x-tools /
xtools
| Total Complexity | 44 |
| Total Lines | 451 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
Complex classes like ArticleInfoApi often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use ArticleInfoApi, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 21 | class ArticleInfoApi extends Model |
||
| 22 | { |
||
| 23 | protected AutomatedEditsHelper $autoEditsHelper; |
||
| 24 | protected I18nHelper $i18n; |
||
| 25 | |||
| 26 | /** @var int Number of revisions that belong to the page. */ |
||
| 27 | protected int $numRevisions; |
||
| 28 | |||
| 29 | /** @var array Prose stats, with keys 'characters', 'words', 'references', 'unique_references', 'sections'. */ |
||
| 30 | protected array $proseStats; |
||
| 31 | |||
| 32 | /** @var array Number of categories, templates and files on the page. */ |
||
| 33 | protected array $transclusionData; |
||
| 34 | |||
| 35 | /** @var array Various statistics about bots that edited the page. */ |
||
| 36 | protected array $bots; |
||
| 37 | |||
| 38 | /** @var int Number of edits made to the page by bots. */ |
||
| 39 | protected int $botRevisionCount; |
||
| 40 | |||
| 41 | /** @var int[] Number of in and outgoing links and redirects to the page. */ |
||
| 42 | protected array $linksAndRedirects; |
||
| 43 | |||
| 44 | /** @var string[] Assessments of the page (see Page::getAssessments). */ |
||
| 45 | protected array $assessments; |
||
| 46 | |||
| 47 | /** @var string[] List of Wikidata and Checkwiki errors. */ |
||
| 48 | protected array $bugs; |
||
| 49 | |||
| 50 | /** |
||
| 51 | * ArticleInfoApi constructor. |
||
| 52 | * @param ArticleInfoRepository $repository |
||
| 53 | * @param I18nHelper $i18n |
||
| 54 | * @param AutomatedEditsHelper $autoEditsHelper |
||
| 55 | * @param Page $page The page to process. |
||
| 56 | * @param false|int $start Start date as Unix timestmap. |
||
| 57 | * @param false|int $end End date as Unix timestamp. |
||
| 58 | */ |
||
| 59 | public function __construct( |
||
| 60 | ArticleInfoRepository $repository, |
||
| 61 | I18nHelper $i18n, |
||
| 62 | AutomatedEditsHelper $autoEditsHelper, |
||
| 63 | Page $page, |
||
| 64 | $start = false, |
||
| 65 | $end = false |
||
| 66 | ) { |
||
| 67 | $this->repository = $repository; |
||
| 68 | $this->i18n = $i18n; |
||
| 69 | $this->autoEditsHelper = $autoEditsHelper; |
||
| 70 | $this->page = $page; |
||
| 71 | $this->start = $start; |
||
| 72 | $this->end = $end; |
||
| 73 | } |
||
| 74 | |||
| 75 | /** |
||
| 76 | * Get the number of revisions belonging to the page. |
||
| 77 | * @return int |
||
| 78 | */ |
||
| 79 | public function getNumRevisions(): int |
||
| 80 | { |
||
| 81 | if (!isset($this->numRevisions)) { |
||
| 82 | $this->numRevisions = $this->page->getNumRevisions(null, $this->start, $this->end); |
||
|
|
|||
| 83 | } |
||
| 84 | return $this->numRevisions; |
||
| 85 | } |
||
| 86 | |||
| 87 | /** |
||
| 88 | * Are there more revisions than we should process, based on the config? |
||
| 89 | * @return bool |
||
| 90 | */ |
||
| 91 | public function tooManyRevisions(): bool |
||
| 92 | { |
||
| 93 | return $this->repository->maxPageRevisions > 0 && |
||
| 94 | $this->getNumRevisions() > $this->repository->maxPageRevisions; |
||
| 95 | } |
||
| 96 | |||
| 97 | /** |
||
| 98 | * Get various basic info used in the API, including the number of revisions, unique authors, initial author |
||
| 99 | * and edit count of the initial author. This is combined into one query for better performance. Caching is |
||
| 100 | * intentionally disabled, because using the gadget, this will get hit for a different page constantly, where |
||
| 101 | * the likelihood of cache benefiting us is slim. |
||
| 102 | * @return string[]|false false if the page was not found. |
||
| 103 | */ |
||
| 104 | public function getBasicEditingInfo() |
||
| 105 | { |
||
| 106 | return $this->repository->getBasicEditingInfo($this->page); |
||
| 107 | } |
||
| 108 | |||
| 109 | /** |
||
| 110 | * Get the top editors to the page by edit count. |
||
| 111 | * @param int $limit Default 20, maximum 1,000. |
||
| 112 | * @param bool $noBots Set to non-false to exclude bots from the result. |
||
| 113 | * @return array |
||
| 114 | */ |
||
| 115 | public function getTopEditorsByEditCount(int $limit = 20, bool $noBots = false): array |
||
| 116 | { |
||
| 117 | // Quick cache, valid only for the same request. |
||
| 118 | static $topEditors = null; |
||
| 119 | if (null !== $topEditors) { |
||
| 120 | return $topEditors; |
||
| 121 | } |
||
| 122 | |||
| 123 | $rows = $this->repository->getTopEditorsByEditCount( |
||
| 124 | $this->page, |
||
| 125 | $this->start, |
||
| 126 | $this->end, |
||
| 127 | min($limit, 1000), |
||
| 128 | $noBots |
||
| 129 | ); |
||
| 130 | |||
| 131 | $topEditors = []; |
||
| 132 | $rank = 0; |
||
| 133 | foreach ($rows as $row) { |
||
| 134 | $topEditors[] = [ |
||
| 135 | 'rank' => ++$rank, |
||
| 136 | 'username' => $row['username'], |
||
| 137 | 'count' => $row['count'], |
||
| 138 | 'minor' => $row['minor'], |
||
| 139 | 'first_edit' => [ |
||
| 140 | 'id' => $row['first_revid'], |
||
| 141 | 'timestamp' => $row['first_timestamp'], |
||
| 142 | ], |
||
| 143 | 'latest_edit' => [ |
||
| 144 | 'id' => $row['latest_revid'], |
||
| 145 | 'timestamp' => $row['latest_timestamp'], |
||
| 146 | ], |
||
| 147 | ]; |
||
| 148 | } |
||
| 149 | |||
| 150 | return $topEditors; |
||
| 151 | } |
||
| 152 | |||
| 153 | /** |
||
| 154 | * Get prose and reference information. |
||
| 155 | * @return array With keys 'characters', 'words', 'references', 'unique_references' |
||
| 156 | */ |
||
| 157 | public function getProseStats(): array |
||
| 158 | { |
||
| 159 | if (isset($this->proseStats)) { |
||
| 160 | return $this->proseStats; |
||
| 161 | } |
||
| 162 | |||
| 163 | $datetime = is_int($this->end) ? new DateTime("@$this->end") : null; |
||
| 164 | $html = $this->page->getHTMLContent($datetime); |
||
| 165 | |||
| 166 | $crawler = new Crawler($html); |
||
| 167 | |||
| 168 | [$chars, $words] = $this->countCharsAndWords($crawler, '#mw-content-text p'); |
||
| 169 | |||
| 170 | $refs = $crawler->filter('#mw-content-text .reference'); |
||
| 171 | $refContent = []; |
||
| 172 | $refs->each(function ($ref) use (&$refContent): void { |
||
| 173 | $refContent[] = $ref->text(); |
||
| 174 | }); |
||
| 175 | $uniqueRefs = count(array_unique($refContent)); |
||
| 176 | |||
| 177 | $sections = count($crawler->filter('#mw-content-text .mw-headline')); |
||
| 178 | |||
| 179 | $this->proseStats = [ |
||
| 180 | 'characters' => $chars, |
||
| 181 | 'words' => $words, |
||
| 182 | 'references' => $refs->count(), |
||
| 183 | 'unique_references' => $uniqueRefs, |
||
| 184 | 'sections' => $sections, |
||
| 185 | ]; |
||
| 186 | return $this->proseStats; |
||
| 187 | } |
||
| 188 | |||
| 189 | /** |
||
| 190 | * Count the number of characters and words of the plain text within the DOM element matched by the given selector. |
||
| 191 | * @param Crawler $crawler |
||
| 192 | * @param string $selector HTML selector. |
||
| 193 | * @return array [num chars, num words] |
||
| 194 | */ |
||
| 195 | private function countCharsAndWords(Crawler $crawler, string $selector): array |
||
| 196 | { |
||
| 197 | $totalChars = 0; |
||
| 198 | $totalWords = 0; |
||
| 199 | $paragraphs = $crawler->filter($selector); |
||
| 200 | $paragraphs->each(function ($node) use (&$totalChars, &$totalWords): void { |
||
| 201 | /** @var Crawler $node */ |
||
| 202 | $text = preg_replace('/\[\d+]/', '', trim($node->text(null, true))); |
||
| 203 | $totalChars += strlen($text); |
||
| 204 | $totalWords += count(explode(' ', $text)); |
||
| 205 | }); |
||
| 206 | |||
| 207 | return [$totalChars, $totalWords]; |
||
| 208 | } |
||
| 209 | |||
| 210 | /** |
||
| 211 | * Get the page assessments of the page. |
||
| 212 | * @see https://www.mediawiki.org/wiki/Extension:PageAssessments |
||
| 213 | * @return string[]|false False if unsupported. |
||
| 214 | * @codeCoverageIgnore |
||
| 215 | */ |
||
| 216 | public function getAssessments() |
||
| 217 | { |
||
| 218 | if (!isset($this->assessments)) { |
||
| 219 | $this->assessments = $this->page |
||
| 220 | ->getProject() |
||
| 221 | ->getPageAssessments() |
||
| 222 | ->getAssessments($this->page); |
||
| 223 | } |
||
| 224 | return $this->assessments; |
||
| 225 | } |
||
| 226 | |||
| 227 | /** |
||
| 228 | * Get the list of page's wikidata and Checkwiki errors. |
||
| 229 | * @see Page::getErrors() |
||
| 230 | * @return string[] |
||
| 231 | */ |
||
| 232 | public function getBugs(): array |
||
| 233 | { |
||
| 234 | if (!isset($this->bugs)) { |
||
| 235 | $this->bugs = $this->page->getErrors(); |
||
| 236 | } |
||
| 237 | return $this->bugs; |
||
| 238 | } |
||
| 239 | |||
| 240 | /** |
||
| 241 | * Get the number of wikidata nad CheckWiki errors. |
||
| 242 | * @return int |
||
| 243 | */ |
||
| 244 | public function numBugs(): int |
||
| 245 | { |
||
| 246 | return count($this->getBugs()); |
||
| 247 | } |
||
| 248 | |||
| 249 | /** |
||
| 250 | * Generate the data structure that will used in the ArticleInfo API response. |
||
| 251 | * @param Project $project |
||
| 252 | * @param Page $page |
||
| 253 | * @return array |
||
| 254 | * @codeCoverageIgnore |
||
| 255 | */ |
||
| 256 | public function getArticleInfoApiData(Project $project, Page $page): array |
||
| 257 | { |
||
| 258 | /** Number of days to query for pageviews */ |
||
| 259 | $pageviewsOffset = 30; |
||
| 260 | |||
| 261 | $data = [ |
||
| 262 | 'project' => $project->getDomain(), |
||
| 263 | 'page' => $page->getTitle(), |
||
| 264 | 'watchers' => (int) $page->getWatchers(), |
||
| 265 | 'pageviews' => $page->getLastPageviews($pageviewsOffset), |
||
| 266 | 'pageviews_offset' => $pageviewsOffset, |
||
| 267 | ]; |
||
| 268 | |||
| 269 | $info = false; |
||
| 270 | |||
| 271 | try { |
||
| 272 | $info = $this->repository->getBasicEditingInfo($page); |
||
| 273 | } catch (ServiceUnavailableHttpException $e) { |
||
| 274 | // No more open database connections. |
||
| 275 | $data['error'] = 'Unable to fetch revision data. Please try again later.'; |
||
| 276 | } catch (HttpException $e) { |
||
| 277 | /** |
||
| 278 | * The query most likely exceeded the maximum query time, |
||
| 279 | * so we'll abort and give only info retrieved by the API. |
||
| 280 | */ |
||
| 281 | $data['error'] = 'Unable to fetch revision data. The query may have timed out.'; |
||
| 282 | } |
||
| 283 | |||
| 284 | if (false !== $info) { |
||
| 285 | $creationDateTime = DateTime::createFromFormat('YmdHis', $info['created_at']); |
||
| 286 | $modifiedDateTime = DateTime::createFromFormat('YmdHis', $info['modified_at']); |
||
| 287 | $secsSinceLastEdit = (new DateTime)->getTimestamp() - $modifiedDateTime->getTimestamp(); |
||
| 288 | |||
| 289 | // Some wikis (such foundation.wikimedia.org) may be missing the creation date. |
||
| 290 | $creationDateTime = false === $creationDateTime |
||
| 291 | ? null |
||
| 292 | : $creationDateTime->format('Y-m-d'); |
||
| 293 | |||
| 294 | $assessment = $page->getProject() |
||
| 295 | ->getPageAssessments() |
||
| 296 | ->getAssessment($page); |
||
| 297 | |||
| 298 | $data = array_merge($data, [ |
||
| 299 | 'revisions' => (int) $info['num_edits'], |
||
| 300 | 'editors' => (int) $info['num_editors'], |
||
| 301 | 'minor_edits' => (int) $info['minor_edits'], |
||
| 302 | 'author' => $info['author'], |
||
| 303 | 'author_editcount' => null === $info['author_editcount'] ? null : (int) $info['author_editcount'], |
||
| 304 | 'created_at' => $creationDateTime, |
||
| 305 | 'created_rev_id' => $info['created_rev_id'], |
||
| 306 | 'modified_at' => $modifiedDateTime->format('Y-m-d H:i'), |
||
| 307 | 'secs_since_last_edit' => $secsSinceLastEdit, |
||
| 308 | 'last_edit_id' => (int) $info['modified_rev_id'], |
||
| 309 | 'assessment' => $assessment, |
||
| 310 | ]); |
||
| 311 | } |
||
| 312 | |||
| 313 | return $data; |
||
| 314 | } |
||
| 315 | |||
| 316 | /************************ Link statistics ************************/ |
||
| 317 | |||
| 318 | /** |
||
| 319 | * Get the number of external links on the page. |
||
| 320 | * @return int |
||
| 321 | */ |
||
| 322 | public function linksExtCount(): int |
||
| 323 | { |
||
| 324 | return $this->getLinksAndRedirects()['links_ext_count']; |
||
| 325 | } |
||
| 326 | |||
| 327 | /** |
||
| 328 | * Get the number of incoming links to the page. |
||
| 329 | * @return int |
||
| 330 | */ |
||
| 331 | public function linksInCount(): int |
||
| 332 | { |
||
| 333 | return $this->getLinksAndRedirects()['links_in_count']; |
||
| 334 | } |
||
| 335 | |||
| 336 | /** |
||
| 337 | * Get the number of outgoing links from the page. |
||
| 338 | * @return int |
||
| 339 | */ |
||
| 340 | public function linksOutCount(): int |
||
| 341 | { |
||
| 342 | return $this->getLinksAndRedirects()['links_out_count']; |
||
| 343 | } |
||
| 344 | |||
| 345 | /** |
||
| 346 | * Get the number of redirects to the page. |
||
| 347 | * @return int |
||
| 348 | */ |
||
| 349 | public function redirectsCount(): int |
||
| 350 | { |
||
| 351 | return $this->getLinksAndRedirects()['redirects_count']; |
||
| 352 | } |
||
| 353 | |||
| 354 | /** |
||
| 355 | * Get the number of external, incoming and outgoing links, along with the number of redirects to the page. |
||
| 356 | * @return int[] |
||
| 357 | * @codeCoverageIgnore |
||
| 358 | */ |
||
| 359 | private function getLinksAndRedirects(): array |
||
| 360 | { |
||
| 361 | if (!isset($this->linksAndRedirects)) { |
||
| 362 | $this->linksAndRedirects = $this->page->countLinksAndRedirects(); |
||
| 363 | } |
||
| 364 | return $this->linksAndRedirects; |
||
| 365 | } |
||
| 366 | |||
| 367 | /** |
||
| 368 | * Fetch transclusion data (categories, templates and files) that are on the page. |
||
| 369 | * @return array With keys 'categories', 'templates' and 'files'. |
||
| 370 | */ |
||
| 371 | public function getTransclusionData(): array |
||
| 372 | { |
||
| 373 | if (!isset($this->transclusionData)) { |
||
| 374 | $this->transclusionData = $this->repository->getTransclusionData($this->page); |
||
| 375 | } |
||
| 376 | return $this->transclusionData; |
||
| 377 | } |
||
| 378 | |||
| 379 | /** |
||
| 380 | * Get the number of categories that are on the page. |
||
| 381 | * @return int |
||
| 382 | */ |
||
| 383 | public function getNumCategories(): int |
||
| 384 | { |
||
| 385 | return $this->getTransclusionData()['categories']; |
||
| 386 | } |
||
| 387 | |||
| 388 | /** |
||
| 389 | * Get the number of templates that are on the page. |
||
| 390 | * @return int |
||
| 391 | */ |
||
| 392 | public function getNumTemplates(): int |
||
| 393 | { |
||
| 394 | return $this->getTransclusionData()['templates']; |
||
| 395 | } |
||
| 396 | |||
| 397 | /** |
||
| 398 | * Get the number of files that are on the page. |
||
| 399 | * @return int |
||
| 400 | */ |
||
| 401 | public function getNumFiles(): int |
||
| 402 | { |
||
| 403 | return $this->getTransclusionData()['files']; |
||
| 404 | } |
||
| 405 | |||
| 406 | /************************ Bot statistics ************************/ |
||
| 407 | |||
| 408 | /** |
||
| 409 | * Number of edits made to the page by current or former bots. |
||
| 410 | * @param string[] $bots Used only in unit tests, where we supply mock data for the bots that will get processed. |
||
| 411 | * @return int |
||
| 412 | */ |
||
| 413 | public function getBotRevisionCount(?array $bots = null): int |
||
| 414 | { |
||
| 415 | if (isset($this->botRevisionCount)) { |
||
| 416 | return $this->botRevisionCount; |
||
| 417 | } |
||
| 418 | |||
| 419 | if (null === $bots) { |
||
| 420 | $bots = $this->getBots(); |
||
| 421 | } |
||
| 422 | |||
| 423 | $count = 0; |
||
| 424 | |||
| 425 | foreach (array_values($bots) as $data) { |
||
| 426 | $count += $data['count']; |
||
| 427 | } |
||
| 428 | |||
| 429 | $this->botRevisionCount = $count; |
||
| 430 | return $count; |
||
| 431 | } |
||
| 432 | |||
| 433 | /** |
||
| 434 | * Get and set $this->bots about bots that edited the page. This is done separately from the main query because |
||
| 435 | * we use this information when computing the top 10 editors in ArticleInfo, where we don't want to include bots. |
||
| 436 | * @return array |
||
| 437 | */ |
||
| 438 | public function getBots(): array |
||
| 439 | { |
||
| 440 | if (isset($this->bots)) { |
||
| 441 | return $this->bots; |
||
| 442 | } |
||
| 443 | |||
| 444 | // Parse the bot edits. |
||
| 445 | $this->bots = []; |
||
| 446 | |||
| 447 | $limit = $this->tooManyRevisions() ? $this->repository->maxPageRevisions : null; |
||
| 448 | |||
| 449 | $botData = $this->repository->getBotData($this->page, $this->start, $this->end, $limit); |
||
| 450 | while ($bot = $botData->fetchAssociative()) { |
||
| 451 | $this->bots[$bot['username']] = [ |
||
| 452 | 'count' => (int)$bot['count'], |
||
| 453 | 'current' => '1' === $bot['current'], |
||
| 454 | ]; |
||
| 455 | } |
||
| 456 | |||
| 457 | // Sort by edit count. |
||
| 458 | uasort($this->bots, function ($a, $b) { |
||
| 459 | return $b['count'] - $a['count']; |
||
| 460 | }); |
||
| 461 | |||
| 462 | return $this->bots; |
||
| 463 | } |
||
| 464 | |||
| 465 | /** |
||
| 466 | * Get the number of bots that edited the page. |
||
| 467 | * @return int |
||
| 468 | */ |
||
| 469 | public function getNumBots(): int |
||
| 472 | } |
||
| 473 | } |
||
| 474 |
Loading history...
This check looks for calls to methods that do not seem to exist on a given type. It looks for the method on the type itself as well as in inherited classes or implemented interfaces.
This is most likely a typographical error or the method has been renamed.