Dispositif /
Wikibot
| Total Complexity | 69 |
| Total Lines | 436 |
| Duplicated Lines | 0 % |
| Changes | 8 | ||
| Bugs | 0 | Features | 0 |
Complex classes like OuvrageComplete 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 OuvrageComplete, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 19 | class OuvrageComplete |
||
| 20 | { |
||
| 21 | const ADD_PRESENTATION_EN_LIGNE = true; |
||
| 22 | |||
| 23 | const WIKI_LANGUAGE = 'fr'; |
||
| 24 | |||
| 25 | /** |
||
| 26 | * @var OuvrageTemplate |
||
| 27 | */ |
||
| 28 | private $origin; |
||
| 29 | |||
| 30 | private $book; |
||
| 31 | |||
| 32 | public $major = false; |
||
| 33 | |||
| 34 | public $notCosmetic = false; |
||
| 35 | |||
| 36 | private $log = []; |
||
| 37 | |||
| 38 | private $sameBook; |
||
| 39 | |||
| 40 | //todo: injection référence base ou mapping ? (Google |
||
| 41 | public function __construct(OuvrageTemplate $origin, OuvrageTemplate $book) |
||
| 42 | { |
||
| 43 | $this->origin = clone $origin; |
||
| 44 | $this->book = $book; |
||
| 45 | } |
||
| 46 | |||
| 47 | public function getLog(): array |
||
| 48 | { |
||
| 49 | return $this->log; |
||
| 50 | } |
||
| 51 | |||
| 52 | /** |
||
| 53 | * @return OuvrageTemplate |
||
| 54 | * @throws Exception |
||
| 55 | */ |
||
| 56 | public function getResult() |
||
| 57 | { |
||
| 58 | $this->complete(); |
||
| 59 | |||
| 60 | return $this->origin; |
||
| 61 | } |
||
| 62 | |||
| 63 | /** |
||
| 64 | * @return bool |
||
| 65 | * @throws Exception |
||
| 66 | */ |
||
| 67 | private function complete() |
||
| 68 | { |
||
| 69 | // si livre suspect, on stoppe |
||
| 70 | $this->sameBook = $this->predictSameBook(); |
||
| 71 | if (!$this->sameBook) { |
||
| 72 | dump('not same book'); |
||
| 73 | |||
| 74 | return false; |
||
| 75 | } |
||
| 76 | |||
| 77 | $skipParam = [ |
||
| 78 | 'isbn invalide', |
||
| 79 | 'auteurs', |
||
| 80 | 'auteur1', |
||
| 81 | 'prénom1', |
||
| 82 | 'nom1', |
||
| 83 | 'auteur2', |
||
| 84 | 'prénom2', |
||
| 85 | 'nom2', |
||
| 86 | 'auteur3', |
||
| 87 | 'prénom3', |
||
| 88 | 'nom3', |
||
| 89 | 'auteur4', |
||
| 90 | 'prénom4', |
||
| 91 | 'nom4', |
||
| 92 | 'lire en ligne', |
||
| 93 | 'présentation en ligne', |
||
| 94 | 'date', |
||
| 95 | 'sous-titre', |
||
| 96 | 'lien auteur1', |
||
| 97 | 'lien auteur2', |
||
| 98 | ]; |
||
| 99 | |||
| 100 | // completion automatique |
||
| 101 | foreach ($this->book->toArray() as $param => $value) { |
||
| 102 | if (empty($this->origin->getParam($param))) { |
||
| 103 | if (in_array($param, $skipParam)) { |
||
| 104 | continue; |
||
| 105 | } |
||
| 106 | // skip 'année' if 'date' not empty |
||
| 107 | if ('année' === $param && !empty($this->origin->getParam('date'))) { |
||
| 108 | continue; |
||
| 109 | } |
||
| 110 | |||
| 111 | $this->origin->setParam($param, $value); |
||
| 112 | |||
| 113 | if ('langue' === $param && static::WIKI_LANGUAGE === $value) { |
||
| 114 | //$this->log('fr'.$param); |
||
| 115 | continue; |
||
| 116 | } |
||
| 117 | |||
| 118 | $this->log('++'.$param); |
||
| 119 | $this->major = true; |
||
| 120 | $this->notCosmetic = true; |
||
| 121 | } |
||
| 122 | } |
||
| 123 | |||
| 124 | $this->processLienAuteur(); |
||
| 125 | $this->googleBookProcess(); |
||
| 126 | $this->processSousTitre(); |
||
| 127 | |||
| 128 | if ($this->notCosmetic && 'BnF' === $this->book->getSource()) { |
||
| 129 | $this->log('(BnF)'); |
||
| 130 | } |
||
| 131 | |||
| 132 | return true; |
||
| 133 | } |
||
| 134 | |||
| 135 | private function log(string $string): void |
||
| 139 | } |
||
| 140 | } |
||
| 141 | |||
| 142 | /** |
||
| 143 | * Complétion 'lien auteur1' d'après Wikidata et BnF. |
||
| 144 | * Logique : faut pas confondre auteur1/auteur2 pour le lien auteur1. |
||
| 145 | * |
||
| 146 | * @throws Exception |
||
| 147 | */ |
||
| 148 | private function processLienAuteur() |
||
| 149 | { |
||
| 150 | $lienAuteur1 = $this->book->getParam('lien auteur1'); |
||
| 151 | if (empty($lienAuteur1)) { |
||
| 152 | return; |
||
| 153 | } |
||
| 154 | if(!empty($this->origin->getParam('lien auteur1'))){ |
||
| 155 | echo "lien auteur1 existe déjà\n"; |
||
| 156 | return; |
||
| 157 | } |
||
| 158 | |||
| 159 | $originAuteur1 = $this->concatParamsAuteur1($this->origin); |
||
| 160 | $bookAuteur1 = $this->concatParamsAuteur1($this->book); |
||
| 161 | |||
| 162 | // WP:"Paul Durand" — Bnf "Paul Durand,..." |
||
| 163 | if (!empty($bookAuteur1) && !empty($originAuteur1) |
||
| 164 | && mb_strpos($bookAuteur1, $originAuteur1) !== false |
||
| 165 | ) { |
||
| 166 | $this->origin->setParam('lien auteur1', $lienAuteur1); |
||
| 167 | $this->log('+lien auteur1'); |
||
| 168 | $this->notCosmetic = true; |
||
| 169 | $this->major = true; |
||
| 170 | } else { |
||
| 171 | echo 'DEBUG : auteur1 pas identifié\n'; |
||
| 172 | } |
||
| 173 | // todo: gérer "not same book" avec inversion auteur1/2 avant d'implémenter +lien auteur2 |
||
| 174 | } |
||
| 175 | |||
| 176 | /** |
||
| 177 | * Concaténation auteur/prénom/nom pour comparaison de wiki-modèles. |
||
| 178 | * |
||
| 179 | * @param OuvrageTemplate $ouvrage |
||
| 180 | * @param int|null $num |
||
| 181 | * |
||
| 182 | * @return string|null |
||
| 183 | * @throws Exception |
||
| 184 | */ |
||
| 185 | private function concatParamsAuteur1(OuvrageTemplate $ouvrage, ?int $num = 1): ?string |
||
| 186 | { |
||
| 187 | $auteur = $ouvrage->getParam('auteur'.$num) ?? ''; |
||
| 188 | $prenom = $ouvrage->getParam('prénom'.$num) ?? ''; |
||
| 189 | $nom = $ouvrage->getParam('nom'.$num) ?? ''; |
||
| 190 | |||
| 191 | return trim($auteur.' '.$prenom.' '.$nom); |
||
| 192 | } |
||
| 193 | |||
| 194 | /** |
||
| 195 | * Complétion lire/présentation en ligne, si vide. |
||
| 196 | * Passe Google Book en accès partiel en 'lire en ligne' (sondage) |
||
| 197 | * |
||
| 198 | * @throws Exception |
||
| 199 | */ |
||
| 200 | private function googleBookProcess() |
||
| 201 | { |
||
| 202 | // si déjà lire/présentation en ligne => on touche à rien |
||
| 203 | if (!empty($this->origin->getParam('lire en ligne')) |
||
| 204 | || !empty($this->origin->getParam('présentation en ligne')) |
||
| 205 | ) { |
||
| 206 | return; |
||
| 207 | } |
||
| 208 | |||
| 209 | // completion basique |
||
| 210 | $booklire = $this->book->getParam('lire en ligne'); |
||
| 211 | if ($booklire) { |
||
| 212 | $this->origin->setParam('lire en ligne', $booklire); |
||
| 213 | $this->log('+lire en ligne'); |
||
| 214 | $this->notCosmetic = true; |
||
| 215 | $this->major = true; |
||
| 216 | |||
| 217 | return; |
||
| 218 | } |
||
| 219 | |||
| 220 | $presentation = $this->book->getParam('présentation en ligne') ?? false; |
||
| 221 | // Ajout du partial Google => mis en lire en ligne |
||
| 222 | // plutôt que 'présentation en ligne' selon sondage |
||
| 223 | if (!empty($presentation) && GoogleLivresTemplate::isGoogleBookValue($presentation)) { |
||
| 224 | $this->origin->setParam('lire en ligne', $presentation); |
||
| 225 | $this->log('+lire en ligne'); |
||
| 226 | $this->notCosmetic = true; |
||
| 227 | $this->major = true; |
||
| 228 | } |
||
| 229 | } |
||
| 230 | |||
| 231 | /** |
||
| 232 | * @return bool |
||
| 233 | * @throws Exception |
||
| 234 | */ |
||
| 235 | private function predictSameBook() |
||
| 236 | { |
||
| 237 | if ($this->hasSameISBN() && ($this->hasSameBookTitles() || $this->hasSameAuthors())) { |
||
| 238 | return true; |
||
| 239 | } |
||
| 240 | if ($this->hasSameBookTitles() && $this->hasSameAuthors()) { |
||
| 241 | return true; |
||
| 242 | } |
||
| 243 | |||
| 244 | return false; |
||
| 245 | } |
||
| 246 | |||
| 247 | /** |
||
| 248 | * @return bool |
||
| 249 | * @throws Exception |
||
| 250 | */ |
||
| 251 | private function hasSameAuthors(): bool |
||
| 252 | { |
||
| 253 | if ($this->authorsFromBook($this->origin) === $this->authorsFromBook($this->book)) { |
||
| 254 | return true; |
||
| 255 | } |
||
| 256 | |||
| 257 | // if there is only 2 char of difference (i.e. typo error) |
||
| 258 | if (levenshtein($this->authorsFromBook($this->origin), $this->authorsFromBook($this->book)) <= 2) { |
||
| 259 | $this->log('typo auteurs?'); |
||
| 260 | |||
| 261 | return true; |
||
| 262 | } |
||
| 263 | |||
| 264 | // Si auteur manquant sur wikipedia |
||
| 265 | if (empty($this->authorsFromBook($this->origin))) { |
||
| 266 | return true; |
||
| 267 | } |
||
| 268 | |||
| 269 | return false; |
||
| 270 | } |
||
| 271 | |||
| 272 | /** |
||
| 273 | * @param OuvrageTemplate $ouv |
||
| 274 | * |
||
| 275 | * @return string |
||
| 276 | * @throws Exception |
||
| 277 | */ |
||
| 278 | private function authorsFromBook(OuvrageTemplate $ouv) |
||
| 279 | { |
||
| 280 | $text = ''; |
||
| 281 | $paramAuteurs = [ |
||
| 282 | 'auteurs', |
||
| 283 | 'auteur1', |
||
| 284 | 'prénom1', |
||
| 285 | 'nom1', |
||
| 286 | 'auteur2', |
||
| 287 | 'prénom2', |
||
| 288 | 'nom2', |
||
| 289 | 'auteur3', |
||
| 290 | 'prénom3', |
||
| 291 | 'nom3', |
||
| 292 | 'auteur4', |
||
| 293 | 'prénom4', |
||
| 294 | 'nom4', |
||
| 295 | ]; |
||
| 296 | foreach ($paramAuteurs as $param) { |
||
| 297 | $value = str_replace(['.', ','], '', $ouv->getParam($param)); |
||
| 298 | // retire wikilien sur auteur |
||
| 299 | if (!empty($value)) { |
||
| 300 | $text .= WikiTextUtil::unWikify($value); |
||
| 301 | } |
||
| 302 | } |
||
| 303 | |||
| 304 | return $this->stripAll($text); |
||
| 305 | } |
||
| 306 | |||
| 307 | /** |
||
| 308 | * @return bool |
||
| 309 | * @throws Exception |
||
| 310 | */ |
||
| 311 | private function hasSameISBN(): bool |
||
| 312 | { |
||
| 313 | if (empty($this->origin->getParam('isbn')) || empty($this->book->getParam('isbn'))) { |
||
| 314 | return false; |
||
| 315 | } |
||
| 316 | // TODO replace with calcul isbn13 |
||
| 317 | $isbn1 = IsbnFacade::isbn2ean($this->origin->getParam('isbn')); |
||
| 318 | $isbn2 = IsbnFacade::isbn2ean($this->book->getParam('isbn')); |
||
| 319 | if ($isbn1 === $isbn2) { |
||
| 320 | return true; |
||
| 321 | } |
||
| 322 | |||
| 323 | return false; |
||
| 324 | } |
||
| 325 | |||
| 326 | /** |
||
| 327 | * Add or extract subtitle like in second book. |
||
| 328 | * |
||
| 329 | * @throws Exception |
||
| 330 | */ |
||
| 331 | private function processSousTitre() |
||
| 366 | } |
||
| 367 | } |
||
| 368 | } |
||
| 369 | |||
| 370 | /** |
||
| 371 | * @return bool |
||
| 372 | * @throws Exception |
||
| 373 | */ |
||
| 374 | private function hasSameBookTitles(): bool |
||
| 375 | { |
||
| 376 | $originBigTitle = $this->charsFromBigTitle($this->origin); |
||
| 377 | $bookBigTitle = $this->charsFromBigTitle($this->book); |
||
| 378 | |||
| 379 | if ($originBigTitle === $bookBigTitle) { |
||
| 380 | return true; |
||
| 381 | } |
||
| 382 | |||
| 383 | // if there is only 2 chars of difference (i.e. typo error) |
||
| 384 | // strlen for resource management |
||
| 385 | if (strlen($originBigTitle) < 40 && strlen($bookBigTitle) < 40 |
||
| 386 | && levenshtein($originBigTitle, $bookBigTitle) <= 2 |
||
| 387 | ) { |
||
| 388 | // $this->log('typo titre?'); // TODO Normalize:: text from external API |
||
| 389 | |||
| 390 | return true; |
||
| 391 | } |
||
| 392 | |||
| 393 | // si l'un des ouvrages ne comporte pas le sous-titre |
||
| 394 | if ($this->stripAll($this->origin->getParam('titre')) === $this->stripAll($this->book->getParam('titre'))) { |
||
| 395 | return true; |
||
| 396 | } |
||
| 397 | |||
| 398 | // sous-titre inclus dans le titre |
||
| 399 | // "Loiret : un département à l'élégance naturelle" <=> "Loiret" |
||
| 400 | if ($this->stripAll($this->mainBookTitle($this->origin->getParam('titre'))) === $this->stripAll( |
||
| 401 | $this->mainBookTitle($this->origin->getParam('titre')) |
||
| 402 | ) |
||
| 403 | ) { |
||
| 404 | return true; |
||
| 405 | } |
||
| 406 | // titre manquant sur wiki |
||
| 407 | if (empty($originBigTitle)) { |
||
| 408 | return true; |
||
| 409 | } |
||
| 410 | |||
| 411 | return false; |
||
| 412 | } |
||
| 413 | |||
| 414 | /** |
||
| 415 | * Give string before ":" (or same string if no ":"). |
||
| 416 | * |
||
| 417 | * @param string $str |
||
| 418 | * |
||
| 419 | * @return string |
||
| 420 | */ |
||
| 421 | private function mainBookTitle(string $str) |
||
| 428 | } |
||
| 429 | |||
| 430 | /** |
||
| 431 | * @param OuvrageTemplate $ouvrage |
||
| 432 | * |
||
| 433 | * @return string |
||
| 434 | * @throws Exception |
||
| 435 | */ |
||
| 436 | private function charsFromBigTitle(OuvrageTemplate $ouvrage): string |
||
| 437 | { |
||
| 438 | $text = $ouvrage->getParam('titre').$ouvrage->getParam('sous-titre'); |
||
| 439 | |||
| 440 | return $this->stripAll(Normalizer::normalize($text)); |
||
| 441 | } |
||
| 442 | |||
| 443 | /** |
||
| 444 | * @param string $text |
||
| 445 | * |
||
| 446 | * @return string |
||
| 447 | */ |
||
| 448 | private function stripAll(string $text): string |
||
| 455 | } |
||
| 456 | } |
||
| 457 |
Loading history...