lib/clive/Video.pm

   1 # -*- coding: ascii -*-
   2 ###########################################################################
   3 # clive, command line video extraction utility.
   4 # Copyright 2007, 2008, 2009 Toni Gundogdu.
   5 #
   6 # This file is part of clive.
   7 #
   8 # clive is free software: you can redistribute it and/or modify it under
   9 # the terms of the GNU General Public License as published by the Free
  10 # Software Foundation, either version 3 of the License, or (at your option)
  11 # any later version.
  12 #
  13 # clive is distributed in the hope that it will be useful, but WITHOUT ANY
  14 # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
  15 # FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
  16 # details.
  17 #
  18 # You should have received a copy of the GNU General Public License along
  19 # with this program. If not, see <http://www.gnu.org/licenses/>.
  20 ###########################################################################
  21 package clive::Video;
  22
  23 use warnings;
  24 use strict;
  25
  26 use Carp;
  27 use POSIX;
  28 use File::Basename qw(basename);
  29 use File::Spec::Functions;
  30 use Cwd qw(getcwd);
  31
  32 use clive::Util;
  33
  34 our $AUTOLOAD;
  35
  36 sub new {
  37     my $class  = shift;
  38     my %fields = (
  39         page_link      => undef,
  40         video_id       => undef,
  41         file_length    => undef,
  42         content_type   => undef,
  43         file_suffix    => undef,
  44         video_link     => undef,
  45         video_host     => undef,
  46         video_format   => undef,
  47         base_filename  => undef,
  48         filename       => undef,
  49         initial_length => undef,
  50         time_stamp     => undef,
  51         nothing_todo   => undef,
  52     );
  53     my $self = {
  54         _permitted => \%fields,
  55         %fields,
  56     };
  57     return bless( $self, $class );
  58 }
  59
  60 sub page_title {
  61     my $self = shift;
  62     if (@_) {
  63         my ( $content, $title ) = @_;
  64         if ( !$title ) {
  65             require HTML::TokeParser;
  66             my $p = HTML::TokeParser->new($content);
  67             $p->get_tag("title");
  68             $self->{page_title} = $p->get_trimmed_text;
  69             _cleanupTitle($self);
  70         }
  71         else {
  72             $self->{page_title} = $title;
  73         }
  74     }
  75     return $self->{page_title};
  76 }
  77
  78 sub printVideo {
  79     my $self = shift;
  80     my $str  = sprintf(
  81         "file: %s  %.1fM  [%s]\n",
  82         $self->{base_filename},
  83         clive::Util::toMB( $self->{file_length} ),
  84         $self->{content_type}
  85     );
  86     clive::Log->instance->out($str);
  87 }
  88
  89 sub emitCSV {
  90     my $self = shift;
  91
  92     require URI::Escape;
  93
  94     my @fields = qw(base_filename file_length video_link);
  95
  96     my $str = "csv:";
  97     $str .= sprintf( qq/"%s",/, $self->$_ ) foreach (@fields);
  98     $str =~ s/,$//;
  99
 100     clive::Log->instance->out("$str\n");
 101 }
 102
 103 sub formatOutputFilename {
 104     my $self = shift;
 105
 106     my $config = clive::Config->instance->config;
 107     my $fname;
 108
 109     if ( !$config->{output_file} ) {
 110
 111         # Apply character-class.
 112         my $title = $self->{page_title};
 113         my $cclass = $config->{cclass} || qr|\w|;
 114
 115         $title = join( '', $self->{page_title} =~ /$cclass/g )
 116             if ( !$config->{no_cclass} );
 117
 118         # Format output filename.
 119         $fname = $config->{filename_format} || "%t.%s";
 120
 121         if ( !$title &&  $fname !~ /%i/ ) {
 122             $title = $self->{video_id};
 123             $title =~ s/-/_/g;
 124         }
 125
 126         $fname =~ s/%t/$title/;
 127         $fname =~ s/%s/$self->{file_suffix}/;
 128         $fname =~ s/%i/$self->{video_id}/;
 129         $fname =~ s/%h/$self->{video_host}/;
 130
 131         my $config = clive::Config->instance->config;
 132         $fname = catfile( $config->{save_dir} || getcwd, $fname );
 133
 134         my $tmp = $fname;
 135
 136         for ( my $i = 1; $i < 9999; ++$i ) {
 137             $self->{initial_length} = clive::Util::fileExists($fname);
 138
 139             if ( $self->{initial_length} == 0 ) {
 140                 last;
 141             }
 142             elsif ( $self->{initial_length} == $self->{file_length} ) {
 143                 $self->{nothing_todo} = 1;
 144                 last;
 145             }
 146             else {
 147                 if ( $config->{continue} ) {
 148                     last;
 149                 }
 150             }
 151             $fname = "$tmp.$i";
 152         }
 153     }
 154     else {
 155         $self->{initial_length}
 156             = clive::Util->fileExists( $config->{output_file} );
 157         if ( $self->{initial_length} == $self->{file_length} ) {
 158             $self->{nothing_todo} = 1;
 159         }
 160         else {
 161             $fname = $config->{output_file};
 162         }
 163     }
 164
 165     if ( !$config->{continue} ) {
 166         $self->{initial_length} = 0;
 167     }
 168
 169     $self->{base_filename} = basename($fname);
 170     $self->{filename}      = $fname;
 171 }
 172
 173 sub fromCacheRecord {
 174     my ( $self, $record ) = @_;
 175
 176     # No need to keep order in sync with clive::Video::toCacheRecord
 177     # or clive::Cache::_mapRecord -- just make sure each item gets
 178     # set here.
 179     $self->{page_title}   = $$record{page_title};
 180     $self->{page_link}    = $$record{page_link};
 181     $self->{video_id}     = $$record{video_id};
 182     $self->{video_link}   = $$record{video_link};
 183     $self->{video_host}   = $$record{video_host};
 184     $self->{video_format} = $$record{video_format};
 185     $self->{file_length}  = $$record{file_length};
 186     $self->{file_suffix}  = $$record{file_suffix};
 187     $self->{content_type} = $$record{content_type};
 188     $self->{time_stamp}   = $$record{time_stamp};
 189
 190     _cleanupTitle($self);
 191 }
 192
 193 sub toCacheRecord {
 194     my $self = shift;
 195
 196     # Should really remove all '#' from the strings
 197     # before storing them. Living on the edge.
 198     $self->{page_title} =~ tr{#}//d;
 199
 200     # Keep the order in sync with clive::Cache::_mapRecord.
 201     my $record
 202         = $self->{page_title} . "#"
 203         . $self->{page_link} . "#"
 204         . $self->{video_id} . "#"
 205         . $self->{video_link} . "#"
 206         . $self->{video_host} . "#"
 207         . $self->{video_format} . "#"
 208         . $self->{file_length} . "#"
 209         . $self->{file_suffix} . "#"
 210         . $self->{content_type} . "#"
 211         . POSIX::strftime( "%F %T", localtime )    # time_stamp
 212         ;
 213     return $record;
 214 }
 215
 216 sub _cleanupTitle {
 217     my $self  = shift;
 218     my $title = $self->{page_title};
 219
 220     $title =~ s/(youtube|video|liveleak.com|sevenload|dailymotion)//gi;
 221     $title =~ s/(cctv.com|redtube)//gi;
 222
 223     $title =~ s/^[-\s]+//;
 224     $title =~ s/\s+$//;
 225
 226     $self->{page_title} = $title;
 227 }
 228
 229 sub AUTOLOAD {
 230     my $self = shift;
 231     my $type = ref($self)
 232         or croak("$self is not an object");
 233     my $name = $AUTOLOAD;
 234     $name =~ s/.*://;
 235     unless ( exists( $self->{_permitted}->{$name} ) ) {
 236         croak("cannot access `$name' field in class $type");
 237     }
 238     if (@_) {
 239         return $self->{$name} = shift;
 240     }
 241     else {
 242         return $self->{$name};
 243     }
 244 }
 245
 246 1;
 247
 248 # Barefoot servants too.