宜搜全站数十万小说爬虫
bigzhangxy
8年前
<p>自从看了师傅爬了顶点全站之后，我也手痒痒的，也想爬一个比较牛逼的小说网看看，于是选了宜搜这个网站，好了，马上开干，这次用的是mogodb数据库，感觉mysql太麻烦了下图是我选择宜搜里面遍历的网站</p> <p><img src="https://simg.open-open.com/show/3c7e927b8acd8caa76b7d2cb382f2be4.png"></p> <p style="text-align:center">Paste_Image.png</p> <p>先看代码框架图</p> <p style="text-align:center"><img src="https://simg.open-open.com/show/92cef2f5b3c0d2813ff2eb9c99853dd2.png"></p> <p style="text-align:center">Paste_Image.png</p> <p>第一个，肯定先提取排行榜里面每个类别的链接啊，然后进入链接进行爬取，先看all_theme文件</p> <pre> <code class="language-python">from bs4 import BeautifulSoup import requests from MogoQueue import MogoQueue spider_queue = MogoQueue('novel_list','crawl_queue')#实例化封装数据库操作这个类，这个表是存每一页书籍的链接的 theme_queue = MogoQueue('novel_list','theme_queue')#这个表是存每一个主题页面的链接的 html = requests.get('http://book.easou.com/w/cat_yanqing.html') soup = BeautifulSoup(html.text,'lxml') all_list = soup.find('div',{'class':'classlist'}).findAll('div',{'class':'tit'}) for list in all_list: title = list.find('span',{'class':'name'}).get_text() book_number = list.find('span',{'class':'count'}).get_text() theme_link = list.find('a')['href'] theme_links='http://book.easou.com/'+theme_link#每个书籍类目的数量 #print(title,book_number,theme_links)找到每个分类的标题和每个类目的链接，然后再下面的links提取出来 theme_queue.push_theme(theme_links,title,book_number) links=['http://book.easou.com//w/cat_yanqing.html', 'http://book.easou.com//w/cat_xuanhuan.html', 'http://book.easou.com//w/cat_dushi.html', 'http://book.easou.com//w/cat_qingxiaoshuo.html', 'http://book.easou.com//w/cat_xiaoyuan.html', 'http://book.easou.com//w/cat_lishi.html', 'http://book.easou.com//w/cat_wuxia.html', 'http://book.easou.com//w/cat_junshi.html', 'http://book.easou.com//w/cat_juqing.html', 'http://book.easou.com//w/cat_wangyou.html', 'http://book.easou.com//w/cat_kehuan.html', 'http://book.easou.com//w/cat_lingyi.html', 'http://book.easou.com//w/cat_zhentan.html', 'http://book.easou.com//w/cat_jishi.html', 'http://book.easou.com//w/cat_mingzhu.html', 'http://book.easou.com//w/cat_qita.html', ] def make_links(number,url):#这里要解释一下，因为每个类目的书页不同，而且最末页是动态数据，源代码没有 #这里采取了手打上最后一页的方法，毕竟感觉抓包花的时间更多 for i in range(int(number)+1): link=url+'?attb=&s=&tpg=500&tp={}'.format(str(i)) spider_queue.push_queue(link)#这里将每一页的书籍链接插进数据库 #print(link) make_links(500,'http://book.easou.com//w/cat_yanqing.html') make_links(500,'http://book.easou.com//w/cat_xuanhuan.html') make_links(500,'http://book.easou.com//w/cat_dushi.html') make_links(5,'http://book.easou.com//w/cat_qingxiaoshuo.html') make_links(500,'http://book.easou.com//w/cat_xiaoyuan.html') make_links(500,'http://book.easou.com//w/cat_lishi.html') make_links(500,'http://book.easou.com//w/cat_wuxia.html') make_links(162,'http://book.easou.com//w/cat_junshi.html') make_links(17,'http://book.easou.com//w/cat_juqing.html') make_links(500,'http://book.easou.com//w/cat_wangyou.html') make_links(474,'http://book.easou.com//w/cat_kehuan.html') make_links(427,'http://book.easou.com//w/cat_lingyi.html') make_links(84,'http://book.easou.com//w/cat_zhentan.html') make_links(9,'http://book.easou.com//w/cat_jishi.html') make_links(93,'http://book.easou.com//w/cat_mingzhu.html') make_links(500,'http://book.easou.com//w/cat_qita.html')</code></pre> <p>看看运行结果，这是书籍类目的</p> <p><img src="https://simg.open-open.com/show/71d0154b64ed516e48d9580ca062723e.png"></p> <p style="text-align:center">Paste_Image.png</p> <p>这是构造出的每一个类目里面所有的页数链接，也是我们爬虫的入口，一共5000多页</p> <p><img src="https://simg.open-open.com/show/20cb4d90eb16ab39dfcfd4cfdde4bfea.png"></p> <p style="text-align:center">Paste_Image.png</p> <p>接下来是封装的数据库操作，因为用到了多进程以及多线程每个进程，他们需要知道那些URL爬取过了、哪些URL需要爬取！我们来给每个URL设置两种状态：</p> <p>outstanding:等待爬取的URL</p> <p>complete:爬取完成的URL</p> <p>processing:正在进行的URL。</p> <p>嗯！当一个所有初始的URL状态都为outstanding；当开始爬取的时候状态改为：processing；爬取完成状态改为：complete；失败的URL重置状态为：outstanding。为了能够处理URL进程被终止的情况、我们设置一个计时参数，当超过这个值时；我们则将状态重置为outstanding。</p> <pre> <code class="language-python">from pymongo import MongoClient,errors from _datetime import datetime,timedelta class MogoQueue(): OUTSIANDING = 1 PROCESSING = 2 COMPLETE = 3 def __init__(self,db,collection,timeout=300): self.client=MongoClient() self.Clinet=self.client[db] self.db=self.Clinet[collection] self.timeout=timeout def __bool__(self): record = self.db.find_one( {'status': {'$ne': self.COMPLETE}} ) return True if record else False def push_theme(self,url,title,number):#这个函数用来添加新的URL以及URL主题名字进去队列 try: self.db.insert({'_id':url,'status':self.OUTSIANDING,'主题':title,'书籍数量':number}) print(title,url,'插入队列成功') except errors.DuplicateKeyError as e:#插入失败则是已经存在于队列了 print(title,url,'已经存在队列中') pass def push_queue(self,url): try: self.db.insert({'_id':url,'status':self.OUTSIANDING}) print(url,'插入队列成功') except errors.DuplicateKeyError as e:#插入失败则是已经存在于队列了 print(url,'已经存在队列中') pass def push_book(self,title,author,book_style,book_introduction,book_url): try: self.db.insert({'_id':book_url,'书籍名称':title,'书籍作者':author,'书籍类型':book_style,'简介':book_introduction}) print(title, '书籍插入队列成功') except errors.DuplicateKeyError as e: print(title, '书籍已经存在队列中') pass def select(self): record = self.db.find_and_modify( query={'status':self.OUTSIANDING}, update={'$set':{'status': self.PROCESSING, 'timestamp':datetime.now() }} ) if record: return record['_id'] else: self.repair() raise KeyError def repair(self): record = self.db.find_and_modify( query={ 'timestamp':{'$lt':datetime.now()-timedelta(seconds=self.timeout)}, 'status':{'$ne':self.COMPLETE} }, update={'$set':{'status':self.OUTSIANDING}}#超时的要更改状态 ) if record: print('重置URL',record['_id']) def complete(self,url): self.db.update({'_id':url},{'$set':{'status':self.COMPLETE}})</code></pre> <p>接下来是爬虫主程序</p> <pre> <code class="language-python">from ip_pool_request import html_request from bs4 import BeautifulSoup import random import multiprocessing import time import threading from ip_pool_request2 import download_request from MogoQueue import MogoQueue def novel_crawl(max_thread=8): crawl_queue = MogoQueue('novel_list','crawl_queue')#实例化数据库操作，链接到数据库，这个是爬虫需要的书籍链接表 book_list = MogoQueue('novel_list','book_list')#爬取的内容放进这里 def pageurl_crawler(): while True: try: url = crawl_queue.select()#从数据库提取链接，开始抓 print(url) except KeyError:#触发这个异常，则是链接都被爬完了 print('队列没有数据，你好坏耶') else: data=html_request.get(url,3) soup = BeautifulSoup(data,'lxml') all_novel = soup.find('div',{'class':'kindContent'}).findAll('li') for novel in all_novel:#提取所需要的所以信息 text_tag =novel.find('div',{'class':'textShow'}) title = text_tag.find('div',{'class':'name'}).find('a').get_text() author = text_tag.find('span',{'class':'author'}).find('a').get_text() book_style = text_tag.find('span',{'class':'kind'}).find('a').get_text() book_introduction= text_tag.find('div',{'class':'desc'}).get_text().strip().replace('\n','') img_tag = novel.find('div',{'class':'imgShow'}).find('a',{'class':'common'}) book_url = 'http://book.easou.com/' + img_tag.attrs['href'] book_list.push_book(title,author,book_style,book_introduction,book_url) crawl_queue.complete(url)#完成之后改变链接的状态 #print(title,author,book_style,book_introduction,book_url) threads=[] while threads or crawl_queue: for thread in threads: if not thread.is_alive(): threads.remove(thread) while len(threads)< max_thread: thread = threading.Thread(target=pageurl_crawler())#创建线程 thread.setDaemon(True)#线程保护 thread.start() threads.append(thread) time.sleep(5) def process_crawler(): process=[] num_cups=multiprocessing.cpu_count() print('将会启动的进程数为',int(num_cups)-2) for i in range(int(num_cups)-2): p=multiprocessing.Process(target=novel_crawl)#创建进程 p.start() process.append(p) for p in process: p.join() if __name__ == '__main__': process_crawler()</code></pre> <p>让我们来看看结果吧</p> <p><img src="https://simg.open-open.com/show/ebe6941e809ca0d0ebe4c557ad219a2a.png"></p> <p style="text-align:center">Paste_Image.png</p> <p>里面因为很多都是重复的，所有去重之后只有十几万本，好失望......</p> <p> </p> <p>来自：http://www.jianshu.com/p/a1c5183f3f4d</p> <p> </p>