• python爬虫篇之 性能相关


    一.背景

      爬虫的本质就是一个socket客户端与服务端的通信过程,如果我们有多个url待爬取,采用串行的方式执行,只能等待爬取一个结束后才能继续下一个,效率会非常低。

      需要强调的是:串行并不意味着低效,如果串行的都是纯计算的任务,那么cpu的利用率仍然会很高,之所以爬虫程序的串行低效,是因为爬虫程序是明显的IO密集型程序。

    二.同步,异步,回调机制

    在编写爬虫是,性能的消耗主要在IO请求中,当单进程单线程模式下 请求URL时,必然会引起等待,从而使得请求整体变慢。

     1.同步:提交一个任务后就在原地等待任务结束,等到拿到任务的结果后在继续下一行代码,效率低下。

    import requests
    
    def fetch_async(url):
        response = requests.get(url)
        return response    #返回信息:response.text...
    
    url_list = ['http://www.jiemian.com','http://www.bing.com']
    
    for url in url_list:
        print(fetch_async(url))
    1.同步执行

    2.简单的解决方案:多线程或多进程

    #在服务器端使用多线程(或多进程)。多线程(或多进程)的目的是让每个连接都拥有独立的线程(或进程),这样任何一个连接的阻塞都不会影响其他的连接。 
    from concurrent.futures import ThreadPoolExecutor
    import requests
    
    
    def fetch_async(url):
        response = requests.get(url)
        return response
    
    
    url_list = ['http://www.github.com', 'http://www.bing.com']
    pool = ThreadPoolExecutor(5)
    for url in url_list:
        pool.submit(fetch_async, url)
    pool.shutdown(wait=True)
    2.多线程执行
    from multiprocessing import Process
    from threading import Thread
    import requests
    
    def get_page(url):
        response=requests.get(url)
        if response.status_code == 200:
            return response.text
    
    
    if __name__ == '__main__':
        urls=['https://www.baidu.com/','http://www.sina.com.cn/','https://www.python.org']
        for url in urls:
            p=Process(target=get_page,args=(url,))
            p.start()
            # t=Thread(target=get_page,args=(url,))
            # t.start()
    2.1

      该方案的局限性:

    #开启多进程或都线程的方式,我们是无法无限制地开启多进程或多线程的:在遇到要同时响应成百上千路的连接请求,则无论多线程还是多进程都会严重占据系统资源,降低系统对外界响应效率,而且线程与进程本身也更容易进入假死状态。

    3.改进方案:线程池或进程池+异步调用:提交一个任务后并不会等待任务结束,而是继续下一行代码

    #很多程序员可能会考虑使用“线程池”或“连接池”。“线程池”旨在减少创建和销毁线程的频率,其维持一定合理数量的线程,并让空闲的线程重新承担新的执行任务。“连接池”维持连接的缓存池,尽量重用已有的连接、减少创建和关闭连接的频率。这两种技术都可以很好的降低系统开销,都被广泛应用很多大型系统,如websphere、tomcat和各种数据库等。
    from concurrent.futures import ProcessPoolExecutor
    import requests
    
    import os
    
    
    def fetch_async(url):
        print('%s GET : %s' % (os.getpid(), url))
        response = requests.get(url)
        return response
    
    
    def callback(future):
        future.result()
        print('%s parsing' %os.getpid())
    
    
    if __name__ == '__main__':
    
        url_list = ['https://www.baidu.com/','http://www.sina.com.cn/','https://www.python.org']
        pool = ProcessPoolExecutor(2)
        for url in url_list:
            v = pool.submit(fetch_async, url)
            v.add_done_callback(callback)
        pool.shutdown(wait=True)
    异步调用+回调机制

    改进后方案其实也存在着问题:

    #“线程池”和“连接池”技术也只是在一定程度上缓解了频繁调用IO接口带来的资源占用。而且,所谓“池”始终有其上限,当请求大大超过上限时,“池”构成的系统对外界的响应并不比没有池的时候效果好多少。所以使用“池”必须考虑其面临的响应规模,并根据响应规模调整“池”的大小。

    对应上例中的所面临的可能同时出现的上千甚至上万次的客户端请求,“线程池”或“连接池”或许可以缓解部分压力,但是不能解决所有问题。总之,多线程模型可以方便高效的解决小规模的服务请求,但面对大规模的服务请求,多线程模型也会遇到瓶颈,可以用非阻塞接口来尝试解决这个问题。

    三.高性能

      上述五路哪种解决方案其实没有解决一个性能相关的问题:IO阻塞,无论是多进程还是多线程,在遇到IO阻塞式都会被操作系统强行剥夺走CPU的执行权限,程序的执行效率因此就降低了下来。

      解决问题的关键在于,我们自己从应用程序级别检测IO阻塞然后切换到我们自己程序的其他任务让其执行,这样吧我们程序的IO降到最低,我们的程序处于就绪态就会增多,以此来迷惑操作系统,操作系统便以为我们的程序是IO比较少的程序,从而会尽可能多的分配CPU给我们,这就就达到提升程序执行效率的目的。

    1、在python3.3之后新增了asyncio模块,可以帮我们检测IO(只能是网络IO),实现应用程序级别的切换

    import asyncio
    
    
    @asyncio.coroutine
    def func1():
        print('before...func1......')
        yield from asyncio.sleep(5)
        print('end...func1......')
    
    
    tasks = [func1(), func1()]
    
    loop = asyncio.get_event_loop()
    loop.run_until_complete(asyncio.gather(*tasks))
    loop.close()
    1.asyncio示例

    2、但asyncio模块只能发tcp级别的请求,不能发http协议,因此,在我们需要发送http请求的时候,需要我们自定义http报头

    #我们爬取一个网页的过程,以https://www.python.org/doc/为例,将关键步骤列举如下
    #步骤一:向www.python.org这台主机发送tcp三次握手,是IO阻塞操作
    #步骤二:封装http协议的报头
    #步骤三:发送http协议的请求包,是IO阻塞操作
    #步骤四:接收http协议的响应包,是IO阻塞操作
    import asyncio
    
    @asyncio.coroutine
    def get_page(host,port=80,url='/'):
        #步骤一(IO阻塞):发起tcp链接,是阻塞操作,因此需要yield from
        recv,send=yield from asyncio.open_connection(host,port)
    
        #步骤二:封装http协议的报头,因为asyncio模块只能封装并发送tcp包,因此这一步需要我们自己封装http协议的包
        requset_headers="""GET %s HTTP/1.0
    Host: %s
    
    """ % (url, host,)
        # requset_headers="""POST %s HTTP/1.0
    Host: %s
    
    name=egon&password=123""" % (url, host,)
        requset_headers=requset_headers.encode('utf-8')
    
        #步骤三(IO阻塞):发送http请求包
        send.write(requset_headers)
        yield from send.drain()
    
        #步骤四(IO阻塞):接收http协议的响应包
        text=yield from recv.read()
    
        #其他处理
        print(host,url,text)
        send.close()
        print('-===>')
        return 1
    
    tasks=[get_page(host='www.python.org',url='/doc'),get_page(host='www.cnblogs.com',url='linhaifeng'),get_page(host='www.openstack.org')]
    
    loop=asyncio.get_event_loop()
    results=loop.run_until_complete(asyncio.gather(*tasks))
    loop.close()
    
    print('=====>',results) #[1, 1, 1]
    asyncio+自定义http协议报头
    import asyncio
    
    
    @asyncio.coroutine
    def fetch_async(host, url='/'):
        print(host, url)
        reader, writer = yield from asyncio.open_connection(host, 80)
    
        request_header_content = """GET %s HTTP/1.0
    Host: %s
    
    """ % (url, host,)
        request_header_content = bytes(request_header_content, encoding='utf-8')
    
        writer.write(request_header_content)
        yield from writer.drain()
        text = yield from reader.read()
        print(host, url, text)
        writer.close()
    
    tasks = [
        fetch_async('www.cnblogs.com', '/wupeiqi/'),
        fetch_async('dig.chouti.com', '/pic/show?nid=4073644713430508&lid=10273091')
    ]
    
    loop = asyncio.get_event_loop()
    results = loop.run_until_complete(asyncio.gather(*tasks))
    loop.close()
    简约版

    3、自定义http报头多少有点麻烦,于是有了aiohttp模块,专门帮我们封装http报头,然后我们还需要用asyncio检测IO实现切换

    import aiohttp
    import asyncio
    
    
    @asyncio.coroutine
    def fetch_async(url):
        print(url)
        response = yield from aiohttp.request('GET', url)
        # data = yield from response.read()
        # print(url, data)
        print(url, response)
        response.close()
    
    
    tasks = [fetch_async('http://www.google.com/'), fetch_async('http://www.chouti.com/')]
    
    event_loop = asyncio.get_event_loop()
    results = event_loop.run_until_complete(asyncio.gather(*tasks))
    event_loop.close()
    asyncio + aiohttp

      4、此外,还可以将requests.get函数传给asyncio,就能够被检测了

    import asyncio
    import requests
    
    
    @asyncio.coroutine
    def fetch_async(func,*args):
        loop = asyncio.get_event_loop()
        future = loop.run_in_executor(None,func,*args)
        response = yield from future
    
    
        print(response.url,response.content)
    
    tasks = [
        fetch_async(requests.get, 'http://www.jiemian.com'),
        fetch_async(requests.get, 'http://www.biying.com')
    ]
    
    loop = asyncio.get_event_loop()
    results = loop.run_until_complete(asyncio.gather(*tasks))
    loop.close()
    asyncio + requests

    5、还有之前在协程时介绍的gevent模块

    import gevent
    
    import requests
    from gevent import monkey
    
    monkey.patch_all()
    
    
    def fetch_async(method, url, req_kwargs):
        print(method, url, req_kwargs)
        response = requests.request(method=method, url=url, **req_kwargs)
        print(response.url, response.content)
    
    # ##### 发送请求 #####
    gevent.joinall([
        gevent.spawn(fetch_async, method='get', url='https://www.python.org/', req_kwargs={}),
        gevent.spawn(fetch_async, method='get', url='https://www.yahoo.com/', req_kwargs={}),
        gevent.spawn(fetch_async, method='get', url='https://github.com/', req_kwargs={}),
    ])
    
    # ##### 发送请求(协程池控制最大协程数量) #####
    # from gevent.pool import Pool
    # pool = Pool(None)
    # gevent.joinall([
    #     pool.spawn(fetch_async, method='get', url='https://www.python.org/', req_kwargs={}),
    #     pool.spawn(fetch_async, method='get', url='https://www.yahoo.com/', req_kwargs={}),
    #     pool.spawn(fetch_async, method='get', url='https://www.github.com/', req_kwargs={}),
    # ])
    gevent + requests

     6、封装了gevent+requests模块的grequests模块

    import grequests
    
    
    request_list = [
        grequests.get('http://httpbin.org/delay/1', timeout=0.001),
        grequests.get('http://fakedomain/'),
        grequests.get('http://httpbin.org/status/500')
    ]
    
    
    # ##### 执行并获取响应列表 #####
    # response_list = grequests.map(request_list)
    # print(response_list)
    
    
    # ##### 执行并获取响应列表(处理异常) #####
    # def exception_handler(request, exception):
    # print(request,exception)
    #     print("Request failed")
    
    # response_list = grequests.map(request_list, exception_handler=exception_handler)
    # print(response_list)
    grequests

     7、twisted:是一个网络框架,其中一个功能是发送异步请求,检测IO并自动切换

    '''
    #问题一:error: Microsoft Visual C++ 14.0 is required. Get it with "Microsoft Visual C++ Build Tools": http://landinghub.visualstudio.com/visual-cpp-build-tools
    https://www.lfd.uci.edu/~gohlke/pythonlibs/#twisted
    pip3 install C:UsersAdministratorDownloadsTwisted-17.9.0-cp36-cp36m-win_amd64.whl
    pip3 install twisted
    
    #问题二:ModuleNotFoundError: No module named 'win32api'
    https://sourceforge.net/projects/pywin32/files/pywin32/
    
    #问题三:openssl
    pip3 install pyopenssl
    '''
    
    #twisted基本用法
    from twisted.web.client import getPage,defer
    from twisted.internet import reactor
    
    def all_done(arg):
        # print(arg)
        reactor.stop()
    
    def callback(res):
        print(res)
        return 1
    
    defer_list=[]
    urls=[
        'http://www.baidu.com',
        'http://www.bing.com',
        'https://www.python.org',
    ]
    for url in urls:
        obj=getPage(url.encode('utf=-8'),)
        obj.addCallback(callback)
        defer_list.append(obj)
    
    defer.DeferredList(defer_list).addBoth(all_done)
    
    reactor.run()
    
    
    
    
    #twisted的getPage的详细用法
    from twisted.internet import reactor
    from twisted.web.client import getPage
    import urllib.parse
    
    
    def one_done(arg):
        print(arg)
        reactor.stop()
    
    post_data = urllib.parse.urlencode({'check_data': 'adf'})
    post_data = bytes(post_data, encoding='utf8')
    headers = {b'Content-Type': b'application/x-www-form-urlencoded'}
    response = getPage(bytes('http://dig.chouti.com/login', encoding='utf8'),
                       method=bytes('POST', encoding='utf8'),
                       postdata=post_data,
                       cookies={},
                       headers=headers)
    response.addBoth(one_done)
    
    reactor.run()
    twisted
    from twisted.web.client import getPage, defer
    from twisted.internet import reactor
    
    
    def all_done(arg):
        reactor.stop()
    
    
    def callback(contents):
        print(contents)
    
    
    deferred_list = []
    
    url_list = ['http://www.bing.com', 'http://www.baidu.com', ]
    for url in url_list:
        deferred = getPage(bytes(url, encoding='utf8'))
        deferred.addCallback(callback)
        deferred_list.append(deferred)
    
    dlist = defer.DeferredList(deferred_list)
    dlist.addBoth(all_done)
    
    reactor.run()
    简约版
    dlist.addBoth(all_done)  #无论正确与否,都会执行 回调函数
    dlist.addCallback(all_done)    #正确的时候,执行
    dlist.addErrback(all_done)      #错误的时候,执行
    执行回调的三种方式
    from twisted.web.client import getPage, defer
    from twisted.internet import reactor
    
    
    
    
    def callback(contents):
        print(contents)
    
    @defer.inlineCallbacks
    def task(url):
        # defer.Deferred()      -------》返回defer对象
        d = getPage(bytes(url, encoding='utf8'))
        d.addCallback(callback)
        yield d
    
      
    url_list = ['http://www.bing.com', 'http://www.baidu.com', ]
    _active = []
    for url in url_list:
        d = task(url)
        _active.append(d)
    
    
    def all_done(arg):
        reactor.stop()
    
    xx = defer.DeferredList(_active)
    xx.addBoth(all_done)
    
    reactor.run()
    @defer.inlineCallbacks

    8.Tornado

    from tornado.httpclient import AsyncHTTPClient
    from tornado.httpclient import HTTPRequest
    from tornado import ioloop
    
    
    def handle_response(response):
        """
        处理返回值内容(需要维护计数器,来停止IO循环),调用 ioloop.IOLoop.current().stop()
        :param response: 
        :return: 
        """
        if response.error:
            print("Error:", response.error)
        else:
            print(response.body)
    
    
    def func():
        url_list = [
            'http://www.baidu.com',
            'http://www.bing.com',
        ]
        for url in url_list:
            print(url)
            http_client = AsyncHTTPClient()
            http_client.fetch(HTTPRequest(url), handle_response)
    
    
    ioloop.IOLoop.current().add_callback(func)
    ioloop.IOLoop.current().start()
    Tornado

     

    from twisted.web.client import getPage,defer
    from twisted.internet import reactor
    
    
    def all_done(arg):
        reactor.stop()
    
    def callback(contents):
        print(contents)
    
    
    deferred_list = []
    
    url_list = ['http://www.jiemian.com', 'http://www.baidu.com',]
    for url in url_list:
        deferred = getPage(bytes(url,encoding='utf8'))
        deferred.addCallback(callback)
        deferred_list.append(deferred)
    
    dlist = defer.DeferredList(deferred_list)
    dlist.addBoth(all_done)
    
    react8、tornado

    以上是Python内置以及第三方提供的异步IO请求模块,使用简便大大提高效率,而对异步IO请求的本质则是 [ 非阻塞Socket ] + [IO多路复用]:

    import select
    import socket
    import time
    
    
    class AsyncTimeoutException(TimeoutError):
        """
        请求超时异常类
        """
    
        def __init__(self, msg):
            self.msg = msg
            super(AsyncTimeoutException, self).__init__(msg)
    
    
    class HttpContext(object):
        """封装请求和相应的基本数据"""
    
        def __init__(self, sock, host, port, method, url, data, callback, timeout=5):
            """
            sock: 请求的客户端socket对象
            host: 请求的主机名
            port: 请求的端口
            port: 请求的端口
            method: 请求方式
            url: 请求的URL
            data: 请求时请求体中的数据
            callback: 请求完成后的回调函数
            timeout: 请求的超时时间
            """
            self.sock = sock
            self.callback = callback
            self.host = host
            self.port = port
            self.method = method
            self.url = url
            self.data = data
    
            self.timeout = timeout
    
            self.__start_time = time.time()
            self.__buffer = []
    
        def is_timeout(self):
            """当前请求是否已经超时"""
            current_time = time.time()
            if (self.__start_time + self.timeout) < current_time:
                return True
    
        def fileno(self):
            """请求sockect对象的文件描述符,用于select监听"""
            return self.sock.fileno()
    
        def write(self, data):
            """在buffer中写入响应内容"""
            self.__buffer.append(data)
    
        def finish(self, exc=None):
            """在buffer中写入响应内容完成,执行请求的回调函数"""
            if not exc:
                response = b''.join(self.__buffer)
                self.callback(self, response, exc)
            else:
                self.callback(self, None, exc)
    
        def send_request_data(self):
            content = """%s %s HTTP/1.0
    Host: %s
    
    %s""" % (
                self.method.upper(), self.url, self.host, self.data,)
    
            return content.encode(encoding='utf8')
    
    
    class AsyncRequest(object):
        def __init__(self):
            self.fds = []
            self.connections = []
    
        def add_request(self, host, port, method, url, data, callback, timeout):
            """创建一个要请求"""
            client = socket.socket()
            client.setblocking(False)
            try:
                client.connect((host, port))
            except BlockingIOError as e:
                pass
                # print('已经向远程发送连接的请求')
            req = HttpContext(client, host, port, method, url, data, callback, timeout)
            self.connections.append(req)
            self.fds.append(req)
    
        def check_conn_timeout(self):
            """检查所有的请求,是否有已经连接超时,如果有则终止"""
            timeout_list = []
            for context in self.connections:
                if context.is_timeout():
                    timeout_list.append(context)
            for context in timeout_list:
                context.finish(AsyncTimeoutException('请求超时'))
                self.fds.remove(context)
                self.connections.remove(context)
    
        def running(self):
            """事件循环,用于检测请求的socket是否已经就绪,从而执行相关操作"""
            while True:
                r, w, e = select.select(self.fds, self.connections, self.fds, 0.05)
    
                if not self.fds:
                    return
    
                for context in r:
                    sock = context.sock
                    while True:
                        try:
                            data = sock.recv(8096)
                            if not data:
                                self.fds.remove(context)
                                context.finish()
                                break
                            else:
                                context.write(data)
                        except BlockingIOError as e:
                            break
                        except TimeoutError as e:
                            self.fds.remove(context)
                            self.connections.remove(context)
                            context.finish(e)
                            break
    
                for context in w:
                    # 已经连接成功远程服务器,开始向远程发送请求数据
                    if context in self.fds:
                        data = context.send_request_data()
                        context.sock.sendall(data)
                        self.connections.remove(context)
    
                self.check_conn_timeout()
    
    
    if __name__ == '__main__':
        def callback_func(context, response, ex):
            """
            :param context: HttpContext对象,内部封装了请求相关信息
            :param response: 请求响应内容
            :param ex: 是否出现异常(如果有异常则值为异常对象;否则值为None)
            :return:
            """
            print(context, response, ex)
    
        obj = AsyncRequest()
        url_list = [
            {'host': 'www.google.com', 'port': 80, 'method': 'GET', 'url': '/', 'data': '', 'timeout': 5,
             'callback': callback_func},
            {'host': 'www.baidu.com', 'port': 80, 'method': 'GET', 'url': '/', 'data': '', 'timeout': 5,
             'callback': callback_func},
            {'host': 'www.bing.com', 'port': 80, 'method': 'GET', 'url': '/', 'data': '', 'timeout': 5,
             'callback': callback_func},
        ]
        for item in url_list:
            print(item)
            obj.add_request(**item)
    
        obj.running()
    武器版的异步IO模块
  • 相关阅读:
    【Thinkphp教程】URL路由功能解析
    MYSQL 错误#145解决方法
    【Thinkphp教程】空模块
    【Thinkphp教程】 如何进行模块分组
    mySQL中删除unique key的语法
    使用php让浏览器刷新
    Spring+Jpa整合的过程中遇到的一个问题。。。纠结了我半天。。。
    关于mule studio的应用
    解决eclipse和myeclipse不能编译项目的问题
    ajax fileupload上传组件的使用感悟
  • 原文地址:https://www.cnblogs.com/zhaochangbo/p/7761432.html
Copyright © 2020-2023  润新知