python 多线程剖析

 1 #!/usr/bin/env python
 2 # -*- coding:utf-8 -*-
 3  
 4 import re
 5 import urllib
 6 import threading
 7 import Queue
 8 import timeit
 9 
10 def getHtml(url):
11     html_page = urllib.urlopen(url).read()
12     return html_page
13  
14 # 提取网页中图片的URL
15 def getUrl(html):
16     pattern = r‘src="(http://img.*?)"‘  # 正则表达式
17     imgre = re.compile(pattern)
18     imglist = re.findall(imgre, html)  # re.findall(pattern,string)  在string中寻找所有匹配成功的字符串，以列表形式返回值
19     return imglist
20  
21 class getImg(threading.Thread):
22     def __init__(self, queue, thread_name=0):  # 线程公用一个队列
23         threading.Thread.__init__(self)
24         self.queue = queue
25         self.thread_name = thread_name
26         self.start()  # 启动线程
27  
28     # 使用队列实现进程间通信
29     def run(self):
30         global count
31         while (True):
32             imgurl = self.queue.get() # 调用队列对象的get()方法从队头删除并返回一个项目
33             urllib.urlretrieve(imgurl, ‘E:\mnt\girls\%s.jpg‘ % count)
34             count += 1
35             if self.queue.empty():
36                 break
37             self.queue.task_done()  # 当使用者线程调用 task_done() 以表示检索了该项目、并完成了所有的工作时，那么未完成的任务的总数就会减少。
38 imglist = []
39 def main():
40     global imglist
41     url = "http://huaban.com/favorite/beauty/"  # 要爬的网页地址
42     html = getHtml(url)
43     imglist = getUrl(html)
44  
45 def main_1():
46     global count
47     threads = []
48     count = 0
49     queue = Queue.Queue()
50     # 将所有任务加入队列
51     for img in imglist:
52         queue.put(img)
53     # 多线程爬去图片
54     for i in range(4):
55         thread = getImg(queue, i)
56         threads.append(thread)
57     # 阻塞线程，直到线程执行完成
58     for thread in threads:
59         thread.join()
60  
61 if __name__ == ‘__main__‘:
62     main()
63     t = timeit.Timer(main_1)
64     print t.timeit(1)

 1 def main_1():
 2     global count
 3     threads = []
 4     count = 0
 5     queue = Queue.Queue()
 6     # 将所有任务加入队列
 7     for img in imglist:
 8         queue.put(img)
 9     # 多线程爬去图片
10     for i in range(1):
11         thread = getImg(queue, i)
12         threads.append(thread)
13     # 阻塞线程，直到线程执行完成
14     for thread in threads:
15         thread.join()

 1 #!/usr/bin/env python
 2 # -*- coding:utf-8 -*-
 3 import threading
 4 import timeit
 5 
 6 def countdown(n):
 7     while n > 0:
 8         n -= 1
 9 
10 def task1():
11     COUNT = 100000000
12     thread1 = threading.Thread(target=countdown, args=(COUNT,))
13     thread1.start()
14     thread1.join()
15 
16 def task2():
17     COUNT = 100000000
18     thread1 = threading.Thread(target=countdown, args=(COUNT // 2,))
19     thread2 = threading.Thread(target=countdown, args=(COUNT // 2,))
20     thread1.start()
21     thread2.start()
22     thread1.join()
23     thread2.join()
24 
25 if __name__ == ‘__main__‘:
26     t1 = timeit.Timer(task1)
27     print "countdown in one thread ", t1.timeit(1)
28     t2 = timeit.Timer(task2)
29     print "countdown in two thread ", t2.timeit(1)