python多线程抓取google搜索链接网页

1)urllib2+beautifulsoup抓取goolge搜索链接

近期,参与的项目需要对google搜索结果进行处理,之前学习了python处理网页相关的工具。实际应用中,使用了urllib2和beautifulsoup来进行网页的抓取,但是在抓取google搜索结果的时候,发现如果是直接对google搜索结果页面的源代码进行处理,会得到很多“脏”链接。

看下图为搜索“titanic james”的结果:

图中红色标记的是不需要的,蓝色标记的是需要抓取处理的。

这种“脏链接”当然可以通过规则过滤的方法来过滤掉,但是这样程序的复杂度就高了。正当自己愁眉苦脸的正在写过滤规则时。同学提醒说google应该提供相关的api,才恍然大明白。

(2)google web search api+多线程

文档中给出使用python进行搜索的例子:

图中红色标记的是不需要的,蓝色标记的是需要抓取处理的。

这种“脏链接”当然可以通过规则过滤的方法来过滤掉,但是这样程序的复杂度就高了。正当自己愁眉苦脸的正在写过滤规则时。同学提醒说google应该提供相关的api,才恍然大明白。

(2)google web search api+多线程

文档中给出使用python进行搜索的例子:

import simplejson
# the request also includes the userip parameter which provides the end
# user’s ip address. doing so will help distinguish this legitimate
# server-side traffic from traffic which doesn’t come from an end-user.
url = (‘https://ajax.googleapis.com/ajax/services/search/web’
‘?v=1.0&q=paris%20hilton&userip=users-ip-address’)
request = urllib2.request(
url, none, {‘referer’: /* enter the url of your site here */})
response = urllib2.urlopen(request)
# process the json string.
results = simplejson.load(response)
# now have some fun with the results…
import simplejson
# the request also includes the userip parameter which provides the end
# user’s ip address. doing so will help distinguish this legitimate
# server-side traffic from traffic which doesn’t come from an end-user.
url = (‘https://ajax.googleapis.com/ajax/services/search/web’
‘?v=1.0&q=paris%20hilton&userip=users-ip-address’)
request = urllib2.request(
url, none, {‘referer’: /* enter the url of your site here */})
response = urllib2.urlopen(request)
# process the json string.
results = simplejson.load(response)
# now have some fun with the results..

实际应用中可能需要抓取google的很多网页,所以还需要使用多线程来分担抓取任务。使用google web search api的参考详细介绍,请看此处(这里介绍了standard url arguments)。另外要特别注意,url中参数rsz必须是8(包括8)以下的值,若大于8,会报错的!

(3)代码实现

代码实现还存在问题,但是能够运行,鲁棒性差,还需要进行改进,希望各路大神指出错误(初学python),不胜感激。

#-*-coding:utf-8-*-
import urllib2,urllib
import simplejson
import os, time,threading
import common, html_filter
#input the keywords
keywords = raw_input(‘enter the keywords: ‘)
#define rnum_perpage, pages
rnum_perpage=8
pages=8
#定义线程函数
def thread_scratch(url, rnum_perpage, page):
url_set = []
try:
request = urllib2.request(url, none, {‘referer’: ‘http://www.sina.com’})
response = urllib2.urlopen(request)
# process the json string.
results = simplejson.load(response)
info = results[‘responsedata’][‘results’]
except exception,e:
print ‘error occured’
print e
else:
for minfo in info:
url_set.append(minfo[‘url’])
print minfo[‘url’]
#处理链接
i = 0
for u in url_set:
try:
request_url = urllib2.request(u, none, {‘referer’: ‘http://www.sina.com’})
request_url.add_header(
‘user-agent’,
‘csc’
)
response_data = urllib2.urlopen(request_url).read()
#过滤文件
#content_data = html_filter.filter_tags(response_data)
#写入文件
filenum = i+page
filename = dir_name+’/related_html_’+str(filenum)
print ‘ write start: related_html_’+str(filenum)
f = open(filename, ‘w+’, -1)
f.write(response_data)
#print content_data
f.close()
print ‘ write down: related_html_’+str(filenum)
except exception, e:
print ‘error occured 2’
print e
i = i+1
return
#创建文件夹
dir_name = ‘related_html_’+urllib.quote(keywords)
if os.path.exists(dir_name):
print ‘exists file’
common.delete_dir_or_file(dir_name)
os.makedirs(dir_name)
#抓取网页
print ‘start to scratch web pages:’
for x in range(pages):
print “page:%s”%(x+1)
page = x * rnum_perpage
url = (‘https://ajax.googleapis.com/ajax/services/search/web’
‘?v=1.0&q=%s&rsz=%s&start=%s’) % (urllib.quote(keywords), rnum_perpage,page)
print url
t = threading.thread(target=thread_scratch, args=(url,rnum_perpage, page))
t.start()
#主线程等待子线程抓取完
main_thread = threading.currentthread()
for t in threading.enumerate():
if t is main_thread:
continue
t.join()

Posted in 未分类

发表评论