import urllib.request
import os
import re
from selenium import webdriver
browser = webdriver.Firefox(executable_path='/home/renpu/pachong/geckodriver')
def open_url(url):
header={'User-Agent':'Mozilla/5.0 (X11; Linux x86_64; rv:57.0) Gecko/20100101 Firefox/57.0'}
result=urllib.request.Request(url = url, headers=header)
print(result.headers)
html = urllib.request.urlopen(result)
return html.read()
def url_webdriver(url,browser):
browser.get(url)
return browser.page_source.encode()
def jindutiao(a,b,c):
pre = 100*a*b/c
if pre>100:
pre =100
print('%.3f%%'%pre)
else:
print('%.3f%%\r'%pre,end='')
def get_img(url,floder):
if os.path.exists(floder)!=True:
os.mkdir(floder)
print('chuangjianchenggong')
else:
print('Have exist!!!!')
os.chdir(floder)
html = open_url(url).decode('utf-8')
imgpath=r'img class="BDE_Image" src="([^"]+\.jpg)"'
imglist=re.findall(imgpath,html)
'''for each in imglist:
print(each)'''
for ea in imglist:
filename=ea.split('/')[-1]
urllib.request.urlretrieve(ea,filename,jindutiao)
def get_ip(url,floder):
if os.path.exists(floder)!=True:
os.mkdir(floder)
print('chuangjianchenggong')
else:
print('Have exist!!!!')
os.chdir(floder)
html = open_url(url).decode('utf-8')
html1 = url_webdriver(url,browser).decode('utf-8')
with open('htmldriver.txt','w') as fht:
fht.write(html1)
with open('html.txt','wb') as fhtml:
fhtml.write(open_url(url))
ip = r'(?:(?:25[0-5]|2[0-4]\d|[01]?\d?\d)\.){3}(?:25[0-5]|2[0-4]\d|[01]?\d?\d)'
iplist=re.findall(ip,html)
for i in iplist:
print(i)
#for iplis in iplist:
with open('ip.txt','w') as f:
f.write(str(iplist))
file = open('ip.txt')
a=file.read()
ip_list=re.findall(ip,a)
print(type(ip_list))
print(ip_list[0])
if __name__ == '__main__':
#url0 = 'http://tieba.baidu.com/p/3563409202'
url1 ='http://www.data5u.com'
floder='tanjingfeng'
#get_img(url0,floder)
get_ip(url1,floder)
import os
import re
from selenium import webdriver
browser = webdriver.Firefox(executable_path='/home/renpu/pachong/geckodriver')
def open_url(url):
header={'User-Agent':'Mozilla/5.0 (X11; Linux x86_64; rv:57.0) Gecko/20100101 Firefox/57.0'}
result=urllib.request.Request(url = url, headers=header)
print(result.headers)
html = urllib.request.urlopen(result)
return html.read()
def url_webdriver(url,browser):
browser.get(url)
return browser.page_source.encode()
def jindutiao(a,b,c):
pre = 100*a*b/c
if pre>100:
pre =100
print('%.3f%%'%pre)
else:
print('%.3f%%\r'%pre,end='')
def get_img(url,floder):
if os.path.exists(floder)!=True:
os.mkdir(floder)
print('chuangjianchenggong')
else:
print('Have exist!!!!')
os.chdir(floder)
html = open_url(url).decode('utf-8')
imgpath=r'img class="BDE_Image" src="([^"]+\.jpg)"'
imglist=re.findall(imgpath,html)
'''for each in imglist:
print(each)'''
for ea in imglist:
filename=ea.split('/')[-1]
urllib.request.urlretrieve(ea,filename,jindutiao)
def get_ip(url,floder):
if os.path.exists(floder)!=True:
os.mkdir(floder)
print('chuangjianchenggong')
else:
print('Have exist!!!!')
os.chdir(floder)
html = open_url(url).decode('utf-8')
html1 = url_webdriver(url,browser).decode('utf-8')
with open('htmldriver.txt','w') as fht:
fht.write(html1)
with open('html.txt','wb') as fhtml:
fhtml.write(open_url(url))
ip = r'(?:(?:25[0-5]|2[0-4]\d|[01]?\d?\d)\.){3}(?:25[0-5]|2[0-4]\d|[01]?\d?\d)'
iplist=re.findall(ip,html)
for i in iplist:
print(i)
#for iplis in iplist:
with open('ip.txt','w') as f:
f.write(str(iplist))
file = open('ip.txt')
a=file.read()
ip_list=re.findall(ip,a)
print(type(ip_list))
print(ip_list[0])
if __name__ == '__main__':
#url0 = 'http://tieba.baidu.com/p/3563409202'
url1 ='http://www.data5u.com'
floder='tanjingfeng'
#get_img(url0,floder)
get_ip(url1,floder)