network.py 11 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314
  1. # SPDX-License-Identifier: AGPL-3.0-or-later
  2. # lint: pylint
  3. # pylint: disable=global-statement
  4. # pylint: disable=missing-module-docstring, missing-class-docstring, missing-function-docstring
  5. import atexit
  6. import asyncio
  7. import ipaddress
  8. from itertools import cycle
  9. import httpx
  10. from .client import new_client, get_loop
  11. DEFAULT_NAME = '__DEFAULT__'
  12. NETWORKS = {}
  13. # requests compatibility when reading proxy settings from settings.yml
  14. PROXY_PATTERN_MAPPING = {
  15. 'http': 'http://',
  16. 'https': 'https://',
  17. 'socks4': 'socks4://',
  18. 'socks5': 'socks5://',
  19. 'socks5h': 'socks5h://',
  20. 'http:': 'http://',
  21. 'https:': 'https://',
  22. 'socks4:': 'socks4://',
  23. 'socks5:': 'socks5://',
  24. 'socks5h:': 'socks5h://',
  25. }
  26. ADDRESS_MAPPING = {
  27. 'ipv4': '0.0.0.0',
  28. 'ipv6': '::'
  29. }
  30. class Network:
  31. __slots__ = (
  32. 'enable_http', 'verify', 'enable_http2',
  33. 'max_connections', 'max_keepalive_connections', 'keepalive_expiry',
  34. 'local_addresses', 'proxies', 'max_redirects', 'retries', 'retry_on_http_error',
  35. '_local_addresses_cycle', '_proxies_cycle', '_clients'
  36. )
  37. def __init__(
  38. # pylint: disable=too-many-arguments
  39. self,
  40. enable_http=True,
  41. verify=True,
  42. enable_http2=False,
  43. max_connections=None,
  44. max_keepalive_connections=None,
  45. keepalive_expiry=None,
  46. proxies=None,
  47. local_addresses=None,
  48. retries=0,
  49. retry_on_http_error=None,
  50. max_redirects=30 ):
  51. self.enable_http = enable_http
  52. self.verify = verify
  53. self.enable_http2 = enable_http2
  54. self.max_connections = max_connections
  55. self.max_keepalive_connections = max_keepalive_connections
  56. self.keepalive_expiry = keepalive_expiry
  57. self.proxies = proxies
  58. self.local_addresses = local_addresses
  59. self.retries = retries
  60. self.retry_on_http_error = retry_on_http_error
  61. self.max_redirects = max_redirects
  62. self._local_addresses_cycle = self.get_ipaddress_cycle()
  63. self._proxies_cycle = self.get_proxy_cycles()
  64. self._clients = {}
  65. self.check_parameters()
  66. def check_parameters(self):
  67. for address in self.iter_ipaddresses():
  68. if '/' in address:
  69. ipaddress.ip_network(address, False)
  70. else:
  71. ipaddress.ip_address(address)
  72. if self.proxies is not None and not isinstance(self.proxies, (str, dict)):
  73. raise ValueError('proxies type has to be str, dict or None')
  74. def iter_ipaddresses(self):
  75. local_addresses = self.local_addresses
  76. if not local_addresses:
  77. return
  78. if isinstance(local_addresses, str):
  79. local_addresses = [local_addresses]
  80. for address in local_addresses:
  81. yield address
  82. def get_ipaddress_cycle(self):
  83. while True:
  84. count = 0
  85. for address in self.iter_ipaddresses():
  86. if '/' in address:
  87. for a in ipaddress.ip_network(address, False).hosts():
  88. yield str(a)
  89. count += 1
  90. else:
  91. a = ipaddress.ip_address(address)
  92. yield str(a)
  93. count += 1
  94. if count == 0:
  95. yield None
  96. def iter_proxies(self):
  97. if not self.proxies:
  98. return
  99. # https://www.python-httpx.org/compatibility/#proxy-keys
  100. if isinstance(self.proxies, str):
  101. yield 'all://', [self.proxies]
  102. else:
  103. for pattern, proxy_url in self.proxies.items():
  104. pattern = PROXY_PATTERN_MAPPING.get(pattern, pattern)
  105. if isinstance(proxy_url, str):
  106. proxy_url = [proxy_url]
  107. yield pattern, proxy_url
  108. def get_proxy_cycles(self):
  109. proxy_settings = {}
  110. for pattern, proxy_urls in self.iter_proxies():
  111. proxy_settings[pattern] = cycle(proxy_urls)
  112. while True:
  113. # pylint: disable=stop-iteration-return
  114. yield tuple((pattern, next(proxy_url_cycle)) for pattern, proxy_url_cycle in proxy_settings.items())
  115. def get_client(self, verify=None, max_redirects=None):
  116. verify = self.verify if verify is None else verify
  117. max_redirects = self.max_redirects if max_redirects is None else max_redirects
  118. local_address = next(self._local_addresses_cycle)
  119. proxies = next(self._proxies_cycle) # is a tuple so it can be part of the key
  120. key = (verify, max_redirects, local_address, proxies)
  121. if key not in self._clients or self._clients[key].is_closed:
  122. self._clients[key] = new_client(
  123. self.enable_http,
  124. verify,
  125. self.enable_http2,
  126. self.max_connections,
  127. self.max_keepalive_connections,
  128. self.keepalive_expiry,
  129. dict(proxies),
  130. local_address,
  131. 0,
  132. max_redirects
  133. )
  134. return self._clients[key]
  135. async def aclose(self):
  136. async def close_client(client):
  137. try:
  138. await client.aclose()
  139. except httpx.HTTPError:
  140. pass
  141. await asyncio.gather(*[close_client(client) for client in self._clients.values()], return_exceptions=False)
  142. @staticmethod
  143. def get_kwargs_clients(kwargs):
  144. kwargs_clients = {}
  145. if 'verify' in kwargs:
  146. kwargs_clients['verify'] = kwargs.pop('verify')
  147. if 'max_redirects' in kwargs:
  148. kwargs_clients['max_redirects'] = kwargs.pop('max_redirects')
  149. return kwargs_clients
  150. def is_valid_respones(self, response):
  151. # pylint: disable=too-many-boolean-expressions
  152. if ((self.retry_on_http_error is True and 400 <= response.status_code <= 599)
  153. or (isinstance(self.retry_on_http_error, list) and response.status_code in self.retry_on_http_error)
  154. or (isinstance(self.retry_on_http_error, int) and response.status_code == self.retry_on_http_error)
  155. ):
  156. return False
  157. return True
  158. async def request(self, method, url, **kwargs):
  159. retries = self.retries
  160. while retries >= 0: # pragma: no cover
  161. kwargs_clients = Network.get_kwargs_clients(kwargs)
  162. client = self.get_client(**kwargs_clients)
  163. try:
  164. response = await client.request(method, url, **kwargs)
  165. if self.is_valid_respones(response) or retries <= 0:
  166. return response
  167. except (httpx.RequestError, httpx.HTTPStatusError) as e:
  168. if retries <= 0:
  169. raise e
  170. retries -= 1
  171. def stream(self, method, url, **kwargs):
  172. retries = self.retries
  173. while retries >= 0: # pragma: no cover
  174. kwargs_clients = Network.get_kwargs_clients(kwargs)
  175. client = self.get_client(**kwargs_clients)
  176. try:
  177. response = client.stream(method, url, **kwargs)
  178. if self.is_valid_respones(response) or retries <= 0:
  179. return response
  180. except (httpx.RequestError, httpx.HTTPStatusError) as e:
  181. if retries <= 0:
  182. raise e
  183. retries -= 1
  184. @classmethod
  185. async def aclose_all(cls):
  186. global NETWORKS
  187. await asyncio.gather(*[network.aclose() for network in NETWORKS.values()], return_exceptions=False)
  188. def get_network(name=None):
  189. global NETWORKS
  190. return NETWORKS.get(name or DEFAULT_NAME)
  191. def initialize(settings_engines=None, settings_outgoing=None):
  192. # pylint: disable=import-outside-toplevel)
  193. from searx.engines import engines
  194. from searx import settings
  195. # pylint: enable=import-outside-toplevel)
  196. global NETWORKS
  197. settings_engines = settings_engines or settings['engines']
  198. settings_outgoing = settings_outgoing or settings['outgoing']
  199. # default parameters for AsyncHTTPTransport
  200. # see https://github.com/encode/httpx/blob/e05a5372eb6172287458b37447c30f650047e1b8/httpx/_transports/default.py#L108-L121 # pylint: disable=line-too-long
  201. default_params = {
  202. 'enable_http': False,
  203. 'verify': True,
  204. 'enable_http2': settings_outgoing['enable_http2'],
  205. 'max_connections': settings_outgoing['pool_connections'],
  206. 'max_keepalive_connections': settings_outgoing['pool_maxsize'],
  207. 'keepalive_expiry': settings_outgoing['keepalive_expiry'],
  208. 'local_addresses': settings_outgoing['source_ips'],
  209. 'proxies': settings_outgoing['proxies'],
  210. 'max_redirects': settings_outgoing['max_redirects'],
  211. 'retries': settings_outgoing['retries'],
  212. 'retry_on_http_error': None,
  213. }
  214. def new_network(params):
  215. nonlocal default_params
  216. result = {}
  217. result.update(default_params)
  218. result.update(params)
  219. return Network(**result)
  220. def iter_networks():
  221. nonlocal settings_engines
  222. for engine_spec in settings_engines:
  223. engine_name = engine_spec['name']
  224. engine = engines.get(engine_name)
  225. if engine is None:
  226. continue
  227. network = getattr(engine, 'network', None)
  228. yield engine_name, engine, network
  229. if NETWORKS:
  230. done()
  231. NETWORKS.clear()
  232. NETWORKS[DEFAULT_NAME] = new_network({})
  233. NETWORKS['ipv4'] = new_network({'local_addresses': '0.0.0.0'})
  234. NETWORKS['ipv6'] = new_network({'local_addresses': '::'})
  235. # define networks from outgoing.networks
  236. for network_name, network in settings_outgoing['networks'].items():
  237. NETWORKS[network_name] = new_network(network)
  238. # define networks from engines.[i].network (except references)
  239. for engine_name, engine, network in iter_networks():
  240. if network is None:
  241. network = {}
  242. for attribute_name, attribute_value in default_params.items():
  243. if hasattr(engine, attribute_name):
  244. network[attribute_name] = getattr(engine, attribute_name)
  245. else:
  246. network[attribute_name] = attribute_value
  247. NETWORKS[engine_name] = new_network(network)
  248. elif isinstance(network, dict):
  249. NETWORKS[engine_name] = new_network(network)
  250. # define networks from engines.[i].network (references)
  251. for engine_name, engine, network in iter_networks():
  252. if isinstance(network, str):
  253. NETWORKS[engine_name] = NETWORKS[network]
  254. @atexit.register
  255. def done():
  256. """Close all HTTP client
  257. Avoid a warning at exit
  258. see https://github.com/encode/httpx/blob/1a6e254f72d9fd5694a1c10a28927e193ab4f76b/httpx/_client.py#L1785
  259. Note: since Network.aclose has to be async, it is not possible to call this method on Network.__del__
  260. So Network.aclose is called here using atexit.register
  261. """
  262. global NETWORKS
  263. try:
  264. loop = get_loop()
  265. if loop:
  266. future = asyncio.run_coroutine_threadsafe(Network.aclose_all(), loop)
  267. # wait 3 seconds to close the HTTP clients
  268. future.result(3)
  269. finally:
  270. NETWORKS.clear()
  271. NETWORKS[DEFAULT_NAME] = Network()