.oO SearXNG Developer Documentation Oo.
Loading...
Searching...
No Matches
network.py
Go to the documentation of this file.
1# SPDX-License-Identifier: AGPL-3.0-or-later
2# pylint: disable=global-statement
3# pylint: disable=missing-module-docstring, missing-class-docstring
4from __future__ import annotations
5
6import typing
7import atexit
8import asyncio
9import ipaddress
10from itertools import cycle
11from typing import Dict
12
13import httpx
14
15from searx import logger, searx_debug
16from searx.extended_types import SXNG_Response
17from .client import new_client, get_loop, AsyncHTTPTransportNoHttp
18from .raise_for_httperror import raise_for_httperror
19
20
21logger = logger.getChild('network')
22DEFAULT_NAME = '__DEFAULT__'
23NETWORKS: Dict[str, 'Network'] = {}
24# requests compatibility when reading proxy settings from settings.yml
25PROXY_PATTERN_MAPPING = {
26 'http': 'http://',
27 'https': 'https://',
28 'socks4': 'socks4://',
29 'socks5': 'socks5://',
30 'socks5h': 'socks5h://',
31 'http:': 'http://',
32 'https:': 'https://',
33 'socks4:': 'socks4://',
34 'socks5:': 'socks5://',
35 'socks5h:': 'socks5h://',
36}
37
38ADDRESS_MAPPING = {'ipv4': '0.0.0.0', 'ipv6': '::'}
39
40
41class Network:
42
43 __slots__ = (
44 'enable_http',
45 'verify',
46 'enable_http2',
47 'max_connections',
48 'max_keepalive_connections',
49 'keepalive_expiry',
50 'local_addresses',
51 'proxies',
52 'using_tor_proxy',
53 'max_redirects',
54 'retries',
55 'retry_on_http_error',
56 '_local_addresses_cycle',
57 '_proxies_cycle',
58 '_clients',
59 '_logger',
60 )
61
62 _TOR_CHECK_RESULT = {}
63
65 # pylint: disable=too-many-arguments
66 self,
67 enable_http=True,
68 verify=True,
69 enable_http2=False,
70 max_connections=None,
71 max_keepalive_connections=None,
72 keepalive_expiry=None,
73 proxies=None,
74 using_tor_proxy=False,
75 local_addresses=None,
76 retries=0,
77 retry_on_http_error=None,
78 max_redirects=30,
79 logger_name=None,
80 ):
81
82 self.enable_http = enable_http
83 self.verify = verify
84 self.enable_http2 = enable_http2
85 self.max_connections = max_connections
86 self.max_keepalive_connections = max_keepalive_connections
87 self.keepalive_expiry = keepalive_expiry
88 self.proxies = proxies
89 self.using_tor_proxy = using_tor_proxy
90 self.local_addresses = local_addresses
91 self.retries = retries
92 self.retry_on_http_error = retry_on_http_error
93 self.max_redirects = max_redirects
96 self._clients = {}
97 self._logger = logger.getChild(logger_name) if logger_name else logger
98 self.check_parameters()
99
101 for address in self.iter_ipaddresses():
102 if '/' in address:
103 ipaddress.ip_network(address, False)
104 else:
105 ipaddress.ip_address(address)
106
107 if self.proxies is not None and not isinstance(self.proxies, (str, dict)):
108 raise ValueError('proxies type has to be str, dict or None')
109
111 local_addresses = self.local_addresses
112 if not local_addresses:
113 return
114 if isinstance(local_addresses, str):
115 local_addresses = [local_addresses]
116 yield from local_addresses
117
119 while True:
120 count = 0
121 for address in self.iter_ipaddresses():
122 if '/' in address:
123 for a in ipaddress.ip_network(address, False).hosts():
124 yield str(a)
125 count += 1
126 else:
127 a = ipaddress.ip_address(address)
128 yield str(a)
129 count += 1
130 if count == 0:
131 yield None
132
133 def iter_proxies(self):
134 if not self.proxies:
135 return
136 # https://www.python-httpx.org/compatibility/#proxy-keys
137 if isinstance(self.proxies, str):
138 yield 'all://', [self.proxies]
139 else:
140 for pattern, proxy_url in self.proxies.items():
141 pattern = PROXY_PATTERN_MAPPING.get(pattern, pattern)
142 if isinstance(proxy_url, str):
143 proxy_url = [proxy_url]
144 yield pattern, proxy_url
145
147 proxy_settings = {}
148 for pattern, proxy_urls in self.iter_proxies():
149 proxy_settings[pattern] = cycle(proxy_urls)
150 while True:
151 # pylint: disable=stop-iteration-return
152 yield tuple((pattern, next(proxy_url_cycle)) for pattern, proxy_url_cycle in proxy_settings.items())
153
154 async def log_response(self, response: httpx.Response):
155 request = response.request
156 status = f"{response.status_code} {response.reason_phrase}"
157 response_line = f"{response.http_version} {status}"
158 content_type = response.headers.get("Content-Type")
159 content_type = f' ({content_type})' if content_type else ''
160 self._logger.debug(f'HTTP Request: {request.method} {request.url} "{response_line}"{content_type}')
161
162 @staticmethod
163 async def check_tor_proxy(client: httpx.AsyncClient, proxies) -> bool:
164 if proxies in Network._TOR_CHECK_RESULT:
165 return Network._TOR_CHECK_RESULT[proxies]
166
167 result = True
168 # ignore client._transport because it is not used with all://
169 for transport in client._mounts.values(): # pylint: disable=protected-access
170 if isinstance(transport, AsyncHTTPTransportNoHttp):
171 continue
172 if getattr(transport, "_pool") and getattr(
173 transport._pool, "_rdns", False # pylint: disable=protected-access
174 ):
175 continue
176 return False
177 response = await client.get("https://check.torproject.org/api/ip", timeout=60)
178 if not response.json()["IsTor"]:
179 result = False
180 Network._TOR_CHECK_RESULT[proxies] = result
181 return result
182
183 async def get_client(self, verify=None, max_redirects=None):
184 verify = self.verify if verify is None else verify
185 max_redirects = self.max_redirects if max_redirects is None else max_redirects
186 local_address = next(self._local_addresses_cycle)
187 proxies = next(self._proxies_cycle) # is a tuple so it can be part of the key
188 key = (verify, max_redirects, local_address, proxies)
189 hook_log_response = self.log_response if searx_debug else None
190 if key not in self._clients or self._clients[key].is_closed:
191 client = new_client(
192 self.enable_http,
193 verify,
194 self.enable_http2,
195 self.max_connections,
197 self.keepalive_expiry,
198 dict(proxies),
199 local_address,
200 0,
201 max_redirects,
202 hook_log_response,
203 )
204 if self.using_tor_proxy and not await self.check_tor_proxy(client, proxies):
205 await client.aclose()
206 raise httpx.ProxyError('Network configuration problem: not using Tor')
207 self._clients[key] = client
208 return self._clients[key]
209
210 async def aclose(self):
211 async def close_client(client):
212 try:
213 await client.aclose()
214 except httpx.HTTPError:
215 pass
216
217 await asyncio.gather(*[close_client(client) for client in self._clients.values()], return_exceptions=False)
218
219 @staticmethod
221 kwargs_clients = {}
222 if 'verify' in kwargs:
223 kwargs_clients['verify'] = kwargs.pop('verify')
224 if 'max_redirects' in kwargs:
225 kwargs_clients['max_redirects'] = kwargs.pop('max_redirects')
226 if 'allow_redirects' in kwargs:
227 # see https://github.com/encode/httpx/pull/1808
228 kwargs['follow_redirects'] = kwargs.pop('allow_redirects')
229 return kwargs_clients
230
231 @staticmethod
233 do_raise_for_httperror = True
234 if 'raise_for_httperror' in kwargs:
235 do_raise_for_httperror = kwargs['raise_for_httperror']
236 del kwargs['raise_for_httperror']
237 return do_raise_for_httperror
238
239 def patch_response(self, response, do_raise_for_httperror) -> SXNG_Response:
240 if isinstance(response, httpx.Response):
241 response = typing.cast(SXNG_Response, response)
242 # requests compatibility (response is not streamed)
243 # see also https://www.python-httpx.org/compatibility/#checking-for-4xx5xx-responses
244 response.ok = not response.is_error
245
246 # raise an exception
247 if do_raise_for_httperror:
248 try:
249 raise_for_httperror(response)
250 except:
251 self._logger.warning(f"HTTP Request failed: {response.request.method} {response.request.url}")
252 raise
253 return response
254
255 def is_valid_response(self, response):
256 # pylint: disable=too-many-boolean-expressions
257 if (
258 (self.retry_on_http_error is True and 400 <= response.status_code <= 599)
259 or (isinstance(self.retry_on_http_error, list) and response.status_code in self.retry_on_http_error)
260 or (isinstance(self.retry_on_http_error, int) and response.status_code == self.retry_on_http_error)
261 ):
262 return False
263 return True
264
265 async def call_client(self, stream, method, url, **kwargs) -> SXNG_Response:
266 retries = self.retries
267 was_disconnected = False
268 do_raise_for_httperror = Network.extract_do_raise_for_httperror(kwargs)
269 kwargs_clients = Network.extract_kwargs_clients(kwargs)
270 while retries >= 0: # pragma: no cover
271 client = await self.get_client(**kwargs_clients)
272 try:
273 if stream:
274 response = client.stream(method, url, **kwargs)
275 else:
276 response = await client.request(method, url, **kwargs)
277 if self.is_valid_response(response) or retries <= 0:
278 return self.patch_response(response, do_raise_for_httperror)
279 except httpx.RemoteProtocolError as e:
280 if not was_disconnected:
281 # the server has closed the connection:
282 # try again without decreasing the retries variable & with a new HTTP client
283 was_disconnected = True
284 await client.aclose()
285 self._logger.warning('httpx.RemoteProtocolError: the server has disconnected, retrying')
286 continue
287 if retries <= 0:
288 raise e
289 except (httpx.RequestError, httpx.HTTPStatusError) as e:
290 if retries <= 0:
291 raise e
292 retries -= 1
293
294 async def request(self, method, url, **kwargs):
295 return await self.call_client(False, method, url, **kwargs)
296
297 async def stream(self, method, url, **kwargs):
298 return await self.call_client(True, method, url, **kwargs)
299
300 @classmethod
301 async def aclose_all(cls):
302 await asyncio.gather(*[network.aclose() for network in NETWORKS.values()], return_exceptions=False)
303
304
305def get_network(name=None):
306 return NETWORKS.get(name or DEFAULT_NAME)
307
308
310 async def check():
311 exception_count = 0
312 for network in NETWORKS.values():
313 if network.using_tor_proxy:
314 try:
315 await network.get_client()
316 except Exception: # pylint: disable=broad-except
317 network._logger.exception('Error') # pylint: disable=protected-access
318 exception_count += 1
319 return exception_count
320
321 future = asyncio.run_coroutine_threadsafe(check(), get_loop())
322 exception_count = future.result()
323 if exception_count > 0:
324 raise RuntimeError("Invalid network configuration")
325
326
327def initialize(settings_engines=None, settings_outgoing=None):
328 # pylint: disable=import-outside-toplevel)
329 from searx.engines import engines
330 from searx import settings
331
332 # pylint: enable=import-outside-toplevel)
333
334 settings_engines = settings_engines or settings['engines']
335 settings_outgoing = settings_outgoing or settings['outgoing']
336
337 # default parameters for AsyncHTTPTransport
338 # see https://github.com/encode/httpx/blob/e05a5372eb6172287458b37447c30f650047e1b8/httpx/_transports/default.py#L108-L121 # pylint: disable=line-too-long
339 default_params = {
340 'enable_http': False,
341 'verify': settings_outgoing['verify'],
342 'enable_http2': settings_outgoing['enable_http2'],
343 'max_connections': settings_outgoing['pool_connections'],
344 'max_keepalive_connections': settings_outgoing['pool_maxsize'],
345 'keepalive_expiry': settings_outgoing['keepalive_expiry'],
346 'local_addresses': settings_outgoing['source_ips'],
347 'using_tor_proxy': settings_outgoing['using_tor_proxy'],
348 'proxies': settings_outgoing['proxies'],
349 'max_redirects': settings_outgoing['max_redirects'],
350 'retries': settings_outgoing['retries'],
351 'retry_on_http_error': None,
352 }
353
354 def new_network(params, logger_name=None):
355 nonlocal default_params
356 result = {}
357 result.update(default_params)
358 result.update(params)
359 if logger_name:
360 result['logger_name'] = logger_name
361 return Network(**result)
362
363 def iter_networks():
364 nonlocal settings_engines
365 for engine_spec in settings_engines:
366 engine_name = engine_spec['name']
367 engine = engines.get(engine_name)
368 if engine is None:
369 continue
370 network = getattr(engine, 'network', None)
371 yield engine_name, engine, network
372
373 if NETWORKS:
374 done()
375 NETWORKS.clear()
376 NETWORKS[DEFAULT_NAME] = new_network({}, logger_name='default')
377 NETWORKS['ipv4'] = new_network({'local_addresses': '0.0.0.0'}, logger_name='ipv4')
378 NETWORKS['ipv6'] = new_network({'local_addresses': '::'}, logger_name='ipv6')
379
380 # define networks from outgoing.networks
381 for network_name, network in settings_outgoing['networks'].items():
382 NETWORKS[network_name] = new_network(network, logger_name=network_name)
383
384 # define networks from engines.[i].network (except references)
385 for engine_name, engine, network in iter_networks():
386 if network is None:
387 network = {}
388 for attribute_name, attribute_value in default_params.items():
389 if hasattr(engine, attribute_name):
390 network[attribute_name] = getattr(engine, attribute_name)
391 else:
392 network[attribute_name] = attribute_value
393 NETWORKS[engine_name] = new_network(network, logger_name=engine_name)
394 elif isinstance(network, dict):
395 NETWORKS[engine_name] = new_network(network, logger_name=engine_name)
396
397 # define networks from engines.[i].network (references)
398 for engine_name, engine, network in iter_networks():
399 if isinstance(network, str):
400 NETWORKS[engine_name] = NETWORKS[network]
401
402 # the /image_proxy endpoint has a dedicated network.
403 # same parameters than the default network, but HTTP/2 is disabled.
404 # It decreases the CPU load average, and the total time is more or less the same
405 if 'image_proxy' not in NETWORKS:
406 image_proxy_params = default_params.copy()
407 image_proxy_params['enable_http2'] = False
408 NETWORKS['image_proxy'] = new_network(image_proxy_params, logger_name='image_proxy')
409
410
411@atexit.register
412def done():
413 """Close all HTTP client
414
415 Avoid a warning at exit
416 See https://github.com/encode/httpx/pull/2026
417
418 Note: since Network.aclose has to be async, it is not possible to call this method on Network.__del__
419 So Network.aclose is called here using atexit.register
420 """
421 try:
422 loop = get_loop()
423 if loop:
424 future = asyncio.run_coroutine_threadsafe(Network.aclose_all(), loop)
425 # wait 3 seconds to close the HTTP clients
426 future.result(3)
427 finally:
428 NETWORKS.clear()
429
430
431NETWORKS[DEFAULT_NAME] = Network()
log_response(self, httpx.Response response)
Definition network.py:154
SXNG_Response patch_response(self, response, do_raise_for_httperror)
Definition network.py:239
SXNG_Response call_client(self, stream, method, url, **kwargs)
Definition network.py:265
stream(self, method, url, **kwargs)
Definition network.py:297
request(self, method, url, **kwargs)
Definition network.py:294
bool check_tor_proxy(httpx.AsyncClient client, proxies)
Definition network.py:163
is_valid_response(self, response)
Definition network.py:255
__init__(self, enable_http=True, verify=True, enable_http2=False, max_connections=None, max_keepalive_connections=None, keepalive_expiry=None, proxies=None, using_tor_proxy=False, local_addresses=None, retries=0, retry_on_http_error=None, max_redirects=30, logger_name=None)
Definition network.py:80
get_client(self, verify=None, max_redirects=None)
Definition network.py:183
::1337x
Definition 1337x.py:1
initialize(settings_engines=None, settings_outgoing=None)
Definition network.py:327
get_network(name=None)
Definition network.py:305