Created
October 4, 2011 15:01
-
-
Save mopemope/1261865 to your computer and use it in GitHub Desktop.
xvideos downloader
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import os | |
from os import path | |
from werkzeug import secure_filename | |
from tobikko.core import * | |
from tobikko.patch import * | |
patch_all() | |
import urllib2 | |
from pyquery import PyQuery as pq | |
import urllib | |
from urlparse import urlparse | |
import re | |
search_urls = [ | |
"http://www.xvideos.com/?k=fad&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=fax&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=nsps&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/?k=sama&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=vema&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=kop&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=tnss&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=bkd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=saq&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=natr&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=sprd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=oksn&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=tnss&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=kbkd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=nade&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=tdmj&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=momj&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=mdyd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=shkd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=vagu&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=anb&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=aed&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=uuru&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=etc&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=vec&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=kk&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=lhby&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=fffd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=yone&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=nade&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=jufd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=twl&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=xkk&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=juc&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=venu&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=hone&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=dse&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=okas&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=emav&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=emaf&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=crc&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=tlso&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=jrzd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=deju&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=atgo&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=mkd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=_cj_&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=scd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=havd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=nxg&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=abs&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=cyf&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=mcsr&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=crpd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=madv&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=bbm&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=rbd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=nhdt&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=lhjf&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=ddb&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=jukd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=gar&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=cmc&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=blk&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=pts&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=emu&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=dasd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=drs&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=mgic&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=axam&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=tnsd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=vnds&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=fpjr&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=ugss&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=sbnr&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=aofa&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=caoy&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=tos&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=hima&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=wnz&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=hthd&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=asw&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=matu&p=%s&durf=10min_more", | |
"http://www.xvideos.com/?k=mdc&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/?k=mama&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/?k=000&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/?k=001&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/?k=002&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/?k=003&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/?k=004&p=%s&durf=10min_more", | |
#"http://www.xvideos.com/tags/japan/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/japanese/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/hotjav/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/jap/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/jpavgod/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/creampie/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/japan/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/jap/%s/s:uploaddate/m:10min_more", | |
#"http://www.xvideos.com/tags/jav/%s/s:uploaddate/m:10min_more", | |
] | |
detail_urls = [] | |
id_mode = True | |
save_path = "/video/xvideo/" | |
#os.mkdir(save_path) | |
pool = Pool(20) | |
q = set() | |
MINSIZE = 1024 * 1024 * 80 | |
download_re = re.compile("flv_url=([\w\d.:/_%]*)&", re.M) | |
def get_pagelist(url, page=1): | |
qurl = [] | |
url = url % page | |
conn = urllib2.urlopen(url) | |
page = conn.read() | |
d = pq(page) | |
for anchor in d(".miniature"): | |
href = pq(anchor).attr.href | |
qurl.append(href) | |
for u in qurl: | |
q.add(u) | |
def get_download_url(url): | |
conn = urllib2.urlopen(url) | |
page = conn.read() | |
url_match = download_re.search(page) | |
if url_match: | |
parsed = urlparse(url) | |
file_name = parsed.path.split("/")[-1] | |
file_id = parsed.path.split("/")[-2] | |
vurl = url_match.group(1) | |
download_url = urllib.unquote(vurl) | |
file_name = "%s_%s" % (file_id, file_name) + ".flv" | |
return url, download_url, file_name | |
return None, None, None | |
def download_flv(url, down_url, file_name): | |
#print "'%s' ---- Try Download ----" % url | |
out_path = path.join(save_path, file_name) | |
if not file_name: | |
#print "'%s' ** Not Found Link ** " % url | |
return | |
partial = False | |
try: | |
conn = urllib2.urlopen(down_url) | |
length = conn.info()['Content-Length'] | |
length = int(length) | |
#name = file_name.find("jpavgod") == -1 | |
if length < MINSIZE or length > 1024 * 1024 * 900: | |
#print "*** '%s' is small! Skip!!!'%s' ***" % (url, length) | |
return | |
#if name : | |
# return | |
if path.exists(out_path): | |
size = path.getsize(out_path) | |
if size < length: | |
r = "bytes=%s-" % size | |
req = urllib2.Request(down_url, headers={"Range":r}) | |
conn = urllib2.urlopen(req) | |
#print "'%s' == Resume!! '%s' ==" % (url, file_name) | |
#print "'%s' == File '%s' Size: %d/%d'" % (url, file_name, size, length) | |
partial = True | |
else: | |
#print "'%s' == Downloaded '%s' ==" % (url, file_name) | |
return | |
except: | |
import traceback | |
print traceback.format_exc() | |
#pool.spawn_n(download, url) | |
return | |
if partial: | |
f = open(out_path, "rb+") | |
f.seek(0, 2) | |
else: | |
f = open(out_path, "wb") | |
print "'%s' == Start '%s' size %s ==" % (url, file_name, length) | |
while True: | |
data = conn.read(1024 * 4 ) | |
if not data: | |
break | |
f.write(data) | |
#per = path.getsize(out_path) / float(length) * 100.0 | |
#print "'%s' == '%s' %d%% done. ==" % (url, file_name, per) | |
print "'%s' == Finish '%s' ==" % (url, file_name) | |
def download(url): | |
try: | |
url, download_url, file_name = get_download_url(url) | |
if url and download_url: | |
download_flv(url, download_url, file_name) | |
except: | |
pass | |
#download(url) | |
def start(url, min_page=0, max_page=10): | |
for i in xrange(min_page, max_page+1): | |
try: | |
pool.spawn(get_pagelist, url, i) | |
except: | |
pool.spawn(get_pagelist, url, i) | |
while q: | |
url = q.pop() | |
pool.spawn(download, url) | |
if __name__ == '__main__': | |
for url in search_urls: | |
start(url=url, min_page=1, max_page=10) | |
pool.join() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
How do I use this source code there? Is it just me copy and paste and that's it? And it will make me able to download the xvideos red?