爬虫案例（二）

1.开心网

(1) 使用cookie登录

#cookie做模拟登录 只需在请求头中添加cookie即可
import  requests
#定义请求头
headers={
   
'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36',
'Cookie': '_ref=5f8f94e847b41; _cpmuid=1296944641; SERVERID=_srv80-67_; BAIDU_SSP_lcr=http://www.baidu.com/link?url=cyev6UfdsEqFmgtrz05yoG6-2ZFyQOEyz33slP-6xs8V4rcZSNH5tSKm8fLwpuBv&wd=&eqid=8360dd6c0042c4cf000000035f8f94e5; Hm_lvt_500f908d39095efce74d0e9c64f55ffb=1603245293; _vid=C91B1909DA9000015E8C157049A699C0; _reg_src=baidu_www_link; _user=d730946172043939d49e3d29322463b6_182006254_1603245412; _preemail=17547065294; onlinenum=c%3A0; noFocusContent=1; Hm_lpvt_500f908d39095efce74d0e9c64f55ffb=1603245461; wpresence=-FnCP4OlWdr2b8C-Oyl4GfYzNGc3qsBuX4-VlQ.ZGZ0MTgyMDA2MjU0'
}
#请求响应
response = requests.get(url='http://www.kaixin001.com/home/?uid=182006254',headers=headers)
#保存文件
with open('kaixin_cookie.html','w',encoding='utf8') as f:
    f.write(response.text)

(2).使用session登录

#模拟登陆 生成一个session对象
import  requests

# 创建一个session对象
zz=requests.session()

#  使用session对象发起请求
# 确定URL：form表单中的action属性
# 确定请求方式：form表单中的method属性
# 定义from表单传递的参数
data={
   
    "loginemail":'17547065294',
    "password":'woaini.520'
}

response = zz.post(url='https://security.kaixin001.com/login/login_post.php',data=data)

# 保存文件
with open('kaixin_session.html','w',encoding='utf8') as f:
    f.write(response.text)

2.百度首页抓取

import  requests

headers={
   
'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36'
}
proxy={
   
    # "http": "http://121.69.97.22:8000"
    "http": "http://113.194.49.146:	9999"
}
# response = requests.get(url=