#__author: han-zhang
#date: 2019/8/16 15:30
from lxml import etree
import urllib.request
import json
#url
url='https://movie.douban.com/j/chart/top_list?type=17&interval_id=100%3A90&action=&start=0&limit=20'
#headers
headers={
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/76.0.3809.100 Safari/537.36'
}
request=urllib.request.Request(url=url,headers=headers)
content=urllib.request.urlopen(request).read().decode('utf8')
#content格式为json
#将其转化为python对象
obj=json.loads(content)
# print(obj)
lt=[]
for movie in obj:
#海报
img_src=movie['cover_url']
print(img_src)
#名称
title = movie['title']
#评分
score = movie['score']
#评价人数
vote_count=movie['vote_count']
#国家
regions=movie['regions']
item={
'电影海报':img_src,
'电影名称':title,
'电影评分':score,
'评价人数':vote_count,
'电影国家':regions,
}
lt.append(item)
#保存
string=json.dumps(lt,ensure_ascii=False)
with open('JSON.json','w')as f:
f.write(string)