Duplicate code is one of the most pungent code smells. A rule that is often used is to re-structure code once it is duplicated in three or more places.
Common duplication problems, and corresponding solutions are:
Complex classes like LogfileCommand 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. You can also have a look at the cohesion graph to spot any un-connected, or weakly-connected components.
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 LogfileCommand, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 70 | class LogfileCommand extends Command |
||
| 71 | { |
||
| 72 | /** |
||
| 73 | * @var array |
||
| 74 | */ |
||
| 75 | private $undefinedClients = array(); |
||
| 76 | |||
| 77 | private $uas = array(); |
||
| 78 | private $uasWithType = array(); |
||
| 79 | |||
| 80 | private $countOk = 0; |
||
| 81 | private $countNok = 0; |
||
| 82 | private $totalCount = 0; |
||
| 83 | |||
| 84 | /** |
||
| 85 | * @var BrowscapCacheInterface |
||
| 86 | */ |
||
| 87 | private $cache = null; |
||
| 88 | |||
| 89 | /** |
||
| 90 | * @var string |
||
| 91 | */ |
||
| 92 | private $defaultCacheFolder; |
||
| 93 | |||
| 94 | /** |
||
| 95 | * @param string $defaultCacheFolder |
||
| 96 | */ |
||
| 97 | 1 | public function __construct($defaultCacheFolder) |
|
| 98 | { |
||
| 99 | 1 | $this->defaultCacheFolder = $defaultCacheFolder; |
|
| 100 | |||
| 101 | 1 | parent::__construct(); |
|
| 102 | 1 | } |
|
| 103 | |||
| 104 | /** |
||
| 105 | * @param \BrowscapPHP\Cache\BrowscapCacheInterface $cache |
||
| 106 | * |
||
| 107 | * @return $this |
||
| 108 | */ |
||
| 109 | 1 | public function setCache(BrowscapCacheInterface $cache) |
|
| 110 | { |
||
| 111 | 1 | $this->cache = $cache; |
|
| 112 | |||
| 113 | 1 | return $this; |
|
| 114 | } |
||
| 115 | |||
| 116 | /** |
||
| 117 | * Configures the current command. |
||
| 118 | */ |
||
| 119 | 1 | protected function configure() |
|
| 120 | { |
||
| 121 | 1 | $this |
|
| 122 | 1 | ->setName('browscap:log') |
|
| 123 | 1 | ->setDescription('Parses the supplied webserver log file.') |
|
| 124 | 1 | ->addArgument( |
|
| 125 | 1 | 'output', |
|
| 126 | 1 | InputArgument::REQUIRED, |
|
| 127 | 1 | 'Path to output log file', |
|
| 128 | null |
||
| 129 | 1 | ) |
|
| 130 | 1 | ->addOption( |
|
| 131 | 1 | 'log-file', |
|
| 132 | 1 | 'f', |
|
| 133 | 1 | InputOption::VALUE_REQUIRED, |
|
| 134 | 'Path to a webserver log file' |
||
| 135 | 1 | ) |
|
| 136 | 1 | ->addOption( |
|
| 137 | 1 | 'log-dir', |
|
| 138 | 1 | 'd', |
|
| 139 | 1 | InputOption::VALUE_REQUIRED, |
|
| 140 | 'Path to webserver log directory' |
||
| 141 | 1 | ) |
|
| 142 | 1 | ->addOption( |
|
| 143 | 1 | 'include', |
|
| 144 | 1 | 'i', |
|
| 145 | 1 | InputOption::VALUE_IS_ARRAY | InputOption::VALUE_REQUIRED, |
|
| 146 | 1 | 'Include glob expressions for log files in the log directory', |
|
| 147 | 1 | array('*.log', '*.log*.gz', '*.log*.bz2') |
|
| 148 | 1 | ) |
|
| 149 | 1 | ->addOption( |
|
| 150 | 1 | 'exclude', |
|
| 151 | 1 | 'e', |
|
| 152 | 1 | InputOption::VALUE_IS_ARRAY | InputOption::VALUE_REQUIRED, |
|
| 153 | 1 | 'Exclude glob expressions for log files in the log directory', |
|
| 154 | 1 | array('*error*') |
|
| 155 | 1 | ) |
|
| 156 | 1 | ->addOption( |
|
| 157 | 1 | 'debug', |
|
| 158 | 1 | null, |
|
| 159 | 1 | InputOption::VALUE_NONE, |
|
| 160 | 'Should the debug mode entered?' |
||
| 161 | 1 | ) |
|
| 162 | 1 | ->addOption( |
|
| 163 | 1 | 'cache', |
|
| 164 | 1 | 'c', |
|
| 165 | 1 | InputOption::VALUE_OPTIONAL, |
|
| 166 | 1 | 'Where the cache files are located', |
|
| 167 | 1 | $this->defaultCacheFolder |
|
| 168 | 1 | ) |
|
| 169 | ; |
||
| 170 | 1 | } |
|
| 171 | |||
| 172 | /** |
||
| 173 | * @param InputInterface $input |
||
| 174 | * @param OutputInterface $output |
||
| 175 | * |
||
| 176 | * @throws \UnexpectedValueException |
||
| 177 | * @throws \BrowscapPHP\Exception\InvalidArgumentException |
||
| 178 | * @return int|null|void |
||
| 179 | */ |
||
| 180 | protected function execute(InputInterface $input, OutputInterface $output) |
||
| 181 | { |
||
| 182 | if (!$input->getOption('log-file') && !$input->getOption('log-dir')) { |
||
| 183 | throw InvalidArgumentException::oneOfCommandArguments('log-file', 'log-dir'); |
||
| 184 | } |
||
| 185 | |||
| 186 | $loggerHelper = new LoggerHelper(); |
||
| 187 | $logger = $loggerHelper->create($input->getOption('debug')); |
||
| 188 | |||
| 189 | $browscap = $this->getBrowscap(); |
||
| 190 | $loader = new IniLoader(); |
||
|
|
|||
| 191 | $collection = ReaderFactory::factory(); |
||
| 192 | $fs = new Filesystem(); |
||
| 193 | |||
| 194 | $browscap |
||
| 195 | ->setLogger($logger) |
||
| 196 | ->setCache($this->getCache($input)) |
||
| 197 | ; |
||
| 198 | |||
| 199 | /** @var $file \Symfony\Component\Finder\SplFileInfo */ |
||
| 200 | foreach ($this->getFiles($input) as $file) { |
||
| 201 | $this->uas = array(); |
||
| 202 | $path = $this->getPath($file); |
||
| 203 | |||
| 204 | $this->countOk = 0; |
||
| 205 | $this->countNok = 0; |
||
| 206 | |||
| 207 | $logger->info('Analyzing file "'.$file->getPathname().'"'); |
||
| 208 | |||
| 209 | $lines = file($path); |
||
| 210 | |||
| 211 | if (empty($lines)) { |
||
| 212 | $logger->info('Skipping empty file "'.$file->getPathname().'"'); |
||
| 213 | continue; |
||
| 214 | } |
||
| 215 | |||
| 216 | $this->totalCount = count($lines); |
||
| 217 | |||
| 218 | foreach ($lines as $line) { |
||
| 219 | $this->handleLine( |
||
| 220 | $output, |
||
| 221 | $collection, |
||
| 222 | $browscap, |
||
| 223 | $line |
||
| 224 | ); |
||
| 225 | } |
||
| 226 | |||
| 227 | $this->outputProgress($output, '', true); |
||
| 228 | |||
| 229 | arsort($this->uas, SORT_NUMERIC); |
||
| 230 | |||
| 231 | try { |
||
| 232 | $fs->dumpFile($input->getArgument('output').'/output.sql', $this->createSqlContent()); |
||
| 233 | } catch (IOException $e) { |
||
| 234 | // do nothing |
||
| 235 | } |
||
| 236 | |||
| 237 | try { |
||
| 238 | $fs->dumpFile( |
||
| 239 | $input->getArgument('output').'/output.txt', |
||
| 240 | implode(PHP_EOL, array_unique($this->undefinedClients)) |
||
| 241 | ); |
||
| 242 | } catch (IOException $e) { |
||
| 243 | // do nothing |
||
| 244 | } |
||
| 245 | |||
| 246 | try { |
||
| 247 | $fs->dumpFile( |
||
| 248 | $input->getArgument('output').'/output-with-amount.txt', |
||
| 249 | $this->createAmountContent() |
||
| 250 | ); |
||
| 251 | } catch (IOException $e) { |
||
| 252 | // do nothing |
||
| 253 | } |
||
| 254 | |||
| 255 | try { |
||
| 256 | $fs->dumpFile( |
||
| 257 | $input->getArgument('output').'/output-with-amount-and-type.txt', |
||
| 258 | $this->createAmountTypeContent() |
||
| 259 | ); |
||
| 260 | } catch (IOException $e) { |
||
| 261 | // do nothing |
||
| 262 | } |
||
| 263 | } |
||
| 264 | |||
| 265 | try { |
||
| 266 | $fs->dumpFile($input->getArgument('output').'/output.sql', $this->createSqlContent()); |
||
| 267 | } catch (IOException $e) { |
||
| 268 | // do nothing |
||
| 269 | } |
||
| 270 | |||
| 271 | $outputFile = $input->getArgument('output').'/output.txt'; |
||
| 272 | |||
| 273 | try { |
||
| 274 | $fs->dumpFile( |
||
| 275 | $outputFile, |
||
| 276 | implode(PHP_EOL, array_unique($this->undefinedClients)) |
||
| 277 | ); |
||
| 278 | } catch (IOException $e) { |
||
| 279 | throw new \UnexpectedValueException('writing to file "'.$outputFile.'" failed', 0, $e); |
||
| 280 | } |
||
| 281 | |||
| 282 | try { |
||
| 283 | $fs->dumpFile( |
||
| 284 | $input->getArgument('output').'/output-with-amount.txt', |
||
| 285 | $this->createAmountContent() |
||
| 286 | ); |
||
| 287 | } catch (IOException $e) { |
||
| 288 | // do nothing |
||
| 289 | } |
||
| 290 | |||
| 291 | try { |
||
| 292 | $fs->dumpFile( |
||
| 293 | $input->getArgument('output').'/output-with-amount-and-type.txt', |
||
| 294 | $this->createAmountTypeContent() |
||
| 295 | ); |
||
| 296 | } catch (IOException $e) { |
||
| 297 | // do nothing |
||
| 298 | } |
||
| 299 | } |
||
| 300 | |||
| 301 | private function createSqlContent() |
||
| 302 | { |
||
| 303 | $content = ''; |
||
| 304 | |||
| 305 | arsort($this->uas, SORT_NUMERIC); |
||
| 306 | |||
| 307 | foreach ($this->uas as $agentOfLine => $count) { |
||
| 308 | $content .= " |
||
| 309 | INSERT INTO `agents` (`agent`, `count`) |
||
| 310 | VALUES ('".addslashes($agentOfLine)."', ".addslashes($count).") |
||
| 311 | ON DUPLICATE KEY UPDATE `count`=`count`+".addslashes($count)."; |
||
| 312 | "; |
||
| 313 | } |
||
| 314 | |||
| 315 | return $content; |
||
| 316 | } |
||
| 317 | |||
| 318 | private function createAmountContent() |
||
| 319 | { |
||
| 320 | $counts = array(); |
||
| 321 | |||
| 322 | foreach ($this->uasWithType as $uas) { |
||
| 323 | foreach ($uas as $userAgentString => $count) { |
||
| 324 | if (isset($counts[$userAgentString])) { |
||
| 325 | $counts[$userAgentString] += $count; |
||
| 326 | } else { |
||
| 327 | $counts[$userAgentString] = $count; |
||
| 328 | } |
||
| 329 | } |
||
| 330 | } |
||
| 331 | |||
| 332 | $content = ''; |
||
| 333 | |||
| 334 | arsort($counts, SORT_NUMERIC); |
||
| 335 | |||
| 336 | foreach ($counts as $agentOfLine => $count) { |
||
| 337 | $content .= "$count\t$agentOfLine\n"; |
||
| 338 | } |
||
| 339 | |||
| 340 | return $content; |
||
| 341 | } |
||
| 342 | |||
| 343 | private function createAmountTypeContent() |
||
| 362 | |||
| 363 | /** |
||
| 364 | * @param \Symfony\Component\Console\Output\OutputInterface $output |
||
| 365 | * @param \BrowscapPHP\Util\Logfile\ReaderCollection $collection |
||
| 366 | * @param \BrowscapPHP\Browscap $browscap |
||
| 367 | * @param integer $line |
||
| 368 | * |
||
| 369 | * @throws UnknownBrowserException |
||
| 370 | * @throws UnknownBrowserTypeException |
||
| 371 | * @throws UnknownDeviceException |
||
| 372 | * @throws UnknownEngineException |
||
| 373 | * @throws UnknownPlatformException |
||
| 374 | * @throws \Exception |
||
| 375 | */ |
||
| 376 | private function handleLine(OutputInterface $output, ReaderCollection $collection, Browscap $browscap, $line) |
||
| 377 | { |
||
| 378 | $userAgentString = ''; |
||
| 379 | |||
| 380 | try { |
||
| 381 | $userAgentString = $collection->read($line); |
||
| 382 | |||
| 383 | try { |
||
| 384 | $this->getResult($browscap->getBrowser($userAgentString)); |
||
| 385 | } catch (\Exception $e) { |
||
| 386 | $this->undefinedClients[] = $userAgentString; |
||
| 387 | |||
| 388 | throw $e; |
||
| 389 | } |
||
| 390 | |||
| 391 | $type = '.'; |
||
| 392 | $this->countOk++; |
||
| 393 | } catch (ReaderException $e) { |
||
| 394 | $type = 'E'; |
||
| 395 | $this->countNok++; |
||
| 396 | } catch (UnknownBrowserTypeException $e) { |
||
| 397 | $type = 'T'; |
||
| 398 | $this->countNok++; |
||
| 399 | } catch (UnknownBrowserException $e) { |
||
| 400 | $type = 'B'; |
||
| 401 | $this->countNok++; |
||
| 402 | } catch (UnknownPlatformException $e) { |
||
| 403 | $type = 'P'; |
||
| 404 | $this->countNok++; |
||
| 405 | } catch (UnknownDeviceException $e) { |
||
| 406 | $type = 'D'; |
||
| 407 | $this->countNok++; |
||
| 408 | } catch (UnknownEngineException $e) { |
||
| 409 | $type = 'N'; |
||
| 410 | $this->countNok++; |
||
| 411 | } catch (\Exception $e) { |
||
| 412 | $type = 'U'; |
||
| 413 | $this->countNok++; |
||
| 414 | } |
||
| 415 | |||
| 416 | $this->outputProgress($output, $type); |
||
| 417 | |||
| 418 | // count all useragents |
||
| 419 | if (isset($this->uas[$userAgentString])) { |
||
| 420 | $this->uas[$userAgentString]++; |
||
| 421 | } else { |
||
| 422 | $this->uas[$userAgentString] = 1; |
||
| 423 | } |
||
| 424 | |||
| 425 | if ('.' !== $type && 'E' !== $type) { |
||
| 426 | // count all undetected useragents grouped by detection error |
||
| 427 | if (!isset($this->uasWithType[$type])) { |
||
| 428 | $this->uasWithType[$type] = array(); |
||
| 429 | } |
||
| 430 | |||
| 431 | if (isset($this->uasWithType[$type][$userAgentString])) { |
||
| 432 | $this->uasWithType[$type][$userAgentString]++; |
||
| 433 | } else { |
||
| 434 | $this->uasWithType[$type][$userAgentString] = 1; |
||
| 435 | } |
||
| 436 | } |
||
| 437 | } |
||
| 438 | |||
| 439 | /** |
||
| 440 | * @param \Symfony\Component\Console\Output\OutputInterface $output |
||
| 441 | * @param string $result |
||
| 442 | * @param bool $end |
||
| 443 | * |
||
| 444 | * @return int |
||
| 445 | */ |
||
| 446 | private function outputProgress(OutputInterface $output, $result, $end = false) |
||
| 447 | { |
||
| 448 | if (($this->totalCount % 70) === 0 || $end) { |
||
| 449 | $formatString = ' %'.strlen($this->countOk).'d OK, %'.strlen($this->countNok).'d NOK, Summary %' |
||
| 450 | .strlen($this->totalCount).'d'; |
||
| 451 | |||
| 452 | if ($end) { |
||
| 453 | $result = str_pad($result, 70 - ($this->totalCount % 70), ' ', STR_PAD_RIGHT); |
||
| 454 | } |
||
| 455 | |||
| 456 | $endString = sprintf($formatString, $this->countOk, $this->countNok, $this->totalCount); |
||
| 457 | |||
| 458 | $output->writeln($result.$endString); |
||
| 459 | |||
| 460 | return; |
||
| 461 | } |
||
| 462 | |||
| 463 | $output->write($result); |
||
| 464 | } |
||
| 465 | |||
| 466 | /** |
||
| 467 | * @param \stdClass $result |
||
| 468 | * |
||
| 469 | * @return string |
||
| 470 | */ |
||
| 471 | private function getResult(\stdClass $result) |
||
| 472 | { |
||
| 473 | if ('Default Browser' === $result->browser) { |
||
| 474 | throw new UnknownBrowserException('unknwon browser found'); |
||
| 475 | } |
||
| 476 | |||
| 477 | if ('unknown' === $result->browser_type) { |
||
| 478 | throw new UnknownBrowserTypeException('unknwon browser type found'); |
||
| 479 | } |
||
| 480 | |||
| 481 | if (in_array($result->browser_type, array('Bot/Crawler', 'Library'))) { |
||
| 482 | return '.'; |
||
| 483 | } |
||
| 484 | |||
| 485 | if ('unknown' === $result->platform) { |
||
| 486 | throw new UnknownPlatformException('unknown platform found'); |
||
| 487 | } |
||
| 488 | |||
| 489 | if ('unknown' === $result->device_type) { |
||
| 490 | throw new UnknownDeviceException('unknwon device type found'); |
||
| 491 | } |
||
| 492 | |||
| 493 | if ('unknown' === $result->renderingengine_name) { |
||
| 494 | throw new UnknownEngineException('unknown rendering engine found'); |
||
| 495 | } |
||
| 496 | |||
| 497 | return '.'; |
||
| 498 | } |
||
| 499 | |||
| 500 | /** |
||
| 501 | * @param \Symfony\Component\Console\Input\InputInterface $input |
||
| 502 | * |
||
| 503 | * @return \Symfony\Component\Finder\Finder |
||
| 504 | */ |
||
| 505 | private function getFiles(InputInterface $input) |
||
| 506 | { |
||
| 507 | $finder = Finder::create(); |
||
| 508 | |||
| 509 | if ($input->getOption('log-file')) { |
||
| 510 | $file = $input->getOption('log-file'); |
||
| 511 | $finder->append(Finder::create()->in(dirname($file))->name(basename($file))); |
||
| 512 | } |
||
| 513 | |||
| 514 | if ($input->getOption('log-dir')) { |
||
| 515 | $dirFinder = Finder::create() |
||
| 516 | ->in($input->getOption('log-dir')); |
||
| 517 | array_map(array($dirFinder, 'name'), $input->getOption('include')); |
||
| 518 | array_map(array($dirFinder, 'notName'), $input->getOption('exclude')); |
||
| 519 | |||
| 520 | $finder->append($dirFinder); |
||
| 521 | } |
||
| 522 | |||
| 523 | return $finder; |
||
| 524 | } |
||
| 525 | |||
| 526 | /** |
||
| 527 | * @param \Symfony\Component\Finder\SplFileInfo $file |
||
| 528 | * |
||
| 529 | * @return string |
||
| 530 | */ |
||
| 531 | private function getPath(SplFileInfo $file) |
||
| 532 | { |
||
| 533 | switch ($file->getExtension()) { |
||
| 534 | case 'gz': |
||
| 535 | $path = 'compress.zlib://'.$file->getPathname(); |
||
| 536 | break; |
||
| 537 | case 'bz2': |
||
| 538 | $path = 'compress.bzip2://'.$file->getPathname(); |
||
| 539 | break; |
||
| 540 | default: |
||
| 541 | $path = $file->getPathname(); |
||
| 542 | break; |
||
| 543 | } |
||
| 544 | |||
| 545 | return $path; |
||
| 546 | } |
||
| 547 | |||
| 548 | private function getBrowscap() |
||
| 552 | |||
| 553 | /** |
||
| 554 | * @param InputInterface $input |
||
| 555 | * |
||
| 556 | * @return BrowscapCacheInterface |
||
| 557 | */ |
||
| 558 | View Code Duplication | private function getCache(InputInterface $input) |
|
| 559 | { |
||
| 560 | if (null === $this->cache) { |
||
| 561 | $cacheAdapter = new File(array(File::DIR => $input->getOption('cache'))); |
||
| 562 | $this->cache = new BrowscapCache($cacheAdapter); |
||
| 563 | } |
||
| 567 | } |
||
| 568 |
This check looks for variable assignements that are either overwritten by other assignments or where the variable is not used subsequently.
Both the
$myVarassignment in line 1 and the$higherassignment in line 2 are dead. The first because$myVaris never used and the second because$higheris always overwritten for every possible time line.