x-tools /
xtools
| Total Complexity | 47 |
| Total Lines | 481 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
Complex classes like PageInfoApi 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 PageInfoApi, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 21 | class PageInfoApi extends Model |
||
| 22 | { |
||
| 23 | /** @var int Number of days of recent data to show for pageviews. */ |
||
| 24 | public const PAGEVIEWS_OFFSET = 30; |
||
| 25 | |||
| 26 | protected AutomatedEditsHelper $autoEditsHelper; |
||
| 27 | protected I18nHelper $i18n; |
||
| 28 | |||
| 29 | /** @var int Number of revisions that belong to the page. */ |
||
| 30 | protected int $numRevisions; |
||
| 31 | |||
| 32 | /** @var array Prose stats, with keys 'characters', 'words', 'references', 'unique_references', 'sections'. */ |
||
| 33 | protected array $proseStats; |
||
| 34 | |||
| 35 | /** @var array Number of categories, templates and files on the page. */ |
||
| 36 | protected array $transclusionData; |
||
| 37 | |||
| 38 | /** @var array Various statistics about bots that edited the page. */ |
||
| 39 | protected array $bots; |
||
| 40 | |||
| 41 | /** @var int Number of edits made to the page by bots. */ |
||
| 42 | protected int $botRevisionCount; |
||
| 43 | |||
| 44 | /** @var int[] Number of in and outgoing links and redirects to the page. */ |
||
| 45 | protected array $linksAndRedirects; |
||
| 46 | |||
| 47 | /** @var string[]|null Assessments of the page (see Page::getAssessments). */ |
||
| 48 | protected ?array $assessments; |
||
| 49 | |||
| 50 | /** @var string[] List of Wikidata and Checkwiki errors. */ |
||
| 51 | protected array $bugs; |
||
| 52 | |||
| 53 | /** |
||
| 54 | * PageInfoApi constructor. |
||
| 55 | * @param PageInfoRepository $repository |
||
| 56 | * @param I18nHelper $i18n |
||
| 57 | * @param AutomatedEditsHelper $autoEditsHelper |
||
| 58 | * @param Page $page The page to process. |
||
| 59 | * @param false|int $start Start date as Unix timestmap. |
||
| 60 | * @param false|int $end End date as Unix timestamp. |
||
| 61 | */ |
||
| 62 | public function __construct( |
||
| 63 | PageInfoRepository $repository, |
||
| 64 | I18nHelper $i18n, |
||
| 65 | AutomatedEditsHelper $autoEditsHelper, |
||
| 66 | Page $page, |
||
| 67 | $start = false, |
||
| 68 | $end = false |
||
| 69 | ) { |
||
| 70 | $this->repository = $repository; |
||
| 71 | $this->i18n = $i18n; |
||
| 72 | $this->autoEditsHelper = $autoEditsHelper; |
||
| 73 | $this->page = $page; |
||
| 74 | $this->start = $start; |
||
| 75 | $this->end = $end; |
||
| 76 | } |
||
| 77 | |||
| 78 | /** |
||
| 79 | * Get the number of revisions belonging to the page. |
||
| 80 | * @return int |
||
| 81 | */ |
||
| 82 | public function getNumRevisions(): int |
||
| 83 | { |
||
| 84 | if (!isset($this->numRevisions)) { |
||
| 85 | $this->numRevisions = $this->page->getNumRevisions(null, $this->start, $this->end); |
||
|
|
|||
| 86 | } |
||
| 87 | return $this->numRevisions; |
||
| 88 | } |
||
| 89 | |||
| 90 | /** |
||
| 91 | * Are there more revisions than we should process, based on the config? |
||
| 92 | * @return bool |
||
| 93 | */ |
||
| 94 | public function tooManyRevisions(): bool |
||
| 95 | { |
||
| 96 | return $this->repository->getMaxPageRevisions() > 0 && |
||
| 97 | $this->getNumRevisions() > $this->repository->getMaxPageRevisions(); |
||
| 98 | } |
||
| 99 | |||
| 100 | /** |
||
| 101 | * Get various basic info used in the API, including the number of revisions, unique authors, initial author |
||
| 102 | * and edit count of the initial author. This is combined into one query for better performance. Caching is |
||
| 103 | * intentionally disabled, because using the gadget, this will get hit for a different page constantly, where |
||
| 104 | * the likelihood of cache benefiting us is slim. |
||
| 105 | * @return string[]|false false if the page was not found. |
||
| 106 | */ |
||
| 107 | public function getBasicEditingInfo() |
||
| 108 | { |
||
| 109 | return $this->repository->getBasicEditingInfo($this->page); |
||
| 110 | } |
||
| 111 | |||
| 112 | /** |
||
| 113 | * Get the top editors to the page by edit count. |
||
| 114 | * @param int $limit Default 20, maximum 1,000. |
||
| 115 | * @param bool $noBots Set to non-false to exclude bots from the result. |
||
| 116 | * @return array |
||
| 117 | */ |
||
| 118 | public function getTopEditorsByEditCount(int $limit = 20, bool $noBots = false): array |
||
| 119 | { |
||
| 120 | // Quick cache, valid only for the same request. |
||
| 121 | static $topEditors = null; |
||
| 122 | if (null !== $topEditors) { |
||
| 123 | return $topEditors; |
||
| 124 | } |
||
| 125 | |||
| 126 | $rows = $this->repository->getTopEditorsByEditCount( |
||
| 127 | $this->page, |
||
| 128 | $this->start, |
||
| 129 | $this->end, |
||
| 130 | min($limit, 1000), |
||
| 131 | $noBots |
||
| 132 | ); |
||
| 133 | |||
| 134 | $topEditors = []; |
||
| 135 | $rank = 0; |
||
| 136 | foreach ($rows as $row) { |
||
| 137 | $topEditors[] = [ |
||
| 138 | 'rank' => ++$rank, |
||
| 139 | 'username' => $row['username'], |
||
| 140 | 'count' => $row['count'], |
||
| 141 | 'minor' => $row['minor'], |
||
| 142 | 'first_edit' => [ |
||
| 143 | 'id' => $row['first_revid'], |
||
| 144 | 'timestamp' => $row['first_timestamp'], |
||
| 145 | ], |
||
| 146 | 'latest_edit' => [ |
||
| 147 | 'id' => $row['latest_revid'], |
||
| 148 | 'timestamp' => $row['latest_timestamp'], |
||
| 149 | ], |
||
| 150 | ]; |
||
| 151 | } |
||
| 152 | |||
| 153 | return $topEditors; |
||
| 154 | } |
||
| 155 | |||
| 156 | /** |
||
| 157 | * Get prose and reference information. |
||
| 158 | * @return array|null With keys 'characters', 'words', 'references', 'unique_references', or null on failure. |
||
| 159 | */ |
||
| 160 | public function getProseStats(): ?array |
||
| 161 | { |
||
| 162 | if (isset($this->proseStats)) { |
||
| 163 | return $this->proseStats; |
||
| 164 | } |
||
| 165 | |||
| 166 | $datetime = is_int($this->end) ? new DateTime("@$this->end") : null; |
||
| 167 | |||
| 168 | try { |
||
| 169 | $html = $this->page->getHTMLContent($datetime); |
||
| 170 | } catch (BadGatewayException $e) { |
||
| 171 | // Prose stats are non-critical, so handle the BadGatewayException gracefully in the views. |
||
| 172 | return null; |
||
| 173 | } |
||
| 174 | |||
| 175 | $crawler = new Crawler($html); |
||
| 176 | $refs = $crawler->filter('[typeof~="mw:Extension/ref"]'); |
||
| 177 | |||
| 178 | [$bytes, $chars, $words] = $this->countCharsAndWords($crawler); |
||
| 179 | |||
| 180 | $refContent = []; |
||
| 181 | $refs->each(function ($ref) use (&$refContent): void { |
||
| 182 | $refContent[] = $ref->text(); |
||
| 183 | }); |
||
| 184 | $uniqueRefs = count(array_unique($refContent)); |
||
| 185 | |||
| 186 | $this->proseStats = [ |
||
| 187 | 'bytes' => $bytes, |
||
| 188 | 'characters' => $chars, |
||
| 189 | 'words' => $words, |
||
| 190 | 'references' => $refs->count(), |
||
| 191 | 'unique_references' => $uniqueRefs, |
||
| 192 | 'sections' => $crawler->filter('section')->count(), |
||
| 193 | ]; |
||
| 194 | return $this->proseStats; |
||
| 195 | } |
||
| 196 | |||
| 197 | /** |
||
| 198 | * Count the number of byes, characters and words of the plain text. |
||
| 199 | * @param Crawler $crawler |
||
| 200 | * @return array [num bytes, num chars, num words] |
||
| 201 | */ |
||
| 202 | private function countCharsAndWords(Crawler $crawler): array |
||
| 203 | { |
||
| 204 | $totalBytes = 0; |
||
| 205 | $totalChars = 0; |
||
| 206 | $totalWords = 0; |
||
| 207 | $paragraphs = $crawler->filter('section > p'); |
||
| 208 | |||
| 209 | // Remove templates, TemplateStyles, math and reference tags. |
||
| 210 | $crawler->filter(implode(',', [ |
||
| 211 | '#coordinates', |
||
| 212 | '[class*="emplate"]', |
||
| 213 | '[typeof~="mw:Extension/templatestyles"]', |
||
| 214 | '[typeof~="mw:Extension/math"]', |
||
| 215 | '[typeof~="mw:Extension/ref"]', |
||
| 216 | ]))->each(function (Crawler $subCrawler) { |
||
| 217 | foreach ($subCrawler as $subNode) { |
||
| 218 | $subNode->parentNode->removeChild($subNode); |
||
| 219 | } |
||
| 220 | }); |
||
| 221 | |||
| 222 | $paragraphs->each(function ($node) use (&$totalBytes, &$totalChars, &$totalWords): void { |
||
| 223 | /** @var Crawler $node */ |
||
| 224 | $text = $node->text(); |
||
| 225 | $totalBytes += strlen($text); |
||
| 226 | $totalChars += mb_strlen($text); |
||
| 227 | $totalWords += count(explode(' ', $text)); |
||
| 228 | }); |
||
| 229 | |||
| 230 | return [$totalBytes, $totalChars, $totalWords]; |
||
| 231 | } |
||
| 232 | |||
| 233 | /** |
||
| 234 | * Get the page assessments of the page. |
||
| 235 | * @see https://www.mediawiki.org/wiki/Special:MyLanguage/Extension:PageAssessments |
||
| 236 | * @return string[]|null null if unsupported. |
||
| 237 | * @codeCoverageIgnore |
||
| 238 | */ |
||
| 239 | public function getAssessments(): ?array |
||
| 240 | { |
||
| 241 | if (!isset($this->assessments)) { |
||
| 242 | $this->assessments = $this->page |
||
| 243 | ->getProject() |
||
| 244 | ->getPageAssessments() |
||
| 245 | ->getAssessments($this->page); |
||
| 246 | } |
||
| 247 | return $this->assessments; |
||
| 248 | } |
||
| 249 | |||
| 250 | /** |
||
| 251 | * Get the list of page's wikidata and Checkwiki errors. |
||
| 252 | * @see Page::getErrors() |
||
| 253 | * @return string[] |
||
| 254 | */ |
||
| 255 | public function getBugs(): array |
||
| 256 | { |
||
| 257 | if (!isset($this->bugs)) { |
||
| 258 | $this->bugs = $this->page->getErrors(); |
||
| 259 | } |
||
| 260 | return $this->bugs; |
||
| 261 | } |
||
| 262 | |||
| 263 | /** |
||
| 264 | * Get the number of wikidata nad CheckWiki errors. |
||
| 265 | * @return int |
||
| 266 | */ |
||
| 267 | public function numBugs(): int |
||
| 268 | { |
||
| 269 | return count($this->getBugs()); |
||
| 270 | } |
||
| 271 | |||
| 272 | /** |
||
| 273 | * Generate the data structure that will used in the PageInfo API response. |
||
| 274 | * @param Project $project |
||
| 275 | * @param Page $page |
||
| 276 | * @return array |
||
| 277 | * @codeCoverageIgnore |
||
| 278 | */ |
||
| 279 | public function getPageInfoApiData(Project $project, Page $page): array |
||
| 280 | { |
||
| 281 | $data = [ |
||
| 282 | 'project' => $project->getDomain(), |
||
| 283 | 'page' => $page->getTitle(), |
||
| 284 | 'watchers' => (int) $page->getWatchers(), |
||
| 285 | 'pageviews' => $page->getLatestPageviews(), |
||
| 286 | 'pageviews_offset' => self::PAGEVIEWS_OFFSET, |
||
| 287 | ]; |
||
| 288 | |||
| 289 | $info = null; |
||
| 290 | |||
| 291 | try { |
||
| 292 | $info = $this->repository->getBasicEditingInfo($page); |
||
| 293 | } catch (ServiceUnavailableHttpException $e) { |
||
| 294 | // No more open database connections. |
||
| 295 | $data['error'] = 'Unable to fetch revision data. Please try again later.'; |
||
| 296 | } catch (HttpException $e) { |
||
| 297 | /** |
||
| 298 | * The query most likely exceeded the maximum query time, |
||
| 299 | * so we'll abort and give only info retrieved by the API. |
||
| 300 | */ |
||
| 301 | $data['error'] = 'Unable to fetch revision data. The query may have timed out.'; |
||
| 302 | } |
||
| 303 | |||
| 304 | if ($info) { |
||
| 305 | $creationDateTime = DateTime::createFromFormat('YmdHis', $info['created_at']); |
||
| 306 | $modifiedDateTime = DateTime::createFromFormat('YmdHis', $info['modified_at']); |
||
| 307 | $secsSinceLastEdit = (new DateTime)->getTimestamp() - $modifiedDateTime->getTimestamp(); |
||
| 308 | |||
| 309 | // Some wikis (such foundation.wikimedia.org) may be missing the creation date. |
||
| 310 | $creationDateTime = false === $creationDateTime |
||
| 311 | ? null |
||
| 312 | : $creationDateTime->format('Y-m-d'); |
||
| 313 | |||
| 314 | $assessment = $page->getProject() |
||
| 315 | ->getPageAssessments() |
||
| 316 | ->getAssessment($page); |
||
| 317 | |||
| 318 | $data = array_merge($data, [ |
||
| 319 | 'revisions' => (int) $info['num_edits'], |
||
| 320 | 'editors' => (int) $info['num_editors'], |
||
| 321 | 'ip_edits' => (int) $info['ip_edits'], |
||
| 322 | 'minor_edits' => (int) $info['minor_edits'], |
||
| 323 | 'author' => $info['author'], |
||
| 324 | 'author_editcount' => null === $info['author_editcount'] ? null : (int) $info['author_editcount'], |
||
| 325 | 'created_at' => $creationDateTime, |
||
| 326 | 'created_rev_id' => $info['created_rev_id'], |
||
| 327 | 'modified_at' => $modifiedDateTime->format('Y-m-d H:i'), |
||
| 328 | 'secs_since_last_edit' => $secsSinceLastEdit, |
||
| 329 | 'last_edit_id' => (int) $info['modified_rev_id'], |
||
| 330 | 'assessment' => $assessment, |
||
| 331 | ]); |
||
| 332 | } |
||
| 333 | |||
| 334 | return $data; |
||
| 335 | } |
||
| 336 | |||
| 337 | /************************ Link statistics ************************/ |
||
| 338 | |||
| 339 | /** |
||
| 340 | * Get the number of external links on the page. |
||
| 341 | * @return int |
||
| 342 | */ |
||
| 343 | public function linksExtCount(): int |
||
| 344 | { |
||
| 345 | return $this->getLinksAndRedirects()['links_ext_count']; |
||
| 346 | } |
||
| 347 | |||
| 348 | /** |
||
| 349 | * Get the number of incoming links to the page. |
||
| 350 | * @return int |
||
| 351 | */ |
||
| 352 | public function linksInCount(): int |
||
| 355 | } |
||
| 356 | |||
| 357 | /** |
||
| 358 | * Get the number of outgoing links from the page. |
||
| 359 | * @return int |
||
| 360 | */ |
||
| 361 | public function linksOutCount(): int |
||
| 362 | { |
||
| 363 | return $this->getLinksAndRedirects()['links_out_count']; |
||
| 364 | } |
||
| 365 | |||
| 366 | /** |
||
| 367 | * Get the number of redirects to the page. |
||
| 368 | * @return int |
||
| 369 | */ |
||
| 370 | public function redirectsCount(): int |
||
| 371 | { |
||
| 372 | return $this->getLinksAndRedirects()['redirects_count']; |
||
| 373 | } |
||
| 374 | |||
| 375 | /** |
||
| 376 | * Get the number of external, incoming and outgoing links, along with the number of redirects to the page. |
||
| 377 | * @return int[] |
||
| 378 | * @codeCoverageIgnore |
||
| 379 | */ |
||
| 380 | private function getLinksAndRedirects(): array |
||
| 381 | { |
||
| 382 | if (!isset($this->linksAndRedirects)) { |
||
| 383 | $this->linksAndRedirects = $this->page->countLinksAndRedirects(); |
||
| 384 | } |
||
| 385 | return $this->linksAndRedirects; |
||
| 386 | } |
||
| 387 | |||
| 388 | /** |
||
| 389 | * Fetch transclusion data (categories, templates and files) that are on the page. |
||
| 390 | * @return array With keys 'categories', 'templates' and 'files'. |
||
| 391 | */ |
||
| 392 | public function getTransclusionData(): array |
||
| 393 | { |
||
| 394 | if (!isset($this->transclusionData)) { |
||
| 395 | $this->transclusionData = $this->repository->getTransclusionData($this->page); |
||
| 396 | } |
||
| 397 | return $this->transclusionData; |
||
| 398 | } |
||
| 399 | |||
| 400 | /** |
||
| 401 | * Get the number of categories that are on the page. |
||
| 402 | * @return int |
||
| 403 | */ |
||
| 404 | public function getNumCategories(): int |
||
| 405 | { |
||
| 406 | return $this->getTransclusionData()['categories']; |
||
| 407 | } |
||
| 408 | |||
| 409 | /** |
||
| 410 | * Get the number of templates that are on the page. |
||
| 411 | * @return int |
||
| 412 | */ |
||
| 413 | public function getNumTemplates(): int |
||
| 414 | { |
||
| 415 | return $this->getTransclusionData()['templates']; |
||
| 416 | } |
||
| 417 | |||
| 418 | /** |
||
| 419 | * Get the number of files that are on the page. |
||
| 420 | * @return int |
||
| 421 | */ |
||
| 422 | public function getNumFiles(): int |
||
| 423 | { |
||
| 424 | return $this->getTransclusionData()['files']; |
||
| 425 | } |
||
| 426 | |||
| 427 | /************************ Bot statistics ************************/ |
||
| 428 | |||
| 429 | /** |
||
| 430 | * Number of edits made to the page by current or former bots. |
||
| 431 | * @param string[][] $bots Used only in unit tests, where we supply mock data for the bots that will get processed. |
||
| 432 | * @return int |
||
| 433 | */ |
||
| 434 | public function getBotRevisionCount(?array $bots = null): int |
||
| 452 | } |
||
| 453 | |||
| 454 | /** |
||
| 455 | * Get and set $this->bots about bots that edited the page. This is done separately from the main query because |
||
| 456 | * we use this information when computing the top 10 editors in PageInfo, where we don't want to include bots. |
||
| 457 | * @return array |
||
| 458 | */ |
||
| 459 | public function getBots(): array |
||
| 460 | { |
||
| 461 | if (isset($this->bots)) { |
||
| 462 | return $this->bots; |
||
| 463 | } |
||
| 464 | |||
| 465 | // Parse the bot edits. |
||
| 466 | $this->bots = []; |
||
| 467 | |||
| 468 | $limit = $this->tooManyRevisions() ? $this->repository->getMaxPageRevisions() : null; |
||
| 469 | |||
| 470 | $botData = $this->repository->getBotData($this->page, $this->start, $this->end, $limit); |
||
| 471 | while ($bot = $botData->fetchAssociative()) { |
||
| 472 | $this->bots[$bot['username']] = [ |
||
| 473 | 'count' => (int)$bot['count'], |
||
| 474 | 'current' => '1' === $bot['current'], |
||
| 475 | ]; |
||
| 476 | } |
||
| 477 | |||
| 478 | // Sort by edit count. |
||
| 479 | uasort($this->bots, function ($a, $b) { |
||
| 480 | return $b['count'] - $a['count']; |
||
| 481 | }); |
||
| 482 | |||
| 483 | return $this->bots; |
||
| 484 | } |
||
| 485 | |||
| 486 | /** |
||
| 487 | * Get the number of bots that edited the page. |
||
| 488 | * @return int |
||
| 489 | */ |
||
| 490 | public function getNumBots(): int |
||
| 491 | { |
||
| 492 | return count($this->getBots()); |
||
| 493 | } |
||
| 494 | |||
| 495 | /** |
||
| 496 | * Get counts of (semi-)automated tools used to edit the page. |
||
| 497 | * @return array |
||
| 498 | */ |
||
| 499 | public function getAutoEditsCounts(): array |
||
| 502 | } |
||
| 503 | } |
||
| 504 |
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.