Changeset 10274

Timestamp:

2005-07-25T11:45:34+12:00 (19 years ago)

Author:

chi

Message:

A modification to allow a secondary plug setting through ConvertToPlug.

File:

• trunk/gsdl/perllib/plugins/PSPlug.pm (modified) (3 diffs)

trunk/gsdl/perllib/plugins/PSPlug.pm

@@ -79 +79 @@
     my $self = (defined $hashArgOptLists)? new ConvertToPlug($pluginlist,$inputargs,$hashArgOptLists): new ConvertToPlug($pluginlist,$inputargs);
 
-    return bless $self, $class;
+
+    my $secondary_plugin_options = $self->{'secondary_plugin_options'};
+    if (!defined $secondary_plugin_options->{'TEXTPlug'}) {
+    $secondary_plugin_options->{'TEXTPlug'} = [];
+    }
+
+    my $text_options = $secondary_plugin_options->{'TEXTPlug'};
+   
+    # following title_sub removes "Page 1" added by ps2ascii, and a leading
+    # "1", which is often the page number at the top of the page. Bad Luck
+    # if your document title actually starts with "1 " - is there a better way?
+    #$self->{'input_encoding'} = "utf8";
+    #$self->{'extract_language'} = 1;
+    #push(@$text_options, "-input_encoding", "utf8");
+    #push(@$text_options,"-extract_language");
+    push(@$text_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
+
+    $self = bless $self, $class;
+
+    $self->load_secondary_plugins($class,$secondary_plugin_options);
+
+    return $self;
 }
 
@@ -95 +116 @@
 }
 
+sub convert_post_process
+{
+    my $self = shift (@_);
+    my ($conv_filename) = @_;
+    
+    my $outhandle=$self->{'outhandle'};
+    
+    my ($language, $encoding) = $self->textcat_get_language_encoding ($conv_filename);
+    
+#    print STDERR "*** Plug encoding = $encoding\n";
+    
+    # read in file ($text will be in utf8)
+    my $text = "";
+    $self->read_file ($conv_filename, $encoding, $language, \$text);
+    
+    # turn any high bytes that aren't valid utf-8 into utf-8.
+    unicode::ensure_utf8(\$text);
+    
+    #print STDERR "*** Text =$text\n";
+    # Write it out again!
+    $self->utf8_write_file (\$text, $conv_filename);
+}
+
 sub extract_metadata_from_postscript {
     my $self = shift (@_);
-    my $filename = shift (@_);
-    my $doc = shift (@_);
+
+    my ($filename,$doc) = @_;
+
     my $section = $doc->get_top_section();
 
@@ -166 +211 @@
 sub process {
     my $self = shift (@_);
-    my ($trash, $trash2, $path, $file, $trash3, $doc, $gli) = @_;
-
-    my $outhandle = $self->{'outhandle'};
-    print $outhandle "PSPlug: passing $_[3] on to $self->{'converted_to'}Plug\n" 
-    if $self->{'verbosity'} > 1;
-    print STDERR "<Processing n='$_[3]' p='PSPlug'>\n" if ($gli);
-
-    &extract_metadata_from_postscript($self,"$path/$file", $doc);
-
-    return ConvertToPlug::process_type($self,"ps",@_);
+    my ($textref, $pluginfo, $base_dir, $file, $metadata, $doc_obj, $gli) = @_;
+
+#    my $outhandle = $self->{'outhandle'};
+
+#    print $outhandle "PSPlug: passing $file on to $self->{'converted_to'}Plug\n" 
+#   if $self->{'verbosity'} > 1;
+#    print STDERR "<Processing n='$file' p='PSPlug'>\n" if ($gli);
+
+    my $filename = &util::filename_cat($base_dir,$file);
+    $self->extract_metadata_from_postscript($filename, $doc_obj);
+
+    return $self->process_type("ps",$base_dir,$file,$doc_obj);
 }