Learn more  » Push, build, and install  RubyGems npm packages Python packages Maven artifacts PHP packages Go Modules Bower components Debian packages RPM packages NuGet packages

hemamaps / Scrapy   python

Repository URL to install this package:

Version: 1.1.1 

/ xlib / tx / client.py

# -*- test-case-name: twisted.web.test.test_webclient,twisted.web.test.test_agent -*-
# Copyright (c) Twisted Matrix Laboratories.
# See LICENSE for details.

"""
HTTP client.
"""

from __future__ import division, absolute_import

import os

try:
    from urlparse import urlunparse
    from urllib import splithost, splittype
except ImportError:
    from urllib.parse import splithost, splittype
    from urllib.parse import urlunparse as _urlunparse

    def urlunparse(parts):
        result = _urlunparse(tuple([p.decode("charmap") for p in parts]))
        return result.encode("charmap")
import zlib

from zope.interface import implementer

from twisted.python import log
from twisted.python.failure import Failure
from twisted.web import http
from twisted.internet import defer, protocol, task, reactor
from twisted.internet.interfaces import IProtocol
from twisted.python import failure
from twisted.python.components import proxyForInterface
from twisted.web import error
from twisted.web.http_headers import Headers

from .endpoints import TCP4ClientEndpoint, SSL4ClientEndpoint
from .iweb import IResponse, UNKNOWN_LENGTH, IBodyProducer


class PartialDownloadError(error.Error):
    """
    Page was only partially downloaded, we got disconnected in middle.

    @ivar response: All of the response body which was downloaded.
    """


class _URL(tuple):
    """
    A parsed URL.

    At some point this should be replaced with a better URL implementation.
    """
    def __new__(self, scheme, host, port, path):
        return tuple.__new__(_URL, (scheme, host, port, path))


    def __init__(self, scheme, host, port, path):
        self.scheme = scheme
        self.host = host
        self.port = port
        self.path = path


def _parse(url, defaultPort=None):
    """
    Split the given URL into the scheme, host, port, and path.

    @type url: C{bytes}
    @param url: An URL to parse.

    @type defaultPort: C{int} or C{None}
    @param defaultPort: An alternate value to use as the port if the URL does
    not include one.

    @return: A four-tuple of the scheme, host, port, and path of the URL.  All
    of these are C{bytes} instances except for port, which is an C{int}.
    """
    url = url.strip()
    parsed = http.urlparse(url)
    scheme = parsed[0]
    path = urlunparse((b'', b'') + parsed[2:])

    if defaultPort is None:
        if scheme == b'https':
            defaultPort = 443
        else:
            defaultPort = 80

    host, port = parsed[1], defaultPort
    if b':' in host:
        host, port = host.split(b':')
        try:
            port = int(port)
        except ValueError:
            port = defaultPort

    if path == b'':
        path = b'/'

    return _URL(scheme, host, port, path)


def _makeGetterFactory(url, factoryFactory, contextFactory=None,
                       *args, **kwargs):
    """
    Create and connect an HTTP page getting factory.

    Any additional positional or keyword arguments are used when calling
    C{factoryFactory}.

    @param factoryFactory: Factory factory that is called with C{url}, C{args}
        and C{kwargs} to produce the getter

    @param contextFactory: Context factory to use when creating a secure
        connection, defaulting to C{None}

    @return: The factory created by C{factoryFactory}
    """
    scheme, host, port, path = _parse(url)
    factory = factoryFactory(url, *args, **kwargs)
    if scheme == b'https':
        from twisted.internet import ssl
        if contextFactory is None:
            contextFactory = ssl.ClientContextFactory()
        reactor.connectSSL(host, port, factory, contextFactory)
    else:
        reactor.connectTCP(host, port, factory)
    return factory


# The code which follows is based on the new HTTP client implementation.  It
# should be significantly better than anything above, though it is not yet
# feature equivalent.

from twisted.web.error import SchemeNotSupported
from ._newclient import Request, Response, HTTP11ClientProtocol
from ._newclient import ResponseDone, ResponseFailed
from ._newclient import RequestNotSent, RequestTransmissionFailed
from ._newclient import (
    ResponseNeverReceived, PotentialDataLoss, _WrapperException)

try:
    from twisted.internet.ssl import ClientContextFactory
except ImportError:
    class WebClientContextFactory(object):
        """
        A web context factory which doesn't work because the necessary SSL
        support is missing.
        """
        def getContext(self, hostname, port):
            raise NotImplementedError("SSL support unavailable")
else:
    class WebClientContextFactory(ClientContextFactory):
        """
        A web context factory which ignores the hostname and port and does no
        certificate verification.
        """
        def getContext(self, hostname, port):
            return ClientContextFactory.getContext(self)



class _WebToNormalContextFactory(object):
    """
    Adapt a web context factory to a normal context factory.

    @ivar _webContext: A web context factory which accepts a hostname and port
        number to its C{getContext} method.

    @ivar _hostname: The hostname which will be passed to
        C{_webContext.getContext}.

    @ivar _port: The port number which will be passed to
        C{_webContext.getContext}.
    """
    def __init__(self, webContext, hostname, port):
        self._webContext = webContext
        self._hostname = hostname
        self._port = port


    def getContext(self):
        """
        Called the wrapped web context factory's C{getContext} method with a
        hostname and port number and return the resulting context object.
        """
        return self._webContext.getContext(self._hostname, self._port)



@implementer(IBodyProducer)
class FileBodyProducer(object):
    """
    L{FileBodyProducer} produces bytes from an input file object incrementally
    and writes them to a consumer.

    Since file-like objects cannot be read from in an event-driven manner,
    L{FileBodyProducer} uses a L{Cooperator} instance to schedule reads from
    the file.  This process is also paused and resumed based on notifications
    from the L{IConsumer} provider being written to.

    The file is closed after it has been read, or if the producer is stopped
    early.

    @ivar _inputFile: Any file-like object, bytes read from which will be
        written to a consumer.

    @ivar _cooperate: A method like L{Cooperator.cooperate} which is used to
        schedule all reads.

    @ivar _readSize: The number of bytes to read from C{_inputFile} at a time.
    """

    # Python 2.4 doesn't have these symbolic constants
    _SEEK_SET = getattr(os, 'SEEK_SET', 0)
    _SEEK_END = getattr(os, 'SEEK_END', 2)

    def __init__(self, inputFile, cooperator=task, readSize=2 ** 16):
        self._inputFile = inputFile
        self._cooperate = cooperator.cooperate
        self._readSize = readSize
        self.length = self._determineLength(inputFile)


    def _determineLength(self, fObj):
        """
        Determine how many bytes can be read out of C{fObj} (assuming it is not
        modified from this point on).  If the determination cannot be made,
        return C{UNKNOWN_LENGTH}.
        """
        try:
            seek = fObj.seek
            tell = fObj.tell
        except AttributeError:
            return UNKNOWN_LENGTH
        originalPosition = tell()
        seek(0, self._SEEK_END)
        end = tell()
        seek(originalPosition, self._SEEK_SET)
        return end - originalPosition


    def stopProducing(self):
        """
        Permanently stop writing bytes from the file to the consumer by
        stopping the underlying L{CooperativeTask}.
        """
        self._inputFile.close()
        self._task.stop()


    def startProducing(self, consumer):
        """
        Start a cooperative task which will read bytes from the input file and
        write them to C{consumer}.  Return a L{Deferred} which fires after all
        bytes have been written.

        @param consumer: Any L{IConsumer} provider
        """
        self._task = self._cooperate(self._writeloop(consumer))
        d = self._task.whenDone()
        def maybeStopped(reason):
            # IBodyProducer.startProducing's Deferred isn't support to fire if
            # stopProducing is called.
            reason.trap(task.TaskStopped)
            return defer.Deferred()
        d.addCallbacks(lambda ignored: None, maybeStopped)
        return d


    def _writeloop(self, consumer):
        """
        Return an iterator which reads one chunk of bytes from the input file
        and writes them to the consumer for each time it is iterated.
        """
        while True:
            bytes = self._inputFile.read(self._readSize)
            if not bytes:
                self._inputFile.close()
                break
            consumer.write(bytes)
            yield None


    def pauseProducing(self):
        """
        Temporarily suspend copying bytes from the input file to the consumer
        by pausing the L{CooperativeTask} which drives that activity.
        """
        self._task.pause()


    def resumeProducing(self):
        """
        Undo the effects of a previous C{pauseProducing} and resume copying
        bytes to the consumer by resuming the L{CooperativeTask} which drives
        the write activity.
        """
        self._task.resume()



class _HTTP11ClientFactory(protocol.Factory):
    """
    A factory for L{HTTP11ClientProtocol}, used by L{HTTPConnectionPool}.

    @ivar _quiescentCallback: The quiescent callback to be passed to protocol
        instances, used to return them to the connection pool.

    @since: 11.1
    """
    def __init__(self, quiescentCallback):
        self._quiescentCallback = quiescentCallback


    def buildProtocol(self, addr):
        return HTTP11ClientProtocol(self._quiescentCallback)



class _RetryingHTTP11ClientProtocol(object):
    """
    A wrapper for L{HTTP11ClientProtocol} that automatically retries requests.

    @ivar _clientProtocol: The underlying L{HTTP11ClientProtocol}.

    @ivar _newConnection: A callable that creates a new connection for a
        retry.
    """

    def __init__(self, clientProtocol, newConnection):
        self._clientProtocol = clientProtocol
        self._newConnection = newConnection


    def _shouldRetry(self, method, exception, bodyProducer):
        """
        Indicate whether request should be retried.

        Only returns C{True} if method is idempotent, no response was
        received, the reason for the failed request was not due to
        user-requested cancellation, and no body was sent. The latter
        requirement may be relaxed in the future, and PUT added to approved
Loading ...