeout is not None:
# No max, but user query: From user query except if above default
actual_timeout = min(default_timeout, query_timeout)
elif max_request_timeout is not None and query_timeout is None:
# Max, no user query: Default except if above max
actual_timeout = min(default_timeout, max_request_timeout)
elif max_request_timeout is not None and query_timeout is not None:
# Max & user query: From user query except if above max
actual_timeout = min(query_timeout, max_request_timeout)
logger.debug("actual_timeout={0} (default_timeout={1}, ?timeout_limit={2}, max_request_timeout={3})"
.format(actual_timeout, default_timeout, query_timeout, max_request_timeout))
return requests, actual_timeout
def search_multiple_requests(self, requests):
search_id = uuid4().__str__()
for engine_name, query, request_params in requests:
th = threading.Thread(
target=processors[engine_name].search,
args=(query, request_params, self.result_container, self.start_time, self.actual_timeout),
name=search_id,
)
th._timeout = False
th._engine_name = engine_name
th.start()
for th in threading.enumerate():
if th.name == search_id:
remaining_time = max(0.0, self.actual_timeout - (time() - self.start_time))
th.join(remaining_time)
if th.is_alive():
th._timeout = True
self.result_container.add_unresponsive_engine(th._engine_name, 'timeout')
logger.warning('engine timeout: {0}'.format(th._engine_name))
def search_standard(self):
"""
Update self.result_container, self.actual_timeout
"""
requests, self.actual_timeout = self._get_requests()
# send all search-request
if requests:
self.search_multiple_requests(requests)
start_new_thread(gc.collect, tuple())
# return results, suggestions, answers and infoboxes
return True
# do search-request
def search(self):
self.start_time = time()
if not self.search_external_bang():
if not self.search_answerers():
self.search_standard()
return self.result_container
class SearchWithPlugins(Search):
"""Similar to the Search class but call the plugins."""
__slots__ = 'ordered_plugin_list', 'request'
def __init__(self, search_query, ordered_plugin_list, request):
super().__init__(search_query)
self.ordered_plugin_list = ordered_plugin_list
self.request = request
def search(self):
if plugins.call(self.ordered_plugin_list, 'pre_search', self.request, self):
super().search()
plugins.call(self.ordered_plugin_list, 'post_search', self.request, self)
results = self.result_container.get_ordered_results()
for result in results:
plugins.call(self.ordered_plugin_list, 'on_result', self.request, self, result)
return self.result_container

View file

@ -0,0 +1,41 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
from .online import OnlineProcessor
from .offline import OfflineProcessor
from .online_dictionary import OnlineDictionaryProcessor
from .online_currency import OnlineCurrencyProcessor
from .abstract import EngineProcessor
from searx import logger
import searx.engines as engines
__all__ = ['EngineProcessor', 'OfflineProcessor', 'OnlineProcessor',
'OnlineDictionaryProcessor', 'OnlineCurrencyProcessor', 'processors']
logger = logger.getChild('search.processors')
processors = {}
def get_processor_class(engine_type):
for c in [OnlineProcessor, OfflineProcessor, OnlineDictionaryProcessor, OnlineCurrencyProcessor]:
if c.engine_type == engine_type:
return c
return None
def get_processor(engine, engine_name):
engine_type = getattr(engine, 'engine_type', 'online')
processor_class = get_processor_class(engine_type)
if processor_class:
return processor_class(engine, engine_name)
else:
return None
def initialize(engine_list):
engines.initialize_engines(engine_list)
for engine_name, engine in engines.engines.items():
processor = get_processor(engine, engine_name)
if processor is None:
logger.error('Error get processor for engine %s', engine_name)
else:
processors[engine_name] = processor

View file

@ -0,0 +1,39 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
from abc import abstractmethod
from searx import logger
logger = logger.getChild('searx.search.processor')
class EngineProcessor:
def __init__(self, engine, engine_name):
self.engine = engine
self.engine_name = engine_name
def get_params(self, search_query, engine_category):
# if paging is not supported, skip
if search_query.pageno > 1 and not self.engine.paging:
return None
# if time_range is not supported, skip
if search_query.time_range and not self.engine.time_range_support:
return None
params = {}
params['category'] = engine_category
params['pageno'] = search_query.pageno
params['safesearch'] = search_query.safesearch
params['time_range'] = search_query.time_range
if hasattr(self.engine, 'language') and self.engine.language:
params['language'] = self.engine.language
else:
params['language'] = search_query.lang
return params
@abstractmethod
def search(self, query, params, result_container, start_time, timeout_limit):
pass

View file

@ -0,0 +1,51 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
import threading
from time import time
from searx import logger
from searx.metrology.error_recorder import record_exception, record_error
from searx.search.processors.abstract import EngineProcessor
logger = logger.getChild('search.processor.offline')
class OfflineProcessor(EngineProcessor):
engine_type = 'offline'
def _record_stats_on_error(self, result_container, start_time):
engine_time = time() - start_time
result_container.add_timing(self.engine_name, engine_time, engine_time)
with threading.RLock():
self.engine.stats['errors'] += 1
def _search_basic(self, query, params):
return self.engine.search(query, params)
def search(self, query, params, result_container, start_time, timeout_limit):
try:
search_results = self._search_basic(query, params)
if search_results:
result_container.extend(self.engine_name, search_results)
engine_time = time() - start_time
result_container.add_timing(self.engine_name, engine_time, engine_time)
with threading.RLock():
self.engine.stats['engine_time'] += engine_time
self.engine.stats['engine_time_count'] += 1
except ValueError as e:
record_exception(self.engine_name, e)
self._record_stats_on_error(result_container, start_time)
logger.exception('engine {0} : invalid input : {1}'.format(self.engine_name, e))
except Exception as e:
record_exception(self.engine_name, e)
self._record_stats_on_error(result_container, start_time)
result_container.add_unresponsive_engine(self.engine_name, 'unexpected crash', str(e))
logger.exception('engine {0} : exception : {1}'.format(self.engine_name, e))
else:
if getattr(threading.current_thread(), '_timeout', False):
record_error(self.engine_name, 'Timeout')

View file

@ -0,0 +1,211 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
from urllib.parse import urlparse
from time import time
import threading
import requests.exceptions
import searx.poolrequests as poolrequests
from searx.engines import settings
from searx import logger
from searx.utils import gen_useragent
from searx.exceptions import (SearxEngineAccessDeniedException, SearxEngineCaptchaException,
SearxEngineTooManyRequestsException,)
from searx.metrology.error_recorder import record_exception, record_error
from searx.search.processors.abstract import EngineProcessor
logger = logger.getChild('search.processor.online')
DEFAULT_PARAMS = {
'method': 'GET',
'headers': {},
'data': {},
'url': '',
'cookies': {},
'verify': True,
'auth': None
}
class OnlineProcessor(EngineProcessor):
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
# skip suspended engines
if self.engine.suspend_end_time >= time():
logger.debug('Engine currently suspended: %s', self.engine_name)
return None
# add default params
params.update(DEFAULT_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']
)
# setting engine based proxies
if hasattr(self.engine, 'proxies'):
request_args['proxies'] = poolrequests.get_proxies(self.engine.proxies)
# max_redirects
max_redirects = params.get('max_redirects')
if max_redirects:
request_args['max_redirects'] = max_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', False)
# specific type of request (GET or POST)
if params['method'] == 'GET':
req = poolrequests.get
else:
req = poolrequests.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 '')
reason = response.reason or ''
hostname = str(urlparse(response.url or '').netloc)
record_error(self.engine_name,
'{} redirects, maximum: {}'.format(len(response.history), soft_max_redirects),
(status_code, reason, hostname))
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
poolrequests.set_timeout_for_thread(timeout_limit, start_time=start_time)
# reset the HTTP total time
poolrequests.reset_time_for_thread()
# suppose everything will be alright
requests_exception = False
suspended_time = None
try:
# send requests and parse the results
search_results = self._search_basic(query, params)
# check if the engine accepted the request
if search_results is not None:
# yes, so add results
result_container.extend(self.engine_name, search_results)
# update engine time when there is no exception
engine_time = time() - start_time
page_load_time = poolrequests.get_time_for_thread()
result_container.add_timing(self.engine_name, engine_time, page_load_time)
with threading.RLock():
self.engine.stats['engine_time'] += engine_time
self.engine.stats['engine_time_count'] += 1
# update stats with the total HTTP time
self.engine.stats['page_load_time'] += page_load_time
self.engine.stats['page_load_count'] += 1
except Exception as e:
record_exception(self.engine_name, e)
# Timing
engine_time = time() - start_time
page_load_time = poolrequests.get_time_for_thread()
result_container.add_timing(self.engine_name, engine_time, page_load_time)
# Record the errors
with threading.RLock():
self.engine.stats['errors'] += 1
if (issubclass(e.__class__, requests.exceptions.Timeout)):
result_container.add_unresponsive_engine(self.engine_name, 'HTTP timeout')
# requests timeout (connect or read)
logger.error("engine {0} : HTTP requests timeout"
"(search duration : {1} s, timeout: {2} s) : {3}"
.format(self.engine_name, engine_time, timeout_limit, e.__class__.__name__))
requests_exception = True
elif (issubclass(e.__class__, requests.exceptions.RequestException)):
result_container.add_unresponsive_engine(self.engine_name, 'HTTP error')
# other requests exception
logger.exception("engine {0} : requests exception"
"(search duration : {1} s, timeout: {2} s) : {3}"
.format(self.engine_name, engine_time, timeout_limit, e))
requests_exception = True
elif (issubclass(e.__class__, SearxEngineCaptchaException)):
result_container.add_unresponsive_engine(self.engine_name, 'CAPTCHA required')
logger.exception('engine {0} : CAPTCHA')
suspended_time = e.suspended_time # pylint: disable=no-member
elif (issubclass(e.__class__, SearxEngineTooManyRequestsException)):
result_container.add_unresponsive_engine(self.engine_name, 'too many requests')
logger.exception('engine {0} : Too many requests')
suspended_time = e.suspended_time # pylint: disable=no-member
elif (issubclass(e.__class__, SearxEngineAccessDeniedException)):
result_container.add_unresponsive_engine(self.engine_name, 'blocked')
logger.exception('engine {0} : Searx is blocked')
suspended_time = e.suspended_time # pylint: disable=no-member
else:
result_container.add_unresponsive_engine(self.engine_name, 'unexpected crash')
# others errors
logger.exception('engine {0} : exception : {1}'.format(self.engine_name, e))
else:
if getattr(threading.current_thread(), '_timeout', False):
record_error(self.engine_name, 'Timeout')
# suspend the engine if there is an HTTP error
# or suspended_time is defined
with threading.RLock():
if requests_exception or suspended_time:
# update continuous_errors / suspend_end_time
self.engine.continuous_errors += 1
if suspended_time is None:
suspended_time = min(settings['search']['max_ban_time_on_fail'],
self.engine.continuous_errors * settings['search']['ban_time_on_fail'])
self.engine.suspend_end_time = time() + suspended_time
else:
# reset the suspend variables
self.engine.continuous_errors = 0
self.engine.suspend_end_time = 0

View file

@ -0,0 +1,57 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
import unicodedata
import re
from searx.data import CURRENCIES
from .online import OnlineProcessor
parser_re = re.compile('.*?(\\d+(?:\\.\\d+)?) ([^.0-9]+) (?:in|to) ([^.0-9]+)', re.I)
def normalize_name(name):
name = name.lower().replace('-', ' ').rstrip('s')
name = re.sub(' +', ' ', name)
return unicodedata.normalize('NFKD', name).lower()
def name_to_iso4217(name):
global CURRENCIES
name = normalize_name(name)
currency = CURRENCIES['names'].get(name, [name])
return currency[0]
def iso4217_to_name(iso4217, language):
global CURRENCIES
return CURRENCIES['iso4217'].get(iso4217, {}).get(language, iso4217)
class OnlineCurrencyProcessor(OnlineProcessor):
engine_type = 'online_currency'
def get_params(self, search_query, engine_category):
params = super().get_params(search_query, engine_category)
if params is None:
return None
m = parser_re.match(search_query.query)
if not m:
return None
amount_str, from_currency, to_currency = m.groups()
try:
amount = float(amount_str)
except ValueError:
return None
from_currency = name_to_iso4217(from_currency.strip())
to_currency = name_to_iso4217(to_currency.strip())
params['amount'] = amount
params['from'] = from_currency
params['to'] = to_currency
params['from_name'] = iso4217_to_name(from_currency, 'en')
params['to_name'] = iso4217_to_name(to_currency, 'en')
return params

View file

@ -0,0 +1,37 @@
# SPDX-License-Identifier: AGPL-3.0-or-later
import re
from searx.utils import is_valid_lang
from .online import OnlineProcessor
parser_re = re.compile('.*?([a-z]+)-([a-z]+) ([^ ]+)$', re.I)
class OnlineDictionaryProcessor(OnlineProcessor):
engine_type = 'online_dictionnary'
def get_params(self, search_query, engine_category):
params = super().get_params(search_query, engine_category)
if params is None:
return None
m = parser_re.match(search_query.query)
if not m:
return None
from_lang, to_lang, query = m.groups()
from_lang = is_valid_lang(from_lang)
to_lang = is_valid_lang(to_lang)
if not from_lang or not to_lang:
return None
params['from_lang'] = from_lang
params['to_lang'] = to_lang
params['query'] = query
return params

View file

@ -60,7 +60,7 @@ from searx import brand, static_path
from searx import settings, searx_dir, searx_debug
from searx.exceptions import SearxParameterException
from searx.engines import (
categories, engines, engine_shortcuts, get_engines_stats, initialize_engines
categories, engines, engine_shortcuts, get_engines_stats
)
from searx.webutils import (
UnicodeWriter, highlight_content, get_resources_directory,
@ -71,7 +71,7 @@ from searx.webadapter import get_search_query_from_webapp, get_selected_categori
from searx.utils import html_to_text, gen_useragent, dict_subset, match_language
from searx.version import VERSION_STRING
from searx.languages import language_codes as languages
from searx.search import SearchWithPlugins
from searx.search import SearchWithPlugins, initialize
from searx.query import RawTextQuery
from searx.autocomplete import searx_bang, backends as autocomplete_backends
from searx.plugins import plugins
@ -131,7 +131,7 @@ werkzeug_reloader = flask_run_development or (searx_debug and __name__ == "__mai
# initialize the engines except on the first run of the werkzeug server.
if not werkzeug_reloader\
or (werkzeug_reloader and os.environ.get("WERKZEUG_RUN_MAIN") == "true"):
initialize_engines(settings['engines'])
initialize()
babel = Babel(app)

View file

@ -3,7 +3,6 @@
from searx.testing import SearxTestCase
from searx.search import SearchQuery, EngineRef
import searx.search
import searx.engines
SAFESEARCH = 0
@ -39,7 +38,7 @@ class SearchTestCase(SearxTestCase):
@classmethod
def setUpClass(cls):
searx.engines.initialize_engines(TEST_ENGINES)
searx.search.initialize(TEST_ENGINES)
def test_timeout_simple(self):
searx.search.max_request_timeout = None

View file

@ -8,8 +8,7 @@ import sys
from mock import Mock, patch
from nose2.tools import params
from searx.search import SearchQuery, EngineRef
from searx.engines import initialize_engines
from searx.search import SearchQuery, EngineRef, initialize
from searx.testing import SearxTestCase
@ -30,7 +29,7 @@ class StandaloneSearx(SearxTestCase):
def setUpClass(cls):
engine_list = [{'engine': 'dummy', 'name': 'engine1', 'shortcut': 'e1'}]
initialize_engines(engine_list)
initialize(engine_list)
def test_parse_argument_no_args(self):
"""Test parse argument without args."""

View file

@ -17,7 +17,7 @@ TEST_ENGINES = [
'categories': 'general',
'shortcut': 'do',
'timeout': 3.0,
'offline': True,
'engine_type': 'offline',
'tokens': ['my-token'],
},
]
@ -28,7 +28,7 @@ class ValidateQueryCase(SearxTestCase):
@classmethod
def setUpClass(cls):
searx.engines.initialize_engines(TEST_ENGINES)
searx.search.initialize(TEST_ENGINES)
def test_query_private_engine_without_token(self):
preferences = Preferences(['oscar'], ['general'], engines, [])

View file

@ -6,8 +6,8 @@ Getting categories without initiate the engine will only return `['general']`
>>> import searx.engines
... list(searx.engines.categories.keys())
['general']
>>> import searx
... searx.engines.initialize_engines(searx.settings['engines'])
>>> import searx.search
... searx.search.initialize()
... list(searx.engines.categories.keys())
['general', 'it', 'science', 'images', 'news', 'videos', 'music', 'files', 'social media', 'map']
@ -22,11 +22,11 @@ Example to run it from python:
>>> import importlib
... import json
... import sys
... import searx
... import searx.engines
... import searx.search
... search_query = 'rain'
... # initialize engines
... searx.engines.initialize_engines(searx.settings['engines'])
... searx.search.initialize()
... # load engines categories once instead of each time the function called
... engine_cs = list(searx.engines.categories.keys())
... # load module
@ -82,7 +82,6 @@ from json import dumps
from typing import Any, Dict, List, Optional
import searx
import searx.engines
import searx.preferences
import searx.query
import searx.search
@ -208,7 +207,7 @@ def parse_argument(
if __name__ == '__main__':
searx.engines.initialize_engines(searx.settings['engines'])
searx.search.initialize()
engine_cs = list(searx.engines.categories.keys())
prog_args = parse_argument(category_choices=engine_cs)
search_q = get_search_query(prog_args, engine_categories=engine_cs)