network.py 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351
  1. # SPDX-License-Identifier: AGPL-3.0-or-later
  2. import atexit
  3. import asyncio
  4. import ipaddress
  5. from itertools import cycle
  6. import httpx
  7. from .client import new_client, get_loop, AsyncHTTPTransportNoHttp
  8. DEFAULT_NAME = '__DEFAULT__'
  9. NETWORKS = {}
  10. # requests compatibility when reading proxy settings from settings.yml
  11. PROXY_PATTERN_MAPPING = {
  12. 'http': 'http://',
  13. 'https': 'https://',
  14. 'socks4': 'socks4://',
  15. 'socks5': 'socks5://',
  16. 'socks5h': 'socks5h://',
  17. 'http:': 'http://',
  18. 'https:': 'https://',
  19. 'socks4:': 'socks4://',
  20. 'socks5:': 'socks5://',
  21. 'socks5h:': 'socks5h://',
  22. }
  23. ADDRESS_MAPPING = {
  24. 'ipv4': '0.0.0.0',
  25. 'ipv6': '::'
  26. }
  27. class Network:
  28. __slots__ = ('enable_http', 'verify', 'enable_http2',
  29. 'max_connections', 'max_keepalive_connections', 'keepalive_expiry',
  30. 'local_addresses', 'proxies', 'using_tor_proxy', 'max_redirects', 'retries', 'retry_on_http_error',
  31. '_local_addresses_cycle', '_proxies_cycle', '_clients')
  32. _TOR_CHECK_RESULT = {}
  33. def __init__(self,
  34. enable_http=True,
  35. verify=True,
  36. enable_http2=False,
  37. max_connections=None,
  38. max_keepalive_connections=None,
  39. keepalive_expiry=None,
  40. proxies=None,
  41. using_tor_proxy=False,
  42. local_addresses=None,
  43. retries=0,
  44. retry_on_http_error=None,
  45. max_redirects=30):
  46. self.enable_http = enable_http
  47. self.verify = verify
  48. self.enable_http2 = enable_http2
  49. self.max_connections = max_connections
  50. self.max_keepalive_connections = max_keepalive_connections
  51. self.keepalive_expiry = keepalive_expiry
  52. self.proxies = proxies
  53. self.using_tor_proxy = using_tor_proxy
  54. self.local_addresses = local_addresses
  55. self.retries = retries
  56. self.retry_on_http_error = retry_on_http_error
  57. self.max_redirects = max_redirects
  58. self._local_addresses_cycle = self.get_ipaddress_cycle()
  59. self._proxies_cycle = self.get_proxy_cycles()
  60. self._clients = {}
  61. self.check_parameters()
  62. def check_parameters(self):
  63. for address in self.iter_ipaddresses():
  64. if '/' in address:
  65. ipaddress.ip_network(address, False)
  66. else:
  67. ipaddress.ip_address(address)
  68. if self.proxies is not None and not isinstance(self.proxies, (str, dict)):
  69. raise ValueError('proxies type has to be str, dict or None')
  70. def iter_ipaddresses(self):
  71. local_addresses = self.local_addresses
  72. if not local_addresses:
  73. return
  74. elif isinstance(local_addresses, str):
  75. local_addresses = [local_addresses]
  76. for address in local_addresses:
  77. yield address
  78. def get_ipaddress_cycle(self):
  79. while True:
  80. count = 0
  81. for address in self.iter_ipaddresses():
  82. if '/' in address:
  83. for a in ipaddress.ip_network(address, False).hosts():
  84. yield str(a)
  85. count += 1
  86. else:
  87. a = ipaddress.ip_address(address)
  88. yield str(a)
  89. count += 1
  90. if count == 0:
  91. yield None
  92. def iter_proxies(self):
  93. if not self.proxies:
  94. return
  95. # https://www.python-httpx.org/compatibility/#proxy-keys
  96. if isinstance(self.proxies, str):
  97. yield 'all://', [self.proxies]
  98. else:
  99. for pattern, proxy_url in self.proxies.items():
  100. pattern = PROXY_PATTERN_MAPPING.get(pattern, pattern)
  101. if isinstance(proxy_url, str):
  102. proxy_url = [proxy_url]
  103. yield pattern, proxy_url
  104. def get_proxy_cycles(self):
  105. proxy_settings = {}
  106. for pattern, proxy_urls in self.iter_proxies():
  107. proxy_settings[pattern] = cycle(proxy_urls)
  108. while True:
  109. yield tuple((pattern, next(proxy_url_cycle)) for pattern, proxy_url_cycle in proxy_settings.items())
  110. @staticmethod
  111. async def check_tor_proxy(client: httpx.AsyncClient, proxies) -> bool:
  112. if proxies in Network._TOR_CHECK_RESULT:
  113. return Network._TOR_CHECK_RESULT[proxies]
  114. result = True
  115. # ignore client._transport because it is not used with all://
  116. for transport in client._mounts.values(): # pylint: disable=protected-access
  117. if isinstance(transport, AsyncHTTPTransportNoHttp):
  118. continue
  119. if not getattr(transport, '_rdns', False):
  120. result = False
  121. break
  122. else:
  123. response = await client.get('https://check.torproject.org/api/ip')
  124. if not response.json()['IsTor']:
  125. result = False
  126. Network._TOR_CHECK_RESULT[proxies] = result
  127. return result
  128. async def get_client(self, verify=None, max_redirects=None):
  129. verify = self.verify if verify is None else verify
  130. max_redirects = self.max_redirects if max_redirects is None else max_redirects
  131. local_address = next(self._local_addresses_cycle)
  132. proxies = next(self._proxies_cycle) # is a tuple so it can be part of the key
  133. key = (verify, max_redirects, local_address, proxies)
  134. if key not in self._clients or self._clients[key].is_closed:
  135. client = new_client(
  136. self.enable_http,
  137. verify,
  138. self.enable_http2,
  139. self.max_connections,
  140. self.max_keepalive_connections,
  141. self.keepalive_expiry,
  142. dict(proxies),
  143. local_address,
  144. 0,
  145. max_redirects
  146. )
  147. if self.using_tor_proxy and not await self.check_tor_proxy(client, proxies):
  148. await client.aclose()
  149. raise httpx.ProxyError('Network configuration problem: not using Tor')
  150. self._clients[key] = client
  151. return self._clients[key]
  152. async def aclose(self):
  153. async def close_client(client):
  154. try:
  155. await client.aclose()
  156. except httpx.HTTPError:
  157. pass
  158. await asyncio.gather(*[close_client(client) for client in self._clients.values()], return_exceptions=False)
  159. @staticmethod
  160. def get_kwargs_clients(kwargs):
  161. kwargs_clients = {}
  162. if 'verify' in kwargs:
  163. kwargs_clients['verify'] = kwargs.pop('verify')
  164. if 'max_redirects' in kwargs:
  165. kwargs_clients['max_redirects'] = kwargs.pop('max_redirects')
  166. return kwargs_clients
  167. def is_valid_respones(self, response):
  168. if (self.retry_on_http_error is True and 400 <= response.status_code <= 599) \
  169. or (isinstance(self.retry_on_http_error, list) and response.status_code in self.retry_on_http_error) \
  170. or (isinstance(self.retry_on_http_error, int) and response.status_code == self.retry_on_http_error):
  171. return False
  172. return True
  173. async def request(self, method, url, **kwargs):
  174. retries = self.retries
  175. while retries >= 0: # pragma: no cover
  176. kwargs_clients = Network.get_kwargs_clients(kwargs)
  177. client = await self.get_client(**kwargs_clients)
  178. try:
  179. response = await client.request(method, url, **kwargs)
  180. if self.is_valid_respones(response) or retries <= 0:
  181. return response
  182. except (httpx.RequestError, httpx.HTTPStatusError) as e:
  183. if retries <= 0:
  184. raise e
  185. retries -= 1
  186. async def stream(self, method, url, **kwargs):
  187. retries = self.retries
  188. while retries >= 0: # pragma: no cover
  189. kwargs_clients = Network.get_kwargs_clients(kwargs)
  190. client = await self.get_client(**kwargs_clients)
  191. try:
  192. response = client.stream(method, url, **kwargs)
  193. if self.is_valid_respones(response) or retries <= 0:
  194. return response
  195. except (httpx.RequestError, httpx.HTTPStatusError) as e:
  196. if retries <= 0:
  197. raise e
  198. retries -= 1
  199. @classmethod
  200. async def aclose_all(cls):
  201. await asyncio.gather(*[network.aclose() for network in NETWORKS.values()], return_exceptions=False)
  202. def get_network(name=None):
  203. global NETWORKS
  204. return NETWORKS[name or DEFAULT_NAME]
  205. def check_network_configuration():
  206. async def check():
  207. exception_count = 0
  208. for network in NETWORKS.values():
  209. if network.using_tor_proxy:
  210. try:
  211. await network.get_client()
  212. except Exception: # pylint: disable=broad-except
  213. exception_count += 1
  214. return exception_count
  215. future = asyncio.run_coroutine_threadsafe(check(), get_loop())
  216. exception_count = future.result()
  217. if exception_count > 0:
  218. raise RuntimeError("Invalid network configuration")
  219. def initialize(settings_engines=None, settings_outgoing=None):
  220. from searx.engines import engines
  221. from searx import settings
  222. global NETWORKS
  223. settings_engines = settings_engines or settings.get('engines')
  224. settings_outgoing = settings_outgoing or settings.get('outgoing')
  225. # default parameters for AsyncHTTPTransport
  226. # see https://github.com/encode/httpx/blob/e05a5372eb6172287458b37447c30f650047e1b8/httpx/_transports/default.py#L108-L121 # noqa
  227. default_params = {
  228. 'enable_http': False,
  229. 'verify': True,
  230. 'enable_http2': settings_outgoing.get('enable_http2', True),
  231. # Magic number kept from previous code
  232. 'max_connections': settings_outgoing.get('pool_connections', 100),
  233. # Picked from constructor
  234. 'max_keepalive_connections': settings_outgoing.get('pool_maxsize', 10),
  235. #
  236. 'keepalive_expiry': settings_outgoing.get('keepalive_expiry', 5.0),
  237. 'local_addresses': settings_outgoing.get('source_ips'),
  238. 'proxies': settings_outgoing.get('proxies'),
  239. 'using_tor_proxy': settings_outgoing.get('using_tor_proxy'),
  240. # default maximum redirect
  241. # from https://github.com/psf/requests/blob/8c211a96cdbe9fe320d63d9e1ae15c5c07e179f8/requests/models.py#L55
  242. 'max_redirects': settings_outgoing.get('max_redirects', 30),
  243. #
  244. 'retries': settings_outgoing.get('retries', 0),
  245. 'retry_on_http_error': None,
  246. }
  247. def new_network(params):
  248. nonlocal default_params
  249. result = {}
  250. result.update(default_params)
  251. result.update(params)
  252. return Network(**result)
  253. def iter_networks():
  254. nonlocal settings_engines
  255. for engine_spec in settings_engines:
  256. engine_name = engine_spec['name']
  257. engine = engines.get(engine_name)
  258. if engine is None:
  259. continue
  260. network = getattr(engine, 'network', None)
  261. yield engine_name, engine, network
  262. if NETWORKS:
  263. done()
  264. NETWORKS.clear()
  265. NETWORKS[DEFAULT_NAME] = new_network({})
  266. NETWORKS['ipv4'] = new_network({'local_addresses': '0.0.0.0'})
  267. NETWORKS['ipv6'] = new_network({'local_addresses': '::'})
  268. # define networks from outgoing.networks
  269. for network_name, network in settings_outgoing.get('networks', {}).items():
  270. NETWORKS[network_name] = new_network(network)
  271. # define networks from engines.[i].network (except references)
  272. for engine_name, engine, network in iter_networks():
  273. if network is None:
  274. network = {}
  275. for attribute_name, attribute_value in default_params.items():
  276. if hasattr(engine, attribute_name):
  277. network[attribute_name] = getattr(engine, attribute_name)
  278. else:
  279. network[attribute_name] = attribute_value
  280. NETWORKS[engine_name] = new_network(network)
  281. elif isinstance(network, dict):
  282. NETWORKS[engine_name] = new_network(network)
  283. # define networks from engines.[i].network (references)
  284. for engine_name, engine, network in iter_networks():
  285. if isinstance(network, str):
  286. NETWORKS[engine_name] = NETWORKS[network]
  287. @atexit.register
  288. def done():
  289. """Close all HTTP client
  290. Avoid a warning at exit
  291. see https://github.com/encode/httpx/blob/1a6e254f72d9fd5694a1c10a28927e193ab4f76b/httpx/_client.py#L1785
  292. Note: since Network.aclose has to be async, it is not possible to call this method on Network.__del__
  293. So Network.aclose is called here using atexit.register
  294. """
  295. try:
  296. loop = get_loop()
  297. if loop:
  298. future = asyncio.run_coroutine_threadsafe(Network.aclose_all(), loop)
  299. # wait 3 seconds to close the HTTP clients
  300. future.result(3)
  301. finally:
  302. NETWORKS.clear()
  303. NETWORKS[DEFAULT_NAME] = Network()