Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 148 |
|
0.00% |
0 / 14 |
CRAP | |
0.00% |
0 / 1 |
| Proposals | |
0.00% |
0 / 148 |
|
0.00% |
0 / 14 |
4692 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
20 | |||
| getLinkList | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| setCollection | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getProposals | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
12 | |||
| hasBans | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| updateLinkList | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| addCollectionArticles | |
0.00% |
0 / 22 |
|
0.00% |
0 / 1 |
90 | |||
| deleteUnusedArticles | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
12 | |||
| resolveRedirects | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| getWeightedLinks | |
0.00% |
0 / 61 |
|
0.00% |
0 / 1 |
506 | |||
| getPropList | |
0.00% |
0 / 30 |
|
0.00% |
0 / 1 |
132 | |||
| searchEntry | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
12 | |||
| checkLink | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
42 | |||
| getPropCount | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Collection Extension for MediaWiki |
| 4 | * |
| 5 | * Copyright (C) 2008-2009, PediaPress GmbH |
| 6 | * |
| 7 | * This program is free software; you can redistribute it and/or modify |
| 8 | * it under the terms of the GNU General Public License as published by |
| 9 | * the Free Software Foundation; either version 2 of the License, or |
| 10 | * (at your option) any later version. |
| 11 | * |
| 12 | * This program is distributed in the hope that it will be useful, |
| 13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 15 | * GNU General Public License for more details. |
| 16 | * |
| 17 | * You should have received a copy of the GNU General Public License along |
| 18 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 19 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 20 | * http://www.gnu.org/copyleft/gpl.html |
| 21 | */ |
| 22 | |
| 23 | namespace MediaWiki\Extension\Collection; |
| 24 | |
| 25 | use Article; |
| 26 | use MediaWiki\Title\Title; |
| 27 | use TextContent; |
| 28 | |
| 29 | /** |
| 30 | * it needs 3 Lists: |
| 31 | * - one with the bookmembers |
| 32 | * - one where it can save the banned articles |
| 33 | * - one where it can save the proposals |
| 34 | * |
| 35 | * the proposallist can be accessed with the public method getProposals() |
| 36 | * |
| 37 | * a list with the bookarticles as first and information about the outgoing |
| 38 | * links of that article as second dimension can be accessed with the method |
| 39 | * getLinkList() |
| 40 | * |
| 41 | * the Class can only sort the proposals, if it can access the function compareProps |
| 42 | */ |
| 43 | class Proposals { |
| 44 | |
| 45 | /** |
| 46 | * @var array |
| 47 | */ |
| 48 | private $mColl; |
| 49 | |
| 50 | /** |
| 51 | * @var array[] |
| 52 | */ |
| 53 | private $mPropList; |
| 54 | |
| 55 | /** |
| 56 | * @var array[] |
| 57 | */ |
| 58 | private $mLinkList; |
| 59 | |
| 60 | /** |
| 61 | * @var string[] |
| 62 | */ |
| 63 | private $mBanList; |
| 64 | |
| 65 | /** |
| 66 | * @param array $coll the collection |
| 67 | * @param string[] $ban the list of the banned articles |
| 68 | * @param array[] $props the list of the proposals |
| 69 | */ |
| 70 | public function __construct( $coll, $ban, $props ) { |
| 71 | $this->mPropList = []; |
| 72 | $this->mColl = is_array( $coll ) ? $coll : []; |
| 73 | $this->mBanList = is_array( $ban ) ? $ban : []; |
| 74 | $this->mLinkList = is_array( $props ) ? $props : []; |
| 75 | } |
| 76 | |
| 77 | /** |
| 78 | * @return array[] |
| 79 | */ |
| 80 | public function getLinkList() { |
| 81 | return $this->mLinkList; |
| 82 | } |
| 83 | |
| 84 | /** |
| 85 | * @param array $collection |
| 86 | */ |
| 87 | public function setCollection( array $collection ) { |
| 88 | $this->mColl = $collection; |
| 89 | } |
| 90 | |
| 91 | /** |
| 92 | * Calculate the new proposals and return it |
| 93 | * |
| 94 | * @param int $num number of proposals to be returned |
| 95 | * 0 or less means, that all proposals will be returned |
| 96 | * this parameter is optional, the method will return |
| 97 | * all proposals by defaulted |
| 98 | * @param bool $doUpdate when true, $linkList will |
| 99 | * updated before calculating the proposals |
| 100 | * default is true |
| 101 | * @return array[] a 2-dimensional array that contains the proposals |
| 102 | * the first dimesion is numeric, the second contains |
| 103 | * 3 entries: |
| 104 | * - 'name': the name of a proposed article |
| 105 | * - 'num' : how often this artikel was linked in the |
| 106 | * bookmembers |
| 107 | * - 'val' : a value between 1 and 1.5, the higher the |
| 108 | * more this article is proposed |
| 109 | */ |
| 110 | public function getProposals( $num = 0, $doUpdate = true ) { |
| 111 | if ( $doUpdate ) { |
| 112 | $this->updateLinkList(); |
| 113 | } |
| 114 | |
| 115 | $this->getPropList(); |
| 116 | |
| 117 | if ( $num > 0 ) { |
| 118 | return array_slice( $this->mPropList, 0, $num ); |
| 119 | } else { |
| 120 | return $this->mPropList; |
| 121 | } |
| 122 | } |
| 123 | |
| 124 | /** |
| 125 | * @return bool |
| 126 | */ |
| 127 | public function hasBans() { |
| 128 | return count( $this->mBanList ) > 0; |
| 129 | } |
| 130 | |
| 131 | private function updateLinkList() { |
| 132 | $this->addCollectionArticles(); |
| 133 | $this->deleteUnusedArticles(); |
| 134 | } |
| 135 | |
| 136 | /** |
| 137 | * Check if all articles form the book are in $mLinkList |
| 138 | */ |
| 139 | private function addCollectionArticles() { |
| 140 | global $wgCollectionSuggestThreshhold; |
| 141 | |
| 142 | if ( !isset( $this->mColl['items'] ) ) { |
| 143 | return; |
| 144 | } |
| 145 | $numItems = count( $this->mColl['items'] ); |
| 146 | if ( $numItems === 0 || $numItems > $wgCollectionSuggestThreshhold ) { |
| 147 | return; |
| 148 | } |
| 149 | |
| 150 | foreach ( $this->mColl['items'] as $item ) { |
| 151 | if ( $this->searchEntry( $item['title'], $this->mLinkList ) === false |
| 152 | && $item['type'] == 'article' |
| 153 | ) { |
| 154 | $articleName = $item['title']; |
| 155 | $title = Title::makeTitleSafe( NS_MAIN, $articleName ); |
| 156 | $article = new Article( $title, $item['revision'] ); |
| 157 | |
| 158 | if ( $article === null ) { |
| 159 | continue; |
| 160 | } |
| 161 | |
| 162 | $content = $article->getPage()->getContent(); |
| 163 | $this->mLinkList[] = [ |
| 164 | 'name' => $articleName, |
| 165 | 'links' => $this->getWeightedLinks( |
| 166 | $numItems, |
| 167 | ( $content instanceof TextContent ) |
| 168 | ? $content->getText() : null |
| 169 | ), |
| 170 | ]; |
| 171 | } |
| 172 | } |
| 173 | } |
| 174 | |
| 175 | /** |
| 176 | * Delete items from $mLinkList that are not in the collection any more |
| 177 | */ |
| 178 | private function deleteUnusedArticles() { |
| 179 | $newList = []; |
| 180 | foreach ( $this->mLinkList as $item ) { |
| 181 | if ( Session::findArticle( $item['name'] ) != -1 ) { |
| 182 | $newList[] = $item; |
| 183 | } |
| 184 | } |
| 185 | $this->mLinkList = $newList; |
| 186 | } |
| 187 | |
| 188 | /** |
| 189 | * @param Title $title |
| 190 | * @return Title |
| 191 | */ |
| 192 | private function resolveRedirects( $title ) { |
| 193 | if ( !$title->isRedirect() ) { |
| 194 | return $title; |
| 195 | } |
| 196 | |
| 197 | $article = new Article( $title, 0 ); |
| 198 | return $article->getPage()->getContent()->getRedirectTarget(); |
| 199 | } |
| 200 | |
| 201 | /** |
| 202 | * Extract & count links from wikitext |
| 203 | * |
| 204 | * @param int $num_articles |
| 205 | * @param string $wikitext article text |
| 206 | * @return float[] with links and their weights |
| 207 | */ |
| 208 | private function getWeightedLinks( $num_articles, $wikitext ) { |
| 209 | global $wgCollectionSuggestCheapWeightThreshhold; |
| 210 | |
| 211 | $allLinks = []; |
| 212 | preg_match_all( |
| 213 | '/\[\[(.+?)\]\]/', |
| 214 | $wikitext, |
| 215 | $allLinks, |
| 216 | PREG_SET_ORDER |
| 217 | ); |
| 218 | |
| 219 | $linkmap = []; |
| 220 | foreach ( $allLinks as $link ) { |
| 221 | $link = $link[1]; |
| 222 | |
| 223 | if ( preg_match( '/[:#]/', $link ) ) { // skip links with ':' and '#' |
| 224 | continue; |
| 225 | } |
| 226 | |
| 227 | // handle links with a displaytitle |
| 228 | if ( preg_match( '/(.+?)\|(.+)/', $link, $matches ) ) { |
| 229 | [ , $link, $alias ] = $matches; |
| 230 | } else { |
| 231 | $alias = $link; |
| 232 | } |
| 233 | |
| 234 | // check & normalize title |
| 235 | $title = Title::makeTitleSafe( NS_MAIN, $link ); |
| 236 | if ( $title === null || !$title->exists() ) { |
| 237 | continue; |
| 238 | } |
| 239 | $resolved = $this->resolveRedirects( $title ); |
| 240 | if ( !$resolved ) { |
| 241 | continue; |
| 242 | } |
| 243 | $link = $resolved->getText(); |
| 244 | |
| 245 | if ( isset( $linkmap[$link] ) ) { |
| 246 | $linkmap[$link][$link] = true; |
| 247 | } else { |
| 248 | $linkmap[$link] = [ $link => true ]; |
| 249 | } |
| 250 | if ( $link != $alias ) { |
| 251 | if ( isset( $linkmap[$alias] ) ) { |
| 252 | $linkmap[$alias][$link] = true; |
| 253 | } else { |
| 254 | $linkmap[$alias] = [ $link => true ]; |
| 255 | } |
| 256 | } |
| 257 | } |
| 258 | |
| 259 | $linkcount = []; |
| 260 | if ( $num_articles < $wgCollectionSuggestCheapWeightThreshhold ) { |
| 261 | // more expensive algorithm: count words |
| 262 | foreach ( $linkmap as $alias => $linked ) { |
| 263 | $matches = []; |
| 264 | preg_match_all( |
| 265 | '/\W' . preg_quote( $alias, '/' ) . '\W/i', |
| 266 | $wikitext, |
| 267 | $matches |
| 268 | ); |
| 269 | $num = count( $matches[0] ); |
| 270 | |
| 271 | foreach ( $linked as $link => $dummy ) { |
| 272 | if ( isset( $linkcount[$link] ) ) { |
| 273 | $linkcount[$link] += $num; |
| 274 | } else { |
| 275 | $linkcount[$link] = $num; |
| 276 | } |
| 277 | } |
| 278 | } |
| 279 | |
| 280 | if ( count( $linkcount ) == 0 ) { |
| 281 | return []; |
| 282 | } |
| 283 | |
| 284 | // normalize: |
| 285 | $lc_max = 0; |
| 286 | foreach ( $linkcount as $count ) { |
| 287 | if ( $count > $lc_max ) { |
| 288 | $lc_max = $count; |
| 289 | } |
| 290 | } |
| 291 | $norm = log( $lc_max ); |
| 292 | $result = []; |
| 293 | if ( $norm > 0 ) { |
| 294 | foreach ( $linkcount as $link => $count ) { |
| 295 | $result[$link] = 1 + 0.5 * log( $count ) / $norm; |
| 296 | } |
| 297 | } else { |
| 298 | foreach ( $linkcount as $link => $count ) { |
| 299 | $result[$link] = 1; |
| 300 | } |
| 301 | } |
| 302 | |
| 303 | return $result; |
| 304 | } else { |
| 305 | // cheaper algorithm: just count links |
| 306 | foreach ( $linkmap as $linked ) { |
| 307 | foreach ( $linked as $link => $dummy ) { |
| 308 | $linkcount[$link] = 1; |
| 309 | } |
| 310 | } |
| 311 | |
| 312 | return $linkcount; |
| 313 | } |
| 314 | } |
| 315 | |
| 316 | /** |
| 317 | * Calculate the $mPropList from $mLinkList and $mBanList |
| 318 | */ |
| 319 | private function getPropList() { |
| 320 | $prop = []; |
| 321 | foreach ( $this->mLinkList as $article ) { |
| 322 | foreach ( $article['links'] as $linkName => $val ) { |
| 323 | if ( !$this->checkLink( $linkName ) ) { |
| 324 | continue; |
| 325 | } |
| 326 | $key = $this->searchEntry( $linkName, $prop ); |
| 327 | if ( $key !== false ) { |
| 328 | $prop[$key]['val'] += $val; |
| 329 | } else { |
| 330 | $prop[] = [ |
| 331 | 'name' => $linkName, |
| 332 | 'val' => $val, |
| 333 | ]; |
| 334 | } |
| 335 | } |
| 336 | } |
| 337 | usort( |
| 338 | $prop, |
| 339 | static function ( $a, $b ) { |
| 340 | if ( $a['val'] == $b['val'] ) { |
| 341 | return strcmp( $a['name'], $b['name'] ); |
| 342 | } |
| 343 | if ( $a['val'] < $b['val'] ) { |
| 344 | return 1; |
| 345 | } else { |
| 346 | return -1; |
| 347 | } |
| 348 | } |
| 349 | ); |
| 350 | $this->mPropList = []; |
| 351 | $have_real_weights = false; |
| 352 | foreach ( $prop as $p ) { |
| 353 | if ( $p['val'] > 1 ) { |
| 354 | $have_real_weights = true; |
| 355 | } |
| 356 | if ( $p['val'] <= 1 && $have_real_weights ) { |
| 357 | break; |
| 358 | } |
| 359 | $this->mPropList[] = $p; |
| 360 | } |
| 361 | } |
| 362 | |
| 363 | /** |
| 364 | * Search an article in an array and returns its key or false |
| 365 | * if the array doesn't contain the article |
| 366 | * |
| 367 | * @param string $entry an articlename |
| 368 | * @param array[] $array to be searched, it has to 2-dimensional |
| 369 | * the 2nd dimension needs the key 'name' |
| 370 | * @return bool|int the key as integer or false |
| 371 | */ |
| 372 | private function searchEntry( $entry, $array ) { |
| 373 | for ( $i = 0, $count = count( $array ); $i < $count; $i++ ) { |
| 374 | if ( $array[$i]['name'] == $entry ) { |
| 375 | return $i; |
| 376 | } |
| 377 | } |
| 378 | return false; |
| 379 | } |
| 380 | |
| 381 | /** |
| 382 | * Check if an article is banned or belongs to the book/collection |
| 383 | * |
| 384 | * @param string $link an articlename |
| 385 | * @return bool true: if the article can be added to the proposals |
| 386 | * false: if the article can't be added to the proposals |
| 387 | */ |
| 388 | private function checkLink( $link ) { |
| 389 | foreach ( $this->mColl['items'] as $item ) { |
| 390 | if ( $item['type'] == 'article' && $item['title'] == $link ) { |
| 391 | return false; |
| 392 | } |
| 393 | } |
| 394 | |
| 395 | if ( $this->hasBans() && in_array( $link, $this->mBanList ) ) { |
| 396 | return false; |
| 397 | } |
| 398 | |
| 399 | return true; |
| 400 | } |
| 401 | |
| 402 | /** |
| 403 | * @return int |
| 404 | */ |
| 405 | private function getPropCount() { |
| 406 | return count( $this->mPropList ); |
| 407 | } |
| 408 | } |