| 1 |
<?php |
| 2 |
|
| 3 |
/** |
| 4 |
* This file is based on code of tecnickcom/TCPDF PDF library. |
| 5 |
* |
| 6 |
* Original author Nicola Asuni (info@tecnick.com) and |
| 7 |
* contributors (https://github.com/tecnickcom/TCPDF/graphs/contributors). |
| 8 |
* |
| 9 |
* @see https://github.com/tecnickcom/TCPDF |
| 10 |
* |
| 11 |
* Original code was licensed on the terms of the LGPL v3. |
| 12 |
* |
| 13 |
* ------------------------------------------------------------------------------ |
| 14 |
* |
| 15 |
* @file This file is part of the PdfParser library. |
| 16 |
* |
| 17 |
* @author Konrad Abicht <k.abicht@gmail.com> |
| 18 |
* |
| 19 |
* @date 2020-01-06 |
| 20 |
* |
| 21 |
* @license LGPLv3 |
| 22 |
* |
| 23 |
* @url <https://github.com/smalot/pdfparser> |
| 24 |
* |
| 25 |
* PdfParser is a pdf library written in PHP, extraction oriented. |
| 26 |
* Copyright (C) 2017 - Sébastien MALOT <sebastien@malot.fr> |
| 27 |
* |
| 28 |
* This program is free software: you can redistribute it and/or modify |
| 29 |
* it under the terms of the GNU Lesser General Public License as published by |
| 30 |
* the Free Software Foundation, either version 3 of the License, or |
| 31 |
* (at your option) any later version. |
| 32 |
* |
| 33 |
* This program is distributed in the hope that it will be useful, |
| 34 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 35 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 36 |
* GNU Lesser General Public License for more details. |
| 37 |
* |
| 38 |
* You should have received a copy of the GNU Lesser General Public License |
| 39 |
* along with this program. |
| 40 |
* If not, see <http://www.pdfparser.org/sites/default/LICENSE.txt>. |
| 41 |
*/ |
| 42 |
|
| 43 |
namespace Smalot\PdfParser\RawData; |
| 44 |
|
| 45 |
use Smalot\PdfParser\Config; |
| 46 |
use Smalot\PdfParser\Exception\EmptyPdfException; |
| 47 |
use Smalot\PdfParser\Exception\MissingPdfHeaderException; |
| 48 |
|
| 49 |
class RawDataParser |
| 50 |
{ |
| 51 |
/** |
| 52 |
* @var Config |
| 53 |
*/ |
| 54 |
private $config; |
| 55 |
|
| 56 |
/** |
| 57 |
* Configuration array. |
| 58 |
* |
| 59 |
* @var array<string,bool> |
| 60 |
*/ |
| 61 |
protected $cfg = [ |
| 62 |
// if `true` ignore filter decoding errors |
| 63 |
'ignore_filter_decoding_errors' => true, |
| 64 |
// if `true` ignore missing filter decoding errors |
| 65 |
'ignore_missing_filter_decoders' => true, |
| 66 |
]; |
| 67 |
|
| 68 |
protected $filterHelper; |
| 69 |
protected $objects; |
| 70 |
|
| 71 |
/** |
| 72 |
* @param array $cfg Configuration array, default is [] |
| 73 |
*/ |
| 74 |
public function __construct($cfg = [], ?Config $config = null) |
| 75 |
{ |
| 76 |
// merge given array with default values |
| 77 |
$this->cfg = array_merge($this->cfg, $cfg); |
| 78 |
|
| 79 |
$this->filterHelper = new FilterHelper(); |
| 80 |
$this->config = $config ?: new Config(); |
| 81 |
} |
| 82 |
|
| 83 |
/** |
| 84 |
* Decode the specified stream. |
| 85 |
* |
| 86 |
* @param string $pdfData PDF data |
| 87 |
* @param array $sdic Stream's dictionary array |
| 88 |
* @param string $stream Stream to decode |
| 89 |
* |
| 90 |
* @return array containing decoded stream data and remaining filters |
| 91 |
* |
| 92 |
* @throws \Exception |
| 93 |
*/ |
| 94 |
protected function decodeStream(string $pdfData, array $xref, array $sdic, string $stream): array |
| 95 |
{ |
| 96 |
// get stream length and filters |
| 97 |
$slength = \strlen($stream); |
| 98 |
if ($slength <= 0) { |
| 99 |
return ['', []]; |
| 100 |
} |
| 101 |
$filters = []; |
| 102 |
foreach ($sdic as $k => $v) { |
| 103 |
if ('/' == $v[0]) { |
| 104 |
if (('Length' == $v[1]) && (isset($sdic[$k + 1])) && ('numeric' == $sdic[$k + 1][0])) { |
| 105 |
// get declared stream length |
| 106 |
$declength = (int) $sdic[$k + 1][1]; |
| 107 |
if ($declength < $slength) { |
| 108 |
$stream = substr($stream, 0, $declength); |
| 109 |
$slength = $declength; |
| 110 |
} |
| 111 |
} elseif (('Filter' == $v[1]) && (isset($sdic[$k + 1]))) { |
| 112 |
// resolve indirect object |
| 113 |
$objval = $this->getObjectVal($pdfData, $xref, $sdic[$k + 1]); |
| 114 |
if ('/' == $objval[0]) { |
| 115 |
// single filter |
| 116 |
$filters[] = $objval[1]; |
| 117 |
} elseif ('[' == $objval[0]) { |
| 118 |
// array of filters |
| 119 |
foreach ($objval[1] as $flt) { |
| 120 |
if ('/' == $flt[0]) { |
| 121 |
$filters[] = $flt[1]; |
| 122 |
} |
| 123 |
} |
| 124 |
} |
| 125 |
} |
| 126 |
} |
| 127 |
} |
| 128 |
|
| 129 |
// decode the stream |
| 130 |
$remaining_filters = []; |
| 131 |
foreach ($filters as $filter) { |
| 132 |
if (\in_array($filter, $this->filterHelper->getAvailableFilters(), true)) { |
| 133 |
try { |
| 134 |
$stream = $this->filterHelper->decodeFilter($filter, $stream, $this->config->getDecodeMemoryLimit()); |
| 135 |
} catch (\Exception $e) { |
| 136 |
$emsg = $e->getMessage(); |
| 137 |
if ((('~' == $emsg[0]) && !$this->cfg['ignore_missing_filter_decoders']) |
| 138 |
|| (('~' != $emsg[0]) && !$this->cfg['ignore_filter_decoding_errors']) |
| 139 |
) { |
| 140 |
throw new \Exception($e->getMessage()); |
| 141 |
} |
| 142 |
} |
| 143 |
} else { |
| 144 |
// add missing filter to array |
| 145 |
$remaining_filters[] = $filter; |
| 146 |
} |
| 147 |
} |
| 148 |
|
| 149 |
return [$stream, $remaining_filters]; |
| 150 |
} |
| 151 |
|
| 152 |
/** |
| 153 |
* Decode the Cross-Reference section |
| 154 |
* |
| 155 |
* @param string $pdfData PDF data |
| 156 |
* @param int $startxref Offset at which the xref section starts (position of the 'xref' keyword) |
| 157 |
* @param array $xref Previous xref array (if any) |
| 158 |
* |
| 159 |
* @return array containing xref and trailer data |
| 160 |
* |
| 161 |
* @throws \Exception |
| 162 |
*/ |
| 163 |
protected function decodeXref(string $pdfData, int $startxref, array $xref = []): array |
| 164 |
{ |
| 165 |
$startxref += 4; // 4 is the length of the word 'xref' |
| 166 |
// skip initial white space chars |
| 167 |
$offset = $startxref + strspn($pdfData, $this->config->getPdfWhitespaces(), $startxref); |
| 168 |
// initialize object number |
| 169 |
$obj_num = 0; |
| 170 |
// search for cross-reference entries or subsection |
| 171 |
while (preg_match('/([0-9]+)[\x20]([0-9]+)[\x20]?([nf]?)(\r\n|[\x20]?[\r\n])/', $pdfData, $matches, \PREG_OFFSET_CAPTURE, $offset) > 0) { |
| 172 |
if ($matches[0][1] != $offset) { |
| 173 |
// we are on another section |
| 174 |
break; |
| 175 |
} |
| 176 |
$offset += \strlen($matches[0][0]); |
| 177 |
if ('n' == $matches[3][0]) { |
| 178 |
// create unique object index: [object number]_[generation number] |
| 179 |
$index = $obj_num.'_'.(int) $matches[2][0]; |
| 180 |
// check if object already exist |
| 181 |
if (!isset($xref['xref'][$index])) { |
| 182 |
// store object offset position |
| 183 |
$xref['xref'][$index] = (int) $matches[1][0]; |
| 184 |
} |
| 185 |
++$obj_num; |
| 186 |
} elseif ('f' == $matches[3][0]) { |
| 187 |
++$obj_num; |
| 188 |
} else { |
| 189 |
// object number (index) |
| 190 |
$obj_num = (int) $matches[1][0]; |
| 191 |
} |
| 192 |
} |
| 193 |
// get trailer data |
| 194 |
if (preg_match('/trailer[\s]*<<(.*)>>/isU', $pdfData, $matches, \PREG_OFFSET_CAPTURE, $offset) > 0) { |
| 195 |
$trailer_data = $matches[1][0]; |
| 196 |
if (!isset($xref['trailer']) || empty($xref['trailer'])) { |
| 197 |
// get only the last updated version |
| 198 |
$xref['trailer'] = []; |
| 199 |
// parse trailer_data |
| 200 |
if (preg_match('/Size[\s]+([0-9]+)/i', $trailer_data, $matches) > 0) { |
| 201 |
$xref['trailer']['size'] = (int) $matches[1]; |
| 202 |
} |
| 203 |
if (preg_match('/Root[\s]+([0-9]+)[\s]+([0-9]+)[\s]+R/i', $trailer_data, $matches) > 0) { |
| 204 |
$xref['trailer']['root'] = (int) $matches[1].'_'.(int) $matches[2]; |
| 205 |
} |
| 206 |
if (preg_match('/Encrypt[\s]+([0-9]+)[\s]+([0-9]+)[\s]+R/i', $trailer_data, $matches) > 0) { |
| 207 |
$xref['trailer']['encrypt'] = (int) $matches[1].'_'.(int) $matches[2]; |
| 208 |
} |
| 209 |
if (preg_match('/Info[\s]+([0-9]+)[\s]+([0-9]+)[\s]+R/i', $trailer_data, $matches) > 0) { |
| 210 |
$xref['trailer']['info'] = (int) $matches[1].'_'.(int) $matches[2]; |
| 211 |
} |
| 212 |
if (preg_match('/ID[\s]*[\[][\s]*[<]([^>]*)[>][\s]*[<]([^>]*)[>]/i', $trailer_data, $matches) > 0) { |
| 213 |
$xref['trailer']['id'] = []; |
| 214 |
$xref['trailer']['id'][0] = $matches[1]; |
| 215 |
$xref['trailer']['id'][1] = $matches[2]; |
| 216 |
} |
| 217 |
} |
| 218 |
if (preg_match('/Prev[\s]+([0-9]+)/i', $trailer_data, $matches) > 0) { |
| 219 |
$offset = (int) $matches[1]; |
| 220 |
if (0 != $offset) { |
| 221 |
// get previous xref |
| 222 |
$xref = $this->getXrefData($pdfData, $offset, $xref); |
| 223 |
} |
| 224 |
} |
| 225 |
} else { |
| 226 |
throw new \Exception('Unable to find trailer'); |
| 227 |
} |
| 228 |
|
| 229 |
return $xref; |
| 230 |
} |
| 231 |
|
| 232 |
/** |
| 233 |
* Decode the Cross-Reference Stream section |
| 234 |
* |
| 235 |
* @param string $pdfData PDF data |
| 236 |
* @param int $startxref Offset at which the xref section starts |
| 237 |
* @param array $xref Previous xref array (if any) |
| 238 |
* |
| 239 |
* @return array containing xref and trailer data |
| 240 |
* |
| 241 |
* @throws \Exception if unknown PNG predictor detected |
| 242 |
*/ |
| 243 |
protected function decodeXrefStream(string $pdfData, int $startxref, array $xref = []): array |
| 244 |
{ |
| 245 |
// try to read Cross-Reference Stream |
| 246 |
$xrefobj = $this->getRawObject($pdfData, $startxref); |
| 247 |
$xrefcrs = $this->getIndirectObject($pdfData, $xref, $xrefobj[1], $startxref, true); |
| 248 |
if (!isset($xref['trailer']) || empty($xref['trailer'])) { |
| 249 |
// get only the last updated version |
| 250 |
$xref['trailer'] = []; |
| 251 |
$filltrailer = true; |
| 252 |
} else { |
| 253 |
$filltrailer = false; |
| 254 |
} |
| 255 |
if (!isset($xref['xref'])) { |
| 256 |
$xref['xref'] = []; |
| 257 |
} |
| 258 |
$valid_crs = false; |
| 259 |
$columns = 0; |
| 260 |
$predictor = null; |
| 261 |
$sarr = $xrefcrs[0][1]; |
| 262 |
if (!\is_array($sarr)) { |
| 263 |
$sarr = []; |
| 264 |
} |
| 265 |
|
| 266 |
$wb = []; |
| 267 |
|
| 268 |
foreach ($sarr as $k => $v) { |
| 269 |
if ( |
| 270 |
('/' == $v[0]) |
| 271 |
&& ('Type' == $v[1]) |
| 272 |
&& ( |
| 273 |
isset($sarr[$k + 1]) |
| 274 |
&& '/' == $sarr[$k + 1][0] |
| 275 |
&& 'XRef' == $sarr[$k + 1][1] |
| 276 |
) |
| 277 |
) { |
| 278 |
$valid_crs = true; |
| 279 |
} elseif (('/' == $v[0]) && ('Index' == $v[1]) && (isset($sarr[$k + 1]))) { |
| 280 |
// initialize list for: first object number in the subsection / number of objects |
| 281 |
$index_blocks = []; |
| 282 |
for ($m = 0; $m < \count($sarr[$k + 1][1]); $m += 2) { |
| 283 |
$index_blocks[] = [$sarr[$k + 1][1][$m][1], $sarr[$k + 1][1][$m + 1][1]]; |
| 284 |
} |
| 285 |
} elseif (('/' == $v[0]) && ('Prev' == $v[1]) && (isset($sarr[$k + 1]) && ('numeric' == $sarr[$k + 1][0]))) { |
| 286 |
// get previous xref offset |
| 287 |
$prevxref = (int) $sarr[$k + 1][1]; |
| 288 |
} elseif (('/' == $v[0]) && ('W' == $v[1]) && (isset($sarr[$k + 1]))) { |
| 289 |
// number of bytes (in the decoded stream) of the corresponding field |
| 290 |
$wb[0] = (int) $sarr[$k + 1][1][0][1]; |
| 291 |
$wb[1] = (int) $sarr[$k + 1][1][1][1]; |
| 292 |
$wb[2] = (int) $sarr[$k + 1][1][2][1]; |
| 293 |
} elseif (('/' == $v[0]) && ('DecodeParms' == $v[1]) && (isset($sarr[$k + 1][1]))) { |
| 294 |
$decpar = $sarr[$k + 1][1]; |
| 295 |
foreach ($decpar as $kdc => $vdc) { |
| 296 |
if ( |
| 297 |
'/' == $vdc[0] |
| 298 |
&& 'Columns' == $vdc[1] |
| 299 |
&& ( |
| 300 |
isset($decpar[$kdc + 1]) |
| 301 |
&& 'numeric' == $decpar[$kdc + 1][0] |
| 302 |
) |
| 303 |
) { |
| 304 |
$columns = (int) $decpar[$kdc + 1][1]; |
| 305 |
} elseif ( |
| 306 |
'/' == $vdc[0] |
| 307 |
&& 'Predictor' == $vdc[1] |
| 308 |
&& ( |
| 309 |
isset($decpar[$kdc + 1]) |
| 310 |
&& 'numeric' == $decpar[$kdc + 1][0] |
| 311 |
) |
| 312 |
) { |
| 313 |
$predictor = (int) $decpar[$kdc + 1][1]; |
| 314 |
} |
| 315 |
} |
| 316 |
} elseif ($filltrailer) { |
| 317 |
if (('/' == $v[0]) && ('Size' == $v[1]) && (isset($sarr[$k + 1]) && ('numeric' == $sarr[$k + 1][0]))) { |
| 318 |
$xref['trailer']['size'] = $sarr[$k + 1][1]; |
| 319 |
} elseif (('/' == $v[0]) && ('Root' == $v[1]) && (isset($sarr[$k + 1]) && ('objref' == $sarr[$k + 1][0]))) { |
| 320 |
$xref['trailer']['root'] = $sarr[$k + 1][1]; |
| 321 |
} elseif (('/' == $v[0]) && ('Info' == $v[1]) && (isset($sarr[$k + 1]) && ('objref' == $sarr[$k + 1][0]))) { |
| 322 |
$xref['trailer']['info'] = $sarr[$k + 1][1]; |
| 323 |
} elseif (('/' == $v[0]) && ('Encrypt' == $v[1]) && (isset($sarr[$k + 1]) && ('objref' == $sarr[$k + 1][0]))) { |
| 324 |
$xref['trailer']['encrypt'] = $sarr[$k + 1][1]; |
| 325 |
} elseif (('/' == $v[0]) && ('ID' == $v[1]) && (isset($sarr[$k + 1]))) { |
| 326 |
$xref['trailer']['id'] = []; |
| 327 |
$xref['trailer']['id'][0] = $sarr[$k + 1][1][0][1]; |
| 328 |
$xref['trailer']['id'][1] = $sarr[$k + 1][1][1][1]; |
| 329 |
} |
| 330 |
} |
| 331 |
} |
| 332 |
|
| 333 |
// decode data |
| 334 |
if ($valid_crs && isset($xrefcrs[1][3][0])) { |
| 335 |
if (null !== $predictor) { |
| 336 |
// number of bytes in a row |
| 337 |
$rowlen = ($columns + 1); |
| 338 |
// convert the stream into an array of integers |
| 339 |
/** @var array<int> */ |
| 340 |
$sdata = unpack('C*', $xrefcrs[1][3][0]); |
| 341 |
// TODO: Handle the case when unpack returns false |
| 342 |
|
| 343 |
// split the rows |
| 344 |
$sdata = array_chunk($sdata, $rowlen); |
| 345 |
|
| 346 |
// initialize decoded array |
| 347 |
$ddata = []; |
| 348 |
// initialize first row with zeros |
| 349 |
$prev_row = array_fill(0, $rowlen, 0); |
| 350 |
// for each row apply PNG unpredictor |
| 351 |
foreach ($sdata as $k => $row) { |
| 352 |
// initialize new row |
| 353 |
$ddata[$k] = []; |
| 354 |
// get PNG predictor value |
| 355 |
$predictor = (10 + $row[0]); |
| 356 |
// for each byte on the row |
| 357 |
for ($i = 1; $i <= $columns; ++$i) { |
| 358 |
// new index |
| 359 |
$j = ($i - 1); |
| 360 |
$row_up = $prev_row[$j]; |
| 361 |
if (1 == $i) { |
| 362 |
$row_left = 0; |
| 363 |
$row_upleft = 0; |
| 364 |
} else { |
| 365 |
$row_left = $row[$i - 1]; |
| 366 |
$row_upleft = $prev_row[$j - 1]; |
| 367 |
} |
| 368 |
switch ($predictor) { |
| 369 |
case 10: // PNG prediction (on encoding, PNG None on all rows) |
| 370 |
$ddata[$k][$j] = $row[$i]; |
| 371 |
break; |
| 372 |
|
| 373 |
case 11: // PNG prediction (on encoding, PNG Sub on all rows) |
| 374 |
$ddata[$k][$j] = (($row[$i] + $row_left) & 0xFF); |
| 375 |
break; |
| 376 |
|
| 377 |
case 12: // PNG prediction (on encoding, PNG Up on all rows) |
| 378 |
$ddata[$k][$j] = (($row[$i] + $row_up) & 0xFF); |
| 379 |
break; |
| 380 |
|
| 381 |
case 13: // PNG prediction (on encoding, PNG Average on all rows) |
| 382 |
$ddata[$k][$j] = (($row[$i] + (($row_left + $row_up) / 2)) & 0xFF); |
| 383 |
break; |
| 384 |
|
| 385 |
case 14: // PNG prediction (on encoding, PNG Paeth on all rows) |
| 386 |
// initial estimate |
| 387 |
$p = ($row_left + $row_up - $row_upleft); |
| 388 |
// distances |
| 389 |
$pa = abs($p - $row_left); |
| 390 |
$pb = abs($p - $row_up); |
| 391 |
$pc = abs($p - $row_upleft); |
| 392 |
$pmin = min($pa, $pb, $pc); |
| 393 |
// return minimum distance |
| 394 |
switch ($pmin) { |
| 395 |
case $pa: |
| 396 |
$ddata[$k][$j] = (($row[$i] + $row_left) & 0xFF); |
| 397 |
break; |
| 398 |
|
| 399 |
case $pb: |
| 400 |
$ddata[$k][$j] = (($row[$i] + $row_up) & 0xFF); |
| 401 |
break; |
| 402 |
|
| 403 |
case $pc: |
| 404 |
$ddata[$k][$j] = (($row[$i] + $row_upleft) & 0xFF); |
| 405 |
break; |
| 406 |
} |
| 407 |
break; |
| 408 |
|
| 409 |
default: // PNG prediction (on encoding, PNG optimum) |
| 410 |
throw new \Exception('Unknown PNG predictor: '.$predictor); |
| 411 |
} |
| 412 |
} |
| 413 |
$prev_row = $ddata[$k]; |
| 414 |
} // end for each row |
| 415 |
// complete decoding |
| 416 |
} else { |
| 417 |
// number of bytes in a row |
| 418 |
$rowlen = array_sum($wb); |
| 419 |
if (0 < $rowlen) { |
| 420 |
// convert the stream into an array of integers |
| 421 |
$sdata = unpack('C*', $xrefcrs[1][3][0]); |
| 422 |
// split the rows |
| 423 |
$ddata = array_chunk($sdata, $rowlen); |
| 424 |
} else { |
| 425 |
// if the row length is zero, $ddata should be an empty array as well |
| 426 |
$ddata = []; |
| 427 |
} |
| 428 |
} |
| 429 |
|
| 430 |
$sdata = []; |
| 431 |
|
| 432 |
// for every row |
| 433 |
foreach ($ddata as $k => $row) { |
| 434 |
// initialize new row |
| 435 |
$sdata[$k] = [0, 0, 0]; |
| 436 |
if (0 == $wb[0]) { |
| 437 |
// default type field |
| 438 |
$sdata[$k][0] = 1; |
| 439 |
} |
| 440 |
$i = 0; // count bytes in the row |
| 441 |
// for every column |
| 442 |
for ($c = 0; $c < 3; ++$c) { |
| 443 |
// for every byte on the column |
| 444 |
for ($b = 0; $b < $wb[$c]; ++$b) { |
| 445 |
if (isset($row[$i])) { |
| 446 |
$sdata[$k][$c] += ($row[$i] << (($wb[$c] - 1 - $b) * 8)); |
| 447 |
} |
| 448 |
++$i; |
| 449 |
} |
| 450 |
} |
| 451 |
} |
| 452 |
|
| 453 |
// fill xref |
| 454 |
if (isset($index_blocks)) { |
| 455 |
// load the first object number of the first /Index entry |
| 456 |
$obj_num = $index_blocks[0][0]; |
| 457 |
} else { |
| 458 |
$obj_num = 0; |
| 459 |
} |
| 460 |
foreach ($sdata as $k => $row) { |
| 461 |
switch ($row[0]) { |
| 462 |
case 0: // (f) linked list of free objects |
| 463 |
break; |
| 464 |
|
| 465 |
case 1: // (n) objects that are in use but are not compressed |
| 466 |
// create unique object index: [object number]_[generation number] |
| 467 |
$index = $obj_num.'_'.$row[2]; |
| 468 |
// check if object already exist |
| 469 |
if (!isset($xref['xref'][$index])) { |
| 470 |
// store object offset position |
| 471 |
$xref['xref'][$index] = $row[1]; |
| 472 |
} |
| 473 |
break; |
| 474 |
|
| 475 |
case 2: // compressed objects |
| 476 |
// $row[1] = object number of the object stream in which this object is stored |
| 477 |
// $row[2] = index of this object within the object stream |
| 478 |
$index = $row[1].'_0_'.$row[2]; |
| 479 |
$xref['xref'][$index] = -1; |
| 480 |
break; |
| 481 |
|
| 482 |
default: // null objects |
| 483 |
break; |
| 484 |
} |
| 485 |
++$obj_num; |
| 486 |
if (isset($index_blocks)) { |
| 487 |
// reduce the number of remaining objects |
| 488 |
--$index_blocks[0][1]; |
| 489 |
if (0 == $index_blocks[0][1]) { |
| 490 |
// remove the actual used /Index entry |
| 491 |
array_shift($index_blocks); |
| 492 |
if (0 < \count($index_blocks)) { |
| 493 |
// load the first object number of the following /Index entry |
| 494 |
$obj_num = $index_blocks[0][0]; |
| 495 |
} else { |
| 496 |
// if there are no more entries, remove $index_blocks to avoid actions on an empty array |
| 497 |
unset($index_blocks); |
| 498 |
} |
| 499 |
} |
| 500 |
} |
| 501 |
} |
| 502 |
} // end decoding data |
| 503 |
if (isset($prevxref)) { |
| 504 |
// get previous xref |
| 505 |
$xref = $this->getXrefData($pdfData, $prevxref, $xref); |
| 506 |
} |
| 507 |
|
| 508 |
return $xref; |
| 509 |
} |
| 510 |
|
| 511 |
protected function getObjectHeaderPattern(array $objRefs): string |
| 512 |
{ |
| 513 |
// consider all whitespace character (PDF specifications) |
| 514 |
return '/'.$objRefs[0].$this->config->getPdfWhitespacesRegex().$objRefs[1].$this->config->getPdfWhitespacesRegex().'obj/'; |
| 515 |
} |
| 516 |
|
| 517 |
protected function getObjectHeaderLen(array $objRefs): int |
| 518 |
{ |
| 519 |
// "4 0 obj" |
| 520 |
// 2 whitespaces + strlen("obj") = 5 |
| 521 |
return 5 + \strlen($objRefs[0]) + \strlen($objRefs[1]); |
| 522 |
} |
| 523 |
|
| 524 |
/** |
| 525 |
* Get content of indirect object. |
| 526 |
* |
| 527 |
* @param string $pdfData PDF data |
| 528 |
* @param string $objRef Object number and generation number separated by underscore character |
| 529 |
* @param int $offset Object offset |
| 530 |
* @param bool $decoding If true decode streams |
| 531 |
* |
| 532 |
* @return array containing object data |
| 533 |
* |
| 534 |
* @throws \Exception if invalid object reference found |
| 535 |
*/ |
| 536 |
protected function getIndirectObject(string $pdfData, array $xref, string $objRef, int $offset = 0, bool $decoding = true): array |
| 537 |
{ |
| 538 |
/* |
| 539 |
* build indirect object header |
| 540 |
*/ |
| 541 |
// $objHeader = "[object number] [generation number] obj" |
| 542 |
$objRefArr = explode('_', $objRef); |
| 543 |
if (2 !== \count($objRefArr)) { |
| 544 |
throw new \Exception('Invalid object reference for $obj.'); |
| 545 |
} |
| 546 |
|
| 547 |
$objHeaderLen = $this->getObjectHeaderLen($objRefArr); |
| 548 |
|
| 549 |
/* |
| 550 |
* check if we are in position |
| 551 |
*/ |
| 552 |
// ignore whitespace characters at offset |
| 553 |
$offset += strspn($pdfData, $this->config->getPdfWhitespaces(), $offset); |
| 554 |
// ignore leading zeros for object number |
| 555 |
$offset += strspn($pdfData, '0', $offset); |
| 556 |
if (0 == preg_match($this->getObjectHeaderPattern($objRefArr), substr($pdfData, $offset, $objHeaderLen))) { |
| 557 |
// an indirect reference to an undefined object shall be considered a reference to the null object |
| 558 |
return ['null', 'null', $offset]; |
| 559 |
} |
| 560 |
|
| 561 |
/* |
| 562 |
* get content |
| 563 |
*/ |
| 564 |
// starting position of object content |
| 565 |
$offset += $objHeaderLen; |
| 566 |
$objContentArr = []; |
| 567 |
$i = 0; // object main index |
| 568 |
$header = null; |
| 569 |
do { |
| 570 |
$oldOffset = $offset; |
| 571 |
// get element |
| 572 |
$element = $this->getRawObject($pdfData, $offset, null != $header ? $header[1] : null); |
| 573 |
$offset = $element[2]; |
| 574 |
// decode stream using stream's dictionary information |
| 575 |
if ($decoding && ('stream' === $element[0]) && null != $header) { |
| 576 |
$element[3] = $this->decodeStream($pdfData, $xref, $header[1], $element[1]); |
| 577 |
} |
| 578 |
$objContentArr[$i] = $element; |
| 579 |
$header = isset($element[0]) && '<<' === $element[0] ? $element : null; |
| 580 |
++$i; |
| 581 |
} while (('endobj' !== $element[0]) && ($offset !== $oldOffset)); |
| 582 |
// remove closing delimiter |
| 583 |
array_pop($objContentArr); |
| 584 |
|
| 585 |
/* |
| 586 |
* return raw object content |
| 587 |
*/ |
| 588 |
return $objContentArr; |
| 589 |
} |
| 590 |
|
| 591 |
/** |
| 592 |
* Get the content of object, resolving indirect object reference if necessary. |
| 593 |
* |
| 594 |
* @param string $pdfData PDF data |
| 595 |
* @param array $obj Object value |
| 596 |
* |
| 597 |
* @return array containing object data |
| 598 |
* |
| 599 |
* @throws \Exception |
| 600 |
*/ |
| 601 |
protected function getObjectVal(string $pdfData, $xref, array $obj): array |
| 602 |
{ |
| 603 |
if ('objref' == $obj[0]) { |
| 604 |
// reference to indirect object |
| 605 |
if (isset($this->objects[$obj[1]])) { |
| 606 |
// this object has been already parsed |
| 607 |
return $this->objects[$obj[1]]; |
| 608 |
} elseif (isset($xref[$obj[1]])) { |
| 609 |
// parse new object |
| 610 |
$this->objects[$obj[1]] = $this->getIndirectObject($pdfData, $xref, $obj[1], $xref[$obj[1]], false); |
| 611 |
|
| 612 |
return $this->objects[$obj[1]]; |
| 613 |
} |
| 614 |
} |
| 615 |
|
| 616 |
return $obj; |
| 617 |
} |
| 618 |
|
| 619 |
/** |
| 620 |
* Get object type, raw value and offset to next object |
| 621 |
* |
| 622 |
* @param int $offset Object offset |
| 623 |
* @param array|null $headerDic obj header's dictionary, parsed by getRawObject. Used for stream parsing optimization |
| 624 |
* |
| 625 |
* @return array containing object type, raw value and offset to next object |
| 626 |
*/ |
| 627 |
protected function getRawObject(string $pdfData, int $offset = 0, ?array $headerDic = null): array |
| 628 |
{ |
| 629 |
$objtype = ''; // object type to be returned |
| 630 |
$objval = ''; // object value to be returned |
| 631 |
|
| 632 |
// skip initial white space chars |
| 633 |
$offset += strspn($pdfData, $this->config->getPdfWhitespaces(), $offset); |
| 634 |
|
| 635 |
// get first char |
| 636 |
$char = $pdfData[$offset]; |
| 637 |
// get object type |
| 638 |
switch ($char) { |
| 639 |
case '%': // \x25 PERCENT SIGN |
| 640 |
// skip comment and search for next token |
| 641 |
$next = strcspn($pdfData, "\r\n", $offset); |
| 642 |
if ($next > 0) { |
| 643 |
$offset += $next; |
| 644 |
|
| 645 |
return $this->getRawObject($pdfData, $offset); |
| 646 |
} |
| 647 |
break; |
| 648 |
|
| 649 |
case '/': // \x2F SOLIDUS |
| 650 |
// name object |
| 651 |
$objtype = $char; |
| 652 |
++$offset; |
| 653 |
$span = strcspn($pdfData, "\x00\x09\x0a\x0c\x0d\x20\n\t\r\v\f\x28\x29\x3c\x3e\x5b\x5d\x7b\x7d\x2f\x25", $offset, 256); |
| 654 |
if ($span > 0) { |
| 655 |
$objval = substr($pdfData, $offset, $span); // unescaped value |
| 656 |
$offset += $span; |
| 657 |
} |
| 658 |
break; |
| 659 |
|
| 660 |
case '(': // \x28 LEFT PARENTHESIS |
| 661 |
case ')': // \x29 RIGHT PARENTHESIS |
| 662 |
// literal string object |
| 663 |
$objtype = $char; |
| 664 |
++$offset; |
| 665 |
$strpos = $offset; |
| 666 |
if ('(' == $char) { |
| 667 |
$open_bracket = 1; |
| 668 |
while ($open_bracket > 0) { |
| 669 |
if (!isset($pdfData[$strpos])) { |
| 670 |
break; |
| 671 |
} |
| 672 |
$ch = $pdfData[$strpos]; |
| 673 |
switch ($ch) { |
| 674 |
case '\\': // REVERSE SOLIDUS (5Ch) (Backslash) |
| 675 |
// skip next character |
| 676 |
++$strpos; |
| 677 |
break; |
| 678 |
|
| 679 |
case '(': // LEFT PARENHESIS (28h) |
| 680 |
++$open_bracket; |
| 681 |
break; |
| 682 |
|
| 683 |
case ')': // RIGHT PARENTHESIS (29h) |
| 684 |
--$open_bracket; |
| 685 |
break; |
| 686 |
} |
| 687 |
++$strpos; |
| 688 |
} |
| 689 |
$objval = substr($pdfData, $offset, $strpos - $offset - 1); |
| 690 |
$offset = $strpos; |
| 691 |
} |
| 692 |
break; |
| 693 |
|
| 694 |
case '[': // \x5B LEFT SQUARE BRACKET |
| 695 |
case ']': // \x5D RIGHT SQUARE BRACKET |
| 696 |
// array object |
| 697 |
$objtype = $char; |
| 698 |
++$offset; |
| 699 |
if ('[' == $char) { |
| 700 |
// get array content |
| 701 |
$objval = []; |
| 702 |
do { |
| 703 |
$oldOffset = $offset; |
| 704 |
// get element |
| 705 |
$element = $this->getRawObject($pdfData, $offset); |
| 706 |
$offset = $element[2]; |
| 707 |
$objval[] = $element; |
| 708 |
} while ((']' != $element[0]) && ($offset != $oldOffset)); |
| 709 |
// remove closing delimiter |
| 710 |
array_pop($objval); |
| 711 |
} |
| 712 |
break; |
| 713 |
|
| 714 |
case '<': // \x3C LESS-THAN SIGN |
| 715 |
case '>': // \x3E GREATER-THAN SIGN |
| 716 |
if (isset($pdfData[$offset + 1]) && ($pdfData[$offset + 1] == $char)) { |
| 717 |
// dictionary object |
| 718 |
$objtype = $char.$char; |
| 719 |
$offset += 2; |
| 720 |
if ('<' == $char) { |
| 721 |
// get array content |
| 722 |
$objval = []; |
| 723 |
do { |
| 724 |
$oldOffset = $offset; |
| 725 |
// get element |
| 726 |
$element = $this->getRawObject($pdfData, $offset); |
| 727 |
$offset = $element[2]; |
| 728 |
$objval[] = $element; |
| 729 |
} while (('>>' != $element[0]) && ($offset != $oldOffset)); |
| 730 |
// remove closing delimiter |
| 731 |
array_pop($objval); |
| 732 |
} |
| 733 |
} else { |
| 734 |
// hexadecimal string object |
| 735 |
$objtype = $char; |
| 736 |
++$offset; |
| 737 |
|
| 738 |
$span = strspn($pdfData, "0123456789abcdefABCDEF\x09\x0a\x0c\x0d\x20", $offset); |
| 739 |
$dataToCheck = $pdfData[$offset + $span] ?? null; |
| 740 |
if ('<' == $char && $span > 0 && '>' == $dataToCheck) { |
| 741 |
// remove white space characters |
| 742 |
$objval = strtr(substr($pdfData, $offset, $span), $this->config->getPdfWhitespaces(), ''); |
| 743 |
$offset += $span + 1; |
| 744 |
} elseif (false !== ($endpos = strpos($pdfData, '>', $offset))) { |
| 745 |
$offset = $endpos + 1; |
| 746 |
} |
| 747 |
} |
| 748 |
break; |
| 749 |
|
| 750 |
default: |
| 751 |
if ('endobj' == substr($pdfData, $offset, 6)) { |
| 752 |
// indirect object |
| 753 |
$objtype = 'endobj'; |
| 754 |
$offset += 6; |
| 755 |
} elseif ('null' == substr($pdfData, $offset, 4)) { |
| 756 |
// null object |
| 757 |
$objtype = 'null'; |
| 758 |
$offset += 4; |
| 759 |
$objval = 'null'; |
| 760 |
} elseif ('true' == substr($pdfData, $offset, 4)) { |
| 761 |
// boolean true object |
| 762 |
$objtype = 'boolean'; |
| 763 |
$offset += 4; |
| 764 |
$objval = 'true'; |
| 765 |
} elseif ('false' == substr($pdfData, $offset, 5)) { |
| 766 |
// boolean false object |
| 767 |
$objtype = 'boolean'; |
| 768 |
$offset += 5; |
| 769 |
$objval = 'false'; |
| 770 |
} elseif ('stream' == substr($pdfData, $offset, 6)) { |
| 771 |
// start stream object |
| 772 |
$objtype = 'stream'; |
| 773 |
$offset += 6; |
| 774 |
if (1 == preg_match('/^( *[\r]?[\n])/isU', substr($pdfData, $offset, 4), $matches)) { |
| 775 |
$offset += \strlen($matches[0]); |
| 776 |
|
| 777 |
// we get stream length here to later help preg_match test less data |
| 778 |
$streamLen = (int) $this->getHeaderValue($headerDic, 'Length', 'numeric', 0); |
| 779 |
$skip = false === $this->config->getRetainImageContent() && 'XObject' == $this->getHeaderValue($headerDic, 'Type', '/') && 'Image' == $this->getHeaderValue($headerDic, 'Subtype', '/'); |
| 780 |
|
| 781 |
$pregResult = preg_match( |
| 782 |
'/(endstream)[\x09\x0a\x0c\x0d\x20]/isU', |
| 783 |
$pdfData, |
| 784 |
$matches, |
| 785 |
\PREG_OFFSET_CAPTURE, |
| 786 |
$offset + $streamLen |
| 787 |
); |
| 788 |
|
| 789 |
if (1 == $pregResult) { |
| 790 |
$objval = $skip ? '' : substr($pdfData, $offset, $matches[0][1] - $offset); |
| 791 |
$offset = $matches[1][1]; |
| 792 |
} |
| 793 |
} |
| 794 |
} elseif ('endstream' == substr($pdfData, $offset, 9)) { |
| 795 |
// end stream object |
| 796 |
$objtype = 'endstream'; |
| 797 |
$offset += 9; |
| 798 |
} elseif (1 == preg_match('/^([0-9]+)[\s]+([0-9]+)[\s]+R/iU', substr($pdfData, $offset, 33), $matches)) { |
| 799 |
// indirect object reference |
| 800 |
$objtype = 'objref'; |
| 801 |
$offset += \strlen($matches[0]); |
| 802 |
$objval = (int) $matches[1].'_'.(int) $matches[2]; |
| 803 |
} elseif (1 == preg_match('/^([0-9]+)[\s]+([0-9]+)[\s]+obj/iU', substr($pdfData, $offset, 33), $matches)) { |
| 804 |
// object start |
| 805 |
$objtype = 'obj'; |
| 806 |
$objval = (int) $matches[1].'_'.(int) $matches[2]; |
| 807 |
$offset += \strlen($matches[0]); |
| 808 |
} elseif (($numlen = strspn($pdfData, '+-.0123456789', $offset)) > 0) { |
| 809 |
// numeric object |
| 810 |
$objtype = 'numeric'; |
| 811 |
$objval = substr($pdfData, $offset, $numlen); |
| 812 |
$offset += $numlen; |
| 813 |
} |
| 814 |
break; |
| 815 |
} |
| 816 |
|
| 817 |
return [$objtype, $objval, $offset]; |
| 818 |
} |
| 819 |
|
| 820 |
/** |
| 821 |
* Get value of an object header's section (obj << YYY >> part ). |
| 822 |
* |
| 823 |
* It is similar to Header::get('...')->getContent(), the only difference is it can be used during the parsing process, |
| 824 |
* when no Smalot\PdfParser\Header objects are created yet. |
| 825 |
* |
| 826 |
* @param string $key header's section name |
| 827 |
* @param string $type type of the section (i.e. 'numeric', '/', '<<', etc.) |
| 828 |
* @param string|array|null $default default value for header's section |
| 829 |
* |
| 830 |
* @return string|array|null value of obj header's section, or default value if none found, or its type doesn't match $type param |
| 831 |
*/ |
| 832 |
private function getHeaderValue(?array $headerDic, string $key, string $type, $default = '') |
| 833 |
{ |
| 834 |
if (false === \is_array($headerDic)) { |
| 835 |
return $default; |
| 836 |
} |
| 837 |
|
| 838 |
/* |
| 839 |
* It recieves dictionary of header fields, as it is returned by RawDataParser::getRawObject, |
| 840 |
* iterates over it, searching for section of type '/' whith requested key. |
| 841 |
* If such a section is found, it tries to receive it's value (next object in dictionary), |
| 842 |
* returning it, if it matches requested type, or default value otherwise. |
| 843 |
*/ |
| 844 |
foreach ($headerDic as $i => $val) { |
| 845 |
$isSectionName = \is_array($val) && 3 == \count($val) && '/' == $val[0]; |
| 846 |
if ( |
| 847 |
$isSectionName |
| 848 |
&& $val[1] == $key |
| 849 |
&& isset($headerDic[$i + 1]) |
| 850 |
) { |
| 851 |
$isSectionValue = \is_array($headerDic[$i + 1]) && 1 < \count($headerDic[$i + 1]); |
| 852 |
|
| 853 |
return $isSectionValue && $type == $headerDic[$i + 1][0] |
| 854 |
? $headerDic[$i + 1][1] |
| 855 |
: $default; |
| 856 |
} |
| 857 |
} |
| 858 |
|
| 859 |
return $default; |
| 860 |
} |
| 861 |
|
| 862 |
/** |
| 863 |
* Get Cross-Reference (xref) table and trailer data from PDF document data. |
| 864 |
* |
| 865 |
* @param int $offset xref offset (if known) |
| 866 |
* @param array $xref previous xref array (if any) |
| 867 |
* |
| 868 |
* @return array containing xref and trailer data |
| 869 |
* |
| 870 |
* @throws \Exception if it was unable to find startxref |
| 871 |
* @throws \Exception if it was unable to find xref |
| 872 |
*/ |
| 873 |
protected function getXrefData(string $pdfData, int $offset = 0, array $xref = []): array |
| 874 |
{ |
| 875 |
// If the $offset is currently pointed at whitespace, bump it |
| 876 |
// forward until it isn't; affects loosely targetted offsets |
| 877 |
// for the 'xref' keyword |
| 878 |
// See: https://github.com/smalot/pdfparser/issues/673 |
| 879 |
$bumpOffset = $offset; |
| 880 |
while (preg_match('/\s/', substr($pdfData, $bumpOffset, 1))) { |
| 881 |
++$bumpOffset; |
| 882 |
} |
| 883 |
|
| 884 |
// Find all startxref tables from this $offset forward |
| 885 |
$startxrefPreg = preg_match_all( |
| 886 |
'/(?<=[\r\n])startxref[\s]*[\r\n]+([0-9]+)[\s]*[\r\n]+%%EOF/i', |
| 887 |
$pdfData, |
| 888 |
$startxrefMatches, |
| 889 |
\PREG_SET_ORDER, |
| 890 |
$offset |
| 891 |
); |
| 892 |
|
| 893 |
if (0 == $startxrefPreg) { |
| 894 |
// No startxref tables were found |
| 895 |
throw new \Exception('Unable to find startxref'); |
| 896 |
} elseif (0 == $offset) { |
| 897 |
// Use the last startxref in the document |
| 898 |
$startxref = (int) $startxrefMatches[\count($startxrefMatches) - 1][1]; |
| 899 |
} elseif (strpos($pdfData, 'xref', $bumpOffset) == $bumpOffset) { |
| 900 |
// Already pointing at the xref table |
| 901 |
$startxref = $bumpOffset; |
| 902 |
} elseif (preg_match('/([0-9]+[\s][0-9]+[\s]obj)/i', $pdfData, $matches, 0, $bumpOffset)) { |
| 903 |
// Cross-Reference Stream object |
| 904 |
$startxref = $bumpOffset; |
| 905 |
} else { |
| 906 |
// Use the next startxref from this $offset |
| 907 |
$startxref = (int) $startxrefMatches[0][1]; |
| 908 |
} |
| 909 |
|
| 910 |
if ($startxref > \strlen($pdfData)) { |
| 911 |
throw new \Exception('Unable to find xref (PDF corrupted?)'); |
| 912 |
} |
| 913 |
|
| 914 |
// check xref position |
| 915 |
if (strpos($pdfData, 'xref', $startxref) == $startxref) { |
| 916 |
// Cross-Reference |
| 917 |
$xref = $this->decodeXref($pdfData, $startxref, $xref); |
| 918 |
} else { |
| 919 |
// Check if the $pdfData might have the wrong line-endings |
| 920 |
$pdfDataUnix = str_replace("\r\n", "\n", $pdfData); |
| 921 |
if ($startxref < \strlen($pdfDataUnix) && strpos($pdfDataUnix, 'xref', $startxref) == $startxref) { |
| 922 |
// Return Unix-line-ending flag |
| 923 |
$xref = ['Unix' => true]; |
| 924 |
} else { |
| 925 |
// Cross-Reference Stream |
| 926 |
$xref = $this->decodeXrefStream($pdfData, $startxref, $xref); |
| 927 |
} |
| 928 |
} |
| 929 |
if (empty($xref)) { |
| 930 |
throw new \Exception('Unable to find xref'); |
| 931 |
} |
| 932 |
|
| 933 |
return $xref; |
| 934 |
} |
| 935 |
|
| 936 |
/** |
| 937 |
* Parses PDF data and returns extracted data as array. |
| 938 |
* |
| 939 |
* @param string $data PDF data to parse |
| 940 |
* |
| 941 |
* @return array array of parsed PDF document objects |
| 942 |
* |
| 943 |
* @throws EmptyPdfException if empty PDF data given |
| 944 |
* @throws MissingPdfHeaderException if PDF data missing `%PDF-` header |
| 945 |
*/ |
| 946 |
public function parseData(string $data): array |
| 947 |
{ |
| 948 |
if (empty($data)) { |
| 949 |
throw new EmptyPdfException('Empty PDF data given.'); |
| 950 |
} |
| 951 |
// find the pdf header starting position |
| 952 |
if (false === ($trimpos = strpos($data, '%PDF-'))) { |
| 953 |
throw new MissingPdfHeaderException('Invalid PDF data: Missing `%PDF-` header.'); |
| 954 |
} |
| 955 |
|
| 956 |
// get PDF content string |
| 957 |
$pdfData = $trimpos > 0 ? substr($data, $trimpos) : $data; |
| 958 |
|
| 959 |
// get xref and trailer data |
| 960 |
$xref = $this->getXrefData($pdfData); |
| 961 |
|
| 962 |
// If we found Unix line-endings |
| 963 |
if (isset($xref['Unix'])) { |
| 964 |
$pdfData = str_replace("\r\n", "\n", $pdfData); |
| 965 |
$xref = $this->getXrefData($pdfData); |
| 966 |
} |
| 967 |
|
| 968 |
// parse all document objects |
| 969 |
$objects = []; |
| 970 |
foreach ($xref['xref'] as $obj => $offset) { |
| 971 |
if (!isset($objects[$obj]) && ($offset > 0)) { |
| 972 |
// decode objects with positive offset |
| 973 |
$objects[$obj] = $this->getIndirectObject($pdfData, $xref, $obj, $offset, true); |
| 974 |
} |
| 975 |
} |
| 976 |
|
| 977 |
return [$xref, $objects]; |
| 978 |
} |
| 979 |
} |
| 980 |
|
