Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
97.59% |
81 / 83 |
|
50.00% |
2 / 4 |
CRAP | |
0.00% |
0 / 1 |
| Word2007 | |
97.59% |
81 / 83 |
|
50.00% |
2 / 4 |
16 | |
0.00% |
0 / 1 |
| load | |
97.67% |
42 / 43 |
|
0.00% |
0 / 1 |
5 | |||
| readPart | |
88.89% |
8 / 9 |
|
0.00% |
0 / 1 |
2.01 | |||
| readRelationships | |
100.00% |
12 / 12 |
|
100.00% |
1 / 1 |
5 | |||
| getRels | |
100.00% |
19 / 19 |
|
100.00% |
1 / 1 |
4 | |||
| 1 | <?php |
| 2 | |
| 3 | /** |
| 4 | * This file is part of PHPWord - A pure PHP library for reading and writing |
| 5 | * word processing documents. |
| 6 | * |
| 7 | * PHPWord is free software distributed under the terms of the GNU Lesser |
| 8 | * General Public License version 3 as published by the Free Software Foundation. |
| 9 | * |
| 10 | * For the full copyright and license information, please read the LICENSE |
| 11 | * file that was distributed with this source code. For the full list of |
| 12 | * contributors, visit https://github.com/PHPOffice/PHPWord/contributors. |
| 13 | * |
| 14 | * @see https://github.com/PHPOffice/PHPWord |
| 15 | * |
| 16 | * @license http://www.gnu.org/licenses/lgpl.txt LGPL version 3 |
| 17 | */ |
| 18 | |
| 19 | namespace PhpOffice\PhpWord\Reader; |
| 20 | |
| 21 | use Exception; |
| 22 | use PhpOffice\PhpWord\Element\AbstractElement; |
| 23 | use PhpOffice\PhpWord\PhpWord; |
| 24 | use PhpOffice\PhpWord\Reader\Word2007\AbstractPart; |
| 25 | use PhpOffice\PhpWord\Shared\XMLReader; |
| 26 | use PhpOffice\PhpWord\Shared\ZipArchive; |
| 27 | |
| 28 | /** |
| 29 | * Reader for Word2007. |
| 30 | * |
| 31 | * @since 0.8.0 |
| 32 | * |
| 33 | * @todo watermark, checkbox, toc |
| 34 | * @todo Partly done: image, object |
| 35 | */ |
| 36 | class Word2007 extends AbstractReader implements ReaderInterface |
| 37 | { |
| 38 | /** |
| 39 | * Loads PhpWord from file. |
| 40 | * |
| 41 | * @param string $docFile |
| 42 | * |
| 43 | * @return PhpWord |
| 44 | */ |
| 45 | public function load($docFile) |
| 46 | { |
| 47 | $phpWord = new PhpWord(); |
| 48 | $relationships = $this->readRelationships($docFile); |
| 49 | $commentRefs = []; |
| 50 | |
| 51 | $steps = [ |
| 52 | [ |
| 53 | 'stepPart' => 'document', |
| 54 | 'stepItems' => [ |
| 55 | 'styles' => 'Styles', |
| 56 | 'numbering' => 'Numbering', |
| 57 | ], |
| 58 | ], |
| 59 | [ |
| 60 | 'stepPart' => 'main', |
| 61 | 'stepItems' => [ |
| 62 | 'officeDocument' => 'Document', |
| 63 | 'core-properties' => 'DocPropsCore', |
| 64 | 'extended-properties' => 'DocPropsApp', |
| 65 | 'custom-properties' => 'DocPropsCustom', |
| 66 | ], |
| 67 | ], |
| 68 | [ |
| 69 | 'stepPart' => 'document', |
| 70 | 'stepItems' => [ |
| 71 | 'endnotes' => 'Endnotes', |
| 72 | 'footnotes' => 'Footnotes', |
| 73 | 'settings' => 'Settings', |
| 74 | 'comments' => 'Comments', |
| 75 | ], |
| 76 | ], |
| 77 | ]; |
| 78 | |
| 79 | foreach ($steps as $step) { |
| 80 | $stepPart = $step['stepPart']; |
| 81 | $stepItems = $step['stepItems']; |
| 82 | if (!isset($relationships[$stepPart])) { |
| 83 | continue; |
| 84 | } |
| 85 | foreach ($relationships[$stepPart] as $relItem) { |
| 86 | $relType = $relItem['type']; |
| 87 | if (isset($stepItems[$relType])) { |
| 88 | $partName = $stepItems[$relType]; |
| 89 | $xmlFile = $relItem['target']; |
| 90 | $part = $this->readPart($phpWord, $relationships, $commentRefs, $partName, $docFile, $xmlFile); |
| 91 | $commentRefs = $part->getCommentReferences(); |
| 92 | } |
| 93 | } |
| 94 | } |
| 95 | |
| 96 | return $phpWord; |
| 97 | } |
| 98 | |
| 99 | /** |
| 100 | * Read document part. |
| 101 | * |
| 102 | * @param array<string, array<string, null|AbstractElement>> $commentRefs |
| 103 | */ |
| 104 | private function readPart(PhpWord $phpWord, array $relationships, array $commentRefs, string $partName, string $docFile, string $xmlFile): AbstractPart |
| 105 | { |
| 106 | $partClass = "PhpOffice\\PhpWord\\Reader\\Word2007\\{$partName}"; |
| 107 | if (!class_exists($partClass)) { |
| 108 | throw new Exception(sprintf('The part "%s" doesn\'t exist', $partClass)); |
| 109 | } |
| 110 | |
| 111 | /** @var AbstractPart $part Type hint */ |
| 112 | $part = new $partClass($docFile, $xmlFile); |
| 113 | $part->setImageLoading($this->hasImageLoading()); |
| 114 | $part->setRels($relationships); |
| 115 | $part->setCommentReferences($commentRefs); |
| 116 | $part->read($phpWord); |
| 117 | |
| 118 | return $part; |
| 119 | } |
| 120 | |
| 121 | /** |
| 122 | * Read all relationship files. |
| 123 | * |
| 124 | * @param string $docFile |
| 125 | * |
| 126 | * @return array |
| 127 | */ |
| 128 | private function readRelationships($docFile) |
| 129 | { |
| 130 | $relationships = []; |
| 131 | |
| 132 | // _rels/.rels |
| 133 | $relationships['main'] = $this->getRels($docFile, '_rels/.rels'); |
| 134 | |
| 135 | // word/_rels/*.xml.rels |
| 136 | $wordRelsPath = 'word/_rels/'; |
| 137 | $zip = new ZipArchive(); |
| 138 | if ($zip->open($docFile) === true) { |
| 139 | for ($i = 0; $i < $zip->numFiles; ++$i) { |
| 140 | $xmlFile = $zip->getNameIndex($i); |
| 141 | if ((substr($xmlFile, 0, strlen($wordRelsPath))) == $wordRelsPath && (substr($xmlFile, -1)) != '/') { |
| 142 | $docPart = str_replace('.xml.rels', '', str_replace($wordRelsPath, '', $xmlFile)); |
| 143 | $relationships[$docPart] = $this->getRels($docFile, $xmlFile, 'word/'); |
| 144 | } |
| 145 | } |
| 146 | $zip->close(); |
| 147 | } |
| 148 | |
| 149 | return $relationships; |
| 150 | } |
| 151 | |
| 152 | /** |
| 153 | * Get relationship array. |
| 154 | * |
| 155 | * @param string $docFile |
| 156 | * @param string $xmlFile |
| 157 | * @param string $targetPrefix |
| 158 | * |
| 159 | * @return array |
| 160 | */ |
| 161 | private function getRels($docFile, $xmlFile, $targetPrefix = '') |
| 162 | { |
| 163 | $metaPrefix = 'http://schemas.openxmlformats.org/package/2006/relationships/metadata/'; |
| 164 | $officePrefix = 'http://schemas.openxmlformats.org/officeDocument/2006/relationships/'; |
| 165 | |
| 166 | $rels = []; |
| 167 | |
| 168 | $xmlReader = new XMLReader(); |
| 169 | $xmlReader->getDomFromZip($docFile, $xmlFile); |
| 170 | $nodes = $xmlReader->getElements('*'); |
| 171 | foreach ($nodes as $node) { |
| 172 | $rId = $xmlReader->getAttribute('Id', $node); |
| 173 | $type = $xmlReader->getAttribute('Type', $node); |
| 174 | $target = $xmlReader->getAttribute('Target', $node); |
| 175 | $mode = $xmlReader->getAttribute('TargetMode', $node); |
| 176 | |
| 177 | // Remove URL prefixes from $type to make it easier to read |
| 178 | $type = str_replace($metaPrefix, '', $type); |
| 179 | $type = str_replace($officePrefix, '', $type); |
| 180 | $docPart = str_replace('.xml', '', $target); |
| 181 | |
| 182 | // Do not add prefix to link source |
| 183 | if ($type != 'hyperlink' && $mode != 'External') { |
| 184 | $target = $targetPrefix . $target; |
| 185 | } |
| 186 | |
| 187 | // Push to return array |
| 188 | $rels[$rId] = ['type' => $type, 'target' => $target, 'docPart' => $docPart, 'targetMode' => $mode]; |
| 189 | } |
| 190 | ksort($rels); |
| 191 | |
| 192 | return $rels; |
| 193 | } |
| 194 | } |