Changeset 3720 for trunk/gsdl/perllib/plugins/PDFPlug.pm
- Timestamp:
- 2003-01-28T16:22:26+13:00 (21 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
trunk/gsdl/perllib/plugins/PDFPlug.pm
r3614 r3720 40 40 'desc' => "Files matching this regular expression will be blocked from being passed to any later plugins in the list. This has no real effect other than to prevent lots of warning messages about input files you don't care about. Each plugin might have a default block_exp. e.g. by default HTMLPlug blocks any files with .gif, .jpg, .jpeg, .png or .css file extensions.", 41 41 'type' => 'string', 42 'deft' => q^^ } 42 'deft' => q^^ }, 43 { 'name' => "noimages", 44 'desc' => "Don't attempt to extract images from PDF.", 45 'type' => "flag" }, 46 { 'name' => "complex", 47 'desc' => "Create more complex output. With this option set the output html will look much more like the original PDF file. For this to function properly you Ghostscript installed (for *nix gs should be on your path while for windows you must have gswin32c.exe on your path).", 48 'type' => "flag" }, 49 { 'name' => "zoom", 50 'desc' => "The factor by which to zoomthe PDF for output (this is only useful if -complex is set).", 51 'type' => "int" } 43 52 ]; 44 53 … … 50 59 sub new { 51 60 my $class = shift (@_); 61 62 my ($noimages, $complex, $zoom); 63 64 if (!parsargv::parse(\@_, 65 q^noimages^, \$noimages, 66 q^complex^, \$complex, 67 q^zoom/\d+/2^, \$zoom, 68 "allow_extra_options")) { 69 70 print STDERR "\nIncorrect options passed to PDFPlug, check your collect.cfg configuration file\n"; 71 &print_usage(); 72 die "\n"; 73 } 52 74 53 75 # following title_sub removes "Page 1" added by pdftohtml, and a leading … … 81 103 my $option_list = $self->{'option_list'}; 82 104 push( @{$option_list}, $options ); 105 106 # these are passed through to gsConvert.pl by ConvertToPlug.pm 107 $self->{'convert_options'} = "-pdf_zoom $zoom"; 108 $self->{'convert_options'} .= " -pdf_complex" if $complex; 109 $self->{'convert_options'} .= " -pdf_ignore_images" if $noimages; 110 111 # pdftohtml will always produce html files encoded as utf-8 112 if ($self->{'input_encoding'} eq "auto") { 113 $self->{'input_encoding'} = "utf8"; 114 $self->{'extract_language'} = 1; 115 } 83 116 84 117 return bless $self, $class; … … 89 122 print STDERR "\n usage: plugin PDFPlug [options]\n\n"; 90 123 print STDERR " options:\n"; 91 print STDERR " -convert_to (html|text) plugin converts to TEXT or HTML\n"; 92 print STDERR " (default html)\n"; 93 print STDERR " -use_sections create a separate section for each page of the PDF file.\n\n"; 124 print STDERR " -convert_to (html|text) Convert to TEXT or HTML (default html)\n"; 125 print STDERR " -use_sections Create a separate section for each page\n"; 126 print STDERR " of the PDF file.\n"; 127 print STDERR " -noimages Don't attempt to extract images from PDF.\n"; 128 print STDERR " -complex Create more complex output. With this option\n"; 129 print STDERR " set the output html will look much more like\n"; 130 print STDERR " the original PDF file. For this to function\n"; 131 print STDERR " properly you Ghostscript installed (for *nix\n"; 132 print STDERR " gs should be on your path while for windows\n"; 133 print STDERR " you must have gswin32c.exe on your path).\n"; 134 print STDERR " -zoom The factor by which to zoomthe PDF for output\n"; 135 print STDERR " (this is only useful if -complex is set).\n\n"; 94 136 } 95 137
Note:
See TracChangeset
for help on using the changeset viewer.