network.py 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315
  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__ = ('enable_http', 'verify', 'enable_http2',
  32. 'max_connections', 'max_keepalive_connections', 'keepalive_expiry',
  33. 'local_addresses', 'proxies', 'max_redirects', 'retries', 'retry_on_http_error',
  34. '_local_addresses_cycle', '_proxies_cycle', '_clients')
  35. def __init__(
  36. # pylint: disable=too-many-arguments
  37. self,
  38. enable_http=True,
  39. verify=True,
  40. enable_http2=False,
  41. max_connections=None,
  42. max_keepalive_connections=None,
  43. keepalive_expiry=None,
  44. proxies=None,
  45. local_addresses=None,
  46. retries=0,
  47. retry_on_http_error=None,
  48. max_redirects=30 ):
  49. self.enable_http = enable_http
  50. self.verify = verify
  51. self.enable_http2 = enable_http2
  52. self.max_connections = max_connections
  53. self.max_keepalive_connections = max_keepalive_connections
  54. self.keepalive_expiry = keepalive_expiry
  55. self.proxies = proxies
  56. self.local_addresses = local_addresses
  57. self.retries = retries
  58. self.retry_on_http_error = retry_on_http_error
  59. self.max_redirects = max_redirects
  60. self._local_addresses_cycle = self.get_ipaddress_cycle()
  61. self._proxies_cycle = self.get_proxy_cycles()
  62. self._clients = {}
  63. self.check_parameters()
  64. def check_parameters(self):
  65. for address in self.iter_ipaddresses():
  66. if '/' in address:
  67. ipaddress.ip_network(address, False)
  68. else:
  69. ipaddress.ip_address(address)
  70. if self.proxies is not None and not isinstance(self.proxies, (str, dict)):
  71. raise ValueError('proxies type has to be str, dict or None')
  72. def iter_ipaddresses(self):
  73. local_addresses = self.local_addresses
  74. if not local_addresses:
  75. return
  76. if isinstance(local_addresses, str):
  77. local_addresses = [local_addresses]
  78. for address in local_addresses:
  79. yield address
  80. def get_ipaddress_cycle(self):
  81. while True:
  82. count = 0
  83. for address in self.iter_ipaddresses():
  84. if '/' in address:
  85. for a in ipaddress.ip_network(address, False).hosts():
  86. yield str(a)
  87. count += 1
  88. else:
  89. a = ipaddress.ip_address(address)
  90. yield str(a)
  91. count += 1
  92. if count == 0:
  93. yield None
  94. def iter_proxies(self):
  95. if not self.proxies:
  96. return
  97. # https://www.python-httpx.org/compatibility/#proxy-keys
  98. if isinstance(self.proxies, str):
  99. yield 'all://', [self.proxies]
  100. else:
  101. for pattern, proxy_url in self.proxies.items():
  102. pattern = PROXY_PATTERN_MAPPING.get(pattern, pattern)
  103. if isinstance(proxy_url, str):
  104. proxy_url = [proxy_url]
  105. yield pattern, proxy_url
  106. def get_proxy_cycles(self):
  107. proxy_settings = {}
  108. for pattern, proxy_urls in self.iter_proxies():
  109. proxy_settings[pattern] = cycle(proxy_urls)
  110. while True:
  111. # pylint: disable=stop-iteration-return
  112. yield tuple((pattern, next(proxy_url_cycle)) for pattern, proxy_url_cycle in proxy_settings.items())
  113. def get_client(self, verify=None, max_redirects=None):
  114. verify = self.verify if verify is None else verify
  115. max_redirects = self.max_redirects if max_redirects is None else max_redirects
  116. local_address = next(self._local_addresses_cycle)
  117. proxies = next(self._proxies_cycle) # is a tuple so it can be part of the key
  118. key = (verify, max_redirects, local_address, proxies)
  119. if key not in self._clients or self._clients[key].is_closed:
  120. self._clients[key] = new_client(self.enable_http,
  121. verify,
  122. self.enable_http2,
  123. self.max_connections,
  124. self.max_keepalive_connections,
  125. self.keepalive_expiry,
  126. dict(proxies),
  127. local_address,
  128. 0,
  129. max_redirects)
  130. return self._clients[key]
  131. async def aclose(self):
  132. async def close_client(client):
  133. try:
  134. await client.aclose()
  135. except httpx.HTTPError:
  136. pass
  137. await asyncio.gather(*[close_client(client) for client in self._clients.values()], return_exceptions=False)
  138. @staticmethod
  139. def get_kwargs_clients(kwargs):
  140. kwargs_clients = {}
  141. if 'verify' in kwargs:
  142. kwargs_clients['verify'] = kwargs.pop('verify')
  143. if 'max_redirects' in kwargs:
  144. kwargs_clients['max_redirects'] = kwargs.pop('max_redirects')
  145. return kwargs_clients
  146. def is_valid_respones(self, response):
  147. # pylint: disable=too-many-boolean-expressions
  148. if (self.retry_on_http_error is True and 400 <= response.status_code <= 599) \
  149. or (isinstance(self.retry_on_http_error, list) and response.status_code in self.retry_on_http_error) \
  150. or (isinstance(self.retry_on_http_error, int) and response.status_code == self.retry_on_http_error):
  151. return False
  152. return True
  153. async def request(self, method, url, **kwargs):
  154. retries = self.retries
  155. while retries >= 0: # pragma: no cover
  156. kwargs_clients = Network.get_kwargs_clients(kwargs)
  157. client = self.get_client(**kwargs_clients)
  158. try:
  159. response = await client.request(method, url, **kwargs)
  160. if self.is_valid_respones(response) or retries <= 0:
  161. return response
  162. except (httpx.RequestError, httpx.HTTPStatusError) as e:
  163. if retries <= 0:
  164. raise e
  165. retries -= 1
  166. def stream(self, method, url, **kwargs):
  167. retries = self.retries
  168. while retries >= 0: # pragma: no cover
  169. kwargs_clients = Network.get_kwargs_clients(kwargs)
  170. client = self.get_client(**kwargs_clients)
  171. try:
  172. response = client.stream(method, url, **kwargs)
  173. if self.is_valid_respones(response) or retries <= 0:
  174. return response
  175. except (httpx.RequestError, httpx.HTTPStatusError) as e:
  176. if retries <= 0:
  177. raise e
  178. retries -= 1
  179. @classmethod
  180. async def aclose_all(cls):
  181. global NETWORKS
  182. await asyncio.gather(*[network.aclose() for network in NETWORKS.values()], return_exceptions=False)
  183. def get_network(name=None):
  184. global NETWORKS
  185. return NETWORKS.get(name or DEFAULT_NAME)
  186. def initialize(settings_engines=None, settings_outgoing=None):
  187. # pylint: disable=import-outside-toplevel)
  188. from searx.engines import engines
  189. from searx import settings
  190. # pylint: enable=import-outside-toplevel)
  191. global NETWORKS
  192. settings_engines = settings_engines or settings.get('engines')
  193. settings_outgoing = settings_outgoing or settings.get('outgoing')
  194. # default parameters for AsyncHTTPTransport
  195. # see https://github.com/encode/httpx/blob/e05a5372eb6172287458b37447c30f650047e1b8/httpx/_transports/default.py#L108-L121 # pylint: disable=line-too-long
  196. default_params = {
  197. 'enable_http': False,
  198. 'verify': True,
  199. 'enable_http2': settings_outgoing.get('enable_http2', True),
  200. # Magic number kept from previous code
  201. 'max_connections': settings_outgoing.get('pool_connections', 100),
  202. # Picked from constructor
  203. 'max_keepalive_connections': settings_outgoing.get('pool_maxsize', 10),
  204. #
  205. 'keepalive_expiry': settings_outgoing.get('keepalive_expiry', 5.0),
  206. 'local_addresses': settings_outgoing.get('source_ips'),
  207. 'proxies': settings_outgoing.get('proxies'),
  208. # default maximum redirect
  209. # from https://github.com/psf/requests/blob/8c211a96cdbe9fe320d63d9e1ae15c5c07e179f8/requests/models.py#L55
  210. 'max_redirects': settings_outgoing.get('max_redirects', 30),
  211. #
  212. 'retries': settings_outgoing.get('retries', 0),
  213. 'retry_on_http_error': None,
  214. }
  215. def new_network(params):
  216. nonlocal default_params
  217. result = {}
  218. result.update(default_params)
  219. result.update(params)
  220. return Network(**result)
  221. def iter_networks():
  222. nonlocal settings_engines
  223. for engine_spec in settings_engines:
  224. engine_name = engine_spec['name']
  225. engine = engines.get(engine_name)
  226. if engine is None:
  227. continue
  228. network = getattr(engine, 'network', None)
  229. yield engine_name, engine, network
  230. if NETWORKS:
  231. done()
  232. NETWORKS.clear()
  233. NETWORKS[DEFAULT_NAME] = new_network({})
  234. NETWORKS['ipv4'] = new_network({'local_addresses': '0.0.0.0'})
  235. NETWORKS['ipv6'] = new_network({'local_addresses': '::'})
  236. # define networks from outgoing.networks
  237. for network_name, network in settings_outgoing.get('networks', {}).items():
  238. NETWORKS[network_name] = new_network(network)
  239. # define networks from engines.[i].network (except references)
  240. for engine_name, engine, network in iter_networks():
  241. if network is None:
  242. network = {}
  243. for attribute_name, attribute_value in default_params.items():
  244. if hasattr(engine, attribute_name):
  245. network[attribute_name] = getattr(engine, attribute_name)
  246. else:
  247. network[attribute_name] = attribute_value
  248. NETWORKS[engine_name] = new_network(network)
  249. elif isinstance(network, dict):
  250. NETWORKS[engine_name] = new_network(network)
  251. # define networks from engines.[i].network (references)
  252. for engine_name, engine, network in iter_networks():
  253. if isinstance(network, str):
  254. NETWORKS[engine_name] = NETWORKS[network]
  255. @atexit.register
  256. def done():
  257. """Close all HTTP client
  258. Avoid a warning at exit
  259. see https://github.com/encode/httpx/blob/1a6e254f72d9fd5694a1c10a28927e193ab4f76b/httpx/_client.py#L1785
  260. Note: since Network.aclose has to be async, it is not possible to call this method on Network.__del__
  261. So Network.aclose is called here using atexit.register
  262. """
  263. global NETWORKS
  264. try:
  265. loop = get_loop()
  266. if loop:
  267. future = asyncio.run_coroutine_threadsafe(Network.aclose_all(), loop)
  268. # wait 3 seconds to close the HTTP clients
  269. future.result(3)
  270. finally:
  271. NETWORKS.clear()
  272. NETWORKS[DEFAULT_NAME] = Network()