ricardosierra /
validate
| Total Complexity | 92 |
| Total Lines | 477 |
| Duplicated Lines | 0 % |
| Changes | 1 | ||
| Bugs | 1 | Features | 0 |
Complex classes like Url 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 Url, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 5 | class Url implements \Validate\Contracts\Validate |
||
| 6 | { |
||
| 7 | public static function toDatabase(string $url) |
||
| 8 | { |
||
| 9 | $url = str_replace('http://', '', $url); |
||
| 10 | $url = str_replace('https://', '', $url); |
||
| 11 | return $url; |
||
| 12 | } |
||
| 13 | |||
| 14 | public static function toUser($url) |
||
| 15 | { |
||
| 16 | return 'https://'.$url; |
||
| 17 | } |
||
| 18 | |||
| 19 | public static function validate($url) |
||
| 20 | { |
||
| 21 | if (strpos($url, ' ') !== false) { |
||
| 22 | return false; |
||
| 23 | } |
||
| 24 | return true; |
||
| 25 | } |
||
| 26 | |||
| 27 | public static function break(string $url) |
||
| 28 | { |
||
| 29 | return self::splitUrl($url); |
||
| 30 | } |
||
| 31 | |||
| 32 | public static function isSame(string $to, string $from) |
||
| 33 | { |
||
| 34 | return (self::toDatabase($to)===self::toDatabase($from)); |
||
| 35 | } |
||
| 36 | |||
| 37 | |||
| 38 | /** |
||
| 39 | * Given a URL calculates the page's directory |
||
| 40 | * |
||
| 41 | * @params string $url target URL |
||
| 42 | * @return string Directory |
||
| 43 | */ |
||
| 44 | public function parseDir($url) |
||
| 45 | { |
||
| 46 | $slash = strrpos($url, '/'); |
||
| 47 | return substr($url, 0, $slash+1); |
||
| 48 | } |
||
| 49 | |||
| 50 | /** |
||
| 51 | * Link Checking Functions |
||
| 52 | */ |
||
| 53 | |||
| 54 | /** |
||
| 55 | * Uniformly cleans a link to avoid duplicates |
||
| 56 | * |
||
| 57 | * 1. Changes relative links to absolute (/bar to http://www.foo.com/bar) |
||
| 58 | * 2. Removes anchor tags (foo.html#bar to foo.html) |
||
| 59 | * 3. Adds trailing slash if directory (foo.com/bar to foo.com/bar/) |
||
| 60 | * 4. Adds www if there is not a subdomain (foo.com to www.foo.com but not bar.foo.com) |
||
| 61 | * |
||
| 62 | * @params string $relativeUrl link to clean |
||
| 63 | * @parmas string $baseUrl directory of parent (linking) page |
||
| 64 | * @return string cleaned link |
||
| 65 | */ |
||
| 66 | public function cleanLink($relativeUrl, $baseUrl) |
||
| 80 | } |
||
| 81 | |||
| 82 | |||
| 83 | /** |
||
| 84 | * Performs a regular expression to see if a given link is an image |
||
| 85 | * |
||
| 86 | * @params string $link target link |
||
| 87 | * @return bool true on image, false on anything else |
||
| 88 | */ |
||
| 89 | public static function isImage($link) |
||
| 90 | { |
||
| 91 | if (preg_match('%\.(gif|jpe?g|png|bmp)$%i', $link)) { |
||
| 92 | return true; |
||
| 93 | } else { |
||
| 94 | return false; |
||
| 95 | } |
||
| 96 | } |
||
| 97 | |||
| 98 | /** |
||
| 99 | * Checks to see that a given link is within the domain/host whitelist |
||
| 100 | * |
||
| 101 | * Improved from original to use regular expression and match hosts. |
||
| 102 | * |
||
| 103 | * @params string $link target link |
||
| 104 | * @return bool true if out of domain, false if on domain whitelist |
||
| 105 | */ |
||
| 106 | public static function outOfDomain($link, $domainArray) |
||
| 107 | { |
||
| 108 | if (!is_array($domainArray)) { |
||
| 109 | $domainArray[] = $domainArray; |
||
| 110 | } |
||
| 111 | |||
| 112 | // get host name from URL |
||
| 113 | preg_match("/^(http:\/\/)?([^\/]+)/i", $link, $matches); |
||
| 114 | $host = $matches[2]; |
||
| 115 | // echo "<br />host: $host"; |
||
| 116 | // get last two segments of host name |
||
| 117 | // preg_match("/[^\.\/]+\.[^\.\/]+$/", $host, $matches); |
||
| 118 | foreach ($domainArray as $domain) { |
||
| 119 | if ($domain == $host) { |
||
| 120 | return false; |
||
| 121 | } |
||
| 122 | } |
||
| 123 | return true; |
||
| 124 | } |
||
| 125 | |||
| 126 | /** |
||
| 127 | * Checks to see that a given link matches a pattern in the exclude list |
||
| 128 | * |
||
| 129 | * @params string $link target link |
||
| 130 | * @return bool true if matches exclude, false if no match |
||
| 131 | */ |
||
| 132 | public function excludeByPattern($link, $excludedArray = []) |
||
| 145 | } |
||
| 146 | |||
| 147 | /** |
||
| 148 | * Checks to see if a given link is in fact a mailto: link |
||
| 149 | * |
||
| 150 | * @params string $link Link to check |
||
| 151 | * @return bool true on mailto:, false on everything else |
||
| 152 | */ |
||
| 153 | public static function isMailto($link) |
||
| 154 | { |
||
| 155 | if (stripos($link, 'mailto:')===false) { |
||
| 156 | return false; |
||
| 157 | } else { |
||
| 158 | return true; |
||
| 159 | } |
||
| 160 | } |
||
| 161 | |||
| 162 | /* Depreciated (I think) |
||
| 163 | |||
| 164 | public function count_slashes($url) { |
||
| 165 | if (strlen($url)<7) return 0; |
||
| 166 | return substr_count($url,'/',7); |
||
| 167 | } |
||
| 168 | |||
| 169 | public function get_slashes($url) { |
||
| 170 | if (preg_match_all('#/#',$url,$matches,PREG_OFFSET_CAPTURE,7)) return $matches[0]; |
||
| 171 | else return array(); |
||
| 172 | } |
||
| 173 | */ |
||
| 174 | |||
| 175 | /** |
||
| 176 | * Converts a relative URL (/bar) to an absolute URL (http://www.foo.com/bar) |
||
| 177 | * |
||
| 178 | * Inspired from code available at http://nadeausoftware.com/node/79, |
||
| 179 | * Code distributed under OSI BSD (http://www.opensource.org/licenses/bsd-license.php) |
||
| 180 | * |
||
| 181 | * @params string $baseUrl Directory of linking page |
||
| 182 | * @params string $relativeURL URL to convert to absolute |
||
| 183 | * @return string Absolute URL |
||
| 184 | */ |
||
| 185 | public static function urlToAbsolute($baseUrl, $relativeUrl) |
||
| 256 | } |
||
| 257 | |||
| 258 | /** |
||
| 259 | * Required public function of URL to absolute |
||
| 260 | * |
||
| 261 | * Inspired from code available at http://nadeausoftware.com/node/79, |
||
| 262 | * Code distributed under OSI BSD (http://www.opensource.org/licenses/bsd-license.php) |
||
| 263 | */ |
||
| 264 | public static function urlRemoveDotSegments($path) |
||
| 265 | { |
||
| 266 | // multi-byte character explode |
||
| 267 | $inSegs = preg_split('!/!u', $path); |
||
| 268 | $outSegs = array( ); |
||
| 269 | foreach ($inSegs as $seg) { |
||
| 270 | if ($seg == '' || $seg == '.') { |
||
| 271 | continue; |
||
| 272 | } |
||
| 273 | if ($seg == '..') { |
||
| 274 | array_pop($outSegs); |
||
| 275 | } else { |
||
| 276 | array_push($outSegs, $seg); |
||
| 277 | } |
||
| 278 | } |
||
| 279 | $outPath = implode('/', $outSegs); |
||
| 280 | if ($path[0] == '/') { |
||
| 281 | $outPath = '/' . $outPath; |
||
| 282 | } |
||
| 283 | |||
| 284 | // compare last multi-byte character against '/' |
||
| 285 | if ($outPath != '/' |
||
| 286 | && (mb_strlen($path)-1) == mb_strrpos($path, '/') |
||
| 287 | ) { |
||
| 288 | $outPath .= '/'; |
||
| 289 | } |
||
| 290 | |||
| 291 | return $outPath; |
||
| 292 | } |
||
| 293 | |||
| 294 | /** |
||
| 295 | * Required public function of URL to absolute |
||
| 296 | * |
||
| 297 | * Inspired from code available at http://nadeausoftware.com/node/79, |
||
| 298 | * Code distributed under OSI BSD (http://www.opensource.org/licenses/bsd-license.php) |
||
| 299 | */ |
||
| 300 | public static function splitUrl(string $url, $decode=true) |
||
| 301 | { |
||
| 302 | $parts = []; |
||
| 303 | $m = []; |
||
| 304 | $xunressub = 'a-zA-Z\d\-._~\!$&\'()*+,;='; |
||
| 305 | $xpchar = $xunressub . ':@%'; |
||
| 306 | |||
| 307 | $xscheme = '([a-zA-Z][a-zA-Z\d+-.]*)'; |
||
| 308 | |||
| 309 | $xuserinfo = '(([' . $xunressub . '%]*)' . |
||
| 310 | '(:([' . $xunressub . ':%]*))?)'; |
||
| 311 | |||
| 312 | $xipv4 = '(\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3})'; |
||
| 313 | |||
| 314 | $xipv6 = '(\[([a-fA-F\d.:]+)\])'; |
||
| 315 | |||
| 316 | $xhost_name = '([a-zA-Z%]+)'; |
||
| 317 | // $xhost_name = '([a-zA-Z\d-.%]+)'; @todo alterado pq tava dando erro nesse parser |
||
| 318 | |||
| 319 | $xhost = '(' . $xhost_name . '|' . $xipv4 . '|' . $xipv6 . ')'; |
||
| 320 | $xport = '(\d*)'; |
||
| 321 | $xauthority = '((' . $xuserinfo . '@)?' . $xhost . |
||
| 322 | '?(:' . $xport . ')?)'; |
||
| 323 | |||
| 324 | $xslash_seg = '(/[' . $xpchar . ']*)'; |
||
| 325 | $xpath_authabs = '((//' . $xauthority . ')((/[' . $xpchar . ']*)*))'; |
||
| 326 | $xpath_rel = '([' . $xpchar . ']+' . $xslash_seg . '*)'; |
||
| 327 | $xpath_abs = '(/(' . $xpath_rel . ')?)'; |
||
| 328 | $xapath = '(' . $xpath_authabs . '|' . $xpath_abs . |
||
| 329 | '|' . $xpath_rel . ')'; |
||
| 330 | |||
| 331 | $xqueryfrag = '([' . $xpchar . '/?' . ']*)'; |
||
| 332 | |||
| 333 | $xurl = '^(' . $xscheme . ':)?' . $xapath . '?' . |
||
| 334 | '(\?' . $xqueryfrag . ')?(#' . $xqueryfrag . ')?$'; |
||
| 335 | |||
| 336 | |||
| 337 | // Split the URL into components. |
||
| 338 | if (!preg_match('!' . $xurl . '!', $url, $m)) { |
||
| 339 | return false; |
||
| 340 | } |
||
| 341 | |||
| 342 | if (!empty($m[2])) { |
||
| 343 | $parts['scheme'] = strtolower($m[2]); |
||
| 344 | } |
||
| 345 | |||
| 346 | if (!empty($m[7])) { |
||
| 347 | if (isset($m[9])) { |
||
| 348 | $parts['user'] = $m[9]; |
||
| 349 | } else { |
||
| 350 | $parts['user'] = ''; |
||
| 351 | } |
||
| 352 | } |
||
| 353 | if (!empty($m[10])) { |
||
| 354 | $parts['pass'] = $m[11]; |
||
| 355 | } |
||
| 356 | |||
| 357 | if (!empty($m[13])) { |
||
| 358 | $h=$parts['host'] = $m[13]; |
||
| 359 | } elseif (!empty($m[14])) { |
||
| 360 | $parts['host'] = $m[14]; |
||
| 361 | } elseif (!empty($m[16])) { |
||
| 362 | $parts['host'] = $m[16]; |
||
| 363 | } elseif (!empty($m[5])) { |
||
| 364 | $parts['host'] = ''; |
||
| 365 | } |
||
| 366 | if (!empty($m[17])) { |
||
| 367 | $parts['port'] = $m[18]; |
||
| 368 | } |
||
| 369 | |||
| 370 | if (!empty($m[19])) { |
||
| 371 | $parts['path'] = $m[19]; |
||
| 372 | } elseif (!empty($m[21])) { |
||
| 373 | $parts['path'] = $m[21]; |
||
| 374 | } elseif (!empty($m[25])) { |
||
| 375 | $parts['path'] = $m[25]; |
||
| 376 | } |
||
| 377 | |||
| 378 | if (!empty($m[27])) { |
||
| 379 | $parts['query'] = $m[28]; |
||
| 380 | } |
||
| 381 | if (!empty($m[29])) { |
||
| 382 | $parts['fragment']= $m[30]; |
||
| 383 | } |
||
| 384 | |||
| 385 | if (!$decode) { |
||
| 386 | return $parts; |
||
| 387 | } |
||
| 388 | if (!empty($parts['user'])) { |
||
| 389 | $parts['user'] = rawurldecode($parts['user']); |
||
| 390 | } |
||
| 391 | if (!empty($parts['pass'])) { |
||
| 392 | $parts['pass'] = rawurldecode($parts['pass']); |
||
| 393 | } |
||
| 394 | if (!empty($parts['path'])) { |
||
| 395 | $parts['path'] = rawurldecode($parts['path']); |
||
| 396 | } |
||
| 397 | if (isset($h)) { |
||
| 398 | $parts['host'] = rawurldecode($parts['host']); |
||
| 399 | } |
||
| 400 | if (!empty($parts['query'])) { |
||
| 401 | $parts['query'] = rawurldecode($parts['query']); |
||
| 402 | } |
||
| 403 | if (!empty($parts['fragment'])) { |
||
| 404 | $parts['fragment'] = rawurldecode($parts['fragment']); |
||
| 405 | } |
||
| 406 | return $parts; |
||
| 407 | } |
||
| 408 | |||
| 409 | /** |
||
| 410 | * Required public function of URL to absolute |
||
| 411 | * |
||
| 412 | * Inspired from code available at http://nadeausoftware.com/node/79, |
||
| 413 | * Code distributed under OSI BSD (http://www.opensource.org/licenses/bsd-license.php) |
||
| 414 | */ |
||
| 415 | public static function joinUrl($parts, $encode=true) |
||
| 482 | } |
||
| 483 | } |
||
| 484 |
Loading history...