`

python crawler(2)

 
阅读更多

以前写过一篇使用python爬虫抓站的一些技巧总结,总结了诸多爬虫使用的方法;那篇东东现在看来还是挺有用的,但是当时很菜(现在也菜,但是比那时进步了不少),很多东西都不是很优,属于”只是能用”这么个层次。这篇进阶篇打算把“能用”提升到“用得省事省心”这个层次。

一、gzip/deflate支持

现在的网页普遍支持gzip压缩,这往往可以解决大量传输时间,以VeryCD的主页为例,未压缩版本247K,压缩了以后45K,为原来的1/5。这就意味着抓取速度会快5倍。

然而python的urllib/urllib2默认都不支持压缩,要返回压缩格式,必须在request的header里面写明’accept- encoding’,然后读取response后更要检查header查看是否有’content-encoding’一项来判断是否需要解码,很繁琐琐 碎。如何让urllib2自动支持gzip, defalte呢?

其实可以继承BaseHanlder类,然后build_opener的方式来处理:

importurllib2fromgzipimportGzipFilefromStringIOimportStringIOclassContentEncodingProcessor<wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">urllib2</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">BaseHandler</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"""A handler to add gzip capabilities to urllib2 requests """</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># add headers to requests</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>http_request<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>, req<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>: req.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">add_header</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"Accept-Encoding"</span>,<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"gzip, deflate"</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>return</strong></span>req <wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># decode</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>http_response<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>, req, resp<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>: old_resp = resp<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># gzip</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>if</strong></span>resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">headers</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">get</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"content-encoding"</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>==<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"gzip"</span>: gz = GzipFile<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>fileobj=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">StringIO</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">read</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>, mode=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"r"</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>resp =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">urllib2</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">addinfourl</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>gz, old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">headers</span>, old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">url</span>, old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">code</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">msg</span>= old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">msg</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># deflate</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>if</strong></span>resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">headers</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">get</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"content-encoding"</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>==<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"deflate"</span>: gz =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">StringIO</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>deflate<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">read</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>resp =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">urllib2</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">addinfourl</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>gz, old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">headers</span>, old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">url</span>, old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">code</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># 'class to add info() and</em></span>resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">msg</span>= old_resp.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">msg</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>return</strong></span>resp <wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># deflate support</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>import</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">zlib</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>deflate<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>data<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># zlib only provides the zlib compress format,<wbr>not the deflate format;</wbr></em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>try</strong></span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># so on top of all there's this workaround:</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>return</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">zlib</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">decompress</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>data, -<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">zlib</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">MAX_WBITS</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>except</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">zlib</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">error</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>return</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">zlib</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">decompress</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>data<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span></wbr></wbr></wbr></wbr>

然后就简单了,

encoding_support = ContentEncodingProcessor<wbr>opener =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">urllib2</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">build_opener</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>encoding_support,<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">urllib2</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">HTTPHandler</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#直接用opener打开网页,如果服务器支持gzip/defalte则自动解压缩</em></span>content = opener.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">open</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">read</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span></wbr></wbr>

二、更方便地多线程

总结一文的确提及了一个简单的多线程模板,但是那个东东真正应用到程序里面去只会让程序变得支离破碎,不堪入目。在怎么更方便地进行多线程方面我也动了一番脑筋。先想想怎么进行多线程调用最方便呢?

1、用twisted进行异步I/O抓取

事实上更高效的抓取并非一定要用多线程,也可以使用异步I/O法:直接用twisted的getPage方法,然后分别加上异步I/O结束时的callback和errback方法即可。例如可以这么干:

fromtwisted.web.clientimportgetPagefromtwisted.internetimportreactor <wbr>links =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">[</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">'http://www.verycd.com/topics/%d/'</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(102,204,102)">%</span>i<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>for</strong></span>i<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>in</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">range</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">5420</span>,<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">5430</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">]</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>parse_page<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>data,url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>print</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">len</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>data<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>,url <wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>fetch_error<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>error,url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>print</strong></span>error.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">getErrorMessage</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>,url <wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># 批量抓取链接</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>for</strong></span>url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>in</strong></span>links: getPage<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>url,timeout=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">5</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>\ .<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">addCallback</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>parse_page,url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>\<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#成功则调用parse_page方法</em></span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">addErrback</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>fetch_error,url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#失败则调用fetch_error方法</em></span><wbr>reactor.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">callLater</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">5</span>, reactor.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">stop</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#5秒钟后通知reactor结束程序</em></span>reactor.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">run</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span></wbr></wbr></wbr></wbr></wbr>

twisted人如其名,写的代码实在是太扭曲了,非正常人所能接受,虽然这个简单的例子看上去还好;每次写twisted的程序整个人都扭曲了,累得不得了,文档等于没有,必须得看源码才知道怎么整,唉不提了。

如果要支持gzip/deflate,甚至做一些登陆的扩展,就得为twisted写个新的HTTPClientFactory类诸如此类,我这眉头真是大皱,遂放弃。有毅力者请自行尝试。

这篇讲怎么用twisted来进行批量网址处理的文章不错,由浅入深,深入浅出,可以一看。

2、设计一个简单的多线程抓取类

还是觉得在urllib之类python“本土”的东东里面折腾起来更舒服。试想一下,如果有个Fetcher类,你可以这么调用

f = Fetcher(threads=10)#设定下载线程数为10forurlinurls: f.push(url)#把所有url推入下载队列whilef.taskleft():#若还有未完成下载的线程content = f.pop()#从下载完成队列中取出结果do_with(content)# 处理content内容

这么个多线程调用简单明了,那么就这么设计吧,首先要有两个队列,用Queue搞定,多线程的基本架构也和“技巧总结”一文类似,push方法和 pop方法都比较好处理,都是直接用Queue的方法,taskleft则是如果有“正在运行的任务”或者”队列中的任务”则为是,也好办,于是代码如 下:

importurllib2fromthreadingimportThread,LockfromQueueimportQueueimporttime<wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>class</strong></span>Fetcher:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,0,205)">__init__</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>,threads<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">opener</span>=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">urllib2</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">build_opener</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">urllib2</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">HTTPHandler</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">lock</span>= Lock<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#线程锁</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_req</span>=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">Queue</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#任务队列</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_ans</span>=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">Queue</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#完成队列</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">threads</span>= threads<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>for</strong></span>i<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>in</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">range</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>threads<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>: t = Thread<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>target=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">threadget</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>t.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">setDaemon</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">True</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>t.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">start</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">running</span>=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">0</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,0,205)">__del__</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#解构时需等待两个队列完成</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">time</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">sleep</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">0.5</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_req</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">join</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_ans</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">join</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>taskleft<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>return</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_req</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">qsize</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>+<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_ans</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">qsize</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>+<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">running</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>push<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>,req<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_req</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">put</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>req<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>pop<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>return</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_ans</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">get</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>def</strong></span>threadget<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>while</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">True</span>: req =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_req</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">get</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>with</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">lock</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em>#要保证该操作的原子性,进入critical area</em></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">running</span>+=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">1</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>try</strong></span>: ans =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">opener</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">open</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>req<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">read</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>except</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">Exception</span>, what: ans =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">''</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>print</strong></span>what<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_ans</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">put</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>req,ans<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>with</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">lock</span>:<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">running</span>-=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">1</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">self</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">q_req</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">task_done</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(220,20,60)">time</span>.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">sleep</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">0.1</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(128,128,128)"><em># don't spam</em></span><wbr><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>if</strong></span>__name__ ==<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">"__main__"</span>: links =<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">[</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(72,61,139)">'http://www.verycd.com/topics/%d/'</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(102,204,102)">%</span>i<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>for</strong></span>i<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>in</strong></span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">range</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">5420</span>,<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">5430</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">]</span>f = Fetcher<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>threads=<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,69,0)">10</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>for</strong></span>url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>in</strong></span>links: f.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">push</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>url<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>while</strong></span>f.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">taskleft</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span>: url,content = f.<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">pop</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(255,119,0)"><strong>print</strong></span>url,<span style="word-wrap:normal; word-break:normal; line-height:21px; color:rgb(0,128,0)">len</span><span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">(</span>content<span style="word-wrap:normal; word-break:normal; line-height:21px; color:black">)</span></wbr></wbr></wbr></wbr></wbr></wbr></wbr>

三、一些琐碎的经验

1、连接池:

opener.open和urllib2.urlopen一样,都会新建一个http请求。通常情况下这不是什么问题,因为线性环境下,一秒钟可能 也就新生成一个请求;然而在多线程环境下,每秒钟可以是几十上百个请求,这么干只要几分钟,正常的有理智的服务器一定会封禁你的。

然而在正常的html请求时,保持同时和服务器几十个连接又是很正常的一件事,所以完全可以手动维护一个HttpConnection的池,然后每次抓取时从连接池里面选连接进行连接即可。

这里有一个取巧的方法,就是利用squid做代理服务器来进行抓取,则squid会自动为你维护连接池,还附带数据缓存功能,而且squid本来就是我每个服务器上面必装的东东,何必再自找麻烦写连接池呢。

2、设定线程的栈大小

栈大小的设定将非常显著地影响python的内存占用,python多线程不设置这个值会导致程序占用大量内存,这对openvz的vps来说非常致命。stack_size必须大于32768,实际上应该总要32768*2以上

fromthreadingimportstack_size stack_size(32768*16)

3、设置失败后自动重试

defget(self,req,retries=3):try: response =self.opener.open(req)data = response.read()exceptException, what:printwhat,reqifretries>0:returnself.get(req,retries-1)else:print'GET Failed',reqreturn''returndata

4、设置超时

importsocketsocket.setdefaulttimeout(10)#设置10秒后连接超时

5、登陆

登陆更加简化了,首先build_opener中要加入cookie支持,参考“总结”一文;如要登陆VeryCD,给Fetcher新增一个空方法login,并在__init__()中调用,然后继承Fetcher类并override login方法:

deflogin(self,username,password):importurllibdata=urllib.urlencode({'username':username,'password':password,'continue':'http://www.verycd.com/','login_submit':u'登录'.encode('utf-8'),'save_cookie':1,})url ='http://www.verycd.com/signin'self.opener.open(url,data).read()

于是在Fetcher初始化时便会自动登录VeryCD网站。

四、总结

如此,把上述所有小技巧都糅合起来就和我目前的私藏最终版的Fetcher类相差不远了,它支持多线程,gzip/deflate压缩,超时设置,自动重试,设置栈大小,自动登录等功能;代码简单,使用方便,性能也不俗,可谓居家旅行,杀人放火,咳咳,之必备工具。

之所以说和最终版差得不远,是因为最终版还有一个保留功能“马甲术”:多代理自动选择。看起来好像仅仅是一个random.choice的区别,其实包含了代理获取,代理验证,代理测速等诸多环节,这就是另一个故事了

分享到:
评论

相关推荐

Global site tag (gtag.js) - Google Analytics