# -*- coding:utf-8 -*- # 仅需修改这个地方https://jn.lianjia.com/ershoufang/pg{}rs/ 将jn换成你所在城市的拼写首字母小写 import requests from lxml import etree import time import random import csv import requests import json class LianjiaSpider(object): def __init__(self): self.url = "https://mobilenext-web.meituan.com/api/newSalesBoard/getSaleBoardDetail?cityId=234&boardType=8&districtId=0&cateId={}&offset=0&limit=50" self.headers = { "User-Agent": "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/535.1 (KHTML, like Gecko) Chrome/14.0.835.163 Safari/535.1"} def get_page(self, url,i): res = requests.get(url=url, headers=self.headers) res.encoding = "utf-8" html = res.text print(i) # print(html) # results_temp = html.replace('{"totalSize":50,"saleBoardDealList":', "").replace("}}]}", "") # results = results_temp + "}}]" # print(results) self.parse_page(html,i) # print(html) # print(i) def parse_page(self, html,i): results = html[35:-20] print(results) print(i) for list in json.loads(results): # print(list) id = list["id"] name = list["name"] weekSaleCount = list["weekSaleCount"] score = list["score"] avgPrice = list["avgPrice"] cateName = list["cateName"] areaName = list["areaName"] distance = list["distance"] rank = list["rank"] frontImg = list["frontImg"] oneSentence = list["oneSentence"] if (i == 1): catId = "火锅" print(catId) elif (i == 2): catId = "自助餐" print(catId) elif (i == 3): catId = "烧烤龙虾" print(catId) elif (i == 4): catId = "地方菜" print(catId) elif (i == 5): catId = "异国料理" print(catId) elif (i == 6): catId = "小吃快餐" print(catId) elif (i == 7): catId = "甜点饮品" print(catId) elif (i == 8): catId = "蛋糕" print(catId) # SaleBoardPoiGroup = list["SaleBoardPoiGroup"] # saleBoardPoiCoupon = list["saleBoardPoiCoupon"] # saleBoardPoiPay = list["saleBoardPoiPay"] # banchList = list["banchList"] with open('meituan.csv', 'a', newline='', encoding='utf-8')as f: write = csv.writer(f) # write.writerow([id,name,weekSaleCount,score,avgPrice,cateName,areaName,distance,rank,frontImg,oneSentence,SaleBoardPoiGroup,saleBoardPoiCoupon,saleBoardPoiPay,banchList]) write.writerow([catId, name, weekSaleCount, score, avgPrice, cateName, areaName, distance, rank, frontImg, oneSentence]) f.close() def main(self): for i in range(1, 9): print(i) time.sleep(random.randint(3, 5)) url = self.url.format(i) # print(url) self.get_page(url,i) # print(j) if __name__ == '__main__': start = time.time() spider = LianjiaSpider() spider.main() end = time.time() print("执行时间:%.2f" % (end - start))
mt店铺榜+餐饮分类
最新推荐文章于 2024-05-08 03:19:42 发布