|
马上注册,结交更多好友,享用更多功能^_^
您需要 登录 才可以下载或查看,没有账号?立即注册
x
# ------------------ladys---------------------- #
import urllib.request
import os
import random
def url_open(url):
UserAgent = "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/31.0.1650.63 Safari/537.36"
req = urllib.request.Request(url)
req.add_header('User-Agent',UserAgent)
response = urllib.request.urlopen(req).read()
return response
#iplist = ['61.38.252.12:3128','58.255.127.104:80','113.31.80.194:8080','218.213.166.218:81'] #代理ip的数组
#proxy_support = urllib.request.ProxyHandler({'http':random.choice(iplist)})
#opener = urllib.request.build_opener(proxy_support)
#opener.addheaders = [('User-Agent','Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/31.0.1650.63 Safari/537.36')]
#urllib.request.install_opener(opener)
#response = urllib.request.urlopen(url).read()
def get_page(url):
html = url_open(url).decode('utf-8')
a = html.find('current-comment-page')+23
b = html.find(']',a)
return html[a:b]
def find_imgs(url):
html = url_open(url).decode('utf-8')
img_addrs = []
a = html.find('img src=')
while a != -1:
b = html.find('.jpg',a,a+255)
if b != -1:
img_addrs.append(html[a+9:b+4])
else:
b = a + 9
a = html.find('img src=',b)
return img_addrs
def save_imgs(folder,img_addrs):
for each in img_addrs:
filename = each.split('/')[-1]
if filename.find('<') != -1 or filename.find('!') != -1:
continue
else:
if os.path.isfile(filename) != True:
with open(filename,'wb') as f:
img = url_open('http:'+each)
f.write(img)
else:
print('图片已经存在了,请明天再试')
continue
def download_mm(folder='hanzi',pages=1):
isdir = os.path.isdir(folder)
if isdir != True:
os.mkdir(folder)
os.chdir(folder)
url = "http://g.jandan.net/n/hanzi"
for i in range(pages):
page_url = url+'?&page='+str(i)
img_addrs = find_imgs(page_url)
save_imgs(folder,img_addrs)
if __name__ == '__main__':
download_mm()
|
|