pymotherless-dl.py

   1 #!/usr/bin/env python
   2
   3 '''
   4     This program is free software; you can redistribute it and/or modify
   5     it under the terms of the Revised BSD License.
   6
   7     This program is distributed in the hope that it will be useful,
   8     but WITHOUT ANY WARRANTY; without even the implied warranty of
   9     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  10     Revised BSD License for more details.
  11
  12     Copyright 2016 Cool Dude 2k - http://idb.berlios.de/
  13     Copyright 2016 Game Maker 2k - http://intdb.sourceforge.net/
  14     Copyright 2016 Kazuki Przyborowski - https://github.com/KazukiPrzyborowski
  15
  16     $FileInfo: pymotherless-dl.py - Last Update: 6/16/2016 Ver. 0.4.7 RC 1 - Author: cooldude2k $
  17 '''
  18
  19 from __future__ import division, absolute_import, print_function;
  20 import re, os, sys, pymotherless, argparse;
  21 import logging as log;
  22
  23 __project__ = pymotherless.__project__;
  24 __program_name__ = pymotherless.__program_name__;
  25 __project_url__ = pymotherless.__project_url__;
  26 __version_info__ = pymotherless.__version_info__;
  27 __version_date_info__ = pymotherless.__version_date_info__;
  28 __version_date__ = pymotherless.__version_date__;
  29 __version_date_plusrc__ = pymotherless.__version_date_plusrc__
  30 __version__ = pymotherless.__version__;
  31 __version_date_plusrc__ = pymotherless.__version_date_plusrc__;
  32
  33 geturls_cj = pymotherless.geturls_cj;
  34 geturls_ua = pymotherless.geturls_ua;
  35 geturls_ua_firefox_windows7 = pymotherless.geturls_ua_firefox_windows7;
  36 geturls_ua_seamonkey_windows7 = pymotherless.geturls_ua_seamonkey_windows7;
  37 geturls_ua_chrome_windows7 = pymotherless.geturls_ua_chrome_windows7;
  38 geturls_ua_chromium_windows7 = pymotherless.geturls_ua_chromium_windows7;
  39 geturls_ua_palemoon_windows7 = pymotherless.geturls_ua_palemoon_windows7;
  40 geturls_ua_opera_windows7 = pymotherless.geturls_ua_opera_windows7;
  41 geturls_ua_vivaldi_windows7 = pymotherless.geturls_ua_chromium_windows7;
  42 geturls_ua_internet_explorer_windows7 = pymotherless.geturls_ua_internet_explorer_windows7;
  43 geturls_ua_microsoft_edge_windows7 = pymotherless.geturls_ua_microsoft_edge_windows7;
  44 geturls_ua_pymotherless_python = pymotherless.geturls_ua_pymotherless_python;
  45 geturls_ua_pymotherless_python_alt = pymotherless.geturls_ua_pymotherless_python_alt;
  46 geturls_ua_googlebot_google = pymotherless.geturls_ua_googlebot_google;
  47 geturls_ua_googlebot_google_old = pymotherless.geturls_ua_googlebot_google_old;
  48 geturls_headers = pymotherless.geturls_headers;
  49 geturls_headers_firefox_windows7 = pymotherless.geturls_headers_firefox_windows7;
  50 geturls_headers_seamonkey_windows7 = pymotherless.geturls_headers_seamonkey_windows7;
  51 geturls_headers_chrome_windows7 = pymotherless.geturls_headers_chrome_windows7;
  52 geturls_headers_chromium_windows7 = pymotherless.geturls_headers_chromium_windows7;
  53 geturls_headers_palemoon_windows7 = pymotherless.geturls_headers_palemoon_windows7;
  54 geturls_headers_opera_windows7 = pymotherless.geturls_headers_opera_windows7;
  55 geturls_headers_vivaldi_windows7 = pymotherless.geturls_headers_vivaldi_windows7;
  56 geturls_headers_internet_explorer_windows7 = pymotherless.geturls_headers_internet_explorer_windows7;
  57 geturls_headers_microsoft_edge_windows7 = pymotherless.geturls_headers_microsoft_edge_windows7;
  58 geturls_headers_pymotherless_python = pymotherless.geturls_headers_pymotherless_python;
  59 geturls_headers_pymotherless_python_alt = pymotherless.geturls_headers_pymotherless_python_alt;
  60 geturls_headers_googlebot_google = pymotherless.geturls_headers_googlebot_google;
  61 geturls_headers_googlebot_google_old = pymotherless.geturls_headers_googlebot_google_old;
  62 geturls_download_sleep = pymotherless.geturls_download_sleep;
  63
  64 parser = argparse.ArgumentParser(description="get urls of images/videos from motherless.com", conflict_handler="resolve", add_help=True);
  65 parser.add_argument("url", help="motherless url");
  66 parser.add_argument("-v", "--version", action="version", version=__program_name__+" "+__version__);
  67 parser.add_argument("-U", "--update", action="store_true", help="update this program to latest version. Make sure that you have sufficient permissions (run with sudo if needed)");
  68 parser.add_argument("-d", "--dump-user-agent", action="store_true", help="display the current browser identification");
  69 parser.add_argument("-u", "--user-agent", default=geturls_ua_firefox_windows7, help="specify a custom user agent");
  70 parser.add_argument("-r", "--referer", default="http://motherless.com/", help="specify a custom referer, use if the video access");
  71 parser.add_argument("-g", "--get-url", action="store_true", help="simulate, quiet but print URL");
  72 parser.add_argument("-p", "--get-pageurl", action="store_true", help="simulate, quiet but print URL");
  73 parser.add_argument("-e", "--get-title", action="store_true", help="simulate, quiet but print title");
  74 parser.add_argument("-i", "--get-id", action="store_true", help="simulate, quiet but print id");
  75 parser.add_argument("-t", "--get-thumbnail", action="store_true", help="simulate, quiet but print thumbnail URL");
  76 parser.add_argument("-f", "--get-filename", action="store_true", help="simulate, quiet but print output filename");
  77 parser.add_argument("-u", "--get-username", action="store_true", help="simulate, quiet but print uploaders username");
  78 parser.add_argument("-i", "--get-views", action="store_true", help="simulate, quiet but print number of views");
  79 parser.add_argument("-f", "--get-favorites", action="store_true", help="simulate, quiet but print number of favorites");
  80 parser.add_argument("-o", "--output-directory", default=os.path.realpath(os.getcwd()), help="specify a directory to output file to");
  81 parser.add_argument("-l", "--use-httplib", default="urllib", help="select library to download file can be urllib or requests or mechanize");
  82 parser.add_argument("-b", "--set-buffersize", default=524288, type=int, help="set how big buffersize is in bytes. how much it will download");
  83 parser.add_argument("-V", "--verbose", action="store_true", help="print various debugging information");
  84 getargs = parser.parse_args();
  85
  86 if(not pymotherless.check_httplib_support(getargs.use_httplib)):
  87  getargs.use_httplib = "urllib";
  88
  89 getargs_cj = geturls_cj;
  90 getargs_headers = {'Referer': getargs.referer, 'User-Agent': getargs.user_agent, 'Accept-Encoding': "gzip, deflate", 'Accept-Language': "en-US,en;q=0.8,en-CA,en-GB;q=0.6", 'Accept-Charset': "ISO-8859-1,ISO-8859-15,utf-8;q=0.7,*;q=0.7", 'Accept': "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8", 'Connection': "close"};
  91
  92 getargs.output_directory = os.path.realpath(getargs.output_directory);
  93
  94 if(getargs.verbose==True):
  95  log.basicConfig(format="%(levelname)s: %(message)s", level=log.DEBUG);
  96
  97 if(getargs.dump_user_agent==True):
  98  print(getargs.user_agent);
  99  sys.exit();
 100
 101 motherless_linktype = pymotherless.get_motherless_link_type(getargs.url);
 102 motherless_links = [];
 103 motherless_links_dir = [];
 104 if(motherless_linktype=="link"):
 105  motherless_links.append(getargs.url);
 106 if(motherless_linktype=="gallery"):
 107  numpages = pymotherless.get_motherless_number_of_pages(getargs.url, getargs_headers, getargs_cj, getargs.use_httplib);
 108  numcount = 1;
 109  while(numcount <= numpages):
 110   getlinks = pymotherless.get_motherless_galleries_links(getargs.url, getargs_headers, getargs_cj, getargs.use_httplib, page=numcount);
 111   innumlinks = getlinks['numoflinks'];
 112   innumcount = 0;
 113   while(innumcount < innumlinks):
 114    inmotherless_linktype = pymotherless.get_motherless_link_type(getlinks[innumcount]['url']);
 115    if(inmotherless_linktype=="link"):
 116     motherless_links_dir.append(getargs.url.rsplit('/', 1)[-1]);
 117     motherless_links.append(getlinks[innumcount]['url']);
 118    if(inmotherless_linktype=="gallery"):
 119     innumpages = pymotherless.get_motherless_number_of_pages(getlinks[innumcount]['url'], getargs_headers, getargs_cj, getargs.use_httplib);
 120     innumcountpages = 1;
 121     while(innumcountpages <= innumpages):
 122      ingetlinks = pymotherless.get_motherless_galleries_links(getlinks[innumcount]['url'], getargs_headers, getargs_cj, getargs.use_httplib, page=innumcountpages);
 123      ininnumlinks = ingetlinks['numoflinks'];
 124      ininnumcount = 0;
 125      while(ininnumcount < ininnumlinks):
 126       ininmotherless_linktype = pymotherless.get_motherless_link_type(ingetlinks[ininnumcount]['url']);
 127       if(ininmotherless_linktype=="link"):
 128        motherless_links_dir.append(getlinks[innumcount]['url'].rsplit('/', 1)[-1]);
 129        motherless_links.append(ingetlinks[ininnumcount]['url']);
 130       ininnumcount = ininnumcount + 1;
 131      innumcountpages = innumcountpages + 1;
 132    innumcount = innumcount + 1;
 133   numcount = numcount + 1;
 134 if(motherless_linktype=="board"):
 135  getlinks = pymotherless.get_motherless_boards_links(getargs.url, getargs_headers, getargs_cj, getargs.use_httplib);
 136  innumlinks = getlinks['numoflinks'];
 137  innumcount = 0;
 138  while(innumcount < innumlinks):
 139   inmotherless_linktype = pymotherless.get_motherless_link_type(getlinks[innumcount]['url']);
 140   if(inmotherless_linktype=="link"):
 141    motherless_links_dir.append(getargs.url.rsplit('/', 1)[-1]);
 142    motherless_links.append(getlinks[innumcount]['url']);
 143   innumcount = innumcount + 1;
 144
 145 if(getargs.get_url==True):
 146  listsize = len(motherless_links);
 147  listcount = 0;
 148  while(listcount < listsize):
 149   print(pymotherless.get_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib)['url']);
 150   listcount = listcount + 1;
 151
 152 if(getargs.get_pageurl==True):
 153  listsize = len(motherless_links);
 154  listcount = 0;
 155  while(listcount < listsize):
 156   print(motherless_links[listcount]);
 157   listcount = listcount + 1;
 158
 159 if(getargs.get_filename==True):
 160  listsize = len(motherless_links);
 161  listcount = 0;
 162  while(listcount < listsize):
 163   print(pymotherless.get_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib)['fullfilename']);
 164   listcount = listcount + 1;
 165
 166 if(getargs.get_title==True):
 167  listsize = len(motherless_links);
 168  listcount = 0;
 169  while(listcount < listsize):
 170   print(pymotherless.get_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib)['title']);
 171   listcount = listcount + 1;
 172
 173 if(getargs.get_username==True):
 174  listsize = len(motherless_links);
 175  listcount = 0;
 176  while(listcount < listsize):
 177   print(pymotherless.get_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib)['username']);
 178   listcount = listcount + 1;
 179
 180 if(getargs.get_views==True):
 181  listsize = len(motherless_links);
 182  listcount = 0;
 183  while(listcount < listsize):
 184   print(pymotherless.get_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib)['numberofviews']);
 185   listcount = listcount + 1;
 186
 187 if(getargs.get_favorites==True):
 188  listsize = len(motherless_links);
 189  listcount = 0;
 190  while(listcount < listsize):
 191   print(pymotherless.get_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib)['numberoffavorites']);
 192   listcount = listcount + 1;
 193
 194 if(getargs.get_url==False and getargs.get_pageurl==False and getargs.get_thumbnail==False and getargs.get_filename==False and getargs.get_title==False and getargs.get_username==False and getargs.get_views==False and getargs.get_favorites==False):
 195  listsize = len(motherless_links);
 196  listcount = 0;
 197  while(listcount < listsize):
 198   listcountalt = listcount + 1;
 199   percentage = str("{0:.2f}".format(float(float(listcountalt / listsize) * 100))).rstrip('0').rstrip('.')+"%";
 200   log.info("Downloading URL Number "+str(listcountalt)+" / "+str(listsize)+" "+str(percentage));
 201   if(motherless_linktype=="link"):
 202    pymotherless.download_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib, buffersize=[getargs.set_buffersize, getargs.set_buffersize], outpath=getargs.output_directory);
 203   if(motherless_linktype=="board" or motherless_linktype=="gallery"):
 204    pymotherless.download_motherless_links(motherless_links[listcount], getargs_headers, getargs_cj, getargs.use_httplib, buffersize=[getargs.set_buffersize, getargs.set_buffersize], outpath=getargs.output_directory+os.path.sep+motherless_links_dir[listcount]);
 205   listcount = listcount + 1;