Merge pull request #1324 from troosan/parse_drawings

Support reading of w:drawing for documents produced by word 2011+

CHANGELOG.md

@@ -37,6 +37,7 @@ v0.15.0 (?? ??? 2018)
 - Fix colspan and rowspan for tables in HTML Writer @mattbolt #1292
 - Fix parsing of Heading and Title formating @troosan @gthomas2 #465
 - Fix Dateformat typo, fix hours casing, add Month-Day-Year formats @ComputerTinker #591
+- Support reading of w:drawing for documents produced by word 2011+ @gthomas2 #464 #1324
 - Fix missing column width in ODText writer @potofcoffee #413
 - Disable entity loader before parsing XML to avoid XXE injection @Tom4t0 #1427
 

composer.json

@@ -61,7 +61,7 @@
         "php": "^5.3.3 || ^7.0",
         "ext-xml": "*",
         "zendframework/zend-escaper": "^2.2",
-        "phpoffice/common": "^0.2"
+        "phpoffice/common": "^0.2.9"
     },
     "require-dev": {
         "ext-zip": "*",

src/PhpWord/Element/AbstractContainer.php

@@ -35,7 +35,7 @@ namespace PhpOffice\PhpWord\Element;
  * @method TOC addTOC(mixed $fontStyle = null, mixed $tocStyle = null, int $minDepth = 1, int $maxDepth = 9)
  * @method PageBreak addPageBreak()
  * @method Table addTable(mixed $style = null)
- * @method Image addImage(string $source, mixed $style = null, bool $isWatermark = false)
+ * @method Image addImage(string $source, mixed $style = null, bool $isWatermark = false, $name = null)
  * @method OLEObject addOLEObject(string $source, mixed $style = null)
  * @method TextBox addTextBox(mixed $style = null)
  * @method Field addField(string $type = null, array $properties = array(), array $options = array(), mixed $text = null)

src/PhpWord/Element/Image.php

@@ -65,6 +65,13 @@ class Image extends AbstractElement
      */
     private $watermark;
 
+    /**
+     * Name of image
+     *
+     * @var string
+     */
+    private $name;
+
     /**
      * Image type
      *
@@ -127,15 +134,17 @@ class Image extends AbstractElement
      * @param string $source
      * @param mixed $style
      * @param bool $watermark
+     * @param string $name
      *
      * @throws \PhpOffice\PhpWord\Exception\InvalidImageException
      * @throws \PhpOffice\PhpWord\Exception\UnsupportedImageTypeException
      */
-    public function __construct($source, $style = null, $watermark = false)
+    public function __construct($source, $style = null, $watermark = false, $name = null)
     {
         $this->source = $source;
-        $this->setIsWatermark($watermark);
         $this->style = $this->setNewStyle(new ImageStyle(), $style, true);
+        $this->setIsWatermark($watermark);
+        $this->setName($name);
 
         $this->checkImage();
     }
@@ -170,6 +179,26 @@ class Image extends AbstractElement
         return $this->sourceType;
     }
 
+    /**
+     * Sets the image name
+     *
+     * @param string $value
+     */
+    public function setName($value)
+    {
+        $this->name = $value;
+    }
+
+    /**
+     * Get image name
+     *
+     * @return null|string
+     */
+    public function getName()
+    {
+        return $this->name;
+    }
+
     /**
      * Get image media ID
      *

src/PhpWord/Reader/Word2007/AbstractPart.php

@@ -261,6 +261,20 @@ abstract class AbstractPart
                 }
                 $parent->addImage($imageSource);
             }
+        } elseif ($node->nodeName == 'w:drawing') {
+            // Office 2011 Image
+            $xmlReader->registerNamespace('wp', 'http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing');
+            $xmlReader->registerNamespace('r', 'http://schemas.openxmlformats.org/officeDocument/2006/relationships');
+            $xmlReader->registerNamespace('pic', 'http://schemas.openxmlformats.org/drawingml/2006/picture');
+            $xmlReader->registerNamespace('a', 'http://schemas.openxmlformats.org/drawingml/2006/main');
+
+            $name = $xmlReader->getAttribute('name', $node, 'wp:inline/a:graphic/a:graphicData/pic:pic/pic:nvPicPr/pic:cNvPr');
+            $embedId = $xmlReader->getAttribute('r:embed', $node, 'wp:inline/a:graphic/a:graphicData/pic:pic/pic:blipFill/a:blip');
+            $target = $this->getMediaTarget($docPart, $embedId);
+            if (!is_null($target)) {
+                $imageSource = "zip://{$this->docFile}#{$target}";
+                $parent->addImage($imageSource, null, false, $name);
+            }
         } elseif ($node->nodeName == 'w:object') {
             // Object
             $rId = $xmlReader->getAttribute('r:id', $node, 'o:OLEObject');

tests/PhpWord/Reader/Word2007/ElementTest.php

@@ -236,4 +236,40 @@ class ElementTest extends AbstractTestReader
         $this->assertEquals('Title', $formattedTitle->getStyle());
         $this->assertInstanceOf('PhpOffice\PhpWord\Element\TextRun', $formattedTitle->getText());
     }
+
+    /**
+     * Test reading Drawing
+     */
+    public function testReadDrawing()
+    {
+        $documentXml = '<w:p>
+            <w:r>
+                <w:drawing xmlns:wp="http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing">
+                    <wp:inline distT="0" distB="0" distL="0" distR="0">
+                        <wp:extent cx="5727700" cy="6621145"/>
+                        <wp:docPr id="1" name="Picture 1"/>
+                        <a:graphic xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main">
+                            <a:graphicData uri="http://schemas.openxmlformats.org/drawingml/2006/picture">
+                                <pic:pic xmlns:pic="http://schemas.openxmlformats.org/drawingml/2006/picture">
+                                    <pic:nvPicPr>
+                                        <pic:cNvPr id="1" name="file_name.jpg"/>
+                                        <pic:cNvPicPr/>
+                                    </pic:nvPicPr>
+                                    <pic:blipFill>
+                                        <a:blip r:embed="rId4" xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships">
+                                        </a:blip>
+                                    </pic:blipFill>
+                                </pic:pic>
+                            </a:graphicData>
+                        </a:graphic>
+                    </wp:inline>
+                </w:drawing>
+            </w:r>
+        </w:p>';
+
+        $phpWord = $this->getDocumentFromString(array('document' => $documentXml));
+
+        $elements = $phpWord->getSection(0)->getElements();
+        $this->assertInstanceOf('PhpOffice\PhpWord\Element\TextRun', $elements[0]);
+    }
 }

tests/PhpWord/Reader/Word2007Test.php

@@ -64,4 +64,18 @@ class Word2007Test extends \PHPUnit\Framework\TestCase
         $doc = TestHelperDOCX::getDocument($phpWord);
         $this->assertFalse($doc->elementExists('/w:document/w:body/w:p/w:r[w:t/node()="italics"]/w:rPr/w:b'));
     }
+
+    /**
+     * Load a Word 2011 file
+     */
+    public function testLoadWord2011()
+    {
+        $filename = __DIR__ . '/../_files/documents/reader-2011.docx';
+        $phpWord = IOFactory::load($filename);
+
+        $this->assertInstanceOf('PhpOffice\\PhpWord\\PhpWord', $phpWord);
+
+        $doc = TestHelperDOCX::getDocument($phpWord);
+        $this->assertTrue($doc->elementExists('/w:document/w:body/w:p[3]/w:r/w:pict/v:shape/v:imagedata'));
+    }
 }