|
1
|
|
|
<?php |
|
2
|
|
|
namespace InlineStyle; |
|
3
|
|
|
|
|
4
|
|
|
/* |
|
5
|
|
|
* InlineStyle MIT License |
|
6
|
|
|
* |
|
7
|
|
|
* Copyright (c) 2012 Christiaan Baartse |
|
8
|
|
|
* |
|
9
|
|
|
* Permission is hereby granted, free of charge, to any person obtaining |
|
10
|
|
|
* a copy of this software and associated documentation files (the |
|
11
|
|
|
* "Software"), to deal in the Software without restriction, including |
|
12
|
|
|
* without limitation the rights to use, copy, modify, merge, publish, |
|
13
|
|
|
* distribute, sublicense, and/or sell copies of the Software, and to |
|
14
|
|
|
* permit persons to whom the Software is furnished to do so, subject to |
|
15
|
|
|
* the following conditions: |
|
16
|
|
|
* |
|
17
|
|
|
* The above copyright notice and this permission notice shall be |
|
18
|
|
|
* included in all copies or substantial portions of the Software. |
|
19
|
|
|
* |
|
20
|
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, |
|
21
|
|
|
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF |
|
22
|
|
|
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND |
|
23
|
|
|
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE |
|
24
|
|
|
* LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION |
|
25
|
|
|
* OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION |
|
26
|
|
|
* WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. |
|
27
|
|
|
*/ |
|
28
|
|
|
|
|
29
|
|
|
use Symfony\Component\CssSelector\Exception\ParseException; |
|
30
|
|
|
|
|
31
|
|
|
/** |
|
32
|
|
|
* Parses a html file and applies all embedded and external stylesheets inline |
|
33
|
|
|
*/ |
|
34
|
|
|
class InlineStyle |
|
35
|
|
|
{ |
|
36
|
|
|
/** |
|
37
|
|
|
* @var \DOMDocument the HTML as DOMDocument |
|
38
|
|
|
*/ |
|
39
|
|
|
private $_dom; |
|
40
|
|
|
|
|
41
|
|
|
/** |
|
42
|
|
|
* Prepare all the necessary objects |
|
43
|
|
|
* |
|
44
|
|
|
* @param string $html |
|
45
|
|
|
*/ |
|
46
|
|
|
public function __construct($html = '') |
|
47
|
|
|
{ |
|
48
|
|
|
if ($html) { |
|
49
|
|
|
if ($html instanceof \DOMDocument) { |
|
50
|
|
|
$this->loadDomDocument(clone $html); |
|
51
|
|
|
} else if (strlen($html) <= PHP_MAXPATHLEN && file_exists($html)) { |
|
52
|
|
|
$this->loadHTMLFile($html); |
|
53
|
|
|
} else { |
|
54
|
|
|
$this->loadHTML($html); |
|
55
|
|
|
} |
|
56
|
|
|
} |
|
57
|
|
|
} |
|
58
|
|
|
|
|
59
|
|
|
/** |
|
60
|
|
|
* Load HTML file |
|
61
|
|
|
* |
|
62
|
|
|
* @param string $filename |
|
63
|
|
|
*/ |
|
64
|
|
|
public function loadHTMLFile($filename) |
|
65
|
|
|
{ |
|
66
|
|
|
$this->loadHTML(file_get_contents($filename)); |
|
67
|
|
|
} |
|
68
|
|
|
|
|
69
|
|
|
/** |
|
70
|
|
|
* Load HTML string (UTF-8 encoding assumed) |
|
71
|
|
|
* |
|
72
|
|
|
* @param string $html |
|
73
|
|
|
*/ |
|
74
|
|
|
public function loadHTML($html) |
|
75
|
|
|
{ |
|
76
|
|
|
$dom = new \DOMDocument(); |
|
77
|
|
|
$dom->formatOutput = true; |
|
78
|
|
|
|
|
79
|
|
|
// strip illegal XML UTF-8 chars |
|
80
|
|
|
// remove all control characters except CR, LF and tab |
|
81
|
|
|
$html = preg_replace('/[\x00-\x08\x0B\x0C\x0E-\x1F\x7F]+/u', '', $html); // 00-09, 11-31, 127 |
|
82
|
|
|
|
|
83
|
|
|
$dom->loadHTML($html); |
|
84
|
|
|
$this->loadDomDocument($dom); |
|
85
|
|
|
} |
|
86
|
|
|
|
|
87
|
|
|
/** |
|
88
|
|
|
* Load the HTML as a DOMDocument directly |
|
89
|
|
|
* |
|
90
|
|
|
* @param \DOMDocument $domDocument |
|
91
|
|
|
*/ |
|
92
|
|
|
public function loadDomDocument(\DOMDocument $domDocument) |
|
93
|
|
|
{ |
|
94
|
|
|
$this->_dom = $domDocument; |
|
95
|
|
|
foreach ($this->_getNodesForCssSelector('[style]') as $node) { |
|
96
|
|
|
$node->setAttribute( |
|
97
|
|
|
'inlinestyle-original-style', |
|
98
|
|
|
$node->getAttribute('style') |
|
99
|
|
|
); |
|
100
|
|
|
} |
|
101
|
|
|
} |
|
102
|
|
|
|
|
103
|
|
|
/** |
|
104
|
|
|
* Applies one or more stylesheets to the current document |
|
105
|
|
|
* |
|
106
|
|
|
* @param string|string[] $stylesheet |
|
107
|
|
|
* @return InlineStyle self |
|
108
|
|
|
*/ |
|
109
|
|
|
public function applyStylesheet($stylesheet) |
|
110
|
|
|
{ |
|
111
|
|
|
$stylesheet = (array) $stylesheet; |
|
112
|
|
|
foreach($stylesheet as $ss) { |
|
113
|
|
|
$parsed = $this->parseStylesheet($ss); |
|
114
|
|
|
$parsed = $this->sortSelectorsOnSpecificity($parsed); |
|
115
|
|
|
foreach($parsed as $arr) { |
|
116
|
|
|
list($selector, $style) = $arr; |
|
117
|
|
|
$this->applyRule($selector, $style); |
|
118
|
|
|
} |
|
119
|
|
|
} |
|
120
|
|
|
|
|
121
|
|
|
return $this; |
|
122
|
|
|
} |
|
123
|
|
|
|
|
124
|
|
|
/** |
|
125
|
|
|
* @param string $sel Css Selector |
|
126
|
|
|
* @return array|\DOMNodeList|\DOMElement[] |
|
127
|
|
|
*/ |
|
128
|
|
|
private function _getNodesForCssSelector($sel) |
|
129
|
|
|
{ |
|
130
|
|
|
try { |
|
131
|
|
|
if (class_exists('Symfony\Component\CssSelector\CssSelectorConverter')) { |
|
132
|
|
|
$converter = new \Symfony\Component\CssSelector\CssSelectorConverter(true); |
|
133
|
|
|
$xpathQuery = $converter->toXPath($sel); |
|
134
|
|
|
} else { |
|
135
|
|
|
$xpathQuery = \Symfony\Component\CssSelector\CssSelector::toXPath($sel); |
|
136
|
|
|
} |
|
137
|
|
|
|
|
138
|
|
|
return $this->_getDomXpath()->query($xpathQuery); |
|
139
|
|
|
} |
|
140
|
|
|
catch(ParseException $e) { |
|
141
|
|
|
// ignore css rule parse exceptions |
|
142
|
|
|
} |
|
143
|
|
|
|
|
144
|
|
|
return array(); |
|
145
|
|
|
} |
|
146
|
|
|
|
|
147
|
|
|
/** |
|
148
|
|
|
* Applies a style rule on the document |
|
149
|
|
|
* @param string $selector |
|
150
|
|
|
* @param string $style |
|
151
|
|
|
* @return InlineStyle self |
|
152
|
|
|
*/ |
|
153
|
|
|
public function applyRule($selector, $style) |
|
154
|
|
|
{ |
|
155
|
|
|
if($selector) { |
|
156
|
|
|
$nodes = $this->_getNodesForCssSelector($selector); |
|
157
|
|
|
$style = $this->_styleToArray($style); |
|
158
|
|
|
|
|
159
|
|
|
foreach($nodes as $node) { |
|
160
|
|
|
$current = $node->hasAttribute("style") ? |
|
161
|
|
|
$this->_styleToArray($node->getAttribute("style")) : |
|
162
|
|
|
array(); |
|
163
|
|
|
|
|
164
|
|
|
$current = $this->_mergeStyles($current, $style); |
|
165
|
|
|
|
|
166
|
|
|
$node->setAttribute("style", $this->_arrayToStyle($current)); |
|
167
|
|
|
} |
|
168
|
|
|
} |
|
169
|
|
|
|
|
170
|
|
|
return $this; |
|
171
|
|
|
} |
|
172
|
|
|
|
|
173
|
|
|
/** |
|
174
|
|
|
* Returns the DOMDocument as html |
|
175
|
|
|
* |
|
176
|
|
|
* @return string the HTML |
|
177
|
|
|
*/ |
|
178
|
|
|
public function getHTML() |
|
179
|
|
|
{ |
|
180
|
|
|
$clone = clone $this; |
|
181
|
|
|
foreach ($clone->_getNodesForCssSelector('[inlinestyle-original-style]') as $node) { |
|
182
|
|
|
$current = $node->hasAttribute("style") ? |
|
183
|
|
|
$this->_styleToArray($node->getAttribute("style")) : |
|
184
|
|
|
array(); |
|
185
|
|
|
$original = $node->hasAttribute("inlinestyle-original-style") ? |
|
186
|
|
|
$this->_styleToArray($node->getAttribute("inlinestyle-original-style")) : |
|
187
|
|
|
array(); |
|
188
|
|
|
|
|
189
|
|
|
$current = $clone->_mergeStyles($current, $original); |
|
190
|
|
|
|
|
191
|
|
|
$node->setAttribute("style", $this->_arrayToStyle($current)); |
|
192
|
|
|
$node->removeAttribute('inlinestyle-original-style'); |
|
193
|
|
|
} |
|
194
|
|
|
|
|
195
|
|
|
return $clone->_dom->saveHTML(); |
|
196
|
|
|
} |
|
197
|
|
|
|
|
198
|
|
|
/** |
|
199
|
|
|
* Recursively extracts the stylesheet nodes from the DOMNode |
|
200
|
|
|
* |
|
201
|
|
|
* This cannot be done with XPath or a CSS selector because the order in |
|
202
|
|
|
* which the elements are found matters |
|
203
|
|
|
* |
|
204
|
|
|
* @param \DOMNode $node leave empty to extract from the whole document |
|
205
|
|
|
* @param string $base The base URI for relative stylesheets |
|
206
|
|
|
* @param array $devices Considered devices |
|
207
|
|
|
* @param boolean $remove Should it remove the original stylesheets |
|
208
|
|
|
* @return array the extracted stylesheets |
|
209
|
|
|
*/ |
|
210
|
|
|
public function extractStylesheets($node = null, $base = '', $devices = array('all', 'screen', 'handheld'), $remove = true) |
|
211
|
|
|
{ |
|
212
|
|
|
if(null === $node) { |
|
213
|
|
|
$node = $this->_dom; |
|
214
|
|
|
} |
|
215
|
|
|
|
|
216
|
|
|
$stylesheets = array(); |
|
217
|
|
|
if($node->hasChildNodes()) { |
|
218
|
|
|
$removeQueue = array(); |
|
219
|
|
|
/** @var $child \DOMElement */ |
|
220
|
|
|
foreach($node->childNodes as $child) { |
|
221
|
|
|
$nodeName = strtolower($child->nodeName); |
|
222
|
|
|
if($nodeName === "style" && $this->isForAllowedMediaDevice($child->getAttribute('media'), $devices)) { |
|
223
|
|
|
$stylesheets[] = $child->nodeValue; |
|
224
|
|
|
$removeQueue[] = $child; |
|
225
|
|
|
} else if($nodeName === "link" && strtolower($child->getAttribute('rel')) === 'stylesheet' && $this->isForAllowedMediaDevice($child->getAttribute('media'), $devices)) { |
|
226
|
|
|
if($child->hasAttribute("href")) { |
|
227
|
|
|
$href = $child->getAttribute("href"); |
|
228
|
|
|
|
|
229
|
|
|
if($base && false === strpos($href, "://")) { |
|
230
|
|
|
$href = "{$base}/{$href}"; |
|
231
|
|
|
} |
|
232
|
|
|
|
|
233
|
|
|
$ext = @file_get_contents($href); |
|
234
|
|
|
|
|
235
|
|
|
if($ext) { |
|
236
|
|
|
$removeQueue[] = $child; |
|
237
|
|
|
$stylesheets[] = $ext; |
|
238
|
|
|
} |
|
239
|
|
|
} |
|
240
|
|
|
} else { |
|
241
|
|
|
$stylesheets = array_merge($stylesheets, |
|
242
|
|
|
$this->extractStylesheets($child, $base, $devices, $remove)); |
|
243
|
|
|
} |
|
244
|
|
|
} |
|
245
|
|
|
|
|
246
|
|
|
if ($remove) { |
|
247
|
|
|
foreach ($removeQueue as $child) { |
|
248
|
|
|
$child->parentNode->removeChild($child); |
|
249
|
|
|
} |
|
250
|
|
|
} |
|
251
|
|
|
} |
|
252
|
|
|
|
|
253
|
|
|
return $stylesheets; |
|
254
|
|
|
} |
|
255
|
|
|
|
|
256
|
|
|
/** |
|
257
|
|
|
* Extracts the stylesheet nodes nodes specified by the xpath |
|
258
|
|
|
* |
|
259
|
|
|
* @param string $xpathQuery xpath query to the desired stylesheet |
|
260
|
|
|
* @return array the extracted stylesheets |
|
261
|
|
|
*/ |
|
262
|
|
|
public function extractStylesheetsWithXpath($xpathQuery) |
|
263
|
|
|
{ |
|
264
|
|
|
$stylesheets = array(); |
|
265
|
|
|
|
|
266
|
|
|
$nodes = $this->_getDomXpath()->query($xpathQuery); |
|
267
|
|
|
foreach ($nodes as $node) |
|
268
|
|
|
{ |
|
269
|
|
|
$stylesheets[] = $node->nodeValue; |
|
270
|
|
|
$node->parentNode->removeChild($node); |
|
271
|
|
|
} |
|
272
|
|
|
|
|
273
|
|
|
return $stylesheets; |
|
274
|
|
|
} |
|
275
|
|
|
|
|
276
|
|
|
/** |
|
277
|
|
|
* Parses a stylesheet to selectors and properties |
|
278
|
|
|
* @param string $stylesheet |
|
279
|
|
|
* @return array |
|
280
|
|
|
*/ |
|
281
|
|
|
public function parseStylesheet($stylesheet) |
|
282
|
|
|
{ |
|
283
|
|
|
$parsed = array(); |
|
284
|
|
|
$stylesheet = $this->_stripStylesheet($stylesheet); |
|
285
|
|
|
$stylesheet = trim(trim($stylesheet), "}"); |
|
286
|
|
|
foreach(explode("}", $stylesheet) as $rule) { |
|
287
|
|
|
//Don't parse empty rules |
|
288
|
|
|
if(!trim($rule))continue; |
|
289
|
|
|
list($selector, $style) = explode("{", $rule, 2); |
|
290
|
|
|
foreach (explode(',', $selector) as $sel) { |
|
291
|
|
|
$parsed[] = array(trim($sel), trim(trim($style), ";")); |
|
292
|
|
|
} |
|
293
|
|
|
} |
|
294
|
|
|
|
|
295
|
|
|
return $parsed; |
|
296
|
|
|
} |
|
297
|
|
|
|
|
298
|
|
|
public function sortSelectorsOnSpecificity($parsed) |
|
299
|
|
|
{ |
|
300
|
|
|
usort($parsed, array($this, 'sortOnSpecificity')); |
|
301
|
|
|
return $parsed; |
|
302
|
|
|
} |
|
303
|
|
|
|
|
304
|
|
|
private function sortOnSpecificity($a, $b) |
|
305
|
|
|
{ |
|
306
|
|
|
$a = $this->getScoreForSelector($a[0]); |
|
307
|
|
|
$b = $this->getScoreForSelector($b[0]); |
|
308
|
|
|
|
|
309
|
|
|
foreach (range(0, 2) as $i) { |
|
310
|
|
|
if ($a[$i] !== $b[$i]) { |
|
311
|
|
|
return $a[$i] < $b[$i] ? -1 : 1; |
|
312
|
|
|
} |
|
313
|
|
|
} |
|
314
|
|
|
return -1; |
|
315
|
|
|
} |
|
316
|
|
|
|
|
317
|
|
|
public function getScoreForSelector($selector) |
|
318
|
|
|
{ |
|
319
|
|
|
return array( |
|
320
|
|
|
preg_match_all('/#\w/i', $selector, $result), // ID's |
|
321
|
|
|
preg_match_all('/\.\w/i', $selector, $result), // Classes |
|
322
|
|
|
preg_match_all('/^\w|\ \w|\(\w|\:[^not]/i', $selector, $result) // Tags |
|
323
|
|
|
); |
|
324
|
|
|
} |
|
325
|
|
|
|
|
326
|
|
|
/** |
|
327
|
|
|
* Parses style properties to a array which can be merged by mergeStyles() |
|
328
|
|
|
* @param string $style |
|
329
|
|
|
* @return array |
|
330
|
|
|
*/ |
|
331
|
|
|
private function _styleToArray($style) |
|
332
|
|
|
{ |
|
333
|
|
|
$styles = array(); |
|
334
|
|
|
$style = trim(trim($style), ";"); |
|
335
|
|
|
if($style) { |
|
336
|
|
|
foreach(explode(";", $style) as $props) { |
|
337
|
|
|
$props = trim(trim($props), ";"); |
|
338
|
|
|
//Don't parse empty props |
|
339
|
|
|
if(!trim($props))continue; |
|
340
|
|
|
//Only match valid CSS rules |
|
341
|
|
|
if (preg_match('#^([-a-z0-9\*]+):(.*)$#i', $props, $matches) && isset($matches[0], $matches[1], $matches[2])) { |
|
342
|
|
|
list(, $prop, $val) = $matches; |
|
343
|
|
|
$styles[$prop] = $val; |
|
344
|
|
|
} |
|
345
|
|
|
} |
|
346
|
|
|
} |
|
347
|
|
|
|
|
348
|
|
|
return $styles; |
|
349
|
|
|
} |
|
350
|
|
|
|
|
351
|
|
|
private function _arrayToStyle($array) |
|
352
|
|
|
{ |
|
353
|
|
|
$st = array(); |
|
354
|
|
|
foreach($array as $prop => $val) { |
|
355
|
|
|
$st[] = "{$prop}:{$val}"; |
|
356
|
|
|
} |
|
357
|
|
|
return \implode(';', $st); |
|
358
|
|
|
} |
|
359
|
|
|
|
|
360
|
|
|
/** |
|
361
|
|
|
* Merges two sets of style properties taking !important into account |
|
362
|
|
|
* @param array $styleA |
|
363
|
|
|
* @param array $styleB |
|
364
|
|
|
* @return array |
|
365
|
|
|
*/ |
|
366
|
|
|
private function _mergeStyles(array $styleA, array $styleB) |
|
367
|
|
|
{ |
|
368
|
|
|
foreach($styleB as $prop => $val) { |
|
369
|
|
|
if(!isset($styleA[$prop]) |
|
370
|
|
|
|| substr(str_replace(" ", "", strtolower($styleA[$prop])), -10) !== "!important") |
|
371
|
|
|
{ |
|
372
|
|
|
$styleA[$prop] = $val; |
|
373
|
|
|
} |
|
374
|
|
|
} |
|
375
|
|
|
|
|
376
|
|
|
return $styleA; |
|
377
|
|
|
} |
|
378
|
|
|
|
|
379
|
|
|
private function _stripStylesheet($s) |
|
380
|
|
|
{ |
|
381
|
|
|
// strip comments |
|
382
|
|
|
$s = preg_replace('!/\*[^*]*\*+([^/][^*]*\*+)*/!','', $s); |
|
383
|
|
|
|
|
384
|
|
|
// strip keyframes rules |
|
385
|
|
|
$s = preg_replace('/@[-|keyframes].*?\{.*?\}[ \r\n]*\}/s', '', $s); |
|
386
|
|
|
|
|
387
|
|
|
return $s; |
|
388
|
|
|
} |
|
389
|
|
|
|
|
390
|
|
|
private function _getDomXpath() |
|
391
|
|
|
{ |
|
392
|
|
|
return new \DOMXPath($this->_dom); |
|
393
|
|
|
} |
|
394
|
|
|
|
|
395
|
|
|
public function __clone() |
|
396
|
|
|
{ |
|
397
|
|
|
$this->_dom = clone $this->_dom; |
|
398
|
|
|
} |
|
399
|
|
|
|
|
400
|
|
|
private function isForAllowedMediaDevice($mediaAttribute, array $devices) |
|
401
|
|
|
{ |
|
402
|
|
|
$mediaAttribute = strtolower($mediaAttribute); |
|
403
|
|
|
$mediaDevices = explode(',', $mediaAttribute); |
|
404
|
|
|
|
|
405
|
|
|
foreach ($mediaDevices as $device) { |
|
406
|
|
|
$device = trim($device); |
|
407
|
|
|
if (!$device || in_array($device, $devices)) { |
|
408
|
|
|
return true; |
|
409
|
|
|
} |
|
410
|
|
|
} |
|
411
|
|
|
return false; |
|
412
|
|
|
} |
|
413
|
|
|
} |
|
414
|
|
|
|