安居客二手房

from lxml import etree
from fake_useragent import UserAgent
import time
import json
import requests
import csv
import random

headers={
            'User-Agent':UserAgent().random
        }

# w=["小区名称","单价","物业类型","物业费","总建面积","总户数","竣工时间","容积率","绿化率","开发商"]
# with open('未央区.csv', 'a', newline='', encoding='utf-8-sig') as fp:
#     writer = csv.writer(fp)
#     writer.writerow(w)
#相关性url
proxies={'http':'125.120.201.120:808',
}
for p in range(1,50):
    
    url='https://xa.anjuke.com/community/weiyangq/p{}'.format(p)
    
    response=requests.get(url=url,headers=headers,proxies=proxies)
    page_text=response.text
    tree=etree.HTML(page_text)
    div_list=tree.xpath('//*[@id="__layout"]/div/section/section[3]/section/div[2]/a')
    print(p,"ok") 
        #爬链接
    for div in div_list:
        li=[]
        name=div.xpath('./div[2]/div[1]//text()')[0]
        # print(name)
        price=div.xpath('./div[3]/div//text()')[0]
        price1=div.xpath('./div[3]/div//text()')[1]
        sum=price+price1
        sum=sum.replace("\n","")
        sum=sum.replace(" ","")
        li.append(name)
        li.append(sum)
        
        url=div.xpath('./@href')[0]
        response=requests.get(url=url,headers=headers,proxies=proxies)
        page_text=response.text
        tree=etree.HTML(page_text)
        div_list1=tree.xpath('//*[@id="basic-infos-box"]/dl')
        for div1 in div_list1:
            name=div1.xpath('.//text()')
            i=1
            for i in range(len(name)):
                a=name[3]
                b=name[7].replace("\t","")
                b=b.replace("\n","")
                c=name[11].replace("\t","")
                c=c.replace("\n","")
                d=name[15]
                e=name[19]
                f=name[23]
                f=name[27]
                g=name[31]
                h=name[35]
                
                i=i+2
            k=[]
            k.append(a)
            k.append(b)
            k.append(c)
            k.append(d)
            k.append(e)
            k.append(f)
            k.append(g)
            k.append(h)
            
        li.extend(k)
        print(li)
        with open('未央区11.csv', 'a', newline='', encoding='utf-8-sig') as fp:
            writer = csv.writer(fp)
            writer.writerow(li)
        time.sleep(random.randint(4,14))