2013-07-30 15:13:28 +00:00
|
|
|
import mechanize, cookielib
|
|
|
|
|
from BeautifulSoup import BeautifulSoup
|
|
|
|
|
import re, os
|
2013-07-30 15:27:43 +00:00
|
|
|
from zipfile import ZipFile
|
2013-07-30 15:13:28 +00:00
|
|
|
|
|
|
|
|
def get_image(tag, folder):
|
|
|
|
|
if not os.path.isdir(folder):
|
|
|
|
|
os.makedirs(folder)
|
|
|
|
|
print image
|
|
|
|
|
filename = os.path.join(folder+'/'+url.split('/')[-1])
|
|
|
|
|
data = br.open(url).read()
|
|
|
|
|
br.back()
|
|
|
|
|
save = open(filename, 'wb')
|
|
|
|
|
save.write(data)
|
|
|
|
|
save.close()
|
2013-07-30 15:27:43 +00:00
|
|
|
return filename
|
|
|
|
|
|
2013-07-30 15:13:28 +00:00
|
|
|
|
|
|
|
|
br=mechanize.Browser()
|
|
|
|
|
cj = cookielib.LWPCookieJar()
|
|
|
|
|
br.set_cookiejar(cj)
|
|
|
|
|
br.set_handle_equiv(True)
|
|
|
|
|
br.set_handle_gzip(True)
|
|
|
|
|
br.set_handle_redirect(True)
|
|
|
|
|
br.set_handle_referer(True)
|
|
|
|
|
br.set_handle_robots(False)
|
|
|
|
|
br.addheaders = [('User-agent', 'Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.1) Gecko/2008071615 Fedora/3.0.1-1.fc9 Firefox/3.0.1')]
|
|
|
|
|
|
2013-07-30 15:27:43 +00:00
|
|
|
tozip = [];
|
|
|
|
|
|
2013-07-30 15:13:28 +00:00
|
|
|
#Individuales
|
|
|
|
|
r = br.open('http://www.etsit.upm.es/fotospromo85.html')
|
|
|
|
|
br.select_form(nr=0)
|
|
|
|
|
br.form['user']='entregadiplomas'
|
|
|
|
|
br.form['pass']='entrega2013'
|
|
|
|
|
r = br.submit()
|
|
|
|
|
r_tags = BeautifulSoup(r)
|
|
|
|
|
link = r_tags(text=re.compile(r'individual'))[0].parent['href']
|
|
|
|
|
galery = BeautifulSoup(br.open(link))
|
|
|
|
|
images_tags = galery(href=re.compile(r'pics/.{10}.jpg'))
|
|
|
|
|
folder = 'individuales'
|
|
|
|
|
for image in images_tags:
|
|
|
|
|
url = image['href']
|
2013-07-30 15:27:43 +00:00
|
|
|
tozip.append(get_image(url, folder))
|
2013-07-30 15:13:28 +00:00
|
|
|
|
|
|
|
|
# Ahora las de grupo
|
|
|
|
|
link=r_tags(text=re.compile(r'grupo'))[0].parent['href']
|
|
|
|
|
galery = BeautifulSoup(br.open(link))
|
|
|
|
|
images_tags = galery(href=re.compile(r'pics/.{10}.jpg'))
|
|
|
|
|
folder = 'grupo'
|
|
|
|
|
for image in images_tags:
|
|
|
|
|
url = image['href']
|
2013-07-30 15:27:43 +00:00
|
|
|
tozip.append(get_image(url, folder))
|
|
|
|
|
|
|
|
|
|
with ZipFile('fotos-85.zip', 'w') as myzip:
|
|
|
|
|
for i in tozip:
|
|
|
|
|
myzip.write(i)
|
