Merge pull request #2450 from PHPOffice/pr2369

Word2007 Reader : Added option to disable loading images

Author: Progi1984

CHANGELOG.md

@@ -11,6 +11,7 @@ This project adheres to [Semantic Versioning](http://semver.org/).
 - Word2007 Reader/Writer : Added noWrap table cell property by @kernusr in #2359
 - HTML Reader : Support for `font-variant: small-caps` by @cambraca in #2117
 - Improved TextDirection for styling a cell by @terryzwt in #2429
+- Word2007 Reader : Added option to disable loading images by @aelliott1485 in #2450
 ### Bug fixes
 
 - Fixed wrong mimetype for docx files by @gamerlv in #2416

src/PhpWord/Reader/AbstractReader.php

@@ -42,6 +42,13 @@ abstract class AbstractReader implements ReaderInterface
      */
     protected $fileHandle;
 
+    /**
+     * Load images.
+     *
+     * @var bool
+     */
+    protected $imageLoading = true;
+
     /**
      * Read data only?
      *
@@ -67,6 +74,18 @@ public function setReadDataOnly($value = true)
         return $this;
     }
 
+    public function hasImageLoading(): bool
+    {
+        return $this->imageLoading;
+    }
+
+    public function setImageLoading(bool $value): self
+    {
+        $this->imageLoading = $value;
+
+        return $this;
+    }
+
     /**
      * Open file for reading.
      *

src/PhpWord/Reader/Word2007.php

@@ -94,6 +94,7 @@ private function readPart(PhpWord $phpWord, $relationships, $partName, $docFile,
         if (class_exists($partClass)) {
             /** @var \PhpOffice\PhpWord\Reader\Word2007\AbstractPart $part Type hint */
             $part = new $partClass($docFile, $xmlFile);
+            $part->setImageLoading($this->hasImageLoading());
             $part->setRels($relationships);
             $part->read($phpWord);
         }

src/PhpWord/Reader/Word2007/AbstractPart.php

@@ -67,6 +67,13 @@ abstract class AbstractPart
      */
     protected $rels = [];
 
+    /**
+     * Image Loading.
+     *
+     * @var bool
+     */
+    protected $imageLoading = true;
+
     /**
      * Read part.
      */
@@ -94,6 +101,18 @@ public function setRels($value): void
         $this->rels = $value;
     }
 
+    public function setImageLoading(bool $value): self
+    {
+        $this->imageLoading = $value;
+
+        return $this;
+    }
+
+    public function hasImageLoading(): bool
+    {
+        return $this->imageLoading;
+    }
+
     /**
      * Read w:p.
      *
@@ -249,7 +268,7 @@ protected function readRunChild(XMLReader $xmlReader, DOMElement $node, Abstract
             // Image
             $rId = $xmlReader->getAttribute('r:id', $node, 'v:shape/v:imagedata');
             $target = $this->getMediaTarget($docPart, $rId);
-            if (null !== $target) {
+            if ($this->hasImageLoading() && null !== $target) {
                 if ('External' == $this->getTargetMode($docPart, $rId)) {
                     $imageSource = $target;
                 } else {
@@ -271,7 +290,7 @@ protected function readRunChild(XMLReader $xmlReader, DOMElement $node, Abstract
                 $embedId = $xmlReader->getAttribute('r:embed', $node, 'wp:anchor/a:graphic/a:graphicData/pic:pic/pic:blipFill/a:blip');
             }
             $target = $this->getMediaTarget($docPart, $embedId);
-            if (null !== $target) {
+            if ($this->hasImageLoading() && null !== $target) {
                 $imageSource = "zip://{$this->docFile}#{$target}";
                 $parent->addImage($imageSource, null, false, $name);
             }

src/PhpWord/Style/Language.php

@@ -256,7 +256,9 @@ private function validateLocale($locale)
         if ($locale !== null && strlen($locale) === 2) {
             return strtolower($locale) . '-' . strtoupper($locale);
         }
-
+        if ($locale === 'und') {
+            return 'en-EN';
+        }
         if ($locale !== null && $locale !== 'zxx' && strstr($locale, '-') === false) {
             throw new InvalidArgumentException($locale . ' is not a valid language code');
         }

tests/PhpWordTests/Reader/Word2007Test.php

@@ -17,7 +17,10 @@
 
 namespace PhpOffice\PhpWordTests\Reader;
 
+use PhpOffice\PhpWord\Element\Image;
+use PhpOffice\PhpWord\Element\TextRun;
 use PhpOffice\PhpWord\IOFactory;
+use PhpOffice\PhpWord\PhpWord;
 use PhpOffice\PhpWord\Reader\Word2007;
 use PhpOffice\PhpWordTests\TestHelperDOCX;
 
@@ -36,8 +39,7 @@ class Word2007Test extends \PHPUnit\Framework\TestCase
     public function testCanRead(): void
     {
         $object = new Word2007();
-        $filename = __DIR__ . '/../_files/documents/reader.docx';
-        self::assertTrue($object->canRead($filename));
+        self::assertTrue($object->canRead(dirname(__DIR__, 1) . '/_files/documents/reader.docx'));
     }
 
     /**
@@ -46,19 +48,17 @@ public function testCanRead(): void
     public function testCanReadFailed(): void
     {
         $object = new Word2007();
-        $filename = __DIR__ . '/../_files/documents/foo.docx';
-        self::assertFalse($object->canRead($filename));
+        self::assertFalse($object->canRead(dirname(__DIR__, 1) . '/_files/documents/foo.docx'));
     }
 
     /**
      * Load.
      */
     public function testLoad(): void
     {
-        $filename = __DIR__ . '/../_files/documents/reader.docx';
-        $phpWord = IOFactory::load($filename);
+        $phpWord = IOFactory::load(dirname(__DIR__, 1) . '/_files/documents/reader.docx', 'Word2007');
 
-        self::assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $phpWord);
+        self::assertInstanceOf(PhpWord::class, $phpWord);
         self::assertTrue($phpWord->getSettings()->hasDoNotTrackMoves());
         self::assertFalse($phpWord->getSettings()->hasDoNotTrackFormatting());
         self::assertEquals(100, $phpWord->getSettings()->getZoom());
@@ -72,12 +72,50 @@ public function testLoad(): void
      */
     public function testLoadWord2011(): void
     {
-        $filename = __DIR__ . '/../_files/documents/reader-2011.docx';
-        $phpWord = IOFactory::load($filename);
+        $reader = new Word2007();
+        $phpWord = $reader->load(dirname(__DIR__, 1) . '/_files/documents/reader-2011.docx');
 
-        self::assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $phpWord);
+        self::assertInstanceOf(PhpWord::class, $phpWord);
 
         $doc = TestHelperDOCX::getDocument($phpWord);
         self::assertTrue($doc->elementExists('/w:document/w:body/w:p[3]/w:r/w:pict/v:shape/v:imagedata'));
     }
+
+    /**
+     * Load a Word without/withoutImages.
+     *
+     * @dataProvider providerSettingsImageLoading
+     */
+    public function testLoadWord2011SettingsImageLoading(bool $hasImageLoading): void
+    {
+        $reader = new Word2007();
+        $reader->setImageLoading($hasImageLoading);
+        $phpWord = $reader->load(dirname(__DIR__, 1) . '/_files/documents/reader-2011.docx');
+
+        self::assertInstanceOf(PhpWord::class, $phpWord);
+
+        $sections = $phpWord->getSections();
+        self::assertCount(1, $sections);
+        $section = $sections[0];
+        $elements = $section->getElements();
+        self::assertCount(3, $elements);
+        $element = $elements[2];
+        self::assertInstanceOf(TextRun::class, $element);
+        $subElements = $element->getElements();
+        if ($hasImageLoading) {
+            self::assertCount(1, $subElements);
+            $subElement = $subElements[0];
+            self::assertInstanceOf(Image::class, $subElement);
+        } else {
+            self::assertCount(0, $subElements);
+        }
+    }
+
+    public function providerSettingsImageLoading(): iterable
+    {
+        return [
+            [true],
+            [false],
+        ];
+    }
 }