最近看了python的scrapy 框架并用其抓取了部分知乎用户数据，代码主要是集中在知乎登陆和抓取时候的逻辑处理上。
1、 首先进入知乎登陆页面zhihu.com/#sigin上， 用xpath提取_xsrf参数， 获取验证码的部分url，完整的url是由当前的时间戳和type参数构成，利用得到的url形成response， 在函数handle_captcha对验证码提取并提示在终端输入验证码，最后再将登陆的url、cookie、用户账号、密码什么的from进去就可以登陆成功了。下面是代码：
# _*_coding:utf-8_*_ from scrapy.spider import CrawlSpider
from scrapy.http import Request, FormRequest
from scrapy.selector import Selector
from zhihu2 import config
from PIL import Image
import time
import json
import re
from zhihu2 import items class ZhiHu_spider(CrawlSpider):
name = 'zhihu2'
allowed_domain = ['https://www.zhihu.com'] def __init__(self, *args, **kwargs):
super(ZhiHu_spider, self).__init__(*args, **kwargs)
self.xsrf = ''
self.headers = config.headers def start_requests(self):
yield Request(
'http://www.zhihu.com/#signin',
meta={
'cookiejar': 1
},
callback=self.post_login
) def post_login(self, response):
print 'parper login in '
sel = Selector(response)
self.xsrf = sel.xpath('//input[@name="_xsrf"]/@value').extract()[0] #验证码的获取 没有自动识别 识别率太低 所以手打
str_time = str(time.time() * 1000)
cap_url = 'https://www.zhihu.com/captcha.gif?r=' + str_time + '&type=login'
print cap_url
yield Request(
cap_url,
meta={'cookiejar': response.meta['cookiejar'],
'_xsrf': self.xsrf,
}, headers=self.headers,
callback=self.handle_captcha ) def handle_captcha(self, response):
with open('E:\\myscrapy\\captcha.gif', 'wb') as gif:
gif.write(response.body)
gif.close()
Im = Image.open('E:\\myscrapy\\captcha.gif')
Im.show()
captcha = raw_input('enter your captcha:') yield FormRequest(
'http://www.zhihu.com/login/phone_num', #s手机号登陆， 对应的可以换成邮箱
method='POST',
meta={'cookiejar': response.meta['cookiejar']},
headers=self.headers,
formdata={
'_xsrf': self.xsrf,
'password': '密码',
'remember_me': 'true',
'phone_num': '账号',
'captcha': captcha
},
callback=self.after_login, )
2、下面是登陆之后获取关注人的信息，由于知乎第一次只会显示20个关注人，剩下的要post数据到www.zhihu.com/node/ProfileFolloweesList2
才能又获取20个，所以在这要获取每个人的关注人数并与20做对比。
# 获取个人主页
def after_login(self, response):
print response.body
print 'login success'
yield Request(
'https://www.zhihu.com/people/你的id需要填写, #自己主页的网址 因为我没获取id 所以要输入自己主页的网址
meta={'cookiejar': response.meta['cookiejar']},
headers=self.headers,
callback=self.parse_people,
)
#获取关注人url
def parse_people(self, response):
# print 'ready'
sel = Selector(response)
follow_url = sel.xpath('//a[@class="item"]/@href').extract_first()
if follow_url:
compete_url = 'https://www.zhihu.com' + follow_url yield Request(
compete_url,
meta={
'cookiejar': response.meta['cookiejar'],
},
headers=self.headers,
callback=self.person_info,
) #处理关注人的url 并获取信息
def person_info(self, response):
item = items.Zhihu2Item()
count = 20
sel = Selector(response) nikname = sel.xpath('//div[@class="title-section"]/a[@class="name"]/text()').extract_first()
location = sel.xpath('//span[@class="location item"]/@title').extract_first()
business = sel.xpath('//span[@class="business item"]/@title').extract_first()
education = sel.xpath('//span[@class="education item"]/@title').extract_first()
education_extra = sel.xpath('//span[@class="education-extra item"]/@title').extract_first()
sex = sel.xpath('//span[@class="item gender"]/i/@class').extract_first().split('-')[-1]
agree = sel.xpath('//span[@class="zm-profile-header-user-agree"]/strong/text()').extract_first()
thanks = sel.xpath('//span[@class="zm-profile-header-user-thanks"]/strong/text()').extract_first() config.try_none(nikname)
config.try_none(location)
config.try_none(business)
config.try_none(education)
config.try_none(education_extra)
config.try_none(sex)
config.try_none(agree)
config.try_none(thanks) peo_num = sel.xpath('/html/body/div[3]/div[2]/div[1]/a[1]/strong/text()').extract_first()
item['nikname'] = nikname
item['business'] = business
item['education_extra'] = education_extra
item['location'] = location
item['education'] =education
item['sex'] = sex
item['agree'] = agree
item['thanks'] = thanks if peo_num: people_urls = sel.xpath('//a[@class="zg-link author-link"]/@href').extract()
for people_url in people_urls:
yield Request(
people_url,
meta={'cookiejar': response.meta['cookiejar']},
headers=self.headers,
callback=self.person_info
) peo_params = sel.xpath('//div[@class="zh-general-list clearfix"]/@data-init').extract_first()
if peo_params:
try:
values = json.loads(str(peo_params))
except ValueError, e:
print e.message
params = {}
params['offset'] = 20
params['order_by'] = 'created'
params['hash_id'] = values['params']['hash_id'] if count < peo_num:
params['offset'] = count
yield FormRequest(
'https://www.zhihu.com/node/ProfileFolloweesListV2',
method='POST',
meta={'cookiejar': response.meta['cookiejar']},
headers=self.headers,
formdata={
'method': 'next',
'params': json.dumps(params),
'_xsrf': self.xsrf,
},
callback=self.foolows_V2
)
count += 20
else:
num = peo_num / 20
params['offset'] = num
yield FormRequest(
'https://www.zhihu.com/node/ProfileFolloweesListV2',
method='POST',
meta={'cookiejar': response.meta['cookiejar']},
headers=self.headers,
formdata={
'method': 'next',
'params': json.dumps(params),
'_xsrf': self.xsrf,
},
callback=self.foolows_V2
)
3、从上面url的response获取关注人的url,得到的url 交由parse_people函数处理，parse_people函数的response交由person_info函数处理，所以就形成了一个循环，不断的有url被提取，也不断的有数据被提取出来，下面是parse_people函数的代码：
def foolows_V2(self, response):
p = re.compile(r'href="https://www\.zhihu\.com/people/(.*?)"') aa = json.loads(response.body)['msg']
for item in aa:
peo = p.search(item).group(1)
followes_url = 'https://www.zhihu.com/people/' + str(peo)
yield Request(
followes_url,
meta={'cookiejar': response.meta['cookiejar']},
headers=self.headers,
callback=self.parse_people
)
下面是一些配置信息：
cofig.py
#_*_coding:utf-8_*_
from settings import USER_AGENT
headers = {
'Host': 'www.zhihu.com',
'Connection': 'keep-alive',
'Pragma': 'no-cache',
'Cache-Control': 'no-cache',
'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',
'Origin': 'https://www.zhihu.com',
'X-Requested-With': 'XMLHttpRequest',
'User-Agent': USER_AGENT,
'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
'Referer': 'https://www.zhihu.com/',
'Accept-Encoding': 'gzip,deflate,sdch',
'Accept-Language': 'zh-CN,zh;q=0.8',
}
def try_none(tag):
try:
tag
except:
tag = 'none'
return tag
items.py:
from scrapy import Item, Field class Zhihu2Item(Item):
nikname = Field()
location = Field()
business = Field()
education = Field()
education_extra = Field()
sex = Field()
thanks = Field()
agree = Field()
代码没有维护已爬取的url和带爬取的url的重复，可能会导致重复抓取，代码的优化也挺烂的。希望大神们多给点意见，如果代码有错误，希望提出，以免给别人误导。