2014-04-27 10:44:26 +07:00
|
|
|
<?php
|
|
|
|
|
/**
|
2014-05-05 13:06:53 +04:00
|
|
|
* This file is part of PHPWord - A pure PHP library for reading and writing
|
|
|
|
|
* word processing documents.
|
|
|
|
|
*
|
|
|
|
|
* PHPWord is free software distributed under the terms of the GNU Lesser
|
|
|
|
|
* General Public License version 3 as published by the Free Software Foundation.
|
|
|
|
|
*
|
|
|
|
|
* For the full copyright and license information, please read the LICENSE
|
|
|
|
|
* file that was distributed with this source code. For the full list of
|
|
|
|
|
* contributors, visit https://github.com/PHPOffice/PHPWord/contributors.
|
2014-04-27 10:44:26 +07:00
|
|
|
*
|
|
|
|
|
* @link https://github.com/PHPOffice/PHPWord
|
2014-05-05 12:38:31 +04:00
|
|
|
* @copyright 2010-2014 PHPWord contributors
|
2014-05-04 21:03:28 +04:00
|
|
|
* @license http://www.gnu.org/licenses/lgpl.txt LGPL version 3
|
2014-04-27 10:44:26 +07:00
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
namespace PhpOffice\PhpWord\Reader\Word2007;
|
|
|
|
|
|
|
|
|
|
use PhpOffice\PhpWord\PhpWord;
|
|
|
|
|
use PhpOffice\PhpWord\Shared\XMLReader;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Abstract part reader
|
|
|
|
|
*/
|
|
|
|
|
abstract class AbstractPart
|
|
|
|
|
{
|
|
|
|
|
/**
|
|
|
|
|
* Document file
|
|
|
|
|
*
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
protected $docFile;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* XML file
|
|
|
|
|
*
|
|
|
|
|
* @var string
|
|
|
|
|
*/
|
|
|
|
|
protected $xmlFile;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Part relationships
|
|
|
|
|
*
|
|
|
|
|
* @var array
|
|
|
|
|
*/
|
|
|
|
|
protected $rels = array();
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Read part
|
|
|
|
|
*/
|
|
|
|
|
abstract public function read(PhpWord &$phpWord);
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Create new instance
|
|
|
|
|
*
|
|
|
|
|
* @param string $docFile
|
|
|
|
|
* @param string $xmlFile
|
|
|
|
|
*/
|
|
|
|
|
public function __construct($docFile, $xmlFile)
|
|
|
|
|
{
|
|
|
|
|
$this->docFile = $docFile;
|
|
|
|
|
$this->xmlFile = $xmlFile;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Set relationships
|
|
|
|
|
*
|
|
|
|
|
* @param array $value
|
|
|
|
|
*/
|
|
|
|
|
public function setRels($value)
|
|
|
|
|
{
|
|
|
|
|
$this->rels = $value;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Read w:r
|
|
|
|
|
*
|
|
|
|
|
* @param \PhpOffice\PhpWord\Shared\XMLReader $xmlReader
|
|
|
|
|
* @param \DOMElement $domNode
|
|
|
|
|
* @param mixed $parent
|
|
|
|
|
* @param string $docPart
|
2014-05-01 14:37:58 +07:00
|
|
|
* @param mixed $paragraphStyle
|
2014-04-27 10:44:26 +07:00
|
|
|
*
|
|
|
|
|
* @todo Footnote paragraph style
|
|
|
|
|
*/
|
2014-05-01 14:37:58 +07:00
|
|
|
protected function readRun(XMLReader $xmlReader, \DOMElement $domNode, &$parent, $docPart, $paragraphStyle = null)
|
2014-04-27 10:44:26 +07:00
|
|
|
{
|
|
|
|
|
if (!in_array($domNode->nodeName, array('w:r', 'w:hyperlink'))) {
|
|
|
|
|
return;
|
|
|
|
|
}
|
2014-05-01 14:37:58 +07:00
|
|
|
$fontStyle = $this->readFontStyle($xmlReader, $domNode);
|
2014-04-27 10:44:26 +07:00
|
|
|
|
|
|
|
|
// Link
|
|
|
|
|
if ($domNode->nodeName == 'w:hyperlink') {
|
|
|
|
|
$rId = $xmlReader->getAttribute('r:id', $domNode);
|
|
|
|
|
$textContent = $xmlReader->getValue('w:r/w:t', $domNode);
|
|
|
|
|
$target = $this->getMediaTarget($docPart, $rId);
|
|
|
|
|
if (!is_null($target)) {
|
2014-05-01 14:37:58 +07:00
|
|
|
$parent->addLink($target, $textContent, $fontStyle, $paragraphStyle);
|
2014-04-27 10:44:26 +07:00
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
// Footnote
|
|
|
|
|
if ($xmlReader->elementExists('w:footnoteReference', $domNode)) {
|
|
|
|
|
$parent->addFootnote();
|
|
|
|
|
|
|
|
|
|
// Endnote
|
|
|
|
|
} elseif ($xmlReader->elementExists('w:endnoteReference', $domNode)) {
|
|
|
|
|
$parent->addEndnote();
|
|
|
|
|
|
|
|
|
|
// Image
|
|
|
|
|
} elseif ($xmlReader->elementExists('w:pict', $domNode)) {
|
|
|
|
|
$rId = $xmlReader->getAttribute('r:id', $domNode, 'w:pict/v:shape/v:imagedata');
|
|
|
|
|
$target = $this->getMediaTarget($docPart, $rId);
|
|
|
|
|
if (!is_null($target)) {
|
|
|
|
|
$imageSource = "zip://{$this->docFile}#{$target}";
|
|
|
|
|
$parent->addImage($imageSource);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// Object
|
|
|
|
|
} elseif ($xmlReader->elementExists('w:object', $domNode)) {
|
|
|
|
|
$rId = $xmlReader->getAttribute('r:id', $domNode, 'w:object/o:OLEObject');
|
|
|
|
|
// $rIdIcon = $xmlReader->getAttribute('r:id', $domNode, 'w:object/v:shape/v:imagedata');
|
|
|
|
|
$target = $this->getMediaTarget($docPart, $rId);
|
|
|
|
|
if (!is_null($target)) {
|
|
|
|
|
$textContent = "<Object: {$target}>";
|
2014-05-01 14:37:58 +07:00
|
|
|
$parent->addText($textContent, $fontStyle, $paragraphStyle);
|
2014-04-27 10:44:26 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
// TextRun
|
|
|
|
|
} else {
|
|
|
|
|
$textContent = $xmlReader->getValue('w:t', $domNode);
|
2014-05-01 14:37:58 +07:00
|
|
|
$parent->addText($textContent, $fontStyle, $paragraphStyle);
|
2014-04-27 10:44:26 +07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Read w:pPr
|
|
|
|
|
*
|
2014-05-08 19:25:29 +07:00
|
|
|
* @return array|null
|
2014-04-27 10:44:26 +07:00
|
|
|
*/
|
|
|
|
|
protected function readParagraphStyle(XMLReader $xmlReader, \DOMElement $domNode)
|
|
|
|
|
{
|
2014-05-08 19:25:29 +07:00
|
|
|
if (!$xmlReader->elementExists('w:pPr', $domNode)) {
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
$style = array();
|
|
|
|
|
$mapping = array(
|
|
|
|
|
'w:pStyle' => 'styleName',
|
|
|
|
|
'w:ind' => 'indent', 'w:spacing' => 'spacing',
|
|
|
|
|
'w:jc' => 'align', 'w:basedOn' => 'basedOn', 'w:next' => 'next',
|
|
|
|
|
'w:widowControl' => 'widowControl', 'w:keepNext' => 'keepNext',
|
|
|
|
|
'w:keepLines' => 'keepLines', 'w:pageBreakBefore' => 'pageBreakBefore',
|
|
|
|
|
);
|
|
|
|
|
|
|
|
|
|
$nodes = $xmlReader->getElements('w:pPr/*', $domNode);
|
|
|
|
|
foreach ($nodes as $node) {
|
|
|
|
|
if (!array_key_exists($node->nodeName, $mapping)) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
$property = $mapping[$node->nodeName];
|
|
|
|
|
switch ($node->nodeName) {
|
|
|
|
|
|
|
|
|
|
case 'w:ind':
|
|
|
|
|
$style['indent'] = $xmlReader->getAttribute('w:left', $node);
|
|
|
|
|
$style['hanging'] = $xmlReader->getAttribute('w:hanging', $node);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:spacing':
|
|
|
|
|
$style['spaceAfter'] = $xmlReader->getAttribute('w:after', $node);
|
|
|
|
|
$style['spaceBefore'] = $xmlReader->getAttribute('w:before', $node);
|
|
|
|
|
// Commented. Need to adjust the number when return value is null
|
|
|
|
|
// $style['spacing'] = $xmlReader->getAttribute('w:line', $node);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:keepNext':
|
|
|
|
|
case 'w:keepLines':
|
|
|
|
|
case 'w:pageBreakBefore':
|
|
|
|
|
$style[$property] = true;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:widowControl':
|
|
|
|
|
$style[$property] = false;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:pStyle':
|
|
|
|
|
case 'w:jc':
|
|
|
|
|
case 'w:basedOn':
|
|
|
|
|
case 'w:next':
|
|
|
|
|
$style[$property] = $xmlReader->getAttribute('w:val', $node);
|
|
|
|
|
break;
|
2014-04-27 10:44:26 +07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return $style;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Read w:rPr
|
|
|
|
|
*
|
2014-05-08 19:25:29 +07:00
|
|
|
* @return array|null
|
2014-04-27 10:44:26 +07:00
|
|
|
*/
|
|
|
|
|
protected function readFontStyle(XMLReader $xmlReader, \DOMElement $domNode)
|
|
|
|
|
{
|
2014-05-08 19:25:29 +07:00
|
|
|
if (is_null($domNode)) {
|
|
|
|
|
return;
|
|
|
|
|
}
|
2014-04-27 10:44:26 +07:00
|
|
|
// Hyperlink has an extra w:r child
|
|
|
|
|
if ($domNode->nodeName == 'w:hyperlink') {
|
|
|
|
|
$domNode = $xmlReader->getElement('w:r', $domNode);
|
|
|
|
|
}
|
2014-05-08 19:25:29 +07:00
|
|
|
if (!$xmlReader->elementExists('w:rPr', $domNode)) {
|
|
|
|
|
return;
|
2014-04-27 10:44:26 +07:00
|
|
|
}
|
2014-05-08 19:25:29 +07:00
|
|
|
|
|
|
|
|
$style = array();
|
|
|
|
|
$mapping = array(
|
|
|
|
|
'w:rStyle' => 'styleName',
|
|
|
|
|
'w:b' => 'bold', 'w:i' => 'italic', 'w:color' => 'color',
|
|
|
|
|
'w:strike' => 'strikethrough', 'w:u' => 'underline',
|
|
|
|
|
'w:highlight' => 'fgColor', 'w:sz' => 'size',
|
|
|
|
|
'w:rFonts' => 'name', 'w:vertAlign' => 'superScript',
|
|
|
|
|
);
|
|
|
|
|
|
|
|
|
|
$nodes = $xmlReader->getElements('w:rPr/*', $domNode);
|
|
|
|
|
foreach ($nodes as $node) {
|
|
|
|
|
if (!array_key_exists($node->nodeName, $mapping)) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
$property = $mapping[$node->nodeName];
|
|
|
|
|
switch ($node->nodeName) {
|
|
|
|
|
|
|
|
|
|
case 'w:rFonts':
|
|
|
|
|
$style['name'] = $xmlReader->getAttribute('w:ascii', $node);
|
|
|
|
|
$style['hint'] = $xmlReader->getAttribute('w:hint', $node);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:b':
|
|
|
|
|
case 'w:i':
|
|
|
|
|
case 'w:strike':
|
|
|
|
|
$style[$property] = true;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:rStyle':
|
|
|
|
|
case 'w:u':
|
|
|
|
|
case 'w:highlight':
|
|
|
|
|
case 'w:color':
|
|
|
|
|
$style[$property] = $xmlReader->getAttribute('w:val', $node);
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:sz':
|
|
|
|
|
$style[$property] = $xmlReader->getAttribute('w:val', $node) / 2;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:vertAlign':
|
|
|
|
|
$style[$property] = $xmlReader->getAttribute('w:val', $node);
|
|
|
|
|
if ($style[$property] == 'superscript') {
|
|
|
|
|
$style['superScript'] = true;
|
|
|
|
|
} else {
|
|
|
|
|
$style['superScript'] = false;
|
|
|
|
|
$style['subScript'] = true;
|
2014-04-27 10:44:26 +07:00
|
|
|
}
|
2014-05-08 19:25:29 +07:00
|
|
|
break;
|
2014-04-27 10:44:26 +07:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return $style;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Read w:tblPr
|
|
|
|
|
*
|
|
|
|
|
* @return string|array|null
|
|
|
|
|
* @todo Capture w:tblStylePr w:type="firstRow"
|
|
|
|
|
*/
|
|
|
|
|
protected function readTableStyle(XMLReader $xmlReader, \DOMElement $domNode)
|
|
|
|
|
{
|
|
|
|
|
$style = null;
|
|
|
|
|
$margins = array('top', 'left', 'bottom', 'right');
|
|
|
|
|
$borders = $margins + array('insideH', 'insideV');
|
|
|
|
|
|
|
|
|
|
if ($xmlReader->elementExists('w:tblPr', $domNode)) {
|
|
|
|
|
if ($xmlReader->elementExists('w:tblPr/w:tblStyle', $domNode)) {
|
|
|
|
|
$style = $xmlReader->getAttribute('w:val', $domNode, 'w:tblPr/w:tblStyle');
|
|
|
|
|
} else {
|
|
|
|
|
$style = array();
|
|
|
|
|
$mapping = array(
|
|
|
|
|
'w:tblCellMar' => 'cellMargin',
|
|
|
|
|
'w:tblBorders' => 'border',
|
|
|
|
|
);
|
|
|
|
|
|
|
|
|
|
$nodes = $xmlReader->getElements('w:tblPr/*', $domNode);
|
|
|
|
|
foreach ($nodes as $node) {
|
|
|
|
|
if (!array_key_exists($node->nodeName, $mapping)) {
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
// $property = $mapping[$node->nodeName];
|
|
|
|
|
switch ($node->nodeName) {
|
|
|
|
|
|
|
|
|
|
case 'w:tblCellMar':
|
|
|
|
|
foreach ($margins as $side) {
|
|
|
|
|
$ucfSide = ucfirst($side);
|
|
|
|
|
$style["cellMargin$ucfSide"] = $xmlReader->getAttribute('w:w', $node, "w:$side");
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case 'w:tblBorders':
|
|
|
|
|
foreach ($borders as $side) {
|
|
|
|
|
$ucfSide = ucfirst($side);
|
|
|
|
|
$style["border{$ucfSide}Size"] = $xmlReader->getAttribute('w:sz', $node, "w:$side");
|
|
|
|
|
$style["border{$ucfSide}Color"] = $xmlReader->getAttribute('w:color', $node, "w:$side");
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return $style;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Returns the target of image, object, or link as stored in ::readMainRels
|
|
|
|
|
*
|
|
|
|
|
* @param string $docPart
|
|
|
|
|
* @param string $rId
|
|
|
|
|
* @return string|null
|
|
|
|
|
*/
|
|
|
|
|
private function getMediaTarget($docPart, $rId)
|
|
|
|
|
{
|
|
|
|
|
$target = null;
|
|
|
|
|
if (array_key_exists($docPart, $this->rels)) {
|
|
|
|
|
if (array_key_exists($rId, $this->rels[$docPart])) {
|
|
|
|
|
$target = $this->rels[$docPart][$rId]['target'];
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return $target;
|
|
|
|
|
}
|
|
|
|
|
}
|