forked from wistbean/learn_python3_spider
-
Notifications
You must be signed in to change notification settings - Fork 0
/
douban_top_250_books.py
74 lines (54 loc) · 2.09 KB
/
douban_top_250_books.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
import requests
from bs4 import BeautifulSoup
import xlwt
def request_douban(url):
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) '
'Chrome/88.0.4324.146 Safari/537.36',
}
try:
response = requests.get(url=url, headers=headers)
if response.status_code == 200:
return response.text
except requests.RequestException:
return None
book = xlwt.Workbook(encoding='utf-8', style_compression=0)
sheet = book.add_sheet('豆瓣电影Top250', cell_overwrite_ok=True)
sheet.write(0, 0, '名称')
sheet.write(0, 1, '图片')
sheet.write(0, 2, '排名')
sheet.write(0, 3, '评分')
sheet.write(0, 4, '作者')
sheet.write(0, 5, '简介')
n = 1
def save_to_excel(soup):
list = soup.find(class_='grid_view').find_all('li')
for item in list:
item_name = item.find(class_='title').string
item_img = item.find('a').find('img').get('src')
item_index = item.find(class_='').string
item_score = item.find(class_='rating_num').string
item_author = item.find('p').text
if item.find(class_='inq') is not None:
item_intr = item.find(class_='inq').string
else:
item_intr = 'NOT AVAILABLE'
# print('爬取电影:' + item_index + ' | ' + item_name +' | ' + item_img +' | ' + item_score +' | ' + item_author +' | ' + item_intr )
print('爬取电影:' + item_index + ' | ' + item_name + ' | ' + item_score + ' | ' + item_intr)
global n
sheet.write(n, 0, item_name)
sheet.write(n, 1, item_img)
sheet.write(n, 2, item_index)
sheet.write(n, 3, item_score)
sheet.write(n, 4, item_author)
sheet.write(n, 5, item_intr)
n = n + 1
def main(page):
url = 'https://movie.douban.com/top250?start=' + str(page * 25) + '&filter='
html = request_douban(url)
soup = BeautifulSoup(html, 'lxml')
save_to_excel(soup)
if __name__ == '__main__':
for i in range(0, 10):
main(i)
book.save(u'豆瓣最受欢迎的250部电影.xlsx')