| 1 | # SPDX-License-Identifier: MIT |
| 2 | from __future__ import absolute_import |
| 3 | import collections |
| 4 | import functools |
| 5 | import logging |
| 6 | |
| 7 | from ._collections import RecentlyUsedContainer |
| 8 | from .connectionpool import HTTPConnectionPool, HTTPSConnectionPool |
| 9 | from .connectionpool import port_by_scheme |
| 10 | from .exceptions import LocationValueError, MaxRetryError, ProxySchemeUnknown |
| 11 | from .packages.six.moves.urllib.parse import urljoin |
| 12 | from .request import RequestMethods |
| 13 | from .util.url import parse_url |
| 14 | from .util.retry import Retry |
| 15 | |
| 16 | |
| 17 | __all__ = ['PoolManager', 'ProxyManager', 'proxy_from_url'] |
| 18 | |
| 19 | |
| 20 | log = logging.getLogger(__name__) |
| 21 | |
| 22 | SSL_KEYWORDS = ('key_file', 'cert_file', 'cert_reqs', 'ca_certs', |
| 23 | 'ssl_version', 'ca_cert_dir', 'ssl_context') |
| 24 | |
| 25 | # All known keyword arguments that could be provided to the pool manager, its |
| 26 | # pools, or the underlying connections. This is used to construct a pool key. |
| 27 | _key_fields = ( |
| 28 | 'key_scheme', # str |
| 29 | 'key_host', # str |
| 30 | 'key_port', # int |
| 31 | 'key_timeout', # int or float or Timeout |
| 32 | 'key_retries', # int or Retry |
| 33 | 'key_strict', # bool |
| 34 | 'key_block', # bool |
| 35 | 'key_source_address', # str |
| 36 | 'key_key_file', # str |
| 37 | 'key_cert_file', # str |
| 38 | 'key_cert_reqs', # str |
| 39 | 'key_ca_certs', # str |
| 40 | 'key_ssl_version', # str |
| 41 | 'key_ca_cert_dir', # str |
| 42 | 'key_ssl_context', # instance of ssl.SSLContext or urllib3.util.ssl_.SSLContext |
| 43 | 'key_maxsize', # int |
| 44 | 'key_headers', # dict |
| 45 | 'key__proxy', # parsed proxy url |
| 46 | 'key__proxy_headers', # dict |
| 47 | 'key_socket_options', # list of (level (int), optname (int), value (int or str)) tuples |
| 48 | 'key__socks_options', # dict |
| 49 | 'key_assert_hostname', # bool or string |
| 50 | 'key_assert_fingerprint', # str |
| 51 | ) |
| 52 | |
| 53 | #: The namedtuple class used to construct keys for the connection pool. |
| 54 | #: All custom key schemes should include the fields in this key at a minimum. |
| 55 | PoolKey = collections.namedtuple('PoolKey', _key_fields) |
| 56 | |
| 57 | |
| 58 | def _default_key_normalizer(key_class, request_context): |
| 59 | """ |
| 60 | Create a pool key out of a request context dictionary. |
| 61 | |
| 62 | According to RFC 3986, both the scheme and host are case-insensitive. |
| 63 | Therefore, this function normalizes both before constructing the pool |
| 64 | key for an HTTPS request. If you wish to change this behaviour, provide |
| 65 | alternate callables to ``key_fn_by_scheme``. |
| 66 | |
| 67 | :param key_class: |
| 68 | The class to use when constructing the key. This should be a namedtuple |
| 69 | with the ``scheme`` and ``host`` keys at a minimum. |
| 70 | :type key_class: namedtuple |
| 71 | :param request_context: |
| 72 | A dictionary-like object that contain the context for a request. |
| 73 | :type request_context: dict |
| 74 | |
| 75 | :return: A namedtuple that can be used as a connection pool key. |
| 76 | :rtype: PoolKey |
| 77 | """ |
| 78 | # Since we mutate the dictionary, make a copy first |
| 79 | context = request_context.copy() |
| 80 | context['scheme'] = context['scheme'].lower() |
| 81 | context['host'] = context['host'].lower() |
| 82 | |
| 83 | # These are both dictionaries and need to be transformed into frozensets |
| 84 | for key in ('headers', '_proxy_headers', '_socks_options'): |
| 85 | if key in context and context[key] is not None: |
| 86 | context[key] = frozenset(context[key].items()) |
| 87 | |
| 88 | # The socket_options key may be a list and needs to be transformed into a |
| 89 | # tuple. |
| 90 | socket_opts = context.get('socket_options') |
| 91 | if socket_opts is not None: |
| 92 | context['socket_options'] = tuple(socket_opts) |
| 93 | |
| 94 | # Map the kwargs to the names in the namedtuple - this is necessary since |
| 95 | # namedtuples can't have fields starting with '_'. |
| 96 | for key in list(context.keys()): |
| 97 | context['key_' + key] = context.pop(key) |
| 98 | |
| 99 | # Default to ``None`` for keys missing from the context |
| 100 | for field in key_class._fields: |
| 101 | if field not in context: |
| 102 | context[field] = None |
| 103 | |
| 104 | return key_class(**context) |
| 105 | |
| 106 | |
| 107 | #: A dictionary that maps a scheme to a callable that creates a pool key. |
| 108 | #: This can be used to alter the way pool keys are constructed, if desired. |
| 109 | #: Each PoolManager makes a copy of this dictionary so they can be configured |
| 110 | #: globally here, or individually on the instance. |
| 111 | key_fn_by_scheme = { |
| 112 | 'http': functools.partial(_default_key_normalizer, PoolKey), |
| 113 | 'https': functools.partial(_default_key_normalizer, PoolKey), |
| 114 | } |
| 115 | |
| 116 | pool_classes_by_scheme = { |
| 117 | 'http': HTTPConnectionPool, |
| 118 | 'https': HTTPSConnectionPool, |
| 119 | } |
| 120 | |
| 121 | |
| 122 | class PoolManager(RequestMethods): |
| 123 | """ |
| 124 | Allows for arbitrary requests while transparently keeping track of |
| 125 | necessary connection pools for you. |
| 126 | |
| 127 | :param num_pools: |
| 128 | Number of connection pools to cache before discarding the least |
| 129 | recently used pool. |
| 130 | |
| 131 | :param headers: |
| 132 | Headers to include with all requests, unless other headers are given |
| 133 | explicitly. |
| 134 | |
| 135 | :param \\**connection_pool_kw: |
| 136 | Additional parameters are used to create fresh |
| 137 | :class:`urllib3.connectionpool.ConnectionPool` instances. |
| 138 | |
| 139 | Example:: |
| 140 | |
| 141 | >>> manager = PoolManager(num_pools=2) |
| 142 | >>> r = manager.request('GET', 'http://google.com/') |
| 143 | >>> r = manager.request('GET', 'http://google.com/mail') |
| 144 | >>> r = manager.request('GET', 'http://yahoo.com/') |
| 145 | >>> len(manager.pools) |
| 146 | 2 |
| 147 | |
| 148 | """ |
| 149 | |
| 150 | proxy = None |
| 151 | |
| 152 | def __init__(self, num_pools=10, headers=None, **connection_pool_kw): |
| 153 | RequestMethods.__init__(self, headers) |
| 154 | self.connection_pool_kw = connection_pool_kw |
| 155 | self.pools = RecentlyUsedContainer(num_pools, |
| 156 | dispose_func=lambda p: p.close()) |
| 157 | |
| 158 | # Locally set the pool classes and keys so other PoolManagers can |
| 159 | # override them. |
| 160 | self.pool_classes_by_scheme = pool_classes_by_scheme |
| 161 | self.key_fn_by_scheme = key_fn_by_scheme.copy() |
| 162 | |
| 163 | def __enter__(self): |
| 164 | return self |
| 165 | |
| 166 | def __exit__(self, exc_type, exc_val, exc_tb): |
| 167 | self.clear() |
| 168 | # Return False to re-raise any potential exceptions |
| 169 | return False |
| 170 | |
| 171 | def _new_pool(self, scheme, host, port, request_context=None): |
| 172 | """ |
| 173 | Create a new :class:`ConnectionPool` based on host, port, scheme, and |
| 174 | any additional pool keyword arguments. |
| 175 | |
| 176 | If ``request_context`` is provided, it is provided as keyword arguments |
| 177 | to the pool class used. This method is used to actually create the |
| 178 | connection pools handed out by :meth:`connection_from_url` and |
| 179 | companion methods. It is intended to be overridden for customization. |
| 180 | """ |
| 181 | pool_cls = self.pool_classes_by_scheme[scheme] |
| 182 | if request_context is None: |
| 183 | request_context = self.connection_pool_kw.copy() |
| 184 | |
| 185 | # Although the context has everything necessary to create the pool, |
| 186 | # this function has historically only used the scheme, host, and port |
| 187 | # in the positional args. When an API change is acceptable these can |
| 188 | # be removed. |
| 189 | for key in ('scheme', 'host', 'port'): |
| 190 | request_context.pop(key, None) |
| 191 | |
| 192 | if scheme == 'http': |
| 193 | for kw in SSL_KEYWORDS: |
| 194 | request_context.pop(kw, None) |
| 195 | |
| 196 | return pool_cls(host, port, **request_context) |
| 197 | |
| 198 | def clear(self): |
| 199 | """ |
| 200 | Empty our store of pools and direct them all to close. |
| 201 | |
| 202 | This will not affect in-flight connections, but they will not be |
| 203 | re-used after completion. |
| 204 | """ |
| 205 | self.pools.clear() |
| 206 | |
| 207 | def connection_from_host(self, host, port=None, scheme='http', pool_kwargs=None): |
| 208 | """ |
| 209 | Get a :class:`ConnectionPool` based on the host, port, and scheme. |
| 210 | |
| 211 | If ``port`` isn't given, it will be derived from the ``scheme`` using |
| 212 | ``urllib3.connectionpool.port_by_scheme``. If ``pool_kwargs`` is |
| 213 | provided, it is merged with the instance's ``connection_pool_kw`` |
| 214 | variable and used to create the new connection pool, if one is |
| 215 | needed. |
| 216 | """ |
| 217 | |
| 218 | if not host: |
| 219 | raise LocationValueError("No host specified.") |
| 220 | |
| 221 | request_context = self._merge_pool_kwargs(pool_kwargs) |
| 222 | request_context['scheme'] = scheme or 'http' |
| 223 | if not port: |
| 224 | port = port_by_scheme.get(request_context['scheme'].lower(), 80) |
| 225 | request_context['port'] = port |
| 226 | request_context['host'] = host |
| 227 | |
| 228 | return self.connection_from_context(request_context) |
| 229 | |
| 230 | def connection_from_context(self, request_context): |
| 231 | """ |
| 232 | Get a :class:`ConnectionPool` based on the request context. |
| 233 | |
| 234 | ``request_context`` must at least contain the ``scheme`` key and its |
| 235 | value must be a key in ``key_fn_by_scheme`` instance variable. |
| 236 | """ |
| 237 | scheme = request_context['scheme'].lower() |
| 238 | pool_key_constructor = self.key_fn_by_scheme[scheme] |
| 239 | pool_key = pool_key_constructor(request_context) |
| 240 | |
| 241 | return self.connection_from_pool_key(pool_key, request_context=request_context) |
| 242 | |
| 243 | def connection_from_pool_key(self, pool_key, request_context=None): |
| 244 | """ |
| 245 | Get a :class:`ConnectionPool` based on the provided pool key. |
| 246 | |
| 247 | ``pool_key`` should be a namedtuple that only contains immutable |
| 248 | objects. At a minimum it must have the ``scheme``, ``host``, and |
| 249 | ``port`` fields. |
| 250 | """ |
| 251 | with self.pools.lock: |
| 252 | # If the scheme, host, or port doesn't match existing open |
| 253 | # connections, open a new ConnectionPool. |
| 254 | pool = self.pools.get(pool_key) |
| 255 | if pool: |
| 256 | return pool |
| 257 | |
| 258 | # Make a fresh ConnectionPool of the desired type |
| 259 | scheme = request_context['scheme'] |
| 260 | host = request_context['host'] |
| 261 | port = request_context['port'] |
| 262 | pool = self._new_pool(scheme, host, port, request_context=request_context) |
| 263 | self.pools[pool_key] = pool |
| 264 | |
| 265 | return pool |
| 266 | |
| 267 | def connection_from_url(self, url, pool_kwargs=None): |
| 268 | """ |
| 269 | Similar to :func:`urllib3.connectionpool.connection_from_url`. |
| 270 | |
| 271 | If ``pool_kwargs`` is not provided and a new pool needs to be |
| 272 | constructed, ``self.connection_pool_kw`` is used to initialize |
| 273 | the :class:`urllib3.connectionpool.ConnectionPool`. If ``pool_kwargs`` |
| 274 | is provided, it is used instead. Note that if a new pool does not |
| 275 | need to be created for the request, the provided ``pool_kwargs`` are |
| 276 | not used. |
| 277 | """ |
| 278 | u = parse_url(url) |
| 279 | return self.connection_from_host(u.host, port=u.port, scheme=u.scheme, |
| 280 | pool_kwargs=pool_kwargs) |
| 281 | |
| 282 | def _merge_pool_kwargs(self, override): |
| 283 | """ |
| 284 | Merge a dictionary of override values for self.connection_pool_kw. |
| 285 | |
| 286 | This does not modify self.connection_pool_kw and returns a new dict. |
| 287 | Any keys in the override dictionary with a value of ``None`` are |
| 288 | removed from the merged dictionary. |
| 289 | """ |
| 290 | base_pool_kwargs = self.connection_pool_kw.copy() |
| 291 | if override: |
| 292 | for key, value in override.items(): |
| 293 | if value is None: |
| 294 | try: |
| 295 | del base_pool_kwargs[key] |
| 296 | except KeyError: |
| 297 | pass |
| 298 | else: |
| 299 | base_pool_kwargs[key] = value |
| 300 | return base_pool_kwargs |
| 301 | |
| 302 | def urlopen(self, method, url, redirect=True, **kw): |
| 303 | """ |
| 304 | Same as :meth:`urllib3.connectionpool.HTTPConnectionPool.urlopen` |
| 305 | with custom cross-host redirect logic and only sends the request-uri |
| 306 | portion of the ``url``. |
| 307 | |
| 308 | The given ``url`` parameter must be absolute, such that an appropriate |
| 309 | :class:`urllib3.connectionpool.ConnectionPool` can be chosen for it. |
| 310 | """ |
| 311 | u = parse_url(url) |
| 312 | conn = self.connection_from_host(u.host, port=u.port, scheme=u.scheme) |
| 313 | |
| 314 | kw['assert_same_host'] = False |
| 315 | kw['redirect'] = False |
| 316 | if 'headers' not in kw: |
| 317 | kw['headers'] = self.headers |
| 318 | |
| 319 | if self.proxy is not None and u.scheme == "http": |
| 320 | response = conn.urlopen(method, url, **kw) |
| 321 | else: |
| 322 | response = conn.urlopen(method, u.request_uri, **kw) |
| 323 | |
| 324 | redirect_location = redirect and response.get_redirect_location() |
| 325 | if not redirect_location: |
| 326 | return response |
| 327 | |
| 328 | # Support relative URLs for redirecting. |
| 329 | redirect_location = urljoin(url, redirect_location) |
| 330 | |
| 331 | # RFC 7231, Section 6.4.4 |
| 332 | if response.status == 303: |
| 333 | method = 'GET' |
| 334 | |
| 335 | retries = kw.get('retries') |
| 336 | if not isinstance(retries, Retry): |
| 337 | retries = Retry.from_int(retries, redirect=redirect) |
| 338 | |
| 339 | try: |
| 340 | retries = retries.increment(method, url, response=response, _pool=conn) |
| 341 | except MaxRetryError: |
| 342 | if retries.raise_on_redirect: |
| 343 | raise |
| 344 | return response |
| 345 | |
| 346 | kw['retries'] = retries |
| 347 | kw['redirect'] = redirect |
| 348 | |
| 349 | log.info("Redirecting %s -> %s", url, redirect_location) |
| 350 | return self.urlopen(method, redirect_location, **kw) |
| 351 | |
| 352 | |
| 353 | class ProxyManager(PoolManager): |
| 354 | """ |
| 355 | Behaves just like :class:`PoolManager`, but sends all requests through |
| 356 | the defined proxy, using the CONNECT method for HTTPS URLs. |
| 357 | |
| 358 | :param proxy_url: |
| 359 | The URL of the proxy to be used. |
| 360 | |
| 361 | :param proxy_headers: |
| 362 | A dictionary contaning headers that will be sent to the proxy. In case |
| 363 | of HTTP they are being sent with each request, while in the |
| 364 | HTTPS/CONNECT case they are sent only once. Could be used for proxy |
| 365 | authentication. |
| 366 | |
| 367 | Example: |
| 368 | >>> proxy = urllib3.ProxyManager('http://localhost:3128/') |
| 369 | >>> r1 = proxy.request('GET', 'http://google.com/') |
| 370 | >>> r2 = proxy.request('GET', 'http://httpbin.org/') |
| 371 | >>> len(proxy.pools) |
| 372 | 1 |
| 373 | >>> r3 = proxy.request('GET', 'https://httpbin.org/') |
| 374 | >>> r4 = proxy.request('GET', 'https://twitter.com/') |
| 375 | >>> len(proxy.pools) |
| 376 | 3 |
| 377 | |
| 378 | """ |
| 379 | |
| 380 | def __init__(self, proxy_url, num_pools=10, headers=None, |
| 381 | proxy_headers=None, **connection_pool_kw): |
| 382 | |
| 383 | if isinstance(proxy_url, HTTPConnectionPool): |
| 384 | proxy_url = '%s://%s:%i' % (proxy_url.scheme, proxy_url.host, |
| 385 | proxy_url.port) |
| 386 | proxy = parse_url(proxy_url) |
| 387 | if not proxy.port: |
| 388 | port = port_by_scheme.get(proxy.scheme, 80) |
| 389 | proxy = proxy._replace(port=port) |
| 390 | |
| 391 | if proxy.scheme not in ("http", "https"): |
| 392 | raise ProxySchemeUnknown(proxy.scheme) |
| 393 | |
| 394 | self.proxy = proxy |
| 395 | self.proxy_headers = proxy_headers or {} |
| 396 | |
| 397 | connection_pool_kw['_proxy'] = self.proxy |
| 398 | connection_pool_kw['_proxy_headers'] = self.proxy_headers |
| 399 | |
| 400 | super(ProxyManager, self).__init__( |
| 401 | num_pools, headers, **connection_pool_kw) |
| 402 | |
| 403 | def connection_from_host(self, host, port=None, scheme='http', pool_kwargs=None): |
| 404 | if scheme == "https": |
| 405 | return super(ProxyManager, self).connection_from_host( |
| 406 | host, port, scheme, pool_kwargs=pool_kwargs) |
| 407 | |
| 408 | return super(ProxyManager, self).connection_from_host( |
| 409 | self.proxy.host, self.proxy.port, self.proxy.scheme, pool_kwargs=pool_kwargs) |
| 410 | |
| 411 | def _set_proxy_headers(self, url, headers=None): |
| 412 | """ |
| 413 | Sets headers needed by proxies: specifically, the Accept and Host |
| 414 | headers. Only sets headers not provided by the user. |
| 415 | """ |
| 416 | headers_ = {'Accept': '*/*'} |
| 417 | |
| 418 | netloc = parse_url(url).netloc |
| 419 | if netloc: |
| 420 | headers_['Host'] = netloc |
| 421 | |
| 422 | if headers: |
| 423 | headers_.update(headers) |
| 424 | return headers_ |
| 425 | |
| 426 | def urlopen(self, method, url, redirect=True, **kw): |
| 427 | "Same as HTTP(S)ConnectionPool.urlopen, ``url`` must be absolute." |
| 428 | u = parse_url(url) |
| 429 | |
| 430 | if u.scheme == "http": |
| 431 | # For proxied HTTPS requests, httplib sets the necessary headers |
| 432 | # on the CONNECT to the proxy. For HTTP, we'll definitely |
| 433 | # need to set 'Host' at the very least. |
| 434 | headers = kw.get('headers', self.headers) |
| 435 | kw['headers'] = self._set_proxy_headers(url, headers) |
| 436 | |
| 437 | return super(ProxyManager, self).urlopen(method, url, redirect=redirect, **kw) |
| 438 | |
| 439 | |
| 440 | def proxy_from_url(url, **kw): |
| 441 | return ProxyManager(proxy_url=url, **kw) |