extract text layer from djvu file (see bug 18046)

diff --git a/includes/DefaultSettings.php b/includes/DefaultSettings.php
index b8f4426..c541371 100644 (file)
--- a/includes/DefaultSettings.php
+++ b/includes/DefaultSettings.php
@@ -1897,6 +1897,11 @@ $wgDiff3 = '/usr/bin/diff3';
  */
 $wgDiff = '/usr/bin/diff';
 
+/**
+ * Path to the GNU sed utility.
+ */
+$wgSed = '/bin/sed';
+
 /**
  * We can also compress text stored in the 'text' table. If this is set on, new
  * revisions will be compressed on page save if zlib support is available. Any
@@ -3532,6 +3537,13 @@ $wgDjvuDump = null;
 # $wgDjvuRenderer = 'ddjvu';
 $wgDjvuRenderer = null;
 
+/**
+ * Path of the djvutxt DJVU text extraction utility
+ * Enable this and $wgDjvuDump to enable text layer extraction from djvu files
+ */
+# $wgDjvuTxt = 'djvutxt';
+$wgDjvuTxt = null;
+
 /**
  * Path of the djvutoxml executable
  * This works like djvudump except much, much slower as of version 3.5.

diff --git a/includes/media/DjVu.php b/includes/media/DjVu.php
index 66e954d..0f282c2 100644 (file)
--- a/includes/media/DjVu.php
+++ b/includes/media/DjVu.php
@@ -52,6 +52,8 @@ class DjVuHandler extends ImageHandler {
                $m = false;
                if ( preg_match( '/^page(\d+)-(\d+)px$/', $str, $m ) ) {
                        return array( 'width' => $m[2], 'page' => $m[1] );
+               } else if ( preg_match( '/^page(\d+)-djvutxt$/', $str, $m ) ) {
+                       return array( 'djvutxt' => 1, 'page' => $m[1] );
                } else {
                        return false;
                }
@@ -64,8 +66,22 @@ class DjVuHandler extends ImageHandler {
                );
        }
 
+
+       function normaliseParams( $image, &$params ) {
+               global $wgDjvuTxt;
+               if( $params['djvutxt'] && $wgDjvuTxt) {
+                       if ( !isset( $params['page'] ) ) {
+                               $params['page'] = 1;
+                       }
+                       $params['width'] = 0;
+                       $params['height'] = 0;
+                       return true;
+               } 
+               else return parent::normaliseParams( $image, $params );
+       }
+
        function doTransform( $image, $dstPath, $dstUrl, $params, $flags = 0 ) {
-               global $wgDjvuRenderer, $wgDjvuPostProcessor;
+               global $wgDjvuRenderer, $wgDjvuPostProcessor, $wgDjvuTxt, $wgSed;
 
                // Fetch XML and check it, to give a more informative error message than the one which
                // normaliseParams will inevitably give.
@@ -96,12 +112,22 @@ class DjVuHandler extends ImageHandler {
 
                # Use a subshell (brackets) to aggregate stderr from both pipeline commands
                # before redirecting it to the overall stdout. This works in both Linux and Windows XP.
-               $cmd = '(' . wfEscapeShellArg( $wgDjvuRenderer ) . " -format=ppm -page={$page} -size={$width}x{$height} " .
-                       wfEscapeShellArg( $srcPath );
-               if ( $wgDjvuPostProcessor ) {
-                       $cmd .= " | {$wgDjvuPostProcessor}";
+
+               if( $params['djvutxt'] && $wgDjvuTxt && $wgSed ) {
+                       #Read text from djvu
+                       $cmd = '(' . wfEscapeShellArg( $wgDjvuTxt ) . " --page={$page} " . wfEscapeShellArg( $srcPath );
+                       #Escape < > & characters 
+                       $cmd .= ' | ' . wfEscapeShellArg( $wgSed ) . ' "s/\&/\&amp;/g ; s/</\&lt;/g ; s/>/\&gt;/g ; s/\"/\&quot;/g "';
+                       $cmd .= ' > ' . wfEscapeShellArg($dstPath) . ') 2>&1';
+               }
+               else {
+                       $cmd = '(' . wfEscapeShellArg( $wgDjvuRenderer ) . " -format=ppm -page={$page} -size={$width}x{$height} " .
+                               wfEscapeShellArg( $srcPath );
+                       if ( $wgDjvuPostProcessor ) {
+                               $cmd .= " | {$wgDjvuPostProcessor}";
+                       }
+                       $cmd .= ' > ' . wfEscapeShellArg($dstPath) . ') 2>&1';
                }
-               $cmd .= ' > ' . wfEscapeShellArg($dstPath) . ') 2>&1';
                wfProfileIn( 'ddjvu' );
                wfDebug( __METHOD__.": $cmd\n" );
                $err = wfShellExec( $cmd, $retval );