Tag (百度)'s result:


最近用python写了一个采集百度url的脚本,一个线程一次爬一个页面的全部url,使用了多线程、队列、bs4,感觉bs4挺强大。 功能描述:采集百度url,自定义页数,线程数,关键词,保存文件名 (注:此脚本使用线程数小于或等于页面数即可,由于使用了队列,即使线程数大于页面数效果也和等于页面数一样) 源代码: #coding: utf-8 import requests,re,threading,time from bs4 import BeautifulSoup as bs from Queue import Queue from argparse import ArgumentParser arg = ArgumentParser(description=’baidu_url_collection’) arg.add_argument(‘keyword’,help=’inurl:.asp?id=1′) arg.add_argument(‘-p’, ‘–page’, help=’page count’, dest=’pagecount’, type=int) arg.add_argument(‘-t’, ‘–thread’, help=’the thread_count’, dest=’thread_count’, type=int, default=10) arg.add_argument(‘-o’, ‘–outfile’, help=’the file save result’, dest=’outfile’, default=’result.txt’) result = arg.parse_args() headers = {‘User-Agent’:’Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; The World)’} class Bd_url(threading.Thread): def __init__(self, que): threading.Thread.__init__(self) self._que = que def run(self): while not self._que.empty(): URL = self._que.get() try: self.bd_url_collect(URL) except Exception,e: print e pass def bd_url_collect(self, url): r = requests.get(url, headers=headers, timeout=5) soup = bs(r.content, ‘lxml’, from_encoding=’utf-8′) bqs = soup.find_all(name=’a’, attrs={‘data-click’:re.compile(r’.’), ‘class’:None}) for bq in bqs: r = requests.get(bq[‘href’], headers=headers, timeout=5) if r.status_code == 200: print r.url with open(result.outfile, ‘a’) as f: f.write(r.url + ‘\n’) def main(): thread = [] thread_count = result.thread_count que = Queue() for i in range(0,(result.pagecount)): que.put(‘https://www.baidu.com/s?wd=’ + result.keyword + ‘&pn=’ + str(i)) for i in range(thread_count): thread.append(Bd_url(que)) for i in thread: i.start() for i in thread: i.join() if __name__ == ‘__main__’: start = time.clock()……