Add support for Reversed Chars instruction in BMC blocs (#402)

smalot · Apr 13, 2021 · b8784eb · b8784eb
1 parent 3205823
commit b8784eb
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 2 deletions.
diff --git a/samples/bugs/Issue398.pdf b/samples/bugs/Issue398.pdf
diff --git a/src/Smalot/PdfParser/PDFObject.php b/src/Smalot/PdfParser/PDFObject.php
@@ -271,7 +271,7 @@ private function getDefaultFont(Page $page = null)
      */
     public function getText(Page $page = null)
     {
-        $text = '';
+        $result = '';
         $sections = $this->getSectionsText($this->content);
         $current_font = $this->getDefaultFont($page);
         $clipped_font = $current_font;
@@ -283,9 +283,17 @@ public function getText(Page $page = null)
 
         foreach ($sections as $section) {
             $commands = $this->getCommandsText($section);
+            $reverse_text = false;
+            $text = '';
 
             foreach ($commands as $command) {
                 switch ($command[self::OPERATOR]) {
+                    case 'BMC':
+                        if ('ReversedChars' == $command[self::COMMAND]) {
+                            $reverse_text = true;
+                        }
+                        break;
+
                     // set character spacing
                     case 'Tc':
                         break;
@@ -453,11 +461,20 @@ public function getText(Page $page = null)
                     default:
                 }
             }
+
+            // Fix Hebrew and other reverse text oriented languages.
+            // @see: https://github.com/smalot/pdfparser/issues/398
+            if ($reverse_text) {
+                $chars = mb_str_split($text, 1, mb_internal_encoding());
+                $text = implode('', array_reverse($chars));
+            }
+
+            $result .= $text;
         }
 
         array_pop(self::$recursionStack);
 
-        return $text.' ';
+        return $result.' ';
     }
 
     /**

diff --git a/tests/Integration/PDFObjectTest.php b/tests/Integration/PDFObjectTest.php
@@ -231,4 +231,20 @@ public function testGetSectionText()
             $sections
         );
     }
+
+    /**
+     * Tests behavior with reversed chars instruction.
+     *
+     * @see: https://github.com/smalot/pdfparser/issues/398
+     */
+    public function testReversedChars()
+    {
+        $filename = $this->rootDir.'/samples/bugs/Issue398.pdf';
+
+        $parser = $this->getParserInstance();
+        $document = $parser->parseFile($filename);
+        $pages = $document->getPages();
+
+        $this->assertStringContainsString('שלומי טסט', $pages[0]->getText());
+    }
 }