Stop Image XObjects being included as empty strings in PDFObject::getTextArray(). (#775)

* Stop Image XObjects being included as empty strings in PDFObject::getTextArray().

* Fix code style issues.

* Add test coverage to PDFObject::getTextArray() to cover the Do command case.

Author: rupertj

src/Smalot/PdfParser/PDFObject.php

@@ -775,16 +775,27 @@ public function getTextArray(?Page $page = null): array
                         break;
 
                     case 'Do':
-                        if (null !== $page) {
-                            $args = preg_split('/\s/s', $command[self::COMMAND]);
-                            $id = trim(array_pop($args), '/ ');
-                            $xobject = $page->getXObject($id);
-
-                            // @todo $xobject could be a ElementXRef object, which would then throw an error
-                            if (\is_object($xobject) && $xobject instanceof self && !\in_array($xobject->getUniqueId(), self::$recursionStack, true)) {
-                                // Not a circular reference.
-                                $text[] = $xobject->getText($page);
-                            }
+                        if (is_null($page)) {
+                            break;
+                        }
+
+                        $args = preg_split('/\s/s', $command[self::COMMAND]);
+                        $id = trim(array_pop($args), '/ ');
+                        $xobject = $page->getXObject($id);
+
+                        // Check we got a PDFObject back.
+                        if (!$xobject instanceof self) {
+                            break;
+                        }
+
+                        // If the PDFObject is an image, do nothing, as images aren't text.
+                        if ($xobject instanceof Image) {
+                            break;
+                        }
+
+                        // Check this is not a circular reference.
+                        if (!\in_array($xobject->getUniqueId(), self::$recursionStack, true)) {
+                            $text[] = $xobject->getText($page);
                         }
                         break;
 

tests/PHPUnit/Unit/PDFObjectTest.php

@@ -6,8 +6,12 @@
 
 use PHPUnitTests\TestCase;
 use Smalot\PdfParser\Document;
+use Smalot\PdfParser\Element;
+use Smalot\PdfParser\Element\ElementArray;
+use Smalot\PdfParser\Header;
 use Smalot\PdfParser\Page;
 use Smalot\PdfParser\PDFObject;
+use Smalot\PdfParser\XObject\Image;
 
 class PDFObjectTest extends TestCase
 {
@@ -22,4 +26,39 @@ public function testGetTextOnPageWithoutContent(): void
 
         static::assertSame(' ', (new PDFObject($document, null, null))->getText(new Page($document)));
     }
+
+    public function testTextArrayObjects(): void
+    {
+        $document = new Document();
+        $document->init();
+
+        $image = new Image($document);
+        $xObject = new PDFObject($document);
+
+        $header1 = new Header([
+            'Resources' => new Header([
+                'XObject' => new Header([
+                    'Im0' => $image,
+                ])
+            ]),
+            'Contents' => new ElementArray([new Element('/Imo Do', $document)], $document),
+        ]);
+        $page1 = new Page($document, $header1);
+
+        $header2 = new Header([
+            'Resources' => new Header([
+                'XObject' => new Header([
+                    'Ps0' => $xObject,
+                ])
+            ]),
+            'Contents' => new ElementArray([new Element('/Ps0 Do', $document)], $document),
+        ]);
+        $page2 = new Page($document, $header2);
+
+        // Page 1 contains an image, which should not appear in the text array.
+        self::assertSame([], $page1->getTextArray());
+
+        // Page 2 contains a non-image object, which should appear in the text array.
+        self::assertSame([' '], $page2->getTextArray());
+    }
 }