vaites /
php-apache-tika
| Total Complexity | 56 |
| Total Lines | 406 |
| Duplicated Lines | 0 % |
| Changes | 8 | ||
| Bugs | 0 | Features | 1 |
Complex classes like CLIClient 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 CLIClient, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 15 | class CLIClient extends Client |
||
| 16 | { |
||
| 17 | protected const MODE = 'cli'; |
||
| 18 | |||
| 19 | /** |
||
| 20 | * Apache Tika app path |
||
| 21 | * |
||
| 22 | * @var string |
||
| 23 | */ |
||
| 24 | protected $path = null; |
||
| 25 | |||
| 26 | /** |
||
| 27 | * Java binary path |
||
| 28 | * |
||
| 29 | * @var string |
||
| 30 | */ |
||
| 31 | protected $java = null; |
||
| 32 | |||
| 33 | /** |
||
| 34 | * Java arguments |
||
| 35 | * |
||
| 36 | * @var string |
||
| 37 | */ |
||
| 38 | protected $javaArgs = null; |
||
| 39 | |||
| 40 | /** |
||
| 41 | * Configure client |
||
| 42 | * |
||
| 43 | * @throws \Exception |
||
| 44 | */ |
||
| 45 | public function __construct(string $path = null, string $java = null, bool $check = true) |
||
| 46 | { |
||
| 47 | parent::__construct(); |
||
| 48 | |||
| 49 | if($path) |
||
| 50 | { |
||
| 51 | $this->setPath($path); |
||
| 52 | } |
||
| 53 | |||
| 54 | if($java) |
||
| 55 | { |
||
| 56 | $this->setJava($java); |
||
| 57 | } |
||
| 58 | |||
| 59 | if($check === true) |
||
| 60 | { |
||
| 61 | $this->check(); |
||
| 62 | } |
||
| 63 | } |
||
| 64 | |||
| 65 | /** |
||
| 66 | * Get the path |
||
| 67 | */ |
||
| 68 | public function getPath(): ?string |
||
| 71 | } |
||
| 72 | |||
| 73 | /** |
||
| 74 | * Set the path |
||
| 75 | */ |
||
| 76 | public function setPath($path): self |
||
| 77 | { |
||
| 78 | $this->path = $path; |
||
| 79 | |||
| 80 | return $this; |
||
| 81 | } |
||
| 82 | |||
| 83 | /** |
||
| 84 | * Get the Java path |
||
| 85 | */ |
||
| 86 | public function getJava(): ?string |
||
| 87 | { |
||
| 88 | return $this->java; |
||
| 89 | } |
||
| 90 | |||
| 91 | /** |
||
| 92 | * Set the Java path |
||
| 93 | */ |
||
| 94 | public function setJava(string $java): self |
||
| 95 | { |
||
| 96 | $this->java = $java; |
||
| 97 | |||
| 98 | return $this; |
||
| 99 | } |
||
| 100 | |||
| 101 | /** |
||
| 102 | * Get the Java arguments |
||
| 103 | */ |
||
| 104 | public function getJavaArgs(): ?string |
||
| 105 | { |
||
| 106 | return $this->javaArgs; |
||
| 107 | } |
||
| 108 | |||
| 109 | /** |
||
| 110 | * Set the Java arguments |
||
| 111 | * |
||
| 112 | * NOTE: to modify child process jvm args, prepend "J" to each argument (-JXmx4g) |
||
| 113 | */ |
||
| 114 | public function setJavaArgs(string $args): self |
||
| 115 | { |
||
| 116 | $this->javaArgs = $args; |
||
| 117 | |||
| 118 | return $this; |
||
| 119 | } |
||
| 120 | |||
| 121 | /** |
||
| 122 | * Returns the supported MIME types |
||
| 123 | * |
||
| 124 | * NOTE: the data provided by the CLI must be parsed: mime type has no spaces, aliases go next prefixed with spaces |
||
| 125 | * |
||
| 126 | * @throws \Exception |
||
| 127 | */ |
||
| 128 | public function getSupportedMIMETypes(): array |
||
| 159 | } |
||
| 160 | |||
| 161 | /** |
||
| 162 | * Returns the available detectors |
||
| 163 | * |
||
| 164 | * @throws \Exception |
||
| 165 | */ |
||
| 166 | public function getAvailableDetectors(): array |
||
| 167 | { |
||
| 168 | $detectors = []; |
||
| 169 | |||
| 170 | $split = preg_split("/\n/", $this->request('detectors')); |
||
| 171 | |||
| 172 | $parent = null; |
||
| 173 | foreach($split as $line) |
||
| 174 | { |
||
| 175 | if(preg_match('/composite/i', $line)) |
||
| 176 | { |
||
| 177 | $parent = trim(preg_replace('/\(.+\):/', '', $line)); |
||
| 178 | $detectors[$parent] = ['children' => [], 'composite' => true, 'name' => $parent]; |
||
| 179 | } |
||
| 180 | else |
||
| 181 | { |
||
| 182 | $child = trim($line); |
||
| 183 | $detectors[$parent]['children'][$child] = ['composite' => false, 'name' => $child]; |
||
| 184 | } |
||
| 185 | } |
||
| 186 | |||
| 187 | return $detectors; |
||
| 188 | } |
||
| 189 | |||
| 190 | /** |
||
| 191 | * Returns the available parsers |
||
| 192 | * |
||
| 193 | * @throws \Exception |
||
| 194 | */ |
||
| 195 | public function getAvailableParsers(): array |
||
| 196 | { |
||
| 197 | $parsers = []; |
||
| 198 | |||
| 199 | $split = preg_split("/\n/", $this->request('parsers')); |
||
| 200 | array_shift($split); |
||
|
Bug
introduced
by
|
|||
| 201 | |||
| 202 | $parent = null; |
||
| 203 | foreach($split as $line) |
||
| 204 | { |
||
| 205 | if(preg_match('/composite/i', $line)) |
||
| 206 | { |
||
| 207 | $parent = trim(preg_replace('/\(.+\):/', '', $line)); |
||
| 208 | |||
| 209 | $parsers[$parent] = ['children' => [], 'composite' => true, 'name' => $parent, 'decorated' => false]; |
||
| 210 | } |
||
| 211 | else |
||
| 212 | { |
||
| 213 | $child = trim($line); |
||
| 214 | |||
| 215 | $parsers[$parent]['children'][$child] = ['composite' => false, 'name' => $child, 'decorated' => false]; |
||
| 216 | } |
||
| 217 | } |
||
| 218 | |||
| 219 | return $parsers; |
||
| 220 | } |
||
| 221 | |||
| 222 | /** |
||
| 223 | * Check Java binary, JAR path or server connection |
||
| 224 | * |
||
| 225 | * @throws \Exception |
||
| 226 | */ |
||
| 227 | public function check(): void |
||
| 228 | { |
||
| 229 | if($this->isChecked() === false) |
||
| 230 | { |
||
| 231 | // Java command must not return an error |
||
| 232 | try |
||
| 233 | { |
||
| 234 | $this->exec(($this->java ?: 'java') . ' -version'); |
||
| 235 | } |
||
| 236 | catch(Exception $exception) |
||
| 237 | { |
||
| 238 | throw new Exception('Java command not found'); |
||
| 239 | } |
||
| 240 | |||
| 241 | // JAR path must exists |
||
| 242 | if(file_exists($this->path) === false) |
||
| 243 | { |
||
| 244 | throw new Exception('Apache Tika app JAR not found'); |
||
| 245 | } |
||
| 246 | |||
| 247 | $this->setChecked(true); |
||
| 248 | } |
||
| 249 | } |
||
| 250 | |||
| 251 | /** |
||
| 252 | * Configure and make a request and return its results |
||
| 253 | * |
||
| 254 | * @throws \Exception |
||
| 255 | */ |
||
| 256 | public function request(string $type, string $file = null): string |
||
| 257 | { |
||
| 258 | // check if not checked |
||
| 259 | $this->check(); |
||
| 260 | |||
| 261 | // check if is cached |
||
| 262 | if($file !== null && $this->isCached($type, $file)) |
||
| 263 | { |
||
| 264 | return $this->getCachedResponse($type, $file); |
||
| 265 | } |
||
| 266 | |||
| 267 | // command arguments |
||
| 268 | $arguments = $this->getArguments($type, $file); |
||
| 269 | |||
| 270 | // check the request |
||
| 271 | $file = $this->checkRequest($type, $file); |
||
| 272 | |||
| 273 | // add last argument |
||
| 274 | if($file) |
||
| 275 | { |
||
| 276 | $arguments[] = escapeshellarg($file); |
||
| 277 | } |
||
| 278 | |||
| 279 | // build command |
||
| 280 | $jar = escapeshellarg($this->path); |
||
| 281 | $command = trim(($this->java ?: 'java') . " -jar $jar " . implode(' ', $arguments) . " {$this->javaArgs}"); |
||
| 282 | |||
| 283 | // run command |
||
| 284 | $response = $this->exec($command); |
||
| 285 | |||
| 286 | // metadata response |
||
| 287 | if(in_array(preg_replace('/\/.+/', '', $type), ['meta', 'rmeta'])) |
||
| 288 | { |
||
| 289 | // fix for invalid? json returned only with images |
||
| 290 | $response = str_replace(basename($file) . '"}{', '", ', $response); |
||
| 291 | |||
| 292 | // on Windows, response must be encoded to UTF8 |
||
| 293 | $response = $this->platform == 'win' ? utf8_encode($response) : $response; |
||
| 294 | } |
||
| 295 | |||
| 296 | // cache certain responses |
||
| 297 | if($this->isCacheable($type)) |
||
| 298 | { |
||
| 299 | $this->cacheResponse($type, $response, $file); |
||
| 300 | } |
||
| 301 | |||
| 302 | return $response; |
||
| 303 | } |
||
| 304 | |||
| 305 | /** |
||
| 306 | * Run the command and return its results |
||
| 307 | * |
||
| 308 | * @throws \Exception |
||
| 309 | */ |
||
| 310 | public function exec(string $command): ?string |
||
| 311 | { |
||
| 312 | // run command |
||
| 313 | $exit = -1; |
||
| 314 | $logfile = sys_get_temp_dir() . DIRECTORY_SEPARATOR . 'tika-error.log'; |
||
| 315 | $descriptors = [['pipe', 'r'], ['pipe', 'w'], ['file', $logfile, 'a']]; |
||
| 316 | $process = proc_open($command, $descriptors, $pipes); |
||
| 317 | $callback = $this->callback; |
||
| 318 | |||
| 319 | // get output if command runs ok |
||
| 320 | if(is_resource($process)) |
||
| 321 | { |
||
| 322 | fclose($pipes[0]); |
||
| 323 | $this->response = ''; |
||
| 324 | while($chunk = stream_get_line($pipes[1], $this->chunkSize)) |
||
| 325 | { |
||
| 326 | if(!is_null($callback)) |
||
| 327 | { |
||
| 328 | $callback($chunk); |
||
| 329 | } |
||
| 330 | |||
| 331 | if($this->callbackAppend === true) |
||
| 332 | { |
||
| 333 | $this->response .= $chunk; |
||
| 334 | } |
||
| 335 | } |
||
| 336 | fclose($pipes[1]); |
||
| 337 | $exit = proc_close($process); |
||
| 338 | } |
||
| 339 | |||
| 340 | // exception if exit value is not zero |
||
| 341 | if($exit > 0) |
||
| 342 | { |
||
| 343 | throw new Exception("Unexpected exit value ($exit) for command $command"); |
||
| 344 | } |
||
| 345 | |||
| 346 | return trim($this->response); |
||
| 347 | } |
||
| 348 | |||
| 349 | /** |
||
| 350 | * Get the arguments to run the command |
||
| 351 | * |
||
| 352 | * @throws Exception |
||
| 353 | */ |
||
| 354 | protected function getArguments(string $type, string $file = null): array |
||
| 421 | } |
||
| 422 | } |
||
| 423 |
Loading history...