add HTML table parsing

samples/Sample_26_Html.php

@@ -14,6 +14,20 @@ $html .= '<ul><li>Item 1</li><li>Item 2</li><ul><li>Item 2.1</li><li>Item 2.1</l
 $html .= '<p>Ordered (numbered) list:</p>';
 $html .= '<ol><li>Item 1</li><li>Item 2</li></ol>';
 
+$html .= '<table style="width: 50%; border: 6px #0000FF double;">
+                <thead>
+                    <tr style="background-color: #FF0000; text-align: center; color: #FFFFFF; font-weight: bold; ">
+                        <th>header a</th>
+                        <th>header b</th>
+                        <th style="background-color: #FFFF00; border-width: 12px"><span style="background-color: #00FF00;">header c</span></th>
+                    </tr>
+                </thead>
+                <tbody>
+                    <tr><td style="border-style: dotted;">1</td><td colspan="2">2</td></tr>
+                    <tr><td>4</td><td>5</td><td>6</td></tr>
+                </tbody>
+            </table>';
+
 \PhpOffice\PhpWord\Shared\Html::addHtml($section, $html);
 
 // Save file

src/PhpWord/Shared/Converter.php

@@ -296,25 +296,40 @@ class Converter
      */
     public static function cssToPoint($value)
     {
-        preg_match('/^[+-]?([0-9]+.?[0-9]+)?(px|em|ex|%|in|cm|mm|pt|pc)$/i', $value, $matches);
+        if ($value == '0') {
-        $size = $matches[1];
+            return 0;
-        $unit = $matches[2];
-
-        switch ($unit) {
-            case 'pt':
-                return $size;
-            case 'px':
-                return self::pixelToPoint($size);
-            case 'cm':
-                return self::cmToPoint($size);
-            case 'mm':
-                return self::cmToPoint($size / 10);
-            case 'in':
-                return self::inchToPoint($size);
-            case 'pc':
-                return self::picaToPoint($size);
-            default:
-                return null;
         }
+        if (preg_match('/^[+-]?([0-9]+\.?[0-9]*)?(px|em|ex|%|in|cm|mm|pt|pc)$/i', $value, $matches)) {
+            $size = $matches[1];
+            $unit = $matches[2];
+
+            switch ($unit) {
+                case 'pt':
+                    return $size;
+                case 'px':
+                    return self::pixelToPoint($size);
+                case 'cm':
+                    return self::cmToPoint($size);
+                case 'mm':
+                    return self::cmToPoint($size / 10);
+                case 'in':
+                    return self::inchToPoint($size);
+                case 'pc':
+                    return self::picaToPoint($size);
+            }
+        }
+
+        return null;
+    }
+
+    /**
+     * Transforms a size in CSS format (eg. 10px, 10px, ...) to twips
+     *
+     * @param string $value
+     * @return float
+     */
+    public static function cssToTwip($value)
+    {
+        return self::pointToTwip(self::cssToPoint($value));
     }
 }

src/PhpWord/Shared/Html.php

@@ -18,6 +18,8 @@
 namespace PhpOffice\PhpWord\Shared;
 
 use PhpOffice\PhpWord\Element\AbstractContainer;
+use PhpOffice\PhpWord\Element\Row;
+use PhpOffice\PhpWord\Element\Table;
 use PhpOffice\PhpWord\SimpleType\Jc;
 
 /**
@@ -99,7 +101,7 @@ class Html
     protected static function parseNode($node, $element, $styles = array(), $data = array())
     {
         // Populate styles array
-        $styleTypes = array('font', 'paragraph', 'list');
+        $styleTypes = array('font', 'paragraph', 'list', 'table', 'row', 'cell');
         foreach ($styleTypes as $styleType) {
             if (!isset($styles[$styleType])) {
                 $styles[$styleType] = array();
@@ -124,10 +126,11 @@ class Html
             'u'         => array('Property',    null,   null,       $styles,    null,   'underline',    'single'),
             'sup'       => array('Property',    null,   null,       $styles,    null,   'superScript',  true),
             'sub'       => array('Property',    null,   null,       $styles,    null,   'subScript',    true),
-            'span'      => array('Property',    null,   null,       $styles,    null,   'span',         $node),
+            'span'      => array('Span',        $node,  null,       $styles,    null,   null,           null),
-            'table'     => array('Table',       $node,  $element,   $styles,    null,   'addTable',     true),
+            'table'     => array('Table',       $node,  $element,   $styles,    null,   null,           null),
-            'tr'        => array('Table',       $node,  $element,   $styles,    null,   'addRow',       true),
+            'tr'        => array('Row',         $node,  $element,   $styles,    null,   null,           null),
-            'td'        => array('Table',       $node,  $element,   $styles,    null,   'addCell',      true),
+            'td'        => array('Cell',        $node,  $element,   $styles,    null,   null,           null),
+            'th'        => array('Cell',        $node,  $element,   $styles,    null,   null,           null),
             'ul'        => array('List',        null,   null,       $styles,    $data,  3,              null),
             'ol'        => array('List',        null,   null,       $styles,    $data,  7,              null),
             'li'        => array('ListItem',    $node,  $element,   $styles,    $data,  null,           null),
@@ -179,7 +182,7 @@ class Html
             $cNodes = $node->childNodes;
             if (count($cNodes) > 0) {
                 foreach ($cNodes as $cNode) {
-                    if ($element instanceof AbstractContainer) {
+                    if ($element instanceof AbstractContainer || $element instanceof Table || $element instanceof Row) {
                         self::parseNode($cNode, $element, $styles, $data);
                     }
                 }
@@ -197,7 +200,7 @@ class Html
      */
     private static function parseParagraph($node, $element, &$styles)
     {
-        $styles['paragraph'] = self::parseInlineStyle($node, $styles['paragraph']);
+        $styles['paragraph'] = self::recursiveParseStylesInHierarchy($node, $styles['paragraph']);
         $newElement = $element->addTextRun($styles['paragraph']);
 
         return $newElement;
@@ -231,7 +234,12 @@ class Html
      */
     private static function parseText($node, $element, &$styles)
     {
-        $styles['font'] = self::parseInlineStyle($node, $styles['font']);
+        $styles['font'] = self::recursiveParseStylesInHierarchy($node, $styles['font']);
+
+        //alignment applies on paragraph, not on font. Let's copy it there
+        if (isset($styles['font']['alignment'])) {
+            $styles['paragraph']['alignment'] = $styles['font']['alignment'];
+        }
 
         if (is_callable(array($element, 'addText'))) {
             $element->addText($node->nodeValue, $styles['font'], $styles['paragraph']);
@@ -247,16 +255,18 @@ class Html
      */
     private static function parseProperty(&$styles, $argument1, $argument2)
     {
-        if ($argument1 !== 'span') {
+        $styles['font'][$argument1] = $argument2;
-            $styles['font'][$argument1] = $argument2;
+    }
-        } else {
+
-            if (!is_null($argument2->attributes)) {
+    /**
-                $nodeAttr = $argument2->attributes->getNamedItem('style');
+     * Parse span node
-                if (!is_null($nodeAttr) && property_exists($nodeAttr, 'value')) {
+     *
-                    $styles['font'] = self::parseStyle($nodeAttr, $styles['font']);
+     * @param \DOMNode $node
-                }
+     * @param array &$styles
-            }
+     */
-        }
+    private static function parseSpan($node, &$styles)
+    {
+        self::parseInlineStyle($node, $styles['font']);
     }
 
     /**
@@ -270,11 +280,11 @@ class Html
      *
      * @todo As soon as TableItem, RowItem and CellItem support relative width and height
      */
-    private static function parseTable($node, $element, &$styles, $argument1)
+    private static function parseTable($node, $element, &$styles)
     {
-        $styles['paragraph'] = self::parseInlineStyle($node, $styles['paragraph']);
+        $elementStyles = self::parseInlineStyle($node, $styles['table']);
 
-        $newElement = $element->$argument1();
+        $newElement = $element->addTable($elementStyles);
 
         // $attributes = $node->attributes;
         // if ($attributes->getNamedItem('width') !== null) {
@@ -291,6 +301,62 @@ class Html
         return $newElement;
     }
 
+    /**
+     * Parse a table row
+     *
+     * @param \DOMNode $node
+     * @param \PhpOffice\PhpWord\Element\Table $element
+     * @param array &$styles
+     * @return \PhpOffice\PhpWord\Element\AbstractContainer $element
+     */
+    private static function parseRow($node, $element, &$styles)
+    {
+        $rowStyles = self::parseInlineStyle($node, $styles['row']);
+        if ($node->parentNode->nodeName == 'thead') {
+            $rowStyles['tblHeader'] = true;
+        }
+
+        return $element->addRow(null, $rowStyles);
+    }
+
+    /**
+     * Parse table cell
+     *
+     * @param \DOMNode $node
+     * @param \PhpOffice\PhpWord\Element\Table $element
+     * @param array &$styles
+     * @return \PhpOffice\PhpWord\Element\AbstractContainer $element
+     */
+    private static function parseCell($node, $element, &$styles)
+    {
+        $cellStyles = self::recursiveParseStylesInHierarchy($node, $styles['cell']);
+
+        $colspan = $node->getAttribute('colspan');
+        if (!empty($colspan)) {
+            $cellStyles['gridSpan'] = $colspan - 0;
+        }
+
+        return $element->addCell(null, $cellStyles);
+    }
+
+    /**
+     * Recursively parses styles on parent nodes
+     * TODO if too slow, add caching of parent nodes, !! everything is static here so watch out for concurrency !!
+     *
+     * @param \DOMNode $node
+     * @param array &$styles
+     */
+    private static function recursiveParseStylesInHierarchy(\DOMNode $node, array $style)
+    {
+        $parentStyle = self::parseInlineStyle($node, array());
+        $style = array_merge($parentStyle, $style);
+        if ($node->parentNode != null && XML_ELEMENT_NODE == $node->parentNode->nodeType) {
+            $style = self::recursiveParseStylesInHierarchy($node->parentNode, $style);
+        }
+
+        return $style;
+    }
+
     /**
      * Parse list node
      *
@@ -400,9 +466,59 @@ class Html
                     }
                     $styles['italic'] = $tValue;
                     break;
+                case 'border-color':
+                    $styles['color'] = trim($cValue, '#');
+                    break;
+                case 'border-width':
+                    $styles['borderSize'] = Converter::cssToPoint($cValue);
+                    break;
+                case 'border-style':
+                    $styles['borderStyle'] = self::mapBorderStyle($cValue);
+                    break;
+                case 'width':
+                    if (preg_match('/([0-9]+[a-z]+)/', $cValue, $matches)) {
+                        $styles['width'] = Converter::cssToTwip($matches[1]);
+                        $styles['unit'] = \PhpOffice\PhpWord\Style\Table::WIDTH_TWIP;
+                    } elseif (preg_match('/([0-9]+)%/', $cValue, $matches)) {
+                        $styles['width'] = $matches[1] * 50;
+                        $styles['unit'] = \PhpOffice\PhpWord\Style\Table::WIDTH_PERCENT;
+                    } elseif (preg_match('/([0-9]+)/', $cValue, $matches)) {
+                        $styles['width'] = $matches[1];
+                        $styles['unit'] = \PhpOffice\PhpWord\Style\Table::WIDTH_AUTO;
+                    }
+                    break;
+                case 'border':
+                    if (preg_match('/([0-9]+[^0-9]*)\s+(\#[a-fA-F0-9]+)\s+([a-z]+)/', $cValue, $matches)) {
+                        $styles['borderSize'] = Converter::cssToPoint($matches[1]);
+                        $styles['borderColor'] = trim($matches[2], '#');
+                        $styles['borderStyle'] = self::mapBorderStyle($matches[3]);
+                    }
+                    break;
             }
         }
 
         return $styles;
     }
+
+    /**
+     * Transforms a CSS border style into a word border style
+     *
+     * @param string $cssBorderStyle
+     * @return null|string
+     */
+    private static function mapBorderStyle($cssBorderStyle)
+    {
+        if ($cssBorderStyle == null) {
+            return null;
+        }
+        switch ($cssBorderStyle) {
+            case 'none':
+            case 'dashed':
+            case 'dotted':
+            case 'double':
+                return $cssBorderStyle;
+            case 'solid':
+                return 'single';
+        }
+    }
 }

src/PhpWord/Style/Border.php

@@ -36,6 +36,13 @@ class Border extends AbstractStyle
      */
     protected $borderTopColor;
 
+    /**
+     * Border Top Style
+     *
+     * @var string
+     */
+    protected $borderTopStyle;
+
     /**
      * Border Left Size
      *
@@ -50,6 +57,13 @@ class Border extends AbstractStyle
      */
     protected $borderLeftColor;
 
+    /**
+     * Border Left Style
+     *
+     * @var string
+     */
+    protected $borderLeftStyle;
+
     /**
      * Border Right Size
      *
@@ -64,6 +78,13 @@ class Border extends AbstractStyle
      */
     protected $borderRightColor;
 
+    /**
+     * Border Right Style
+     *
+     * @var string
+     */
+    protected $borderRightStyle;
+
     /**
      * Border Bottom Size
      *
@@ -78,6 +99,13 @@ class Border extends AbstractStyle
      */
     protected $borderBottomColor;
 
+    /**
+     * Border Bottom Style
+     *
+     * @var string
+     */
+    protected $borderBottomStyle;
+
     /**
      * Get border size
      *
@@ -140,6 +168,37 @@ class Border extends AbstractStyle
         return $this;
     }
 
+    /**
+     * Get border style
+     *
+     * @return string[]
+     */
+    public function getBorderStyle()
+    {
+        return array(
+            $this->getBorderTopStyle(),
+            $this->getBorderLeftStyle(),
+            $this->getBorderRightStyle(),
+            $this->getBorderBottomStyle(),
+        );
+    }
+
+    /**
+     * Set border style
+     *
+     * @param string $value
+     * @return self
+     */
+    public function setBorderStyle($value = null)
+    {
+        $this->setBorderTopStyle($value);
+        $this->setBorderLeftStyle($value);
+        $this->setBorderRightStyle($value);
+        $this->setBorderBottomStyle($value);
+
+        return $this;
+    }
+
     /**
      * Get border top size
      *
@@ -186,6 +245,29 @@ class Border extends AbstractStyle
         return $this;
     }
 
+    /**
+     * Get border top style
+     *
+     * @return string
+     */
+    public function getBorderTopStyle()
+    {
+        return $this->borderTopStyle;
+    }
+
+    /**
+     * Set border top Style
+     *
+     * @param string $value
+     * @return self
+     */
+    public function setBorderTopStyle($value = null)
+    {
+        $this->borderTopStyle = $value;
+
+        return $this;
+    }
+
     /**
      * Get border left size
      *
@@ -232,6 +314,29 @@ class Border extends AbstractStyle
         return $this;
     }
 
+    /**
+     * Get border left style
+     *
+     * @return string
+     */
+    public function getBorderLeftStyle()
+    {
+        return $this->borderLeftStyle;
+    }
+
+    /**
+     * Set border left style
+     *
+     * @param string $value
+     * @return self
+     */
+    public function setBorderLeftStyle($value = null)
+    {
+        $this->borderLeftStyle = $value;
+
+        return $this;
+    }
+
     /**
      * Get border right size
      *
@@ -278,6 +383,29 @@ class Border extends AbstractStyle
         return $this;
     }
 
+    /**
+     * Get border right style
+     *
+     * @return string
+     */
+    public function getBorderRightStyle()
+    {
+        return $this->borderRightStyle;
+    }
+
+    /**
+     * Set border right style
+     *
+     * @param string $value
+     * @return self
+     */
+    public function setBorderRightStyle($value = null)
+    {
+        $this->borderRightStyle = $value;
+
+        return $this;
+    }
+
     /**
      * Get border bottom size
      *
@@ -324,6 +452,29 @@ class Border extends AbstractStyle
         return $this;
     }
 
+    /**
+     * Get border bottom style
+     *
+     * @return string
+     */
+    public function getBorderBottomStyle()
+    {
+        return $this->borderBottomStyle;
+    }
+
+    /**
+     * Set border bottom style
+     *
+     * @param string $value
+     * @return self
+     */
+    public function setBorderBottomStyle($value = null)
+    {
+        $this->borderBottomStyle = $value;
+
+        return $this;
+    }
+
     /**
      * Check if any of the border is not null
      *

src/PhpWord/Style/Cell.php

@@ -32,13 +32,31 @@ class Cell extends Border
     const VALIGN_BOTTOM = 'bottom';
     const VALIGN_BOTH = 'both';
 
+    //Text direction constants
     /**
-     * Text direction constants
+     * Left to Right, Top to Bottom
-     *
+     */
-     * @const string
+    const TEXT_DIR_LRTB = 'lrTb';
+    /**
+     * Top to Bottom, Right to Left
+     */
+    const TEXT_DIR_TBRL = 'tbRl';
+    /**
+     * Bottom to Top, Left to Right
      */
     const TEXT_DIR_BTLR = 'btLr';
-    const TEXT_DIR_TBRL = 'tbRl';
+    /**
+     * Left to Right, Top to Bottom Rotated
+     */
+    const TEXT_DIR_LRTBV = 'lrTbV';
+    /**
+     * Top to Bottom, Right to Left Rotated
+     */
+    const TEXT_DIR_TBRLV = 'tbRlV';
+    /**
+     * Top to Bottom, Left to Right Rotated
+     */
+    const TEXT_DIR_TBLRV = 'tbLrV';
 
     /**
      * Vertical merge (rowspan) constants
@@ -93,6 +111,20 @@ class Cell extends Border
      */
     private $shading;
 
+    /**
+     * Width
+     *
+     * @var int
+     */
+    private $width;
+
+    /**
+     * Width type
+     *
+     * @var string
+     */
+    private $widthType = Table::WIDTH_TWIP;
+
     /**
      * Get vertical align.
      *
@@ -236,6 +268,51 @@ class Cell extends Border
         return $this;
     }
 
+    /**
+     * Get cell width
+     *
+     * @return int
+     */
+    public function getWidth()
+    {
+        return $this->width;
+    }
+
+    /**
+     * Set cell width
+     *
+     * @param int $value
+     * @return self
+     */
+    public function setWidth($value)
+    {
+        $this->setIntVal($value);
+
+        return $this;
+    }
+
+    /**
+     * Get width type
+     *
+     * @return string
+     */
+    public function getWidthType()
+    {
+        return $this->widthType;
+    }
+
+    /**
+     * Set width type
+     *
+     * @param string $value
+     */
+    public function setWidthType($value)
+    {
+        $this->widthType = $this->setEnumVal($value, array(Table::WIDTH_AUTO, Table::WIDTH_PERCENT, Table::WIDTH_TWIP), Table::WIDTH_TWIP);
+
+        return $this;
+    }
+
     /**
      * Get default border color
      *

src/PhpWord/Writer/Word2007/Style/Cell.php

@@ -65,6 +65,7 @@ class Cell extends AbstractStyle
             $styleWriter = new MarginBorder($xmlWriter);
             $styleWriter->setSizes($style->getBorderSize());
             $styleWriter->setColors($style->getBorderColor());
+            $styleWriter->setStyles($style->getBorderStyle());
             $styleWriter->setAttributes(array('defaultColor' => CellStyle::DEFAULT_BORDER_COLOR));
             $styleWriter->write();
 

src/PhpWord/Writer/Word2007/Style/MarginBorder.php

@@ -40,6 +40,13 @@ class MarginBorder extends AbstractStyle
      */
     private $colors = array();
 
+    /**
+     * Border styles
+     *
+     * @var string[]
+     */
+    private $styles = array();
+
     /**
      * Other attributes
      *
@@ -62,7 +69,8 @@ class MarginBorder extends AbstractStyle
                 if (isset($this->colors[$i])) {
                     $color = $this->colors[$i];
                 }
-                $this->writeSide($xmlWriter, $sides[$i], $this->sizes[$i], $color);
+                $style = isset($this->styles[$i]) ? $this->styles[$i] : 'single';
+                $this->writeSide($xmlWriter, $sides[$i], $this->sizes[$i], $color, $style);
             }
         }
     }
@@ -74,8 +82,9 @@ class MarginBorder extends AbstractStyle
      * @param string $side
      * @param int $width
      * @param string $color
+     * @param string $borderStyle
      */
-    private function writeSide(XMLWriter $xmlWriter, $side, $width, $color = null)
+    private function writeSide(XMLWriter $xmlWriter, $side, $width, $color = null, $borderStyle = 'solid')
     {
         $xmlWriter->startElement('w:' . $side);
         if (!empty($this->colors)) {
@@ -84,9 +93,9 @@ class MarginBorder extends AbstractStyle
                     $color = $this->attributes['defaultColor'];
                 }
             }
-            $xmlWriter->writeAttribute('w:val', 'single');
+            $xmlWriter->writeAttribute('w:val', $borderStyle);
             $xmlWriter->writeAttribute('w:sz', $width);
-            $xmlWriter->writeAttribute('w:color', $color);
+            $xmlWriter->writeAttributeIf($color != null, 'w:color', $color);
             if (!empty($this->attributes)) {
                 if (isset($this->attributes['space'])) {
                     $xmlWriter->writeAttribute('w:space', $this->attributes['space']);
@@ -119,6 +128,16 @@ class MarginBorder extends AbstractStyle
         $this->colors = $value;
     }
 
+    /**
+     * Set border styles.
+     *
+     * @param string[] $value
+     */
+    public function setStyles($value)
+    {
+        $this->styles = $value;
+    }
+
     /**
      * Set attributes.
      *

src/PhpWord/Writer/Word2007/Style/Shading.php

@@ -36,9 +36,9 @@ class Shading extends AbstractStyle
         $xmlWriter = $this->getXmlWriter();
 
         $xmlWriter->startElement('w:shd');
-        $xmlWriter->writeAttribute('w:val', $style->getPattern());
+        $xmlWriter->writeAttributeIf(!is_null($style->getPattern()), 'w:val', $style->getPattern());
-        $xmlWriter->writeAttribute('w:color', $style->getColor());
+        $xmlWriter->writeAttributeIf(!is_null($style->getColor()), 'w:color', $style->getColor());
-        $xmlWriter->writeAttribute('w:fill', $style->getFill());
+        $xmlWriter->writeAttributeIf(!is_null($style->getFill()), 'w:fill', $style->getFill());
         $xmlWriter->endElement();
     }
 }

tests/PhpWord/Shared/HtmlTest.php

@@ -187,25 +187,23 @@ class HtmlTest extends \PHPUnit\Framework\TestCase
     {
         $phpWord = new \PhpOffice\PhpWord\PhpWord();
         $section = $phpWord->addSection();
-        $html = '
+        $html = '<table style="width: 50%; border: 6px #0000FF solid;">
-            <table style="width: 50%; border: 6px #0000FF solid;">
                 <thead>
                     <tr style="background-color: #FF0000; text-align: center; color: #FFFFFF; font-weight: bold; ">
-                        <th>a</th>
+                        <th>header a</th>
-                        <th>b</th>
+                        <th>header b</th>
-                        <th>c</th>
+                        <th style="border-color: #00FF00; border-width: 3px">header c</th>
                     </tr>
                 </thead>
                 <tbody>
-                    <tr><td>1</td><td colspan="2">2</td></tr>
+                    <tr><td style="border-style: dotted;">1</td><td colspan="2">2</td></tr>
                     <tr><td>4</td><td>5</td><td>6</td></tr>
                 </tbody>
             </table>';
         Html::addHtml($section, $html);
 
         $doc = TestHelperDOCX::getDocument($phpWord, 'Word2007');
-//         echo $doc->printXml();
+        $this->assertTrue($doc->elementExists('/w:document/w:body/w:tbl'));
-        $this->assertTrue($doc->elementExists('/w:document/w:body/w:p'));
+        $this->assertTrue($doc->elementExists('/w:document/w:body/w:tbl/w:tr/w:tc'));
-//         $this->assertTrue($doc->elementExists('/w:document/w:body/w:p/w:tbl/w:tr/w:tc'));
     }
 }

tests/PhpWord/Writer/Word2007/ElementTest.php

@@ -45,7 +45,7 @@ class ElementTest extends \PHPUnit\Framework\TestCase
         $elements = array(
             'CheckBox', 'Container', 'Footnote', 'Image', 'Link', 'ListItem', 'ListItemRun',
             'Object', 'PreserveText', 'Table', 'Text', 'TextBox', 'TextBreak', 'Title', 'TOC',
-            'Field', 'Line', 'Shape', 'Chart', 'FormField', 'SDT',
+            'Field', 'Line', 'Shape', 'Chart', 'FormField', 'SDT', 'Bookmark',
         );
         foreach ($elements as $element) {
             $objectClass = 'PhpOffice\\PhpWord\\Writer\\Word2007\\Element\\' . $element;