NNTmux /
newznab-tmux
| Total Complexity | 98 |
| Total Lines | 398 |
| Duplicated Lines | 0 % |
| Changes | 1 | ||
| Bugs | 0 | Features | 0 |
Complex classes like XxxCategorizer 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 XxxCategorizer, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 12 | class XxxCategorizer extends AbstractCategorizer |
||
| 13 | { |
||
| 14 | protected int $priority = 10; // High priority - should run early |
||
| 15 | |||
| 16 | // Known adult studios/sites - comprehensive list |
||
| 17 | protected const KNOWN_STUDIOS = 'Brazzers|NaughtyAmerica|RealityKings|Bangbros|BangBros18|TeenFidelity|PornPros|SexArt|WowGirls|Vixen|Blacked|Tushy|Deeper|Bellesa|Defloration|MetArt|MetArtX|TheLifeErotic|VivThomas|JoyMii|Nubiles|NubileFilms|Anilos|FamilyStrokes|X-Art|Babes|Twistys|WetAndPuffy|WowPorn|MomsTeachSex|Mofos|BangBus|Passion-HD|EvilAngel|DorcelClub|Private|Hustler|CherryPimps|PureTaboo|LadyLyne|TeamSkeet|GirlsWay|SweetSinner|NewSensations|Digital[._ -]?Playground|Wicked|Penthouse|Playboy|Kink|HardX|ArchAngel|JulesJordan|ManuelFerrara|LesbianX|AllAnal|DarkX|Elegant[._ -]?Angel|ZeroTolerance|Score|PornFidelity|Kelly[._ -]?Madison|DDF[._ -]?Network|21Sextury|21Naturals|Colette|SexMex|Bang|SpankBang|PornWorld|LegalPorno|AnalVids|GonzoXXX|RoccoSiffredi|Fake[._ -]?Hub|FakeAgent|FakeTaxi|FakeHostel|PublicAgent|StrandedTeens|Property[._ -]?Sex|Dane[._ -]?Jones|Lets[._ -]?Doe[._ -]?It|Office[._ -]?Obsession|SexyHub|Massage[._ -]?Rooms|Fitness[._ -]?Rooms|Female[._ -]?Agent|MissaX|All[._ -]?Girl[._ -]?Massage|Fantasy[._ -]?Massage|Nurumassage|Soapymassage|Reality[._ -]?Junkies|Perv[._ -]?Mom|Bad[._ -]?Milfs|Milf[._ -]?Body|Step[._ -]?Siblings|Sis[._ -]?Loves[._ -]?Me|Brother[._ -]??Crush|Dad[._ -]?Crush|Mom[._ -]?Knows[._ -]?Best|Bratty[._ -]?Sis|My[._ -]?Family[._ -]?Pies|Family[._ -]?Therapy|Nubiles[._ -]?Porn|Step[._ -]?Fantasy|Caught[._ -]?Fapping|She[._ -]?Will[._ -]?Cheat|Dirty[._ -]?Wives[._ -]?Club|Big[._ -]?Tits[._ -]?Round[._ -]?Asses|Ass[._ -]?Parade|Monsters[._ -]?Of[._ -]?Cock|Brown[._ -]?Bunnies|Teens[._ -]?Love[._ -]?Huge[._ -]?Cocks|Ass[._ -]?Masterpiece|Bang[._ -]?Casting|Holed|Tiny4K|Lubed|POVD|Exotic4K|CastingCouch[._ -]?X|Casting[._ -]?Couch|Creampie[._ -]?Angels|Digital[._ -]?Desire|Femjoy|Hegre|Joymii|Met[._ -]?Art|MPL[._ -]?Studios|Rylsky[._ -]?Art|Showy[._ -]?Beauty|Stunning18|Photodromm|Watch4Beauty|Wow[._ -]?Girls|Yonitale|Mommys[._ -]?Boy|AllOver30|MyFirst|10musume|Caribbeancom|Heyzo|Pacopacomama|1Pondo|TokyoHot'; |
||
| 18 | |||
| 19 | // Adult keywords |
||
| 20 | protected const ADULT_KEYWORDS = 'Anal|Ass|BBW|BDSM|Blow|Boob|Bukkake|Casting|Couch|Cock|Compilation|Creampie|Cum|Dick|Dildo|Facial|Fetish|Fuck|Gang|Hardcore|Homemade|Horny|Interracial|Lesbian|MILF|Masturbat|Nympho|Oral|Orgasm|Penetrat|Pornstar|POV|Pussy|Riding|Seduct|Sex|Shaved|Slut|Squirt|Suck|Swallow|Threesome|Tits|Titty|Toy|Virgin|Whore'; |
||
| 21 | |||
| 22 | // VR sites |
||
| 23 | protected const VR_SITES = 'SexBabesVR|LittleCapriceVR|VRoomed|VRMagic|TonightsGirlfriend|NaughtyAmericaVR|BaDoinkVR|WankzVR|VRBangers|StripzVR|RealJamVR|TmwVRnet|MilfVR|KinkVR|CzechVR(?:Fetish)?|HoloGirlsVR|WetVR|XSinsVR|VRCosplayX|BIBIVR|SLR|SexLikeReal'; |
||
| 24 | |||
| 25 | public function getName(): string |
||
| 28 | } |
||
| 29 | |||
| 30 | public function categorize(ReleaseContext $context): CategorizationResult |
||
| 97 | } |
||
| 98 | |||
| 99 | /** |
||
| 100 | * Check if release name looks like adult content. |
||
| 101 | */ |
||
| 102 | protected function looksLikeXxx(string $name): bool |
||
| 103 | { |
||
| 104 | // Check for XXX marker |
||
| 105 | if (preg_match('/\bXXX\b/i', $name)) { |
||
| 106 | return true; |
||
| 107 | } |
||
| 108 | |||
| 109 | // Check for known studios/sites |
||
| 110 | if (preg_match('/\b(' . self::KNOWN_STUDIOS . ')\b/i', $name)) { |
||
| 111 | return true; |
||
| 112 | } |
||
| 113 | |||
| 114 | // Check for known VR sites |
||
| 115 | if (preg_match('/\b(' . self::VR_SITES . ')\b/i', $name)) { |
||
| 116 | return true; |
||
| 117 | } |
||
| 118 | |||
| 119 | // Check for adult content indicators combined with video markers |
||
| 120 | if (preg_match('/\b(' . self::ADULT_KEYWORDS . ')\b/i', $name) && |
||
| 121 | preg_match('/\b(720p|1080p|2160p|4k|mp4|mkv|avi|wmv)\b/i', $name)) { |
||
| 122 | return true; |
||
| 123 | } |
||
| 124 | |||
| 125 | // Check for JAV/AV marker (common in Japanese adult releases) |
||
| 126 | if (preg_match('/\b(AV|JAV)\b/', $name) && preg_match('/\b(' . self::KNOWN_STUDIOS . ')\b/i', $name)) { |
||
| 127 | return true; |
||
| 128 | } |
||
| 129 | |||
| 130 | // Site with date pattern: sitename.YYYY.MM.DD or sitename.YY.MM.DD |
||
| 131 | // This pattern is very common for adult sites but rare for regular content |
||
| 132 | if (preg_match('/^[A-Za-z]+[.\-_ ](19|20)?\d{2}[.\-_ ]\d{2}[.\-_ ]\d{2}[.\-_ ][A-Za-z]/i', $name)) { |
||
| 133 | // Check it's not a TV daily show by checking for adult keywords or specific patterns |
||
| 134 | if (preg_match('/\b(' . self::ADULT_KEYWORDS . ')\b/i', $name)) { |
||
| 135 | return true; |
||
| 136 | } |
||
| 137 | // Check for performer name patterns (firstname.lastname) after the date |
||
| 138 | if (preg_match('/\d{2}[.\-_ ]([a-z]+)[.\-_ ]([a-z]+)[.\-_ ]/i', $name)) { |
||
| 139 | // Has a "firstname.lastname" pattern after date - likely adult |
||
| 140 | // But exclude obvious TV patterns |
||
| 141 | if (!preg_match('/\b(S\d{1,2}E\d{1,2}|Episode|Season|HDTV|PDTV)\b/i', $name)) { |
||
| 142 | return true; |
||
| 143 | } |
||
| 144 | } |
||
| 145 | } |
||
| 146 | |||
| 147 | return false; |
||
| 148 | } |
||
| 149 | |||
| 150 | protected function checkOnlyFans(string $name): ?CategorizationResult |
||
| 151 | { |
||
| 152 | // Skip photo packs unless there's a video hint |
||
| 153 | if (preg_match('/\b(photo(set)?|image(set)?|pics?|wallpapers?|collection|pack)\b/i', $name) && |
||
| 154 | !preg_match('/\b(mp4|mkv|mov|wmv|avi|webm|h\.?264|x264|h\.?265|x265)\b/i', $name)) { |
||
| 155 | return null; |
||
| 156 | } |
||
| 157 | |||
| 158 | if (preg_match('/\bOnly[-_ ]?Fans\b|^OF\./i', $name)) { |
||
| 159 | return $this->matched(Category::XXX_ONLYFANS, 0.95, 'onlyfans'); |
||
| 160 | } |
||
| 161 | |||
| 162 | return null; |
||
| 163 | } |
||
| 164 | |||
| 165 | protected function checkVR(string $name): ?CategorizationResult |
||
| 166 | { |
||
| 167 | if (stripos($name, 'vr') === false && stripos($name, 'oculus') === false && stripos($name, 'quest') === false) { |
||
| 168 | return null; |
||
| 169 | } |
||
| 170 | |||
| 171 | // Check for known VR site |
||
| 172 | $hasVRSite = preg_match('/\b(' . self::VR_SITES . ')\b/i', $name); |
||
| 173 | |||
| 174 | // Require either a VR site token, explicit VR180/VR360, or VR device |
||
| 175 | if (!preg_match('/\bVR(?:180|360)\b/i', $name) && |
||
| 176 | !$hasVRSite && |
||
| 177 | !preg_match('/\b(?:GearVR|Oculus|Quest[123]?|PSVR|Vive|Index|Pimax)\b/i', $name)) { |
||
| 178 | return null; |
||
| 179 | } |
||
| 180 | |||
| 181 | // VR pattern matching - includes VR devices |
||
| 182 | $vrPattern = '/\b(' . self::VR_SITES . ')\b|\bVR(?:180|360)\b|\b(?:5K|6K|7K|8K)\b.*\bVR\b|\b(?:GearVR|Oculus|Quest[123]?|PSVR|Vive|Index|Pimax)\b/i'; |
||
| 183 | |||
| 184 | if (preg_match($vrPattern, $name)) { |
||
| 185 | // VR sites are definitively adult content |
||
| 186 | if ($hasVRSite) { |
||
| 187 | return $this->matched(Category::XXX_VR, 0.95, 'vr_site'); |
||
| 188 | } |
||
| 189 | // VR device with adult keywords |
||
| 190 | if (preg_match('/\bXXX\b/i', $name) || preg_match('/\b(' . self::ADULT_KEYWORDS . ')\b/i', $name)) { |
||
| 191 | return $this->matched(Category::XXX_VR, 0.9, 'vr_device'); |
||
| 192 | } |
||
| 193 | } |
||
| 194 | |||
| 195 | return null; |
||
| 196 | } |
||
| 197 | |||
| 198 | protected function checkUHD(string $name): ?CategorizationResult |
||
| 199 | { |
||
| 200 | if (!preg_match('/\b(2160p|4k|UHD|Ultra[._ -]?HD)\b/i', $name)) { |
||
| 201 | return null; |
||
| 202 | } |
||
| 203 | |||
| 204 | // Check for adult markers |
||
| 205 | $hasAdultMarker = preg_match('/\bXXX\b/i', $name) || |
||
| 206 | preg_match('/\b(' . self::KNOWN_STUDIOS . ')\b/i', strtolower($name)) || |
||
| 207 | preg_match('/\b(Hardcore|Porn|Sex|Anal|Creampie|MILF|Lesbian|Teen|Interracial)\b/i', $name); |
||
| 208 | |||
| 209 | if (!$hasAdultMarker) { |
||
| 210 | return null; |
||
| 211 | } |
||
| 212 | |||
| 213 | // Known UHD release groups |
||
| 214 | if (preg_match('/XXX.+2160p[\w\-.]+M[PO][V4]-(KTR|GUSH|FaiLED|SEXORS|hUSHhUSH|YAPG|WRB|NBQ|FETiSH)/i', $name)) { |
||
| 215 | return $this->matched(Category::XXX_UHD, 0.95, 'uhd_group'); |
||
| 216 | } |
||
| 217 | |||
| 218 | return $this->matched(Category::XXX_UHD, 0.9, 'uhd'); |
||
| 219 | } |
||
| 220 | |||
| 221 | protected function checkClipHD(string $name): ?CategorizationResult |
||
| 285 | } |
||
| 286 | |||
| 287 | protected function checkPack(string $name): ?CategorizationResult |
||
| 288 | { |
||
| 289 | if (preg_match('/[ .]PACK[ .]/i', $name)) { |
||
| 290 | return $this->matched(Category::XXX_PACK, 0.85, 'pack'); |
||
| 291 | } |
||
| 292 | |||
| 293 | return null; |
||
| 294 | } |
||
| 295 | |||
| 296 | protected function checkClipSD(string $name, string $poster): ?CategorizationResult |
||
| 307 | } |
||
| 308 | |||
| 309 | protected function checkSD(string $name): ?CategorizationResult |
||
| 310 | { |
||
| 311 | if (preg_match('/SDX264XXX|XXX\.HR\./i', $name)) { |
||
| 312 | return $this->matched(Category::XXX_SD, 0.85, 'sd'); |
||
| 313 | } |
||
| 314 | |||
| 315 | return null; |
||
| 316 | } |
||
| 317 | |||
| 318 | protected function checkWebDL(string $name): ?CategorizationResult |
||
| 319 | { |
||
| 320 | // Exclude TV shows |
||
| 321 | if (preg_match('/\b(S\d{1,2}E\d{1,2})\b/i', $name)) { |
||
| 322 | return null; |
||
| 323 | } |
||
| 324 | |||
| 325 | if (preg_match('/web[._ -]dl|web-?rip/i', $name) && |
||
| 326 | (preg_match('/\b(' . self::ADULT_KEYWORDS . ')\b/i', $name) || |
||
| 327 | preg_match('/\b(' . self::KNOWN_STUDIOS . ')\b/i', $name) || |
||
| 328 | preg_match('/\b(XXX|Porn|Adult|JAV|Hentai)\b/i', $name))) { |
||
| 329 | return $this->matched(Category::XXX_WEBDL, 0.85, 'webdl'); |
||
| 330 | } |
||
| 331 | |||
| 332 | return null; |
||
| 333 | } |
||
| 334 | |||
| 335 | protected function checkX264(string $name): ?CategorizationResult |
||
| 336 | { |
||
| 337 | // Exclude HEVC/x265 |
||
| 338 | if (preg_match('/\b(x265|hevc)\b/i', $name)) { |
||
| 339 | return null; |
||
| 340 | } |
||
| 341 | |||
| 342 | // Require H.264/x264/AVC |
||
| 343 | if (!preg_match('/\b((x|h)[\.\-_ ]?264|AVC)\b/i', $name)) { |
||
| 344 | return null; |
||
| 345 | } |
||
| 346 | |||
| 347 | // Reject obvious non-targets |
||
| 348 | if (preg_match('/\bwmv\b|S\d{1,2}E\d{1,2}|\d+x\d+/i', $name)) { |
||
| 349 | return null; |
||
| 350 | } |
||
| 351 | |||
| 352 | // Check for adult content |
||
| 353 | $adultPattern = '/\bXXX\b|a\.b\.erotica|BangBros|Cum|Defloration|Err?oticax?|JoyMii|MetArt|Nubiles|Porn|SexArt|Tushy|Vixen|JAV|Brazzers|NaughtyAmerica|RealityKings/i'; |
||
| 354 | |||
| 355 | if (preg_match($adultPattern, $name)) { |
||
| 356 | return $this->matched(Category::XXX_X264, 0.85, 'x264'); |
||
| 357 | } |
||
| 358 | |||
| 359 | return null; |
||
| 360 | } |
||
| 361 | |||
| 362 | protected function checkXvid(string $name): ?CategorizationResult |
||
| 363 | { |
||
| 364 | if (preg_match('/(b[dr]|dvd)rip|detoxication|divx|nympho|pornolation|swe6|tesoro|xvid/i', $name)) { |
||
| 365 | return $this->matched(Category::XXX_XVID, 0.8, 'xvid'); |
||
| 366 | } |
||
| 367 | |||
| 368 | return null; |
||
| 369 | } |
||
| 370 | |||
| 371 | protected function checkImageset(string $name): ?CategorizationResult |
||
| 372 | { |
||
| 373 | if (preg_match('/IMAGESET|PICTURESET|ABPEA/i', $name)) { |
||
| 374 | return $this->matched(Category::XXX_IMAGESET, 0.9, 'imageset'); |
||
| 375 | } |
||
| 376 | |||
| 377 | return null; |
||
| 378 | } |
||
| 379 | |||
| 380 | protected function checkWMV(string $name): ?CategorizationResult |
||
| 381 | { |
||
| 382 | // Exclude modern formats |
||
| 383 | if (preg_match('/\b(720p|1080p|2160p|x264|x265|h264|h265|hevc|XviD|MP4-|\.mp4)[._ -]/i', $name)) { |
||
| 384 | return null; |
||
| 385 | } |
||
| 386 | |||
| 387 | if (preg_match('/\b(WMV|Windows\s?Media\s?Video)\b|\.wmv$|[._ -]wmv[._ -]/i', $name)) { |
||
| 388 | return $this->matched(Category::XXX_WMV, 0.8, 'wmv'); |
||
| 389 | } |
||
| 390 | |||
| 391 | return null; |
||
| 392 | } |
||
| 393 | |||
| 394 | protected function checkDVD(string $name): ?CategorizationResult |
||
| 401 | } |
||
| 402 | |||
| 403 | protected function checkOther(string $name): ?CategorizationResult |
||
| 404 | { |
||
| 405 | if (preg_match('/[._ -]Brazzers|Creampie|[._ -]JAV[._ -]|North\.Pole|She[._ -]?Male|Transsexual|OLDER ANGELS/i', $name)) { |
||
| 406 | return $this->matched(Category::XXX_OTHER, 0.7, 'other'); |
||
| 407 | } |
||
| 408 | |||
| 409 | return null; |
||
| 410 | } |
||
| 411 | } |
||
| 412 | |||
| 413 |
