SpecialLinkSearch: clean up munged query variable handling
[mediawiki.git] / includes / HtmlFormatter.php
blobb2926d17bc923c6f10f4e7a9c2ad5f528b0ea89f
1 <?php
2 /**
3 * Performs transformations of HTML by wrapping around libxml2 and working
4 * around its countless bugs.
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published by
8 * the Free Software Foundation; either version 2 of the License, or
9 * (at your option) any later version.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU General Public License for more details.
16 * You should have received a copy of the GNU General Public License along
17 * with this program; if not, write to the Free Software Foundation, Inc.,
18 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
19 * http://www.gnu.org/copyleft/gpl.html
21 * @file
23 class HtmlFormatter {
24 /**
25 * @var DOMDocument
27 private $doc;
29 private $html;
30 private $itemsToRemove = array();
31 private $elementsToFlatten = array();
32 protected $removeMedia = false;
34 /**
35 * Constructor
37 * @param string $html Text to process
39 public function __construct( $html ) {
40 $this->html = $html;
43 /**
44 * Turns a chunk of HTML into a proper document
45 * @param string $html
46 * @return string
48 public static function wrapHTML( $html ) {
49 return '<!doctype html><html><head></head><body>' . $html . '</body></html>';
52 /**
53 * Override this in descendant class to modify HTML after it has been converted from DOM tree
54 * @param string $html HTML to process
55 * @return string Processed HTML
57 protected function onHtmlReady( $html ) {
58 return $html;
61 /**
62 * @return DOMDocument DOM to manipulate
64 public function getDoc() {
65 if ( !$this->doc ) {
66 $html = mb_convert_encoding( $this->html, 'HTML-ENTITIES', 'UTF-8' );
68 // Workaround for bug that caused spaces before references
69 // to disappear during processing:
70 // https://bugzilla.wikimedia.org/show_bug.cgi?id=53086
72 // Please replace with a better fix if one can be found.
73 $html = str_replace( ' <', '&#32;<', $html );
75 libxml_use_internal_errors( true );
76 $loader = libxml_disable_entity_loader();
77 $this->doc = new DOMDocument();
78 $this->doc->strictErrorChecking = false;
79 $this->doc->loadHTML( $html );
80 libxml_disable_entity_loader( $loader );
81 libxml_use_internal_errors( false );
82 $this->doc->encoding = 'UTF-8';
84 return $this->doc;
87 /**
88 * Sets whether images/videos/sounds should be removed from output
89 * @param bool $flag
91 public function setRemoveMedia( $flag = true ) {
92 $this->removeMedia = $flag;
95 /**
96 * Adds one or more selector of content to remove. A subset of CSS selector
97 * syntax is supported:
99 * <tag>
100 * <tag>.class
101 * .<class>
102 * #<id>
104 * @param array|string $selectors Selector(s) of stuff to remove
106 public function remove( $selectors ) {
107 $this->itemsToRemove = array_merge( $this->itemsToRemove, (array)$selectors );
111 * Adds one or more element name to the list to flatten (remove tag, but not its content)
112 * Can accept undelimited regexes
114 * Note this interface may fail in surprising unexpected ways due to usage of regexes,
115 * so should not be relied on for HTML markup security measures.
117 * @param array|string $elements Name(s) of tag(s) to flatten
119 public function flatten( $elements ) {
120 $this->elementsToFlatten = array_merge( $this->elementsToFlatten, (array)$elements );
124 * Instructs the formatter to flatten all tags
126 public function flattenAllTags() {
127 $this->flatten( '[?!]?[a-z0-9]+' );
131 * Removes content we've chosen to remove. The text of the removed elements can be
132 * extracted with the getText method.
133 * @return array Array of removed DOMElements
135 public function filterContent() {
136 $removals = $this->parseItemsToRemove();
138 // Bail out early if nothing to do
139 if ( array_reduce( $removals,
140 function ( $carry, $item ) {
141 return $carry && !$item;
143 true
144 ) ) {
145 return array();
148 $doc = $this->getDoc();
150 // Remove tags
152 // You can't remove DOMNodes from a DOMNodeList as you're iterating
153 // over them in a foreach loop. It will seemingly leave the internal
154 // iterator on the foreach out of wack and results will be quite
155 // strange. Though, making a queue of items to remove seems to work.
156 $domElemsToRemove = array();
157 foreach ( $removals['TAG'] as $tagToRemove ) {
158 $tagToRemoveNodes = $doc->getElementsByTagName( $tagToRemove );
159 foreach ( $tagToRemoveNodes as $tagToRemoveNode ) {
160 if ( $tagToRemoveNode ) {
161 $domElemsToRemove[] = $tagToRemoveNode;
165 $removed = $this->removeElements( $domElemsToRemove );
167 // Elements with named IDs
168 $domElemsToRemove = array();
169 foreach ( $removals['ID'] as $itemToRemove ) {
170 $itemToRemoveNode = $doc->getElementById( $itemToRemove );
171 if ( $itemToRemoveNode ) {
172 $domElemsToRemove[] = $itemToRemoveNode;
175 $removed = array_merge( $removed, $this->removeElements( $domElemsToRemove ) );
177 // CSS Classes
178 $domElemsToRemove = array();
179 $xpath = new DOMXPath( $doc );
180 foreach ( $removals['CLASS'] as $classToRemove ) {
181 $elements = $xpath->query( '//*[contains(@class, "' . $classToRemove . '")]' );
183 /** @var $element DOMElement */
184 foreach ( $elements as $element ) {
185 $classes = $element->getAttribute( 'class' );
186 if ( preg_match( "/\b$classToRemove\b/", $classes ) && $element->parentNode ) {
187 $domElemsToRemove[] = $element;
191 $removed = array_merge( $removed, $this->removeElements( $domElemsToRemove ) );
193 // Tags with CSS Classes
194 foreach ( $removals['TAG_CLASS'] as $classToRemove ) {
195 $parts = explode( '.', $classToRemove );
197 $elements = $xpath->query(
198 '//' . $parts[0] . '[@class="' . $parts[1] . '"]'
200 $removed = array_merge( $removed, $this->removeElements( $elements ) );
203 return $removed;
207 * Removes a list of elelments from DOMDocument
208 * @param array|DOMNodeList $elements
209 * @return array Array of removed elements
211 private function removeElements( $elements ) {
212 $list = $elements;
213 if ( $elements instanceof DOMNodeList ) {
214 $list = array();
215 foreach ( $elements as $element ) {
216 $list[] = $element;
219 /** @var $element DOMElement */
220 foreach ( $list as $element ) {
221 if ( $element->parentNode ) {
222 $element->parentNode->removeChild( $element );
225 return $list;
229 * libxml in its usual pointlessness converts many chars to entities - this function
230 * perfoms a reverse conversion
231 * @param string $html
232 * @return string
234 private function fixLibXML( $html ) {
235 static $replacements;
236 if ( !$replacements ) {
237 // We don't include rules like '&#34;' => '&amp;quot;' because entities had already been
238 // normalized by libxml. Using this function with input not sanitized by libxml is UNSAFE!
239 $replacements = new ReplacementArray( array(
240 '&quot;' => '&amp;quot;',
241 '&amp;' => '&amp;amp;',
242 '&lt;' => '&amp;lt;',
243 '&gt;' => '&amp;gt;',
244 ) );
246 $html = $replacements->replace( $html );
247 $html = mb_convert_encoding( $html, 'UTF-8', 'HTML-ENTITIES' );
248 return $html;
252 * Performs final transformations and returns resulting HTML. Note that if you want to call this
253 * both without an element and with an element you should call it without an element first. If you
254 * specify the $element in the method it'll change the underlying dom and you won't be able to get
255 * it back.
257 * @param DOMElement|string|null $element ID of element to get HTML from or
258 * false to get it from the whole tree
259 * @return string Processed HTML
261 public function getText( $element = null ) {
263 if ( $this->doc ) {
264 if ( $element !== null && !( $element instanceof DOMElement ) ) {
265 $element = $this->doc->getElementById( $element );
267 if ( $element ) {
268 $body = $this->doc->getElementsByTagName( 'body' )->item( 0 );
269 $nodesArray = array();
270 foreach ( $body->childNodes as $node ) {
271 $nodesArray[] = $node;
273 foreach ( $nodesArray as $nodeArray ) {
274 $body->removeChild( $nodeArray );
276 $body->appendChild( $element );
278 $html = $this->doc->saveHTML();
280 $html = $this->fixLibXml( $html );
281 if ( wfIsWindows() ) {
282 // Cleanup for CRLF misprocessing of unknown origin on Windows.
284 // If this error continues in the future, please track it down in the
285 // XML code paths if possible and fix there.
286 $html = str_replace( '&#13;', '', $html );
288 } else {
289 $html = $this->html;
291 // Remove stuff added by wrapHTML()
292 $html = preg_replace( '/<!--.*?-->|^.*?<body>|<\/body>.*$/s', '', $html );
293 $html = $this->onHtmlReady( $html );
295 if ( $this->elementsToFlatten ) {
296 $elements = implode( '|', $this->elementsToFlatten );
297 $html = preg_replace( "#</?($elements)\\b[^>]*>#is", '', $html );
300 return $html;
304 * Helper function for parseItemsToRemove(). This function extracts the selector type
305 * and the raw name of a selector from a CSS-style selector string and assigns those
306 * values to parameters passed by reference. For example, if given '#toc' as the
307 * $selector parameter, it will assign 'ID' as the $type and 'toc' as the $rawName.
308 * @param string $selector CSS selector to parse
309 * @param string $type The type of selector (ID, CLASS, TAG_CLASS, or TAG)
310 * @param string $rawName The raw name of the selector
311 * @return bool Whether the selector was successfully recognised
312 * @throws MWException
314 protected function parseSelector( $selector, &$type, &$rawName ) {
315 if ( strpos( $selector, '.' ) === 0 ) {
316 $type = 'CLASS';
317 $rawName = substr( $selector, 1 );
318 } elseif ( strpos( $selector, '#' ) === 0 ) {
319 $type = 'ID';
320 $rawName = substr( $selector, 1 );
321 } elseif ( strpos( $selector, '.' ) !== 0 && strpos( $selector, '.' ) !== false ) {
322 $type = 'TAG_CLASS';
323 $rawName = $selector;
324 } elseif ( strpos( $selector, '[' ) === false && strpos( $selector, ']' ) === false ) {
325 $type = 'TAG';
326 $rawName = $selector;
327 } else {
328 throw new MWException( __METHOD__ . "(): unrecognized selector '$selector'" );
331 return true;
335 * Transforms CSS-style selectors into an internal representation suitable for
336 * processing by filterContent()
337 * @return array
339 protected function parseItemsToRemove() {
340 $removals = array(
341 'ID' => array(),
342 'TAG' => array(),
343 'CLASS' => array(),
344 'TAG_CLASS' => array(),
347 foreach ( $this->itemsToRemove as $itemToRemove ) {
348 $type = '';
349 $rawName = '';
350 if ( $this->parseSelector( $itemToRemove, $type, $rawName ) ) {
351 $removals[$type][] = $rawName;
355 if ( $this->removeMedia ) {
356 $removals['TAG'][] = 'img';
357 $removals['TAG'][] = 'audio';
358 $removals['TAG'][] = 'video';
361 return $removals;