aboutsummaryrefslogtreecommitdiffhomepage
path: root/inc/3rdparty/libraries/content-extractor
diff options
context:
space:
mode:
authortcitworld <thomas.citharet@gmail.com>2014-01-04 12:30:31 -0800
committertcitworld <thomas.citharet@gmail.com>2014-01-04 12:30:31 -0800
commit7f667839764621b5aa01c9db8ce5dde2a29ef18f (patch)
tree93d8241ee81c87e18494325ae02f0589a8e328a2 /inc/3rdparty/libraries/content-extractor
parenta84f77d6ba15a64ff00453f5d5190c021ce460ed (diff)
parent2abcccb37180c17318f5226f5d4bc28f30b621ea (diff)
downloadwallabag-7f667839764621b5aa01c9db8ce5dde2a29ef18f.tar.gz
wallabag-7f667839764621b5aa01c9db8ce5dde2a29ef18f.tar.zst
wallabag-7f667839764621b5aa01c9db8ce5dde2a29ef18f.zip
Merge pull request #1 from inthepoche/dev
Dev
Diffstat (limited to 'inc/3rdparty/libraries/content-extractor')
-rw-r--r--inc/3rdparty/libraries/content-extractor/ContentExtractor.php728
-rw-r--r--inc/3rdparty/libraries/content-extractor/SiteConfig.php338
2 files changed, 1066 insertions, 0 deletions
diff --git a/inc/3rdparty/libraries/content-extractor/ContentExtractor.php b/inc/3rdparty/libraries/content-extractor/ContentExtractor.php
new file mode 100644
index 00000000..ddd33bb5
--- /dev/null
+++ b/inc/3rdparty/libraries/content-extractor/ContentExtractor.php
@@ -0,0 +1,728 @@
1<?php
2/**
3 * Content Extractor
4 *
5 * Uses patterns specified in site config files and auto detection (hNews/PHP Readability)
6 * to extract content from HTML files.
7 *
8 * @version 1.0
9 * @date 2013-02-05
10 * @author Keyvan Minoukadeh
11 * @copyright 2013 Keyvan Minoukadeh
12 * @license http://www.gnu.org/licenses/agpl-3.0.html AGPL v3
13 */
14
15class ContentExtractor
16{
17 protected static $tidy_config = array(
18 'clean' => true,
19 'output-xhtml' => true,
20 'logical-emphasis' => true,
21 'show-body-only' => false,
22 'new-blocklevel-tags' => 'article, aside, footer, header, hgroup, menu, nav, section, details, datagrid',
23 'new-inline-tags' => 'mark, time, meter, progress, data',
24 'wrap' => 0,
25 'drop-empty-paras' => true,
26 'drop-proprietary-attributes' => false,
27 'enclose-text' => true,
28 'enclose-block-text' => true,
29 'merge-divs' => true,
30 'merge-spans' => true,
31 'char-encoding' => 'utf8',
32 'hide-comments' => true
33 );
34 protected $html;
35 protected $config;
36 protected $title;
37 protected $author = array();
38 protected $language;
39 protected $date;
40 protected $body;
41 protected $success = false;
42 protected $nextPageUrl;
43 public $allowedParsers = array('libxml', 'html5lib');
44 public $fingerprints = array();
45 public $readability;
46 public $debug = false;
47 public $debugVerbose = false;
48
49 function __construct($path, $fallback=null) {
50 SiteConfig::set_config_path($path, $fallback);
51 }
52
53 protected function debug($msg) {
54 if ($this->debug) {
55 $mem = round(memory_get_usage()/1024, 2);
56 $memPeak = round(memory_get_peak_usage()/1024, 2);
57 echo '* ',$msg;
58 if ($this->debugVerbose) echo ' - mem used: ',$mem," (peak: $memPeak)";
59 echo "\n";
60 ob_flush();
61 flush();
62 }
63 }
64
65 public function reset() {
66 $this->html = null;
67 $this->readability = null;
68 $this->config = null;
69 $this->title = null;
70 $this->body = null;
71 $this->author = array();
72 $this->language = null;
73 $this->date = null;
74 $this->nextPageUrl = null;
75 $this->success = false;
76 }
77
78 public function findHostUsingFingerprints($html) {
79 $this->debug('Checking fingerprints...');
80 $head = substr($html, 0, 8000);
81 foreach ($this->fingerprints as $_fp => $_fphost) {
82 $lookin = 'html';
83 if (is_array($_fphost)) {
84 if (isset($_fphost['head']) && $_fphost['head']) {
85 $lookin = 'head';
86 }
87 $_fphost = $_fphost['hostname'];
88 }
89 if (strpos($$lookin, $_fp) !== false) {
90 $this->debug("Found match: $_fphost");
91 return $_fphost;
92 }
93 }
94 $this->debug('No fingerprint matches');
95 return false;
96 }
97
98 // returns SiteConfig instance (joined in order: exact match, wildcard, fingerprint, global, default)
99 public function buildSiteConfig($url, $html='', $add_to_cache=true) {
100 // extract host name
101 $host = @parse_url($url, PHP_URL_HOST);
102 $host = strtolower($host);
103 if (substr($host, 0, 4) == 'www.') $host = substr($host, 4);
104 // is merged version already cached?
105 if (SiteConfig::is_cached("$host.merged")) {
106 $this->debug("Returning cached and merged site config for $host");
107 return SiteConfig::build("$host.merged");
108 }
109 // let's build from site_config/custom/ and standard/
110 $config = SiteConfig::build($host);
111 if ($add_to_cache && $config && !SiteConfig::is_cached("$host")) {
112 SiteConfig::add_to_cache($host, $config);
113 }
114 // if no match, use defaults
115 if (!$config) $config = new SiteConfig();
116 // load fingerprint config?
117 if ($config->autodetect_on_failure()) {
118 // check HTML for fingerprints
119 if (!empty($this->fingerprints) && ($_fphost = $this->findHostUsingFingerprints($html))) {
120 if ($config_fingerprint = SiteConfig::build($_fphost)) {
121 $this->debug("Appending site config settings from $_fphost (fingerprint match)");
122 $config->append($config_fingerprint);
123 if ($add_to_cache && !SiteConfig::is_cached($_fphost)) {
124 //$config_fingerprint->cache_in_apc = true;
125 SiteConfig::add_to_cache($_fphost, $config_fingerprint);
126 }
127 }
128 }
129 }
130 // load global config?
131 if ($config->autodetect_on_failure()) {
132 if ($config_global = SiteConfig::build('global', true)) {
133 $this->debug('Appending site config settings from global.txt');
134 $config->append($config_global);
135 if ($add_to_cache && !SiteConfig::is_cached('global')) {
136 //$config_global->cache_in_apc = true;
137 SiteConfig::add_to_cache('global', $config_global);
138 }
139 }
140 }
141 // store copy of merged config
142 if ($add_to_cache) {
143 // do not store in APC if wildcard match
144 $use_apc = ($host == $config->cache_key);
145 $config->cache_key = null;
146 SiteConfig::add_to_cache("$host.merged", $config, $use_apc);
147 }
148 return $config;
149 }
150
151 // returns true on success, false on failure
152 // $smart_tidy indicates that if tidy is used and no results are produced, we will
153 // try again without it. Tidy helps us deal with PHP's patchy HTML parsing most of the time
154 // but it has problems of its own which we try to avoid with this option.
155 public function process($html, $url, $smart_tidy=true) {
156 $this->reset();
157 $this->config = $this->buildSiteConfig($url, $html);
158
159 // do string replacements
160 if (!empty($this->config->find_string)) {
161 if (count($this->config->find_string) == count($this->config->replace_string)) {
162 $html = str_replace($this->config->find_string, $this->config->replace_string, $html, $_count);
163 $this->debug("Strings replaced: $_count (find_string and/or replace_string)");
164 } else {
165 $this->debug('Skipped string replacement - incorrect number of find-replace strings in site config');
166 }
167 unset($_count);
168 }
169
170 // use tidy (if it exists)?
171 // This fixes problems with some sites which would otherwise
172 // trouble DOMDocument's HTML parsing. (Although sometimes it
173 // makes matters worse, which is why you can override it in site config files.)
174 $tidied = false;
175 if ($this->config->tidy() && function_exists('tidy_parse_string') && $smart_tidy) {
176 $this->debug('Using Tidy');
177 $tidy = tidy_parse_string($html, self::$tidy_config, 'UTF8');
178 if (tidy_clean_repair($tidy)) {
179 $original_html = $html;
180 $tidied = true;
181 $html = $tidy->value;
182 }
183 unset($tidy);
184 }
185
186 // load and parse html
187 $_parser = $this->config->parser();
188 if (!in_array($_parser, $this->allowedParsers)) {
189 $this->debug("HTML parser $_parser not listed, using libxml instead");
190 $_parser = 'libxml';
191 }
192 $this->debug("Attempting to parse HTML with $_parser");
193 $this->readability = new Readability($html, $url, $_parser);
194
195 // we use xpath to find elements in the given HTML document
196 // see http://en.wikipedia.org/wiki/XPath_1.0
197 $xpath = new DOMXPath($this->readability->dom);
198
199 // try to get next page link
200 foreach ($this->config->next_page_link as $pattern) {
201 $elems = @$xpath->evaluate($pattern, $this->readability->dom);
202 if (is_string($elems)) {
203 $this->nextPageUrl = trim($elems);
204 break;
205 } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
206 foreach ($elems as $item) {
207 if ($item instanceof DOMElement && $item->hasAttribute('href')) {
208 $this->nextPageUrl = $item->getAttribute('href');
209 break 2;
210 } elseif ($item instanceof DOMAttr && $item->value) {
211 $this->nextPageUrl = $item->value;
212 break 2;
213 }
214 }
215 }
216 }
217
218 // try to get title
219 foreach ($this->config->title as $pattern) {
220 // $this->debug("Trying $pattern");
221 $elems = @$xpath->evaluate($pattern, $this->readability->dom);
222 if (is_string($elems)) {
223 $this->title = trim($elems);
224 $this->debug('Title expression evaluated as string: '.$this->title);
225 $this->debug("...XPath match: $pattern");
226 break;
227 } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
228 $this->title = $elems->item(0)->textContent;
229 $this->debug('Title matched: '.$this->title);
230 $this->debug("...XPath match: $pattern");
231 // remove title from document
232 try {
233 $elems->item(0)->parentNode->removeChild($elems->item(0));
234 } catch (DOMException $e) {
235 // do nothing
236 }
237 break;
238 }
239 }
240
241 // try to get author (if it hasn't already been set)
242 if (empty($this->author)) {
243 foreach ($this->config->author as $pattern) {
244 $elems = @$xpath->evaluate($pattern, $this->readability->dom);
245 if (is_string($elems)) {
246 if (trim($elems) != '') {
247 $this->author[] = trim($elems);
248 $this->debug('Author expression evaluated as string: '.trim($elems));
249 $this->debug("...XPath match: $pattern");
250 break;
251 }
252 } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
253 foreach ($elems as $elem) {
254 if (!isset($elem->parentNode)) continue;
255 $this->author[] = trim($elem->textContent);
256 $this->debug('Author matched: '.trim($elem->textContent));
257 }
258 if (!empty($this->author)) {
259 $this->debug("...XPath match: $pattern");
260 break;
261 }
262 }
263 }
264 }
265
266 // try to get language
267 $_lang_xpath = array('//html[@lang]/@lang', '//meta[@name="DC.language"]/@content');
268 foreach ($_lang_xpath as $pattern) {
269 $elems = @$xpath->evaluate($pattern, $this->readability->dom);
270 if (is_string($elems)) {
271 if (trim($elems) != '') {
272 $this->language = trim($elems);
273 $this->debug('Language matched: '.$this->language);
274 break;
275 }
276 } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
277 foreach ($elems as $elem) {
278 if (!isset($elem->parentNode)) continue;
279 $this->language = trim($elem->textContent);
280 $this->debug('Language matched: '.$this->language);
281 }
282 if ($this->language) break;
283 }
284 }
285
286 // try to get date
287 foreach ($this->config->date as $pattern) {
288 $elems = @$xpath->evaluate($pattern, $this->readability->dom);
289 if (is_string($elems)) {
290 $this->date = strtotime(trim($elems, "; \t\n\r\0\x0B"));
291 } elseif ($elems instanceof DOMNodeList && $elems->length > 0) {
292 $this->date = $elems->item(0)->textContent;
293 $this->date = strtotime(trim($this->date, "; \t\n\r\0\x0B"));
294 // remove date from document
295 // $elems->item(0)->parentNode->removeChild($elems->item(0));
296 }
297 if (!$this->date) {
298 $this->date = null;
299 } else {
300 $this->debug('Date matched: '.date('Y-m-d H:i:s', $this->date));
301 $this->debug("...XPath match: $pattern");
302 break;
303 }
304 }
305
306 // strip elements (using xpath expressions)
307 foreach ($this->config->strip as $pattern) {
308 $elems = @$xpath->query($pattern, $this->readability->dom);
309 // check for matches
310 if ($elems && $elems->length > 0) {
311 $this->debug('Stripping '.$elems->length.' elements (strip)');
312 for ($i=$elems->length-1; $i >= 0; $i--) {
313 $elems->item($i)->parentNode->removeChild($elems->item($i));
314 }
315 }
316 }
317
318 // strip elements (using id and class attribute values)
319 foreach ($this->config->strip_id_or_class as $string) {
320 $string = strtr($string, array("'"=>'', '"'=>''));
321 $elems = @$xpath->query("//*[contains(@class, '$string') or contains(@id, '$string')]", $this->readability->dom);
322 // check for matches
323 if ($elems && $elems->length > 0) {
324 $this->debug('Stripping '.$elems->length.' elements (strip_id_or_class)');
325 for ($i=$elems->length-1; $i >= 0; $i--) {
326 $elems->item($i)->parentNode->removeChild($elems->item($i));
327 }
328 }
329 }
330
331 // strip images (using src attribute values)
332 foreach ($this->config->strip_image_src as $string) {
333 $string = strtr($string, array("'"=>'', '"'=>''));
334 $elems = @$xpath->query("//img[contains(@src, '$string')]", $this->readability->dom);
335 // check for matches
336 if ($elems && $elems->length > 0) {
337 $this->debug('Stripping '.$elems->length.' image elements');
338 for ($i=$elems->length-1; $i >= 0; $i--) {
339 $elems->item($i)->parentNode->removeChild($elems->item($i));
340 }
341 }
342 }
343 // strip elements using Readability.com and Instapaper.com ignore class names
344 // .entry-unrelated and .instapaper_ignore
345 // See https://www.readability.com/publishers/guidelines/#view-plainGuidelines
346 // and http://blog.instapaper.com/post/730281947
347 $elems = @$xpath->query("//*[contains(concat(' ',normalize-space(@class),' '),' entry-unrelated ') or contains(concat(' ',normalize-space(@class),' '),' instapaper_ignore ')]", $this->readability->dom);
348 // check for matches
349 if ($elems && $elems->length > 0) {
350 $this->debug('Stripping '.$elems->length.' .entry-unrelated,.instapaper_ignore elements');
351 for ($i=$elems->length-1; $i >= 0; $i--) {
352 $elems->item($i)->parentNode->removeChild($elems->item($i));
353 }
354 }
355
356 // strip elements that contain style="display: none;"
357 $elems = @$xpath->query("//*[contains(@style,'display:none')]", $this->readability->dom);
358 // check for matches
359 if ($elems && $elems->length > 0) {
360 $this->debug('Stripping '.$elems->length.' elements with inline display:none style');
361 for ($i=$elems->length-1; $i >= 0; $i--) {
362 $elems->item($i)->parentNode->removeChild($elems->item($i));
363 }
364 }
365
366 // try to get body
367 foreach ($this->config->body as $pattern) {
368 $elems = @$xpath->query($pattern, $this->readability->dom);
369 // check for matches
370 if ($elems && $elems->length > 0) {
371 $this->debug('Body matched');
372 $this->debug("...XPath match: $pattern");
373 if ($elems->length == 1) {
374 $this->body = $elems->item(0);
375 // prune (clean up elements that may not be content)
376 if ($this->config->prune()) {
377 $this->debug('...pruning content');
378 $this->readability->prepArticle($this->body);
379 }
380 break;
381 } else {
382 $this->body = $this->readability->dom->createElement('div');
383 $this->debug($elems->length.' body elems found');
384 foreach ($elems as $elem) {
385 if (!isset($elem->parentNode)) continue;
386 $isDescendant = false;
387 foreach ($this->body->childNodes as $parent) {
388 if ($this->isDescendant($parent, $elem)) {
389 $isDescendant = true;
390 break;
391 }
392 }
393 if ($isDescendant) {
394 $this->debug('...element is child of another body element, skipping.');
395 } else {
396 // prune (clean up elements that may not be content)
397 if ($this->config->prune()) {
398 $this->debug('Pruning content');
399 $this->readability->prepArticle($elem);
400 }
401 $this->debug('...element added to body');
402 $this->body->appendChild($elem);
403 }
404 }
405 if ($this->body->hasChildNodes()) break;
406 }
407 }
408 }
409
410 // auto detect?
411 $detect_title = $detect_body = $detect_author = $detect_date = false;
412 // detect title?
413 if (!isset($this->title)) {
414 if (empty($this->config->title) || $this->config->autodetect_on_failure()) {
415 $detect_title = true;
416 }
417 }
418 // detect body?
419 if (!isset($this->body)) {
420 if (empty($this->config->body) || $this->config->autodetect_on_failure()) {
421 $detect_body = true;
422 }
423 }
424 // detect author?
425 if (empty($this->author)) {
426 if (empty($this->config->author) || $this->config->autodetect_on_failure()) {
427 $detect_author = true;
428 }
429 }
430 // detect date?
431 if (!isset($this->date)) {
432 if (empty($this->config->date) || $this->config->autodetect_on_failure()) {
433 $detect_date = true;
434 }
435 }
436
437 // check for hNews
438 if ($detect_title || $detect_body) {
439 // check for hentry
440 $elems = @$xpath->query("//*[contains(concat(' ',normalize-space(@class),' '),' hentry ')]", $this->readability->dom);
441 if ($elems && $elems->length > 0) {
442 $this->debug('hNews: found hentry');
443 $hentry = $elems->item(0);
444
445 if ($detect_title) {
446 // check for entry-title
447 $elems = @$xpath->query(".//*[contains(concat(' ',normalize-space(@class),' '),' entry-title ')]", $hentry);
448 if ($elems && $elems->length > 0) {
449 $this->title = $elems->item(0)->textContent;
450 $this->debug('hNews: found entry-title: '.$this->title);
451 // remove title from document
452 $elems->item(0)->parentNode->removeChild($elems->item(0));
453 $detect_title = false;
454 }
455 }
456
457 if ($detect_date) {
458 // check for time element with pubdate attribute
459 $elems = @$xpath->query(".//time[@pubdate] | .//abbr[contains(concat(' ',normalize-space(@class),' '),' published ')]", $hentry);
460 if ($elems && $elems->length > 0) {
461 $this->date = strtotime(trim($elems->item(0)->textContent));
462 // remove date from document
463 //$elems->item(0)->parentNode->removeChild($elems->item(0));
464 if ($this->date) {
465 $this->debug('hNews: found publication date: '.date('Y-m-d H:i:s', $this->date));
466 $detect_date = false;
467 } else {
468 $this->date = null;
469 }
470 }
471 }
472
473 if ($detect_author) {
474 // check for time element with pubdate attribute
475 $elems = @$xpath->query(".//*[contains(concat(' ',normalize-space(@class),' '),' vcard ') and (contains(concat(' ',normalize-space(@class),' '),' author ') or contains(concat(' ',normalize-space(@class),' '),' byline '))]", $hentry);
476 if ($elems && $elems->length > 0) {
477 $author = $elems->item(0);
478 $fn = @$xpath->query(".//*[contains(concat(' ',normalize-space(@class),' '),' fn ')]", $author);
479 if ($fn && $fn->length > 0) {
480 foreach ($fn as $_fn) {
481 if (trim($_fn->textContent) != '') {
482 $this->author[] = trim($_fn->textContent);
483 $this->debug('hNews: found author: '.trim($_fn->textContent));
484 }
485 }
486 } else {
487 if (trim($author->textContent) != '') {
488 $this->author[] = trim($author->textContent);
489 $this->debug('hNews: found author: '.trim($author->textContent));
490 }
491 }
492 $detect_author = empty($this->author);
493 }
494 }
495
496 // check for entry-content.
497 // according to hAtom spec, if there are multiple elements marked entry-content,
498 // we include all of these in the order they appear - see http://microformats.org/wiki/hatom#Entry_Content
499 if ($detect_body) {
500 $elems = @$xpath->query(".//*[contains(concat(' ',normalize-space(@class),' '),' entry-content ')]", $hentry);
501 if ($elems && $elems->length > 0) {
502 $this->debug('hNews: found entry-content');
503 if ($elems->length == 1) {
504 // what if it's empty? (some sites misuse hNews - place their content outside an empty entry-content element)
505 $e = $elems->item(0);
506 if (($e->tagName == 'img') || (trim($e->textContent) != '')) {
507 $this->body = $elems->item(0);
508 // prune (clean up elements that may not be content)
509 if ($this->config->prune()) {
510 $this->debug('Pruning content');
511 $this->readability->prepArticle($this->body);
512 }
513 $detect_body = false;
514 } else {
515 $this->debug('hNews: skipping entry-content - appears not to contain content');
516 }
517 unset($e);
518 } else {
519 $this->body = $this->readability->dom->createElement('div');
520 $this->debug($elems->length.' entry-content elems found');
521 foreach ($elems as $elem) {
522 if (!isset($elem->parentNode)) continue;
523 $isDescendant = false;
524 foreach ($this->body->childNodes as $parent) {
525 if ($this->isDescendant($parent, $elem)) {
526 $isDescendant = true;
527 break;
528 }
529 }
530 if ($isDescendant) {
531 $this->debug('Element is child of another body element, skipping.');
532 } else {
533 // prune (clean up elements that may not be content)
534 if ($this->config->prune()) {
535 $this->debug('Pruning content');
536 $this->readability->prepArticle($elem);
537 }
538 $this->debug('Element added to body');
539 $this->body->appendChild($elem);
540 }
541 }
542 $detect_body = false;
543 }
544 }
545 }
546 }
547 }
548
549 // check for elements marked with instapaper_title
550 if ($detect_title) {
551 // check for instapaper_title
552 $elems = @$xpath->query("//*[contains(concat(' ',normalize-space(@class),' '),' instapaper_title ')]", $this->readability->dom);
553 if ($elems && $elems->length > 0) {
554 $this->title = $elems->item(0)->textContent;
555 $this->debug('Title found (.instapaper_title): '.$this->title);
556 // remove title from document
557 $elems->item(0)->parentNode->removeChild($elems->item(0));
558 $detect_title = false;
559 }
560 }
561 // check for elements marked with instapaper_body
562 if ($detect_body) {
563 $elems = @$xpath->query("//*[contains(concat(' ',normalize-space(@class),' '),' instapaper_body ')]", $this->readability->dom);
564 if ($elems && $elems->length > 0) {
565 $this->debug('body found (.instapaper_body)');
566 $this->body = $elems->item(0);
567 // prune (clean up elements that may not be content)
568 if ($this->config->prune()) {
569 $this->debug('Pruning content');
570 $this->readability->prepArticle($this->body);
571 }
572 $detect_body = false;
573 }
574 }
575
576 // Find author in rel="author" marked element
577 // We only use this if there's exactly one.
578 // If there's more than one, it could indicate more than
579 // one author, but it could also indicate that we're processing
580 // a page listing different articles with different authors.
581 if ($detect_author) {
582 $elems = @$xpath->query("//a[contains(concat(' ',normalize-space(@rel),' '),' author ')]", $this->readability->dom);
583 if ($elems && $elems->length == 1) {
584 $author = trim($elems->item(0)->textContent);
585 if ($author != '') {
586 $this->debug("Author found (rel=\"author\"): $author");
587 $this->author[] = $author;
588 $detect_author = false;
589 }
590 }
591 }
592
593 // Find date in pubdate marked time element
594 // For the same reason given above, we only use this
595 // if there's exactly one element.
596 if ($detect_date) {
597 $elems = @$xpath->query("//time[@pubdate]", $this->readability->dom);
598 if ($elems && $elems->length == 1) {
599 $this->date = strtotime(trim($elems->item(0)->textContent));
600 // remove date from document
601 //$elems->item(0)->parentNode->removeChild($elems->item(0));
602 if ($this->date) {
603 $this->debug('Date found (pubdate marked time element): '.date('Y-m-d H:i:s', $this->date));
604 $detect_date = false;
605 } else {
606 $this->date = null;
607 }
608 }
609 }
610
611 // still missing title or body, so we detect using Readability
612 if ($detect_title || $detect_body) {
613 $this->debug('Using Readability');
614 // clone body if we're only using Readability for title (otherwise it may interfere with body element)
615 if (isset($this->body)) $this->body = $this->body->cloneNode(true);
616 $success = $this->readability->init();
617 }
618 if ($detect_title) {
619 $this->debug('Detecting title');
620 $this->title = $this->readability->getTitle()->textContent;
621 }
622 if ($detect_body && $success) {
623 $this->debug('Detecting body');
624 $this->body = $this->readability->getContent();
625 if ($this->body->childNodes->length == 1 && $this->body->firstChild->nodeType === XML_ELEMENT_NODE) {
626 $this->body = $this->body->firstChild;
627 }
628 // prune (clean up elements that may not be content)
629 if ($this->config->prune()) {
630 $this->debug('Pruning content');
631 $this->readability->prepArticle($this->body);
632 }
633 }
634 if (isset($this->body)) {
635 // remove scripts
636 $this->readability->removeScripts($this->body);
637 // remove any h1-h6 elements that appear as first thing in the body
638 // and which match our title
639 if (isset($this->title) && ($this->title != '')) {
640 $firstChild = $this->body->firstChild;
641 while ($firstChild->nodeType && ($firstChild->nodeType !== XML_ELEMENT_NODE)) {
642 $firstChild = $firstChild->nextSibling;
643 }
644 if (($firstChild->nodeType === XML_ELEMENT_NODE)
645 && in_array(strtolower($firstChild->tagName), array('h1', 'h2', 'h3', 'h4', 'h5', 'h6'))
646 && (strtolower(trim($firstChild->textContent)) == strtolower(trim($this->title)))) {
647 $this->body->removeChild($firstChild);
648 }
649 }
650 // prevent self-closing iframes
651 $elems = $this->body->getElementsByTagName('iframe');
652 for ($i = $elems->length-1; $i >= 0; $i--) {
653 $e = $elems->item($i);
654 if (!$e->hasChildNodes()) {
655 $e->appendChild($this->body->ownerDocument->createTextNode('[embedded content]'));
656 }
657 }
658 // remove image lazy loading - WordPress plugin http://wordpress.org/extend/plugins/lazy-load/
659 // the plugin replaces the src attribute to point to a 1x1 gif and puts the original src
660 // inside the data-lazy-src attribute. It also places the original image inside a noscript element
661 // next to the amended one.
662 $elems = @$xpath->query("//img[@data-lazy-src]", $this->body);
663 for ($i = $elems->length-1; $i >= 0; $i--) {
664 $e = $elems->item($i);
665 // let's see if we can grab image from noscript
666 if ($e->nextSibling !== null && $e->nextSibling->nodeName === 'noscript') {
667 $_new_elem = $e->ownerDocument->createDocumentFragment();
668 @$_new_elem->appendXML($e->nextSibling->innerHTML);
669 $e->nextSibling->parentNode->replaceChild($_new_elem, $e->nextSibling);
670 $e->parentNode->removeChild($e);
671 } else {
672 // Use data-lazy-src as src value
673 $e->setAttribute('src', $e->getAttribute('data-lazy-src'));
674 $e->removeAttribute('data-lazy-src');
675 }
676 }
677
678 $this->success = true;
679 }
680
681 // if we've had no success and we've used tidy, there's a chance
682 // that tidy has messed up. So let's try again without tidy...
683 if (!$this->success && $tidied && $smart_tidy) {
684 $this->debug('Trying again without tidy');
685 $this->process($original_html, $url, false);
686 }
687
688 return $this->success;
689 }
690
691 private function isDescendant(DOMElement $parent, DOMElement $child) {
692 $node = $child->parentNode;
693 while ($node != null) {
694 if ($node->isSameNode($parent)) return true;
695 $node = $node->parentNode;
696 }
697 return false;
698 }
699
700 public function getContent() {
701 return $this->body;
702 }
703
704 public function getTitle() {
705 return $this->title;
706 }
707
708 public function getAuthors() {
709 return $this->author;
710 }
711
712 public function getLanguage() {
713 return $this->language;
714 }
715
716 public function getDate() {
717 return $this->date;
718 }
719
720 public function getSiteConfig() {
721 return $this->config;
722 }
723
724 public function getNextPageUrl() {
725 return $this->nextPageUrl;
726 }
727}
728?> \ No newline at end of file
diff --git a/inc/3rdparty/libraries/content-extractor/SiteConfig.php b/inc/3rdparty/libraries/content-extractor/SiteConfig.php
new file mode 100644
index 00000000..c5e300d7
--- /dev/null
+++ b/inc/3rdparty/libraries/content-extractor/SiteConfig.php
@@ -0,0 +1,338 @@
1<?php
2/**
3 * Site Config
4 *
5 * Each instance of this class should hold extraction patterns and other directives
6 * for a website. See ContentExtractor class to see how it's used.
7 *
8 * @version 0.7
9 * @date 2012-08-27
10 * @author Keyvan Minoukadeh
11 * @copyright 2012 Keyvan Minoukadeh
12 * @license http://www.gnu.org/licenses/agpl-3.0.html AGPL v3
13 */
14
15class SiteConfig
16{
17 // Use first matching element as title (0 or more xpath expressions)
18 public $title = array();
19
20 // Use first matching element as body (0 or more xpath expressions)
21 public $body = array();
22
23 // Use first matching element as author (0 or more xpath expressions)
24 public $author = array();
25
26 // Use first matching element as date (0 or more xpath expressions)
27 public $date = array();
28
29 // Strip elements matching these xpath expressions (0 or more)
30 public $strip = array();
31
32 // Strip elements which contain these strings (0 or more) in the id or class attribute
33 public $strip_id_or_class = array();
34
35 // Strip images which contain these strings (0 or more) in the src attribute
36 public $strip_image_src = array();
37
38 // Additional HTTP headers to send
39 // NOT YET USED
40 public $http_header = array();
41
42 // Process HTML with tidy before creating DOM (bool or null if undeclared)
43 public $tidy = null;
44
45 protected $default_tidy = true; // used if undeclared
46
47 // Autodetect title/body if xpath expressions fail to produce results.
48 // Note that this applies to title and body separately, ie.
49 // * if we get a body match but no title match, this option will determine whether we autodetect title
50 // * if neither match, this determines whether we autodetect title and body.
51 // Also note that this only applies when there is at least one xpath expression in title or body, ie.
52 // * if title and body are both empty (no xpath expressions), this option has no effect (both title and body will be auto-detected)
53 // * if there's an xpath expression for title and none for body, body will be auto-detected and this option will determine whether we auto-detect title if the xpath expression for it fails to produce results.
54 // Usage scenario: you want to extract something specific from a set of URLs, e.g. a table, and if the table is not found, you want to ignore the entry completely. Auto-detection is unlikely to succeed here, so you construct your patterns and set this option to false. Another scenario may be a site where auto-detection has proven to fail (or worse, picked up the wrong content).
55 // bool or null if undeclared
56 public $autodetect_on_failure = null;
57 protected $default_autodetect_on_failure = true; // used if undeclared
58
59 // Clean up content block - attempt to remove elements that appear to be superfluous
60 // bool or null if undeclared
61 public $prune = null;
62 protected $default_prune = true; // used if undeclared
63
64 // Test URL - if present, can be used to test the config above
65 public $test_url = array();
66
67 // Single-page link - should identify a link element or URL pointing to the page holding the entire article
68 // This is useful for sites which split their articles across multiple pages. Links to such pages tend to
69 // display the first page with links to the other pages at the bottom. Often there is also a link to a page
70 // which displays the entire article on one page (e.g. 'print view').
71 // This should be an XPath expression identifying the link to that page. If present and we find a match,
72 // we will retrieve that page and the rest of the options in this config will be applied to the new page.
73 public $single_page_link = array();
74
75 public $next_page_link = array();
76
77 // Single-page link in feed? - same as above, but patterns applied to item description HTML taken from feed
78 public $single_page_link_in_feed = array();
79
80 // Which parser to use for turning raw HTML into a DOMDocument (either 'libxml' or 'html5lib')
81 // string or null if undeclared
82 public $parser = null;
83 protected $default_parser = 'libxml'; // used if undeclared
84
85 // Strings to search for in HTML before processing begins (used with $replace_string)
86 public $find_string = array();
87 // Strings to replace those found in $find_string before HTML processing begins
88 public $replace_string = array();
89
90 // the options below cannot be set in the config files which this class represents
91
92 //public $cache_in_apc = false; // used to decide if we should cache in apc or not
93 public $cache_key = null;
94 public static $debug = false;
95 protected static $apc = false;
96 protected static $config_path;
97 protected static $config_path_fallback;
98 protected static $config_cache = array();
99 const HOSTNAME_REGEX = '/^(([a-zA-Z0-9-]*[a-zA-Z0-9])\.)*([A-Za-z0-9-]*[A-Za-z0-9])$/';
100
101 protected static function debug($msg) {
102 if (self::$debug) {
103 //$mem = round(memory_get_usage()/1024, 2);
104 //$memPeak = round(memory_get_peak_usage()/1024, 2);
105 echo '* ',$msg;
106 //echo ' - mem used: ',$mem," (peak: $memPeak)\n";
107 echo "\n";
108 ob_flush();
109 flush();
110 }
111 }
112
113 // enable APC caching of certain site config files?
114 // If enabled the following site config files will be
115 // cached in APC cache (when requested for first time):
116 // * anything in site_config/custom/ and its corresponding file in site_config/standard/
117 // * the site config files associated with HTML fingerprints
118 // * the global site config file
119 // returns true if enabled, false otherwise
120 public static function use_apc($apc=true) {
121 if (!function_exists('apc_add')) {
122 if ($apc) self::debug('APC will not be used (function apc_add does not exist)');
123 return false;
124 }
125 self::$apc = $apc;
126 return $apc;
127 }
128
129 // return bool or null
130 public function tidy($use_default=true) {
131 if ($use_default) return (isset($this->tidy)) ? $this->tidy : $this->default_tidy;
132 return $this->tidy;
133 }
134
135 // return bool or null
136 public function prune($use_default=true) {
137 if ($use_default) return (isset($this->prune)) ? $this->prune : $this->default_prune;
138 return $this->prune;
139 }
140
141 // return string or null
142 public function parser($use_default=true) {
143 if ($use_default) return (isset($this->parser)) ? $this->parser : $this->default_parser;
144 return $this->parser;
145 }
146
147 // return bool or null
148 public function autodetect_on_failure($use_default=true) {
149 if ($use_default) return (isset($this->autodetect_on_failure)) ? $this->autodetect_on_failure : $this->default_autodetect_on_failure;
150 return $this->autodetect_on_failure;
151 }
152
153 public static function set_config_path($path, $fallback=null) {
154 self::$config_path = $path;
155 self::$config_path_fallback = $fallback;
156 }
157
158 public static function add_to_cache($key, SiteConfig $config, $use_apc=true) {
159 $key = strtolower($key);
160 if (substr($key, 0, 4) == 'www.') $key = substr($key, 4);
161 if ($config->cache_key) $key = $config->cache_key;
162 self::$config_cache[$key] = $config;
163 if (self::$apc && $use_apc) {
164 self::debug("Adding site config to APC cache with key sc.$key");
165 apc_add("sc.$key", $config);
166 }
167 self::debug("Cached site config with key $key");
168 }
169
170 public static function is_cached($key) {
171 $key = strtolower($key);
172 if (substr($key, 0, 4) == 'www.') $key = substr($key, 4);
173 if (array_key_exists($key, self::$config_cache)) {
174 return true;
175 } elseif (self::$apc && (bool)apc_fetch("sc.$key")) {
176 return true;
177 }
178 return false;
179 }
180
181 public function append(SiteConfig $newconfig) {
182 // check for commands where we accept multiple statements (no test_url)
183 foreach (array('title', 'body', 'author', 'date', 'strip', 'strip_id_or_class', 'strip_image_src', 'single_page_link', 'single_page_link_in_feed', 'next_page_link', 'http_header', 'find_string', 'replace_string') as $var) {
184 // append array elements for this config variable from $newconfig to this config
185 //$this->$var = $this->$var + $newconfig->$var;
186 $this->$var = array_unique(array_merge($this->$var, $newconfig->$var));
187 }
188 // check for single statement commands
189 // we do not overwrite existing non null values
190 foreach (array('tidy', 'prune', 'parser', 'autodetect_on_failure') as $var) {
191 if ($this->$var === null) $this->$var = $newconfig->$var;
192 }
193 }
194
195 // returns SiteConfig instance if an appropriate one is found, false otherwise
196 // if $exact_host_match is true, we will not look for wildcard config matches
197 // by default if host is 'test.example.org' we will look for and load '.example.org.txt' if it exists
198 public static function build($host, $exact_host_match=false) {
199 $host = strtolower($host);
200 if (substr($host, 0, 4) == 'www.') $host = substr($host, 4);
201 if (!$host || (strlen($host) > 200) || !preg_match(self::HOSTNAME_REGEX, ltrim($host, '.'))) return false;
202 // check for site configuration
203 $try = array($host);
204 // should we look for wildcard matches
205 if (!$exact_host_match) {
206 $split = explode('.', $host);
207 if (count($split) > 1) {
208 array_shift($split);
209 $try[] = '.'.implode('.', $split);
210 }
211 }
212
213 // look for site config file in primary folder
214 self::debug(". looking for site config for $host in primary folder");
215 foreach ($try as $h) {
216 if (array_key_exists($h, self::$config_cache)) {
217 self::debug("... site config for $h already loaded in this request");
218 return self::$config_cache[$h];
219 } elseif (self::$apc && ($sconfig = apc_fetch("sc.$h"))) {
220 self::debug("... site config for $h in APC cache");
221 return $sconfig;
222 } elseif (file_exists(self::$config_path."/$h.txt")) {
223 self::debug("... found site config ($h.txt)");
224 $file_primary = self::$config_path."/$h.txt";
225 $matched_name = $h;
226 break;
227 }
228 }
229
230 // if we found site config, process it
231 if (isset($file_primary)) {
232 $config_lines = file($file_primary, FILE_IGNORE_NEW_LINES | FILE_SKIP_EMPTY_LINES);
233 if (!$config_lines || !is_array($config_lines)) return false;
234 $config = self::build_from_array($config_lines);
235 // if APC caching is available and enabled, mark this for cache
236 //$config->cache_in_apc = true;
237 $config->cache_key = $matched_name;
238
239 // if autodetec on failure is off (on by default) we do not need to look
240 // in secondary folder
241 if (!$config->autodetect_on_failure()) {
242 self::debug('... autodetect on failure is disabled (no other site config files will be loaded)');
243 return $config;
244 }
245 }
246
247 // look for site config file in secondary folder
248 if (isset(self::$config_path_fallback)) {
249 self::debug(". looking for site config for $host in secondary folder");
250 foreach ($try as $h) {
251 if (file_exists(self::$config_path_fallback."/$h.txt")) {
252 self::debug("... found site config in secondary folder ($h.txt)");
253 $file_secondary = self::$config_path_fallback."/$h.txt";
254 $matched_name = $h;
255 break;
256 }
257 }
258 if (!isset($file_secondary)) {
259 self::debug("... no site config match in secondary folder");
260 }
261 }
262
263 // return false if no config file found
264 if (!isset($file_primary) && !isset($file_secondary)) {
265 self::debug("... no site config match for $host");
266 return false;
267 }
268
269 // return primary config if secondary not found
270 if (!isset($file_secondary) && isset($config)) {
271 return $config;
272 }
273
274 // process secondary config file
275 $config_lines = file($file_secondary, FILE_IGNORE_NEW_LINES | FILE_SKIP_EMPTY_LINES);
276 if (!$config_lines || !is_array($config_lines)) {
277 // failed to process secondary
278 if (isset($config)) {
279 // return primary config
280 return $config;
281 } else {
282 return false;
283 }
284 }
285
286 // merge with primary and return
287 if (isset($config)) {
288 self::debug('. merging config files');
289 $config->append(self::build_from_array($config_lines));
290 return $config;
291 } else {
292 // return just secondary
293 $config = self::build_from_array($config_lines);
294 // if APC caching is available and enabled, mark this for cache
295 //$config->cache_in_apc = true;
296 $config->cache_key = $matched_name;
297 return $config;
298 }
299 }
300
301 public static function build_from_array(array $lines) {
302 $config = new SiteConfig();
303 foreach ($lines as $line) {
304 $line = trim($line);
305
306 // skip comments, empty lines
307 if ($line == '' || $line[0] == '#') continue;
308
309 // get command
310 $command = explode(':', $line, 2);
311 // if there's no colon ':', skip this line
312 if (count($command) != 2) continue;
313 $val = trim($command[1]);
314 $command = trim($command[0]);
315 if ($command == '' || $val == '') continue;
316
317 // check for commands where we accept multiple statements
318 if (in_array($command, array('title', 'body', 'author', 'date', 'strip', 'strip_id_or_class', 'strip_image_src', 'single_page_link', 'single_page_link_in_feed', 'next_page_link', 'http_header', 'test_url', 'find_string', 'replace_string'))) {
319 array_push($config->$command, $val);
320 // check for single statement commands that evaluate to true or false
321 } elseif (in_array($command, array('tidy', 'prune', 'autodetect_on_failure'))) {
322 $config->$command = ($val == 'yes');
323 // check for single statement commands stored as strings
324 } elseif (in_array($command, array('parser'))) {
325 $config->$command = $val;
326 // check for replace_string(find): replace
327 } elseif ((substr($command, -1) == ')') && preg_match('!^([a-z0-9_]+)\((.*?)\)$!i', $command, $match)) {
328 if (in_array($match[1], array('replace_string'))) {
329 $command = $match[1];
330 array_push($config->find_string, $match[2]);
331 array_push($config->$command, $val);
332 }
333 }
334 }
335 return $config;
336 }
337}
338?> \ No newline at end of file
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-10 10:30:36 +0000