一.爬虫基础
1.1 requests类
1.1.1 request的7个方法
requests.request() 实例化一个对象,拥有以下方法
requests.get(url, *args)
requests.head() 头信息
requests.post()
requests.put()
requests.patch() 修改一部分内容
requests.delete()
url = "http://quanben5.com/n/doupocangqiong/6.html"headers={"User-Agent":"Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.81 Safari/537.36"}
r= requests.get(url, headers=headers)
data={"pinyin": "doupocangqiong","content_id": "4",
}
r= requests.post(url, data=data, headers=headers)
1.1.2*arg里面的参数
params 字典或者字节序列,作为参数增加到url中
data 字典字节序列文件对象, 放在url里面对应的地方
json 作为requests的内容
headers 字典 模拟服务头
cookies 字典 cookieJar
auth 元组
files 字典类型,传输文件
timeout 设定的超时时间
proxies 字典类型,设定访问代理服务器,可以增加登录认证 pxs={"http":"http://user:pass@10.10.1.1234"}
allow_redirects 默认True 是否允许重定向
stream 默认TRUE 获得数据立刻下载
verity 默认True 认证SSL证书开关
cert 本地SSL证书路径
1.2 BeautifulSoup类
soup = BeautifulSoup("","html.parser")
soup.prettify()
soup.find_all(name, attrs, recursive, string, **kwargs)
name: 标签名称
attrs: 属性
string: 检索字符串
soup.head
soup.head.contents 列表
soup.body.contents[1]
soup.body.children
soup.body.descendants
.parent
.parents
.next_sibling 下一个
.previous_sibling 上一个
.next_siblings 下一个所有 迭代
.previous_siblings 上一个所有
1.3 selenium
1 from selenium importwebdriver2 from selenium.webdriver.chrome.options importOptions3 importtime4
5
6 chrome_options =Options()7 chrome_options.add_argument('--headless')8 chrome_options.add_argument('--disable-gpu')9
10 d = webdriver.Chrome(chrome_options=chrome_options) #设置成不显示浏览器的模式
11 d.get('http://quanben5.com/n/dazhuzai/23241.html')12 time.sleep(2)13 print(d.page_source) #显示出的是加载完之后的内容
查找单个和多个元素
d.find_element_by_id
d.find_elements_by_id
元素交互
from selenium importwebdriver
d=webdriver.Chrome()
d.get('http://www.baidu.com')
inputText= d.find_element_by_id("kw")
inputText.send_keys("萝莉")
button= d.find_element_by_id("su")
button.click()
二.实战
首先要找到可以在线看小说的网页
这里我随便百度了一下,首先选择了一个全本5200小说网("https://www.qb5200.tw")
打开某小说章节目录表
https://www.qb5200.tw/xiaoshuo/0/357/
查看源代码
发现正文卷是在class为listmain的div下面的第二个dt标签里
之后的路径标签为a
1 url_text = soup.find_all('div', 'listmain')[0] #找到第一个class=listmain的div标签
2 main_text = url_text.find_all('dt')[1] #找到下面的第二个dt标签
3 for tag inmain_text.next_siblings:4 try:5 url = ''.join(['https://', host, tag.a.attrs['href']])6 print('parsering', url)7 except:8 continue
找到每个章节的url
在随便打开一章
查看源代码为:
1 def getHTMLText(url, headers={}):2 """
3 获取网站源码4 :param url:5 :return: class response6 """
7 if headers !={}:8 headers = {"User-Agent":"Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.81 Safari/537.36"}9
10 #proxies = get_random_ip()
11 proxies ={}12 try:13 #print("start", url)
14 r = requests.get(url, proxies=proxies, headers=headers)15 r.raise_for_status()16 #r.encoding = r.apparent_encoding
17 #print('end', url)
18 returnr19 except:20 returnr.status_code21
22 defparseByQB5200(soup, host, f):23 """
24 在全本小说网的源码下爬取小说25 :param soup, host:26 :param f:27 :return:28 """
29 url_text = soup.find_all('div', 'listmain')[0]30 main_text = url_text.find_all('dt')[1]31 x =032 for tag inmain_text.next_siblings:33 time.sleep(1)34 try:35 url = ''.join(['https://', host, tag.a.attrs['href']])36 print('parsering', url)37 soup = BeautifulSoup(getHTMLText(url).text, "html.parser")38 passage = soup.find_all("div", "content")[0]39 title =passage.h1.string40 f.writelines(''.join([title, '\n']))41 passage = soup.find_all("div", "showtxt")[0]42 for i inpassage.descendants:43 if i.name != "br":44 st =i.string45 if st.strip().startswith('http') or st.strip().startswith('请记住本书'):46 continue
47 f.writelines(''.join([' ', st.strip(), '\n']))48 x += 1
49 print('%d.%s 下载完成' %(x, title))50 except:51 continue
52
53 defgetNovelUrls(url):54 """
55 通过小说的目录网址判断小说所在的网站56 并调用属于该网站的爬虫语句57 :param url:58 :return:59 """
60
61 response =getHTMLText(url)62 host = url.split('//')[1].split('/')[0]63 host_list ={64 "www.qb5200.tw": parseByQB5200,65 #"www.qu.la": parseByQuLa,
66 "quanben5.com": parseByQB567 }68 print(response)69 soup = BeautifulSoup(response.text, 'html.parser')70 with open('1.txt', 'w', encoding='utf8') as f:71 host_list[host](soup, host, f)72
73 if __name__ == '__main__':74 getNovelUrls("https://www.qb5200.tw/xiaoshuo/0/357/")
在全本5200爬取小说txt
问题在于
全本小说网("www.qb5200.tw")在这样的暴力获取下只允许爬3次,之后就403错误
本来以为是同一IP限制访问次数, 使用了IP代理之后发现问题依旧
猜测应该是跟请求头有关
因此加上了访问该网站时浏览器的所有请求头,并且把User-Agent设置为随机
USER_AGENT_LIST =['MSIE (MSIE 6.0; X11; Linux; i686) Opera 7.23','Opera/9.20 (Macintosh; Intel Mac OS X; U; en)','Opera/9.0 (Macintosh; PPC Mac OS X; U; en)','iTunes/9.0.3 (Macintosh; U; Intel Mac OS X 10_6_2; en-ca)','Mozilla/4.76 [en_jp] (X11; U; SunOS 5.8 sun4u)','iTunes/4.2 (Macintosh; U; PPC Mac OS X 10.2)','Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:5.0) Gecko/20100101 Firefox/5.0','Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:9.0) Gecko/20100101 Firefox/9.0','Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:16.0) Gecko/20120813 Firefox/16.0','Mozilla/4.77 [en] (X11; I; IRIX;64 6.5 IP30)','Mozilla/4.8 [en] (X11; U; SunOS; 5.7 sun4u)']
headers= {"User-Agent": random.choice(USER_AGENT_LIST),"Host": "www.qb5200.tw","Connection": "keep-alive","Pragma": "no-cache","Cache-Control": "no-cache","Upgrade-Insecure-Requests": "1","Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8","Referer": "https://www.qb5200.tw/xiaoshuo/0/355/","Accept-Encoding": "gzip, deflate, br","Accept-Language": "zh-CN,zh;q=0.9","Cookie": "UM_distinctid=16736c47b1d8b-04bd85d2b5e8ab-50422618-144000-16736c47b1e1f2; bcolor=; font=; size=; fontcolor=; width=; CNZZDATA1260750615=1497982366-1542811927-%7C1542882690; fikker-7ZUK-qXIL=qDaKEBihatSNEFgtMlGVIKubCHYLi89J; fikker-7ZUK-qXIL=qDaKEBihatSNEFgtMlGVIKubCHYLi89J; fikker-Sbwr-GN9F=GbMX3HOhwvoDMxopLMGt3VWliXQIK0SP; fikker-Sbwr-GN9F=GbMX3HOhwvoDMxopLMGt3VWliXQIK0SP; fikker-yJ3O-W61D=UfinETMnCR38ADWZEl1KNHQRU2m81Fwb; fikker-yJ3O-W61D=UfinETMnCR38ADWZEl1KNHQRU2m81Fwb; fikker-rWK3-6KHs=T9T5b7lYVJReTQviPm2IdLPyHu83RwFM; fikker-rWK3-6KHs=T9T5b7lYVJReTQviPm2IdLPyHu83RwFM"}
随机请求头
解决问题.
三.使用ajax动态加载的实例:
全本5小说网("quanben5.com")
同样的方法搜索源代码
然而发现了问题
给出的html页面只有一半的源码
因此按F12打开检查
发现所有的文本存在这个xhr里
点击查看请求头信息
发现是post请求
请求的url是/index.php?c=book&a=ajax_content
请求的数据在最下面的form表单里
打开网页源文件和js文件,搜索这些表单信息
分别在ajax.js里和源文件里找到了这些
源文件里面的可以直接生成data数据表单
在ajax.js里可以知道rndval字段是当前时间,精确到毫秒
四.优化
采用了gvent进行异步IO处理,每一张网页保存在temp里面,最后将文件合成一个txt
加入了搜索功能,目前仅支持一个小说网站
代码如下:
1 #!/usr/bin/env python
2 #encoding: utf-8
3
4 """
5 @version:6 @author: Wish Chen7 @contact: 986859110@qq.com8 @file: get_novels.py9 @time: 2018/11/21 19:4310
11 """
12 importgevent13 from gevent importmonkey14 monkey.patch_all()15 importrequests, time, random, re, os16 from bs4 importBeautifulSoup17
18 dir = os.path.dirname(os.path.abspath(__file__))19
20 def getHTMLText(url, data=[], method='get'):21 """
22 获取网站的源代码23 请求默认为get24 :param url:25 :param data:26 :param method:27 :return:28 """
29 headers = {"User-Agent":"Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.81 Safari/537.36"}30 proxies ={}31 try:32 #print("start", url)
33 r = requests.request(method, url, proxies=proxies, headers=headers, data=data)34 r.raise_for_status()35 #r.encoding = r.apparent_encoding
36 #print('end', url)
37 returnr38 except:39 returnr.status_code40
41
42 deffetch_async(x, url):43 """
44 异步IO所需要执行的操作45 获取源文件46 模拟向ajax请求获取完整文字47 每一章输入到temp文件夹下48 :param x:49 :param url:50 :return:51 """
52 url_main = "http://quanben5.com/index.php?c=book&a=ajax_content"
53 r = getHTMLText(url) #获取每一章的源文件
54 title = re.search(r'
(.*)
', r.text).group(1)55 result = re.search(r'', r.text).group(1)56 num_list = result.split("','")57 num_list[9] = num_list[9][:-1]58 content = {} #开始模拟post请求发送的表单59 for i in range(1, 5):60 content[num_list[i * 2]] = num_list[i * 2 + 1]61 content['_type'] = "ajax"
62 content['rndval'] = int(time.time() * 1000)63 r = getHTMLText(url_main, data=content, method='post') #模拟post请求
64 soup = BeautifulSoup(r.text, "lxml")65 with open(os.path.join(dir, 'temp', "%s.txt" % x), "w", encoding='utf8') as f:66 f.writelines(''.join([str(x), '.', title, '\n\n']))67 for tag insoup.body.children:68 if tag.name == 'p':69 f.writelines(''.join([' ', tag.string.strip(), '\n\n']))70 print('%d.%s 下载完成' %(x, title))71
72
73 defget_together(name, author, x):74 """
75 将temp目录下的各网页下载下来的txt76 合并在一起77 并删除temp文件78 :param name:79 :param author:80 :return:81 """
82 with open(os.path.join(dir, "%s.txt" % name), "w", encoding='utf8') as f:83 f.writelines(''.join([name, '\n\n作者:', author, '\n\n']))84
85 for i inrange(x):86 try:87 f.write(open(os.path.join(dir, 'temp', "%s.txt" % (i+1)), "r", encoding='utf8').read())88 f.write('\n\n')89 #os.remove(os.path.join(dir, 'temp', "%s.txt" % (i+1)))
90 except:91 continue
92
93
94 defparseByQB5(response, host):95 """
96 在全本5小说网的源码下爬取小说97 获得书名和作者98 采用gevent异步IO优化99 :param response:100 :param host:101 :return:102 """
103 soup = BeautifulSoup(response.text, 'html.parser')104 url_text = soup.find_all('div', 'box')[2]105 main_text = url_text.find_all('h2')[0].next_sibling106 url_list =[]107 for tag inmain_text.descendants:108 if tag.name == 'li':109 url = ''.join(['http://', host, tag.a.attrs['href']])110 url_list.append(url)111 from gevent.pool importPool112 pool = Pool(100)113
114 gevent.joinall([pool.spawn(fetch_async, i+1, url=url_list[i]) for i inrange(len(url_list))])115
116 name = re.search(r"
(.*)
", response.text).group(1)117 author = re.search(r' ', response.text).group(1)118 print("%d文档已下载,正在合并..." %len(url_list))119 get_together(name, author, len(url_list))120121
122 defgetNovelUrls(url):123 """
124 通过小说的目录网址判断小说所在的网站125 并调用属于该网站的爬虫语句126 :param url:127 :return:128 """
129
130 response =getHTMLText(url)131 host = url.split('//')[1].split('/')[0]132 host_list ={133 "quanben5.com": parseByQB5134 }135 host_list[host](response, host)136
137
138 defget_url():139 input_name = input('>>')140 r = getHTMLText("http://quanben5.com//index.php?c=book&a=search&keywords=%s" %input_name)141 soup = BeautifulSoup(r.text, "html.parser")142 main_book = soup.find_all("div", "pic_txt_list")143 for i inrange(len(main_book)):144 tag =main_book[i].h3145 print("%s.%s %s" %(i, tag.span.text, tag.next_sibling.next_sibling.text))146 choice = int(input(">>"))147 if choice inrange(len(main_book)):148 return ''.join(["http://quanben5.com", main_book[choice].h3.a["href"], "xiaoshuo.html"])149
150
151 if __name__ == '__main__':152 #url_list = [
153 #"https://www.qu.la/book/365/",
154 #"https://www.qb5200.tw/xiaoshuo/0/357/",
155 #"http://quanben5.com/n/doupocangqiong/xiaoshuo.html",
156 #"http://quanben5.com/n/dazhuzai/xiaoshuo.html",
157 #"http://quanben5.com/n/douluodalu/xiaoshuo.html",
158 #"http://quanben5.com/n/renxingderuodian/xiaoshuo.html"
159 #]
160 #if not os.path.exists('temp'):
161 #os.mkdir('temp')
162 #getNovelUrls(url_list[5])
163 whileTrue:164 url =get_url()165 time_start =time.time()166 getNovelUrls(url)167 print("成功爬取! 用时:%ds" % (int((time.time()-time_start)*100)/100))
异步IO+搜索
封装性还不够好
最好一个网站用一个类来封装
采用scrapy框架
正在设计中...
五.未解决问题:
代理IP问题:
目前只有免费的代理IP网
生成随机IP并使用代理IP访问
多网站定制
要观察各个网站的目录源代码结构以及文章源代码结构
每一个网站都可以用一个parse函数来解析其内容