chunk.php
13 years ago
config.php
13 years ago
download.php
13 years ago
helper.php
13 years ago
input.php
13 years ago
chunk.php
359 lines
| 1 | <?php |
| 2 | /** |
| 3 | * Chunk |
| 4 | * |
| 5 | * Reads a large file in as chunks for easier parsing. |
| 6 | * |
| 7 | * The chunks returned are whole <$this->options['element']/>s found within file. |
| 8 | * |
| 9 | * Each call to read() returns the whole element including start and end tags. |
| 10 | * |
| 11 | * Tested with a 1.8MB file, extracted 500 elements in 0.11s |
| 12 | * (with no work done, just extracting the elements) |
| 13 | * |
| 14 | * Usage: |
| 15 | * <code> |
| 16 | * // initialize the object |
| 17 | * $file = new Chunk('chunk-test.xml', array('element' => 'Chunk')); |
| 18 | * |
| 19 | * // loop through the file until all lines are read |
| 20 | * while ($xml = $file->read()) { |
| 21 | * // do whatever you want with the string |
| 22 | * $o = simplexml_load_string($xml); |
| 23 | * } |
| 24 | * </code> |
| 25 | * |
| 26 | * @package default |
| 27 | * @author Dom Hastings |
| 28 | */ |
| 29 | class PMXI_Chunk { |
| 30 | /** |
| 31 | * options |
| 32 | * |
| 33 | * @var array Contains all major options |
| 34 | * @access public |
| 35 | */ |
| 36 | public $options = array( |
| 37 | 'path' => './', // string The path to check for $file in |
| 38 | 'element' => '', // string The XML element to return |
| 39 | 'chunkSize' => 4096, // integer The amount of bytes to retrieve in each chunk |
| 40 | 'type' => 'upload' |
| 41 | ); |
| 42 | |
| 43 | /** |
| 44 | * file |
| 45 | * |
| 46 | * @var string The filename being read |
| 47 | * @access public |
| 48 | */ |
| 49 | public $file = ''; |
| 50 | /** |
| 51 | * pointer |
| 52 | * |
| 53 | * @var integer The current position the file is being read from |
| 54 | * @access public |
| 55 | */ |
| 56 | public $pointer = 0; |
| 57 | |
| 58 | public $cloud = array(); |
| 59 | |
| 60 | /** |
| 61 | * handle |
| 62 | * |
| 63 | * @var resource The fopen() resource |
| 64 | * @access private |
| 65 | */ |
| 66 | private $handle = null; |
| 67 | /** |
| 68 | * reading |
| 69 | * |
| 70 | * @var boolean Whether the script is currently reading the file |
| 71 | * @access private |
| 72 | */ |
| 73 | private $reading = false; |
| 74 | /** |
| 75 | * readBuffer |
| 76 | * |
| 77 | * @var string Used to make sure start tags aren't missed |
| 78 | * @access private |
| 79 | */ |
| 80 | private $readBuffer = ''; |
| 81 | |
| 82 | /** |
| 83 | * __construct |
| 84 | * |
| 85 | * Builds the Chunk object |
| 86 | * |
| 87 | * @param string $file The filename to work with |
| 88 | * @param array $options The options with which to parse the file |
| 89 | * @author Dom Hastings |
| 90 | * @access public |
| 91 | */ |
| 92 | public function __construct($file, $options = array(), $pointer = 0) { |
| 93 | // merge the options together |
| 94 | $this->options = array_merge($this->options, (is_array($options) ? $options : array())); |
| 95 | |
| 96 | // check that the path ends with a / |
| 97 | if (substr($this->options['path'], -1) != '/') { |
| 98 | $this->options['path'] .= '/'; |
| 99 | } |
| 100 | |
| 101 | // normalize the filename |
| 102 | $file_base = basename($file); |
| 103 | |
| 104 | // make sure chunkSize is an int |
| 105 | $this->options['chunkSize'] = intval($this->options['chunkSize']); |
| 106 | |
| 107 | // check it's valid |
| 108 | if ($this->options['chunkSize'] < 64) { |
| 109 | $this->options['chunkSize'] = 1024; |
| 110 | } |
| 111 | |
| 112 | $this->pointer = $pointer; |
| 113 | |
| 114 | // set the filename |
| 115 | $this->file = ($this->options['path'] != './') ? realpath($this->options['path'].$file_base) : $file; |
| 116 | |
| 117 | // check the file exists |
| 118 | if (!file_exists($this->file)){ |
| 119 | throw new Exception('File doesn\'t exist'); |
| 120 | } |
| 121 | |
| 122 | // open the file |
| 123 | $this->handle = fopen($this->file, 'rb'); |
| 124 | |
| 125 | // check the file opened successfully |
| 126 | if (!$this->handle) { |
| 127 | throw new Exception('Error opening file for reading'); |
| 128 | } |
| 129 | |
| 130 | } |
| 131 | |
| 132 | /** |
| 133 | * __destruct |
| 134 | * |
| 135 | * Cleans up |
| 136 | * |
| 137 | * @return void |
| 138 | * @author Dom Hastings |
| 139 | * @access public |
| 140 | */ |
| 141 | public function __destruct() { |
| 142 | // close the file resource |
| 143 | fclose($this->handle); |
| 144 | } |
| 145 | |
| 146 | /** |
| 147 | * read |
| 148 | * |
| 149 | * Reads the first available occurence of the XML element $this->options['element'] |
| 150 | * |
| 151 | * @return string The XML string from $this->file |
| 152 | * @author Dom Hastings |
| 153 | * @access public |
| 154 | */ |
| 155 | public function read() { |
| 156 | // check we have an element specified |
| 157 | if (!empty($this->options['element'])) { |
| 158 | // trim it |
| 159 | $element = trim($this->options['element']); |
| 160 | |
| 161 | } else { |
| 162 | $element = ''; |
| 163 | } |
| 164 | |
| 165 | // initialize the buffer |
| 166 | $buffer = false; |
| 167 | |
| 168 | // if the element is empty, then start auto detect root element tag name |
| 169 | if (empty($element)) { |
| 170 | // let the script know we're reading |
| 171 | $this->reading = true; |
| 172 | $founded_tags = array(); |
| 173 | // read in the whole doc, cos we don't know what's wanted |
| 174 | while ($this->reading) { |
| 175 | $c = fread($this->handle, $this->options['chunkSize']); |
| 176 | if ( preg_match_all("/<\\w+\\s*[^<]*\\s*\/?>/i", $c, $matches, PREG_PATTERN_ORDER) ){ |
| 177 | foreach ($matches[0] as $tag) { |
| 178 | $tag = explode(" ",trim(str_replace(array('<','>','/'), '', $tag))); |
| 179 | array_push($founded_tags, $tag[0]); |
| 180 | } |
| 181 | } |
| 182 | $this->reading = (!feof($this->handle)); |
| 183 | } |
| 184 | |
| 185 | // we must be looking for a specific element |
| 186 | } |
| 187 | |
| 188 | if (empty($element) and !empty($founded_tags)) { |
| 189 | |
| 190 | $element_counts = array_count_values($founded_tags); |
| 191 | |
| 192 | if (!empty($element_counts)){ |
| 193 | |
| 194 | $this->cloud = array_slice($element_counts, 0, 2); |
| 195 | |
| 196 | foreach ($element_counts as $tag => $count) { |
| 197 | if ($count > 1 and empty($this->options['element'])) { |
| 198 | $this->options['element'] = $element = $tag; |
| 199 | } |
| 200 | elseif ($count > 1){ |
| 201 | $this->cloud[$tag] = $count; |
| 202 | } |
| 203 | } |
| 204 | } |
| 205 | |
| 206 | } |
| 207 | |
| 208 | // return it all if element doesn't founded |
| 209 | if (empty($element)) |
| 210 | return false; |
| 211 | |
| 212 | // we must be looking for a specific element |
| 213 | //} |
| 214 | |
| 215 | // initialize the buffer |
| 216 | $buffer = false; |
| 217 | |
| 218 | // set up the strings to find |
| 219 | $open = '<'.$element; |
| 220 | $close = '</'.$element.'>'; |
| 221 | |
| 222 | // let the script know we're reading |
| 223 | $this->reading = true; |
| 224 | |
| 225 | // reset the global buffer |
| 226 | $this->readBuffer = ''; |
| 227 | |
| 228 | // this is used to ensure all data is read, and to make sure we don't send the start data again by mistake |
| 229 | $store = false; |
| 230 | |
| 231 | // seek to the position we need in the file |
| 232 | fseek($this->handle, $this->pointer); |
| 233 | |
| 234 | // start reading |
| 235 | while ($this->reading && !feof($this->handle)) { |
| 236 | |
| 237 | // store the chunk in a temporary variable |
| 238 | $tmp = fread($this->handle, $this->options['chunkSize']); |
| 239 | |
| 240 | // update the global buffer |
| 241 | $this->readBuffer .= $tmp; |
| 242 | |
| 243 | // check for the open string |
| 244 | $checkOpen = strpos($tmp, $open." "); |
| 245 | if (!$checkOpen) $checkOpen = strpos($tmp, $open.">"); |
| 246 | |
| 247 | // if it wasn't in the new buffer |
| 248 | if (!$checkOpen && !($store)) { |
| 249 | // check the full buffer (in case it was only half in this buffer) |
| 250 | $checkOpen = strpos($this->readBuffer, $open." "); |
| 251 | if (!$checkOpen) $checkOpen = strpos($this->readBuffer, $open.">"); |
| 252 | |
| 253 | // if it was in there |
| 254 | if ($checkOpen) { |
| 255 | // set it to the remainder |
| 256 | $checkOpen = $checkOpen % $this->options['chunkSize']; |
| 257 | } |
| 258 | } |
| 259 | |
| 260 | // check for the close string |
| 261 | $checkClose = strpos($tmp, $close); |
| 262 | $withoutcloseelement = false; |
| 263 | if (!$checkClose){ |
| 264 | $checkClose = (preg_match_all("/\/>\s*".$open."\s*/", $this->readBuffer, $matches)) ? strpos($this->readBuffer, $matches[0][0]) : false; |
| 265 | if ($checkClose) |
| 266 | $withoutcloseelement = true; |
| 267 | else{ |
| 268 | /*$checkClose = (preg_match_all("/\s*\/>\s*<\/", $this->readBuffer, $matches)) ? strpos($this->readBuffer, $matches[0][0]) : false; |
| 269 | if ($checkClose) |
| 270 | $withoutcloseelement = true;*/ |
| 271 | } |
| 272 | } |
| 273 | |
| 274 | // if it wasn't in the new buffer |
| 275 | if (!$checkClose && ($store)) { |
| 276 | // check the full buffer (in case it was only half in this buffer) |
| 277 | $checkClose = strpos($this->readBuffer, $close); |
| 278 | |
| 279 | $withoutcloseelement = false; |
| 280 | if (!$checkClose){ |
| 281 | $checkClose = (preg_match_all("/\/>\s*".$open."\s*/", $this->readBuffer, $matches)) ? strpos($this->readBuffer, $matches[0][0]) : false; |
| 282 | if ($checkClose) |
| 283 | $withoutcloseelement = true; |
| 284 | else{ |
| 285 | /*$checkClose = (preg_match_all("//>\\s*<\//", $this->readBuffer, $matches)) ? strpos($this->readBuffer, $matches[0][0]) : false; |
| 286 | if ($checkClose) |
| 287 | $withoutcloseelement = true;*/ |
| 288 | } |
| 289 | } |
| 290 | // if it was in there |
| 291 | if ($checkClose) { |
| 292 | // set it to the remainder plus the length of the close string itself |
| 293 | if (!$withoutcloseelement){ |
| 294 | $checkClose = ($checkClose + strlen($close)) % $this->options['chunkSize']; |
| 295 | }else{ |
| 296 | $checkClose = ($checkClose + strlen("/>")) % $this->options['chunkSize']; |
| 297 | } |
| 298 | } |
| 299 | |
| 300 | // if it was |
| 301 | } elseif ($checkClose) { |
| 302 | // add the length of the close string itself |
| 303 | if ( ! $withoutcloseelement) |
| 304 | $checkClose += strlen($close); |
| 305 | else |
| 306 | $checkClose += strlen("/>"); // "/>" symbols |
| 307 | } |
| 308 | |
| 309 | // if we've found the opening string and we're not already reading another element |
| 310 | if ($checkOpen !== false && !($store)) { |
| 311 | // if we're found the end element too |
| 312 | if ($checkClose !== false) { |
| 313 | // append the string only between the start and end element |
| 314 | $buffer .= substr($tmp, $checkOpen, ($checkClose - $checkOpen)); |
| 315 | |
| 316 | // update the pointer |
| 317 | $this->pointer += $checkClose; |
| 318 | |
| 319 | // let the script know we're done |
| 320 | $this->reading = false; |
| 321 | |
| 322 | } else { |
| 323 | // append the data we know to be part of this element |
| 324 | $buffer .= substr($tmp, $checkOpen); |
| 325 | |
| 326 | // update the pointer |
| 327 | $this->pointer += $this->options['chunkSize']; |
| 328 | |
| 329 | // let the script know we're gonna be storing all the data until we find the close element |
| 330 | $store = true; |
| 331 | } |
| 332 | |
| 333 | // if we've found the closing element |
| 334 | } elseif ($checkClose !== false) { |
| 335 | // update the buffer with the data upto and including the close tag |
| 336 | $buffer .= substr($tmp, 0, $checkClose); |
| 337 | |
| 338 | // update the pointer |
| 339 | $this->pointer += $checkClose; |
| 340 | |
| 341 | // let the script know we're done |
| 342 | $this->reading = false; |
| 343 | |
| 344 | // if we've found the closing element, but half in the previous chunk |
| 345 | } elseif ($store) { |
| 346 | // update the buffer |
| 347 | $buffer .= $tmp; |
| 348 | |
| 349 | // and the pointer |
| 350 | $this->pointer += $this->options['chunkSize']; |
| 351 | } |
| 352 | |
| 353 | } |
| 354 | |
| 355 | // return the element (or the whole file if we're not looking for elements) |
| 356 | return $buffer; |
| 357 | } |
| 358 | } |
| 359 |