作为概述,您将需要执行四个主要任务:
- 向网站提交请求,
- 从站点检索响应
- 解析这些响应
- 在上面的任务中有一些逻辑可以迭代,带有与导航关联的参数(到结果列表中的“下一页”)
http请求和响应处理是通过Python标准库的urllib和urllib2中的方法和类完成的。可以使用Python的标准库的HTMLParser
或其他模块(例如Beautiful
Soup)来完成html页面的解析。
以下代码段演示了在问题指示的站点上请求和接收搜索的过程。该站点是ASP驱动的,因此,我们需要确保发送多个表单字段,其中一些表单字段具有“可怕”的值,因为ASP逻辑使用这些字段来维护状态并在某种程度上对请求进行身份验证。确实是提交。必须使用
http POST方法
发送请求,因为这是该ASP应用程序所期望的。主要困难在于确定ASP期望的表单字段和关联值(使用Python获取页面是最容易的部分)。
直到我删除了大部分的VSTATE值,并可能通过添加注释引入了一两个拼写错误,这段代码才起作用,或者更确切地说, 是 起作用的。
import urllibimport urllib2uri = 'http://legistar.council.nyc.gov/Legislation.aspx'#the http headers are useful to simulate a particular browser (some sites deny#access to non-browsers (bots, etc.)#also needed to pass the content type. headers = { 'HTTP_USER_AGENT': 'Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.13) Gecko/2009073022 Firefox/3.0.13', 'HTTP_ACCEPT': 'text/html,application/xhtml+xml,application/xml; q=0.9,*/*; q=0.8', 'Content-Type': 'application/x-www-form-urlenpred'}# we group the form fields and their values in a list (any# iterable, actually) of name-value tuples. This helps# with clarity and also makes it easy to later encoding of them.formFields = ( # the viewstate is actualy 800+ characters in length! I truncated it # for this sample pre. It can be lifted from the first page # obtained from the site. It may be ok to hardpre this value, or # it may have to be refreshed each time / each day, by essentially # running an extra page request and parse, for this specific value. (r'__VSTATE', r'7TzretNIlrZiKb7EOB3AQE ... ...2qd6g5xD8CGXm5EftXtNPt+H8B'), # following are more of these ASP form fields (r'__VIEWSTATE', r''), (r'__EVENTVALIDATION', r'/wEWDwL+raDpAgKnpt8nAs3q+pQOAs3q/pQOAs3qgpUOAs3qhpUOAoPE36ANAve684YCAoOs79EIAoOs89EIAoOs99EIAoOs39EIAoOs49EIAoOs09EIAoSs99EI6IQ74SEV9n4XbtWm1rEbB6Ic3/M='), (r'ctl00_RadscriptManager1_HiddenField', ''), (r'ctl00_tabTop_ClientState', ''), (r'ctl00_ContentPlaceHolder1_menuMain_ClientState', ''), (r'ctl00_ContentPlaceHolder1_gridMain_ClientState', ''), #but then we come to fields of interest: the search #criteria the collections to search from etc.# Check boxes (r'ctl00$ContentPlaceHolder1$chkOptions', 'on'), # file number (r'ctl00$ContentPlaceHolder1$chkOptions', 'on'), # Legislative text (r'ctl00$ContentPlaceHolder1$chkOptions', 'on'), # attachement# etc. (not all listed) (r'ctl00$ContentPlaceHolder1$txtSearch', 'york'), # Search text (r'ctl00$ContentPlaceHolder1$lstYears', 'All Years'), # Years to include (r'ctl00$ContentPlaceHolder1$lstTypeBasic', 'All Types'), #types to include (r'ctl00$ContentPlaceHolder1$btnSearch', 'Search Legislation') # Search button itself)# these have to be enpred enpredFields = urllib.urlenpre(formFields)req = urllib2.Request(uri, enpredFields, headers)f= urllib2.urlopen(req) #that's the actual call to the http site.# *** here would normally be the in-memory parsing of f # contents, but instead I store this to file# this is useful during design, allowing to have a# sample of what is to be parsed in a text editor, for analysis.try: fout = open('tmp.htm', 'w')except: print('Could not open output filen')fout.writelines(f.readlines())fout.close()
就是为了获得初始页面。如上所述,然后需要解析页面,即找到感兴趣的部分并适当地收集它们,并将它们存储到文件/数据库/任何地方。可以通过很多方法来完成这项工作:使用html解析器或XSLT类型的技术(实际上是将html解析为xml之后),甚至对于简单的工作,也可以使用简单的正则表达式。同样,一个通常提取的项目之一是“下一个信息”,即各种链接,可用于对服务器的新请求中以获取后续页面。
这应该使您大致了解“长手”
html刮擦的含义。还有许多其他方法,例如专用工具,Mozilla(FireFox)GreaseMonkey插件中的脚本,XSLT …
欢迎分享,转载请注明来源:内存溢出
评论列表(0条)