Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
Total | |
12.28% |
28 / 228 |
|
5.00% |
2 / 40 |
CRAP | |
0.00% |
0 / 1 |
DOMDataUtils | |
12.28% |
28 / 228 |
|
5.00% |
2 / 40 |
5314.98 | |
0.00% |
0 / 1 |
getBag | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
getCodec | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
prepareDoc | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
2 | |||
prepareChildDoc | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
stashObjectInDoc | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
noAttrs | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
20 | |||
getNodeData | |
0.00% |
0 / 16 |
|
0.00% |
0 / 1 |
20 | |||
setNodeData | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
getDataParsoid | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
setDataParsoid | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
getDataMwI18n | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
setDataMwI18n | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
getDataNodeI18n | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
setDataNodeI18n | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
getDataAttrI18n | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
setDataAttrI18n | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
getDataAttrI18nNames | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
getDataParsoidDiff | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
setDataParsoidDiff | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
getDataMw | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
setDataMw | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
validDataMw | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
validDataMwI18n | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
getJSONAttribute | |
0.00% |
0 / 9 |
|
0.00% |
0 / 1 |
12 | |||
setJSONAttribute | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
6 | |||
setShadowInfo | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
2 | |||
setShadowInfoIfModified | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
42 | |||
addNormalizedAttribute | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
getPageBundle | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
storeInPageBundle | |
80.95% |
17 / 21 |
|
0.00% |
0 / 1 |
6.25 | |||
getCodecHints | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
6 | |||
injectPageBundle | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
1 | |||
extractPageBundle | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
2 | |||
visitAndLoadDataAttribs | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
loadDataAttribs | |
0.00% |
0 / 25 |
|
0.00% |
0 / 1 |
30 | |||
usedIdIndex | |
0.00% |
0 / 11 |
|
0.00% |
0 / 1 |
12 | |||
visitAndStoreDataAttribs | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
6 | |||
storeDataAttribs | |
0.00% |
0 / 43 |
|
0.00% |
0 / 1 |
210 | |||
cloneNode | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
fixClonedData | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
20 |
1 | <?php |
2 | declare( strict_types = 1 ); |
3 | |
4 | namespace Wikimedia\Parsoid\Utils; |
5 | |
6 | use Composer\Semver\Semver; |
7 | use stdClass; |
8 | use Wikimedia\Assert\Assert; |
9 | use Wikimedia\Assert\UnreachableException; |
10 | use Wikimedia\JsonCodec\Hint; |
11 | use Wikimedia\JsonCodec\JsonCodec; |
12 | use Wikimedia\Parsoid\Config\Env; |
13 | use Wikimedia\Parsoid\Core\PageBundle; |
14 | use Wikimedia\Parsoid\DOM\Document; |
15 | use Wikimedia\Parsoid\DOM\Element; |
16 | use Wikimedia\Parsoid\DOM\Node; |
17 | use Wikimedia\Parsoid\NodeData\DataBag; |
18 | use Wikimedia\Parsoid\NodeData\DataMw; |
19 | use Wikimedia\Parsoid\NodeData\DataMwI18n; |
20 | use Wikimedia\Parsoid\NodeData\DataParsoid; |
21 | use Wikimedia\Parsoid\NodeData\I18nInfo; |
22 | use Wikimedia\Parsoid\NodeData\NodeData; |
23 | use Wikimedia\Parsoid\NodeData\TempData; |
24 | |
25 | /** |
26 | * These helpers pertain to HTML and data attributes of a node. |
27 | */ |
28 | class DOMDataUtils { |
29 | public const DATA_OBJECT_ATTR_NAME = 'data-object-id'; |
30 | |
31 | /** |
32 | * Return the dynamic "bag" property of a Document. |
33 | * @param Document $doc |
34 | * @return DataBag |
35 | */ |
36 | public static function getBag( Document $doc ): DataBag { |
37 | // This is a dynamic property; it is not declared. |
38 | // All references go through here so we can suppress phan's complaint. |
39 | // @phan-suppress-next-line PhanUndeclaredProperty |
40 | return $doc->bag; |
41 | } |
42 | |
43 | /** |
44 | * Return the JsonCodec used for rich attributes in a Document. |
45 | * @param Document $doc |
46 | * @return JsonCodec |
47 | */ |
48 | public static function getCodec( Document $doc ): JsonCodec { |
49 | // This is a dynamic property; it is not declared. |
50 | // All references go through here so we can suppress phan's complaint. |
51 | // @phan-suppress-next-line PhanUndeclaredProperty |
52 | return $doc->codec; |
53 | } |
54 | |
55 | public static function prepareDoc( Document $doc ): void { |
56 | // `bag` is a deliberate dynamic property; see DOMDataUtils::getBag() |
57 | // @phan-suppress-next-line PhanUndeclaredProperty dynamic property |
58 | $doc->bag = new DataBag(); |
59 | // `codec` is a deliberate dynamic property; see DOMDataUtils::getCodec() |
60 | // @phan-suppress-next-line PhanUndeclaredProperty dynamic property |
61 | $doc->codec = new JsonCodec(); |
62 | |
63 | // Cache the head and body. |
64 | DOMCompat::getHead( $doc ); |
65 | DOMCompat::getBody( $doc ); |
66 | } |
67 | |
68 | /** |
69 | * @param Document $topLevelDoc |
70 | * @param Document $childDoc |
71 | */ |
72 | public static function prepareChildDoc( Document $topLevelDoc, Document $childDoc ) { |
73 | // @phan-suppress-next-line PhanUndeclaredProperty dynamic property |
74 | Assert::invariant( $topLevelDoc->bag instanceof DataBag, 'doc bag not set' ); |
75 | // @phan-suppress-next-line PhanUndeclaredProperty dynamic property |
76 | $childDoc->bag = $topLevelDoc->bag; |
77 | } |
78 | |
79 | /** |
80 | * Stash $obj in $doc and return an id for later retrieval |
81 | * @param Document $doc |
82 | * @param NodeData $obj |
83 | * @return int |
84 | */ |
85 | public static function stashObjectInDoc( Document $doc, NodeData $obj ): int { |
86 | return self::getBag( $doc )->stashObject( $obj ); |
87 | } |
88 | |
89 | /** |
90 | * Does this node have any attributes? |
91 | * @param Element $node |
92 | * @return bool |
93 | */ |
94 | public static function noAttrs( Element $node ): bool { |
95 | // The 'xmlns' attribute is "invisible" T235295 |
96 | if ( $node->hasAttribute( 'xmlns' ) ) { |
97 | return false; |
98 | } |
99 | $numAttrs = count( $node->attributes ); |
100 | return $numAttrs === 0 || |
101 | ( $numAttrs === 1 && $node->hasAttribute( self::DATA_OBJECT_ATTR_NAME ) ); |
102 | } |
103 | |
104 | /** |
105 | * Get data object from a node. |
106 | * |
107 | * @param Element $node node |
108 | * @return NodeData |
109 | */ |
110 | public static function getNodeData( Element $node ): NodeData { |
111 | if ( !$node->hasAttribute( self::DATA_OBJECT_ATTR_NAME ) ) { |
112 | // Initialized on first request |
113 | $dataObject = new NodeData; |
114 | self::setNodeData( $node, $dataObject ); |
115 | return $dataObject; |
116 | } |
117 | |
118 | $nodeId = DOMCompat::getAttribute( $node, self::DATA_OBJECT_ATTR_NAME ); |
119 | if ( $nodeId !== null ) { |
120 | $dataObject = self::getBag( $node->ownerDocument )->getObject( (int)$nodeId ); |
121 | } else { |
122 | $dataObject = null; // Make phan happy |
123 | } |
124 | Assert::invariant( isset( $dataObject ), 'Bogus nodeId given!' ); |
125 | if ( isset( $dataObject->storedId ) ) { |
126 | throw new UnreachableException( |
127 | 'Trying to fetch node data without loading!' . |
128 | // If this node's data-object id is different from storedId, |
129 | // it will indicate that the data-parsoid object was shared |
130 | // between nodes without getting cloned. Useful for debugging. |
131 | 'Node id: ' . $nodeId . |
132 | 'Stored data: ' . PHPUtils::jsonEncode( $dataObject ) |
133 | ); |
134 | } |
135 | return $dataObject; |
136 | } |
137 | |
138 | /** |
139 | * Set node data. |
140 | * |
141 | * @param Element $node node |
142 | * @param NodeData $data data |
143 | */ |
144 | public static function setNodeData( Element $node, NodeData $data ): void { |
145 | $nodeId = self::stashObjectInDoc( $node->ownerDocument, $data ); |
146 | $node->setAttribute( self::DATA_OBJECT_ATTR_NAME, (string)$nodeId ); |
147 | } |
148 | |
149 | /** |
150 | * Get data parsoid info from a node. |
151 | * |
152 | * @param Element $node node |
153 | * @return DataParsoid |
154 | */ |
155 | public static function getDataParsoid( Element $node ): DataParsoid { |
156 | $data = self::getNodeData( $node ); |
157 | $data->parsoid ??= new DataParsoid; |
158 | return $data->parsoid; |
159 | } |
160 | |
161 | /** |
162 | * Set data parsoid info on a node. |
163 | * |
164 | * @param Element $node node |
165 | * @param DataParsoid $dp data-parsoid |
166 | */ |
167 | public static function setDataParsoid( Element $node, DataParsoid $dp ): void { |
168 | $data = self::getNodeData( $node ); |
169 | $data->parsoid = $dp; |
170 | } |
171 | |
172 | /** |
173 | * Returns the i18n information of a node. This is in private access because it shouldn't |
174 | * typically be used directly; instead getDataNodeI18n and getDataAttrI18n should be used. |
175 | * @param Element $node |
176 | * @return DataMwI18n|null |
177 | */ |
178 | private static function getDataMwI18n( Element $node ): ?DataMwI18n { |
179 | $data = self::getNodeData( $node ); |
180 | // We won't set a default value for this property |
181 | return $data->i18n ?? null; |
182 | } |
183 | |
184 | /** |
185 | * Sets the i18n information of a node. This is in private access because it shouldn't |
186 | * typically be used directly; instead setDataNodeI18n and setDataAttrI18n should be used. |
187 | */ |
188 | private static function setDataMwI18n( Element $node, DataMwI18n $i18n ) { |
189 | $data = self::getNodeData( $node ); |
190 | $data->i18n = $i18n; |
191 | } |
192 | |
193 | /** |
194 | * Retrieves internationalization (i18n) information of a node (typically for localization) |
195 | * @param Element $node |
196 | * @return ?I18nInfo |
197 | */ |
198 | public static function getDataNodeI18n( Element $node ): ?I18nInfo { |
199 | $data = self::getNodeData( $node ); |
200 | // We won't set a default value for this property |
201 | if ( !isset( $data->i18n ) ) { |
202 | return null; |
203 | } |
204 | return $data->i18n->getSpanInfo(); |
205 | } |
206 | |
207 | /** |
208 | * Sets internationalization (i18n) information of a node, used for later localization |
209 | */ |
210 | public static function setDataNodeI18n( Element $node, I18nInfo $i18n ) { |
211 | $data = self::getNodeData( $node ); |
212 | $data->i18n ??= new DataMwI18n(); |
213 | $data->i18n->setSpanInfo( $i18n ); |
214 | } |
215 | |
216 | /** |
217 | * Retrieves internationalization (i18n) information of an attribute value (typically for |
218 | * localization) |
219 | * @param Element $node |
220 | * @param string $name |
221 | * @return ?I18nInfo |
222 | */ |
223 | public static function getDataAttrI18n( Element $node, string $name ): ?I18nInfo { |
224 | $data = self::getNodeData( $node ); |
225 | // We won't set a default value for this property |
226 | if ( !isset( $data->i18n ) ) { |
227 | return null; |
228 | } |
229 | return $data->i18n->getAttributeInfo( $name ); |
230 | } |
231 | |
232 | /** |
233 | * Sets internationalization (i18n) information of a attribute value, used for later |
234 | * localization |
235 | * @param Element $node |
236 | * @param string $name |
237 | * @param I18nInfo $i18n |
238 | */ |
239 | public static function setDataAttrI18n( Element $node, string $name, I18nInfo $i18n ) { |
240 | $data = self::getNodeData( $node ); |
241 | $data->i18n ??= new DataMwI18n(); |
242 | $data->i18n->setAttributeInfo( $name, $i18n ); |
243 | } |
244 | |
245 | /** |
246 | * @param Element $node |
247 | * @return array |
248 | */ |
249 | public static function getDataAttrI18nNames( Element $node ): array { |
250 | $data = self::getNodeData( $node ); |
251 | // We won't set a default value for this property |
252 | if ( !isset( $data->i18n ) ) { |
253 | return []; |
254 | } |
255 | return $data->i18n->getAttributeNames(); |
256 | } |
257 | |
258 | /** |
259 | * Get data diff info from a node. |
260 | * |
261 | * @param Element $node node |
262 | * @return ?stdClass |
263 | */ |
264 | public static function getDataParsoidDiff( Element $node ): ?stdClass { |
265 | $data = self::getNodeData( $node ); |
266 | // We won't set a default value for this property |
267 | return $data->parsoid_diff ?? null; |
268 | } |
269 | |
270 | /** |
271 | * Set data diff info on a node. |
272 | * |
273 | * @param Element $node node |
274 | * @param ?stdClass $diffObj data-parsoid-diff object |
275 | */ |
276 | public static function setDataParsoidDiff( Element $node, ?stdClass $diffObj ): void { |
277 | $data = self::getNodeData( $node ); |
278 | $data->parsoid_diff = $diffObj; |
279 | } |
280 | |
281 | /** |
282 | * Get data meta wiki info from a node. |
283 | * |
284 | * @param Element $node node |
285 | * @return DataMw |
286 | */ |
287 | public static function getDataMw( Element $node ): DataMw { |
288 | $data = self::getNodeData( $node ); |
289 | $data->mw ??= new DataMw; |
290 | return $data->mw; |
291 | } |
292 | |
293 | /** |
294 | * Set data meta wiki info from a node. |
295 | * |
296 | * @param Element $node node |
297 | * @param ?DataMw $dmw data-mw |
298 | */ |
299 | public static function setDataMw( Element $node, ?DataMw $dmw ): void { |
300 | $data = self::getNodeData( $node ); |
301 | $data->mw = $dmw; |
302 | } |
303 | |
304 | /** |
305 | * Check if there is meta wiki info in a node. |
306 | * |
307 | * @param Element $node node |
308 | * @return bool |
309 | */ |
310 | public static function validDataMw( Element $node ): bool { |
311 | return (array)self::getDataMw( $node ) !== []; |
312 | } |
313 | |
314 | /** |
315 | * Check if there is i18n info on a node (for the node or its attributes) |
316 | * @param Element $node |
317 | * @return bool |
318 | */ |
319 | public static function validDataMwI18n( Element $node ): bool { |
320 | return self::getDataMwI18n( $node ) !== null; |
321 | } |
322 | |
323 | /** |
324 | * Get an object from a JSON-encoded XML attribute on a node. |
325 | * |
326 | * @param Element $node node |
327 | * @param string $name name |
328 | * @param mixed $defaultVal |
329 | * @return mixed |
330 | */ |
331 | public static function getJSONAttribute( Element $node, string $name, $defaultVal ) { |
332 | $attVal = DOMCompat::getAttribute( $node, $name ); |
333 | if ( $attVal === null ) { |
334 | return $defaultVal; |
335 | } |
336 | $decoded = PHPUtils::jsonDecode( $attVal, false ); |
337 | if ( $decoded !== null ) { |
338 | return $decoded; |
339 | } else { |
340 | error_log( 'ERROR: Could not decode attribute-val ' . $attVal . |
341 | ' for ' . $name . ' on node ' . DOMCompat::nodeName( $node ) ); |
342 | return $defaultVal; |
343 | } |
344 | } |
345 | |
346 | /** |
347 | * Set a attribute on a node with a JSON-encoded object. |
348 | * |
349 | * @param Element $node node |
350 | * @param string $name Name of the attribute. |
351 | * @param mixed $obj value of the attribute to |
352 | */ |
353 | public static function setJSONAttribute( Element $node, string $name, $obj ): void { |
354 | $val = $obj === [] ? '{}' : PHPUtils::jsonEncode( $obj ); |
355 | $node->setAttribute( $name, $val ); |
356 | } |
357 | |
358 | /** |
359 | * Set shadow info on a node; similar to the method on tokens. |
360 | * Records a key = value pair in data-parsoid['a'] property. |
361 | * |
362 | * This is effectively a call of 'setShadowInfoIfModified' except |
363 | * there is no original value, so by definition, $val is modified. |
364 | * |
365 | * @param Element $node node |
366 | * @param string $name Name of the attribute. |
367 | * @param mixed $val val |
368 | */ |
369 | public static function setShadowInfo( Element $node, string $name, $val ): void { |
370 | $dp = self::getDataParsoid( $node ); |
371 | $dp->a ??= []; |
372 | $dp->sa ??= []; |
373 | $dp->a[$name] = $val; |
374 | } |
375 | |
376 | /** |
377 | * Set shadow info on a node; similar to the method on tokens. |
378 | * |
379 | * If the new value ($val) for the key ($name) is different from the |
380 | * original value ($origVal): |
381 | * - the new value is recorded in data-parsoid->a and |
382 | * - the original value is recorded in data-parsoid->sa |
383 | * |
384 | * @param Element $node node |
385 | * @param string $name Name of the attribute. |
386 | * @param mixed $val val |
387 | * @param mixed $origVal original value (null is a valid value) |
388 | * @param bool $skipOrig |
389 | */ |
390 | public static function setShadowInfoIfModified( |
391 | Element $node, string $name, $val, $origVal, bool $skipOrig = false |
392 | ): void { |
393 | if ( !$skipOrig && ( $val === $origVal || $origVal === null ) ) { |
394 | return; |
395 | } |
396 | $dp = self::getDataParsoid( $node ); |
397 | $dp->a ??= []; |
398 | $dp->sa ??= []; |
399 | // FIXME: This is a hack to not overwrite already shadowed info. |
400 | // We should either fix the call site that depends on this |
401 | // behaviour to do an explicit check, or double down on this |
402 | // by porting it to the token method as well. |
403 | if ( !$skipOrig && !array_key_exists( $name, $dp->a ) ) { |
404 | $dp->sa[$name] = $origVal; |
405 | } |
406 | $dp->a[$name] = $val; |
407 | } |
408 | |
409 | /** |
410 | * Set an attribute and shadow info to a node. |
411 | * Similar to the method on tokens |
412 | * |
413 | * @param Element $node node |
414 | * @param string $name Name of the attribute. |
415 | * @param mixed $val value |
416 | * @param mixed $origVal original value |
417 | * @param bool $skipOrig |
418 | */ |
419 | public static function addNormalizedAttribute( |
420 | Element $node, string $name, $val, $origVal, bool $skipOrig = false |
421 | ): void { |
422 | if ( $name === 'id' ) { |
423 | DOMCompat::setIdAttribute( $node, $val ); |
424 | } else { |
425 | $node->setAttribute( $name, $val ); |
426 | } |
427 | self::setShadowInfoIfModified( $node, $name, $val, $origVal, $skipOrig ); |
428 | } |
429 | |
430 | /** |
431 | * Get this document's pagebundle object |
432 | * @param Document $doc |
433 | * @return PageBundle |
434 | */ |
435 | public static function getPageBundle( Document $doc ): PageBundle { |
436 | return self::getBag( $doc )->getPageBundle(); |
437 | } |
438 | |
439 | /** |
440 | * Removes the `data-*` attribute from a node, and migrates the data to the |
441 | * document's JSON store. Generates a unique id with the following format: |
442 | * ``` |
443 | * mw<base64-encoded counter> |
444 | * ``` |
445 | * but attempts to keep user defined ids. |
446 | * |
447 | * TODO: Note that $data is effective a partial PageBundle containing |
448 | * only the 'parsoid' and 'mw' properties. |
449 | * |
450 | * @param Element $node node |
451 | * @param Env $env environment |
452 | * @param stdClass $data data |
453 | * @param array $idIndex Index of used id attributes in the DOM |
454 | */ |
455 | public static function storeInPageBundle( |
456 | Element $node, Env $env, stdClass $data, array $idIndex |
457 | ): void { |
458 | $hints = self::getCodecHints(); |
459 | $uid = DOMCompat::getAttribute( $node, 'id' ); |
460 | $document = $node->ownerDocument; |
461 | $pb = self::getPageBundle( $document ); |
462 | $codec = self::getCodec( $document ); |
463 | $docDp = &$pb->parsoid; |
464 | $origId = $uid; |
465 | if ( $uid !== null && array_key_exists( $uid, $docDp['ids'] ) ) { |
466 | $uid = null; |
467 | $env->log( 'info', 'Wikitext for this page has duplicate ids: ' . $origId ); |
468 | } |
469 | if ( $uid === '' ) { |
470 | $uid = null; |
471 | $env->log( 'info', 'Bogus empty id' ); |
472 | } |
473 | if ( $uid === null ) { |
474 | do { |
475 | $docDp['counter'] += 1; |
476 | // PORT-FIXME: NOTE that we aren't updating the idIndex here because |
477 | // we are generating unique ids that will not conflict. In any case, |
478 | // the idIndex is a workaround for the PHP DOM's issues and we might |
479 | // switch out of this in the future anyway. |
480 | $uid = 'mw' . PHPUtils::counterToBase64( $docDp['counter'] ); |
481 | } while ( isset( $idIndex[$uid] ) ); |
482 | self::addNormalizedAttribute( $node, 'id', $uid, $origId ); |
483 | } |
484 | // Convert from DataParsoid/DataMw objects to associative array |
485 | $docDp['ids'][$uid] = $codec->toJsonArray( $data->parsoid, $hints['data-parsoid'] ); |
486 | if ( isset( $data->mw ) ) { |
487 | $pb->mw['ids'][$uid] = $codec->toJsonArray( $data->mw, $hints['data-mw'] ); |
488 | } |
489 | } |
490 | |
491 | /** |
492 | * Helper function to create static Hint objects for JsonCodec. |
493 | * @return array<Hint> |
494 | */ |
495 | public static function getCodecHints(): array { |
496 | static $hints = null; |
497 | if ( $hints === null ) { |
498 | $hints = [ |
499 | 'data-parsoid' => Hint::build( DataParsoid::class, Hint::ALLOW_OBJECT ), |
500 | 'data-mw' => Hint::build( DataMw::class, Hint::ALLOW_OBJECT ), |
501 | ]; |
502 | } |
503 | return $hints; |
504 | } |
505 | |
506 | /** |
507 | * @param Document $doc doc |
508 | * @param PageBundle $pb object |
509 | */ |
510 | public static function injectPageBundle( Document $doc, PageBundle $pb ): void { |
511 | $script = DOMUtils::appendToHead( $doc, 'script', [ |
512 | 'id' => 'mw-pagebundle', |
513 | 'type' => 'application/x-mw-pagebundle', |
514 | ] ); |
515 | $script->appendChild( $doc->createTextNode( $pb->encodeForHeadElement() ) ); |
516 | } |
517 | |
518 | /** |
519 | * @param Document $doc doc |
520 | * @return stdClass|null |
521 | */ |
522 | public static function extractPageBundle( Document $doc ): ?stdClass { |
523 | $pb = null; |
524 | $dpScriptElt = DOMCompat::getElementById( $doc, 'mw-pagebundle' ); |
525 | if ( $dpScriptElt ) { |
526 | $dpScriptElt->parentNode->removeChild( $dpScriptElt ); |
527 | // we actually want arrays in the page bundle rather than stdClasses; but we still |
528 | // want to access the object properties |
529 | $pb = (object)PHPUtils::jsonDecode( $dpScriptElt->textContent ); |
530 | } |
531 | return $pb; |
532 | } |
533 | |
534 | /** |
535 | * Walk DOM from node downward calling loadDataAttribs |
536 | * |
537 | * @param Node $node node |
538 | * @param array $options options |
539 | */ |
540 | public static function visitAndLoadDataAttribs( Node $node, array $options = [] ): void { |
541 | DOMUtils::visitDOM( $node, [ self::class, 'loadDataAttribs' ], $options ); |
542 | } |
543 | |
544 | /** |
545 | * These are intended be used on a document after post-processing, so that |
546 | * the underlying .dataobject is transparently applied (in the store case) |
547 | * and reloaded (in the load case), rather than worrying about keeping |
548 | * the attributes up-to-date throughout that phase. For the most part, |
549 | * using this.ppTo* should be sufficient and using these directly should be |
550 | * avoided. |
551 | * |
552 | * @param Node $node node |
553 | * @param array $options options |
554 | */ |
555 | public static function loadDataAttribs( Node $node, array $options ): void { |
556 | if ( !( $node instanceof Element ) ) { |
557 | return; |
558 | } |
559 | // Reset the node data object's stored state, since we're reloading it |
560 | self::setNodeData( $node, new NodeData ); |
561 | $codec = self::getCodec( $node->ownerDocument ); |
562 | $dataParsoidAttr = DOMCompat::getAttribute( $node, 'data-parsoid' ); |
563 | $dp = $codec->newFromJsonString( |
564 | $dataParsoidAttr ?? '{}', self::getCodecHints()['data-parsoid'] |
565 | ); |
566 | if ( !empty( $options['markNew'] ) ) { |
567 | $dp->setTempFlag( TempData::IS_NEW, $dataParsoidAttr === null ); |
568 | } |
569 | self::setDataParsoid( $node, $dp ); |
570 | $node->removeAttribute( 'data-parsoid' ); |
571 | |
572 | $dataMwAttr = DOMCompat::getAttribute( $node, 'data-mw' ); |
573 | $dmw = $dataMwAttr === null ? null : |
574 | $codec->newFromJsonString( $dataMwAttr, self::getCodecHints()['data-mw'] ); |
575 | self::setDataMw( $node, $dmw ); |
576 | $node->removeAttribute( 'data-mw' ); |
577 | |
578 | $dpd = self::getJSONAttribute( $node, 'data-parsoid-diff', null ); |
579 | self::setDataParsoidDiff( $node, $dpd ); |
580 | $node->removeAttribute( 'data-parsoid-diff' ); |
581 | $dataI18n = DOMCompat::getAttribute( $node, 'data-mw-i18n' ); |
582 | if ( $dataI18n !== null ) { |
583 | $i18n = DataMwI18n::fromJson( PHPUtils::jsonDecode( $dataI18n, true ) ); |
584 | self::setDataMwI18n( $node, $i18n ); |
585 | $node->removeAttribute( 'data-mw-i18n' ); |
586 | } |
587 | } |
588 | |
589 | /** |
590 | * Builds an index of id attributes seen in the DOM |
591 | * @param Node $node |
592 | * @return array |
593 | */ |
594 | public static function usedIdIndex( Node $node ): array { |
595 | $index = []; |
596 | DOMUtils::visitDOM( DOMCompat::getBody( $node->ownerDocument ), |
597 | static function ( Node $n, ?array $options = null ) use ( &$index ) { |
598 | if ( $n instanceof Element ) { |
599 | $id = DOMCompat::getAttribute( $n, 'id' ); |
600 | if ( $id !== null ) { |
601 | $index[$id] = true; |
602 | } |
603 | } |
604 | }, |
605 | [] |
606 | ); |
607 | return $index; |
608 | } |
609 | |
610 | /** |
611 | * Walk DOM from node downward calling storeDataAttribs |
612 | * |
613 | * @param Node $node node |
614 | * @param array $options options |
615 | */ |
616 | public static function visitAndStoreDataAttribs( Node $node, array $options = [] ): void { |
617 | // PORT-FIXME: storeDataAttribs calls storeInPageBundle which calls getElementById. |
618 | // PHP's `getElementById` implementation is broken, and we work around that by |
619 | // using Zest which uses XPath. So, getElementById call can be O(n) and calling it |
620 | // on on every element of the DOM via vistDOM here makes it O(n^2) instead of O(n). |
621 | // So, we work around that by building an index and avoiding getElementById entirely |
622 | // in storeInPageBundle. |
623 | if ( !empty( $options['storeInPageBundle'] ) ) { |
624 | $options['idIndex'] = self::usedIdIndex( $node ); |
625 | } |
626 | DOMUtils::visitDOM( $node, [ self::class, 'storeDataAttribs' ], $options ); |
627 | } |
628 | |
629 | /** |
630 | * Copy data attributes from the bag to either JSON-encoded attributes on |
631 | * each node, or the page bundle, erasing the data-object-id attributes. |
632 | * |
633 | * @param Node $node node |
634 | * @param ?array $options options |
635 | * - discardDataParsoid: Discard DataParsoid objects instead of storing them |
636 | * - keepTmp: Preserve DataParsoid::$tmp |
637 | * - storeInPageBundle: If true, data will be stored in the page bundle |
638 | * instead of data-parsoid and data-mw. |
639 | * - env: The Env object required for various features |
640 | * - idIndex: Array of used ID attributes |
641 | */ |
642 | public static function storeDataAttribs( Node $node, ?array $options = null ): void { |
643 | $hints = self::getCodecHints(); |
644 | $options ??= []; |
645 | if ( !( $node instanceof Element ) ) { |
646 | return; |
647 | } |
648 | Assert::invariant( empty( $options['discardDataParsoid'] ) || empty( $options['keepTmp'] ), |
649 | 'Conflicting options: discardDataParsoid and keepTmp are both enabled.' ); |
650 | $codec = self::getCodec( $node->ownerDocument ); |
651 | $dp = self::getDataParsoid( $node ); |
652 | $discardDataParsoid = !empty( $options['discardDataParsoid'] ); |
653 | if ( $dp->getTempFlag( TempData::IS_NEW ) ) { |
654 | // Only necessary to support the cite extension's getById, |
655 | // that's already been loaded once. |
656 | // |
657 | // This is basically a hack to ensure that DOMUtils.isNewElt |
658 | // continues to work since we effectively rely on the absence |
659 | // of data-parsoid to identify new elements. But, loadDataAttribs |
660 | // creates an empty {} if one doesn't exist. So, this hack |
661 | // ensures that a loadDataAttribs + storeDataAttribs pair don't |
662 | // dirty the node by introducing an empty data-parsoid attribute |
663 | // where one didn't exist before. |
664 | // |
665 | // Ideally, we'll find a better solution for this edge case later. |
666 | $discardDataParsoid = true; |
667 | } |
668 | $data = null; |
669 | if ( !$discardDataParsoid ) { |
670 | if ( empty( $options['keepTmp'] ) ) { |
671 | // @phan-suppress-next-line PhanTypeObjectUnsetDeclaredProperty |
672 | unset( $dp->tmp ); |
673 | } |
674 | |
675 | if ( !empty( $options['storeInPageBundle'] ) ) { |
676 | $data ??= new stdClass; |
677 | $data->parsoid = $dp; |
678 | } else { |
679 | $node->setAttribute( |
680 | 'data-parsoid', |
681 | PHPUtils::jsonEncode( |
682 | $codec->toJsonArray( $dp, $hints['data-parsoid'] ) |
683 | ) |
684 | ); |
685 | } |
686 | } |
687 | |
688 | // Strip invalid data-mw attributes |
689 | if ( self::validDataMw( $node ) ) { |
690 | if ( |
691 | !empty( $options['storeInPageBundle'] ) && isset( $options['env'] ) && |
692 | // The pagebundle didn't have data-mw before 999.x |
693 | Semver::satisfies( $options['env']->getOutputContentVersion(), '^999.0.0' ) |
694 | ) { |
695 | $data ??= new stdClass; |
696 | $data->mw = self::getDataMw( $node ); |
697 | } else { |
698 | $node->setAttribute( |
699 | 'data-mw', |
700 | PHPUtils::jsonEncode( |
701 | $codec->toJsonArray( self::getDataMw( $node ), $hints['data-mw'] ) |
702 | ) |
703 | ); |
704 | } |
705 | } |
706 | |
707 | if ( self::validDataMwI18n( $node ) ) { |
708 | self::setJSONAttribute( $node, 'data-mw-i18n', self::getDataMwI18n( $node ) ); |
709 | } |
710 | |
711 | // Store pagebundle |
712 | if ( $data !== null ) { |
713 | self::storeInPageBundle( $node, $options['env'], $data, $options['idIndex'] ); |
714 | } |
715 | |
716 | // Indicate that this node's data has been stored so that if we try |
717 | // to access it after the fact we're aware and remove the attribute |
718 | // since it's no longer needed. |
719 | $nd = self::getNodeData( $node ); |
720 | $id = DOMCompat::getAttribute( $node, self::DATA_OBJECT_ATTR_NAME ); |
721 | $nd->storedId = $id !== null ? intval( $id ) : null; |
722 | $node->removeAttribute( self::DATA_OBJECT_ATTR_NAME ); |
723 | } |
724 | |
725 | /** |
726 | * Clones a node and its data bag |
727 | * @param Element $elt |
728 | * @param bool $deep |
729 | * @return Element |
730 | */ |
731 | public static function cloneNode( Element $elt, bool $deep ): Element { |
732 | $clone = $elt->cloneNode( $deep ); |
733 | '@phan-var Element $clone'; // @var Element $clone |
734 | // We do not need to worry about $deep because a shallow clone does not have child nodes, |
735 | // so it's always cloning data on the cloned tree (which may be empty). |
736 | self::fixClonedData( $clone ); |
737 | return $clone; |
738 | } |
739 | |
740 | /** |
741 | * Recursively fixes cloned data from $elt: to avoid conflicts of element IDs, we clone the |
742 | * data and set it in the node with a new element ID (which setNodeData does). |
743 | * @param Element $elt |
744 | */ |
745 | private static function fixClonedData( Element $elt ) { |
746 | if ( $elt->hasAttribute( self::DATA_OBJECT_ATTR_NAME ) ) { |
747 | self::setNodeData( $elt, self::getNodeData( $elt )->cloneNodeData() ); |
748 | } |
749 | foreach ( $elt->childNodes as $child ) { |
750 | if ( $child instanceof Element ) { |
751 | self::fixClonedData( $child ); |
752 | } |
753 | } |
754 | } |
755 | } |