PageRenderTime 42ms CodeModel.GetById 14ms RepoModel.GetById 0ms app.codeStats 0ms

/Pdf/Resource/Font/CidFont.php

https://bitbucket.org/gkawka/zend-framework
PHP | 492 lines | 192 code | 54 blank | 246 comment | 34 complexity | 028e953229d84200c6eb990d31fbaeb1 MD5 | raw file
  1. <?php
  2. /**
  3. * Zend Framework
  4. *
  5. * LICENSE
  6. *
  7. * This source file is subject to the new BSD license that is bundled
  8. * with this package in the file LICENSE.txt.
  9. * It is also available through the world-wide-web at this URL:
  10. * http://framework.zend.com/license/new-bsd
  11. * If you did not receive a copy of the license and are unable to
  12. * obtain it through the world-wide-web, please send an email
  13. * to license@zend.com so we can send you a copy immediately.
  14. *
  15. * @category Zend
  16. * @package Zend_Pdf
  17. * @subpackage Fonts
  18. * @copyright Copyright (c) 2005-2012 Zend Technologies USA Inc. (http://www.zend.com)
  19. * @license http://framework.zend.com/license/new-bsd New BSD License
  20. * @version $Id: CidFont.php 24664 2012-02-26 16:36:51Z adamlundrigan $
  21. */
  22. /** Internally used classes */
  23. require_once 'Zend/Pdf/Element/Array.php';
  24. require_once 'Zend/Pdf/Element/Dictionary.php';
  25. require_once 'Zend/Pdf/Element/Name.php';
  26. require_once 'Zend/Pdf/Element/Numeric.php';
  27. require_once 'Zend/Pdf/Element/String.php';
  28. /** Zend_Pdf_Resource_Font */
  29. require_once 'Zend/Pdf/Resource/Font.php';
  30. /**
  31. * Adobe PDF CIDFont font object implementation
  32. *
  33. * A CIDFont program contains glyph descriptions that are accessed using a CID as
  34. * the character selector. There are two types of CIDFont. A Type 0 CIDFont contains
  35. * glyph descriptions based on Adobe’s Type 1 font format, whereas those in a
  36. * Type 2 CIDFont are based on the TrueType font format.
  37. *
  38. * A CIDFont dictionary is a PDF object that contains information about a CIDFont program.
  39. * Although its Type value is Font, a CIDFont is not actually a font. It does not have an Encoding
  40. * entry, it cannot be listed in the Font subdictionary of a resource dictionary, and it cannot be
  41. * used as the operand of the Tf operator. It is used only as a descendant of a Type 0 font.
  42. * The CMap in the Type 0 font is what defines the encoding that maps character codes to CIDs
  43. * in the CIDFont.
  44. *
  45. * Font objects should be normally be obtained from the factory methods
  46. * {@link Zend_Pdf_Font::fontWithName} and {@link Zend_Pdf_Font::fontWithPath}.
  47. *
  48. * @package Zend_Pdf
  49. * @subpackage Fonts
  50. * @copyright Copyright (c) 2005-2012 Zend Technologies USA Inc. (http://www.zend.com)
  51. * @license http://framework.zend.com/license/new-bsd New BSD License
  52. */
  53. abstract class Zend_Pdf_Resource_Font_CidFont extends Zend_Pdf_Resource_Font
  54. {
  55. /**
  56. * Object representing the font's cmap (character to glyph map).
  57. * @var Zend_Pdf_Cmap
  58. */
  59. protected $_cmap = null;
  60. /**
  61. * Array containing the widths of each character that have entries in used character map.
  62. *
  63. * @var array
  64. */
  65. protected $_charWidths = null;
  66. /**
  67. * Width for characters missed in the font
  68. *
  69. * @var integer
  70. */
  71. protected $_missingCharWidth = 0;
  72. /**
  73. * Object constructor
  74. *
  75. * @param Zend_Pdf_FileParser_Font_OpenType $fontParser Font parser object
  76. * containing OpenType file.
  77. * @param integer $embeddingOptions Options for font embedding.
  78. * @throws Zend_Pdf_Exception
  79. */
  80. public function __construct(Zend_Pdf_FileParser_Font_OpenType $fontParser)
  81. {
  82. parent::__construct();
  83. $fontParser->parse();
  84. /* Object properties */
  85. $this->_fontNames = $fontParser->names;
  86. $this->_isBold = $fontParser->isBold;
  87. $this->_isItalic = $fontParser->isItalic;
  88. $this->_isMonospaced = $fontParser->isMonospaced;
  89. $this->_underlinePosition = $fontParser->underlinePosition;
  90. $this->_underlineThickness = $fontParser->underlineThickness;
  91. $this->_strikePosition = $fontParser->strikePosition;
  92. $this->_strikeThickness = $fontParser->strikeThickness;
  93. $this->_unitsPerEm = $fontParser->unitsPerEm;
  94. $this->_ascent = $fontParser->ascent;
  95. $this->_descent = $fontParser->descent;
  96. $this->_lineGap = $fontParser->lineGap;
  97. $this->_cmap = $fontParser->cmap;
  98. /* Resource dictionary */
  99. $baseFont = $this->getFontName(Zend_Pdf_Font::NAME_POSTSCRIPT, 'en', 'UTF-8');
  100. $this->_resource->BaseFont = new Zend_Pdf_Element_Name($baseFont);
  101. /**
  102. * Prepare widths array.
  103. */
  104. /* Constract characters widths array using font CMap and glyphs widths array */
  105. $glyphWidths = $fontParser->glyphWidths;
  106. $charGlyphs = $this->_cmap->getCoveredCharactersGlyphs();
  107. $charWidths = array();
  108. foreach ($charGlyphs as $charCode => $glyph) {
  109. if(isset($glyphWidths[$glyph]) && !is_null($glyphWidths[$glyph])) {
  110. $charWidths[$charCode] = $glyphWidths[$glyph];
  111. }
  112. }
  113. $this->_charWidths = $charWidths;
  114. $this->_missingCharWidth = $glyphWidths[0];
  115. /* Width array optimization. Step1: extract default value */
  116. $widthFrequencies = array_count_values($charWidths);
  117. $defaultWidth = null;
  118. $defaultWidthFrequency = -1;
  119. foreach ($widthFrequencies as $width => $frequency) {
  120. if ($frequency > $defaultWidthFrequency) {
  121. $defaultWidth = $width;
  122. $defaultWidthFrequency = $frequency;
  123. }
  124. }
  125. // Store default value in the font dictionary
  126. $this->_resource->DW = new Zend_Pdf_Element_Numeric($this->toEmSpace($defaultWidth));
  127. // Remove characters which corresponds to default width from the widths array
  128. $defWidthChars = array_keys($charWidths, $defaultWidth);
  129. foreach ($defWidthChars as $charCode) {
  130. unset($charWidths[$charCode]);
  131. }
  132. // Order cheracter widths aray by character codes
  133. ksort($charWidths, SORT_NUMERIC);
  134. /* Width array optimization. Step2: Compact character codes sequences */
  135. $lastCharCode = -1;
  136. $widthsSequences = array();
  137. foreach ($charWidths as $charCode => $width) {
  138. if ($lastCharCode == -1) {
  139. $charCodesSequense = array();
  140. $sequenceStartCode = $charCode;
  141. } else if ($charCode != $lastCharCode + 1) {
  142. // New chracters sequence detected
  143. $widthsSequences[$sequenceStartCode] = $charCodesSequense;
  144. $charCodesSequense = array();
  145. $sequenceStartCode = $charCode;
  146. }
  147. $charCodesSequense[] = $width;
  148. $lastCharCode = $charCode;
  149. }
  150. // Save last sequence, if widths array is not empty (it may happens for monospaced fonts)
  151. if (count($charWidths) != 0) {
  152. $widthsSequences[$sequenceStartCode] = $charCodesSequense;
  153. }
  154. $pdfCharsWidths = array();
  155. foreach ($widthsSequences as $startCode => $widthsSequence) {
  156. /* Width array optimization. Step3: Compact widths sequences */
  157. $pdfWidths = array();
  158. $lastWidth = -1;
  159. $widthsInSequence = 0;
  160. foreach ($widthsSequence as $width) {
  161. if ($lastWidth != $width) {
  162. // New width is detected
  163. if ($widthsInSequence != 0) {
  164. // Previous width value was a part of the widths sequence. Save it as 'c_1st c_last w'.
  165. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($startCode); // First character code
  166. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($startCode + $widthsInSequence - 1); // Last character code
  167. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($this->toEmSpace($lastWidth)); // Width
  168. // Reset widths sequence
  169. $startCode = $startCode + $widthsInSequence;
  170. $widthsInSequence = 0;
  171. }
  172. // Collect new width
  173. $pdfWidths[] = new Zend_Pdf_Element_Numeric($this->toEmSpace($width));
  174. $lastWidth = $width;
  175. } else {
  176. // Width is equal to previous
  177. if (count($pdfWidths) != 0) {
  178. // We already have some widths collected
  179. // So, we've just detected new widths sequence
  180. // Remove last element from widths list, since it's a part of widths sequence
  181. array_pop($pdfWidths);
  182. // and write the rest if it's not empty
  183. if (count($pdfWidths) != 0) {
  184. // Save it as 'c_1st [w1 w2 ... wn]'.
  185. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($startCode); // First character code
  186. $pdfCharsWidths[] = new Zend_Pdf_Element_Array($pdfWidths); // Widths array
  187. // Reset widths collection
  188. $startCode += count($pdfWidths);
  189. $pdfWidths = array();
  190. }
  191. $widthsInSequence = 2;
  192. } else {
  193. // Continue widths sequence
  194. $widthsInSequence++;
  195. }
  196. }
  197. }
  198. // Check if we have widths collection or widths sequence to wite it down
  199. if (count($pdfWidths) != 0) {
  200. // We have some widths collected
  201. // Save it as 'c_1st [w1 w2 ... wn]'.
  202. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($startCode); // First character code
  203. $pdfCharsWidths[] = new Zend_Pdf_Element_Array($pdfWidths); // Widths array
  204. } else if ($widthsInSequence != 0){
  205. // We have widths sequence
  206. // Save it as 'c_1st c_last w'.
  207. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($startCode); // First character code
  208. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($startCode + $widthsInSequence - 1); // Last character code
  209. $pdfCharsWidths[] = new Zend_Pdf_Element_Numeric($this->toEmSpace($lastWidth)); // Width
  210. }
  211. }
  212. /* Create the Zend_Pdf_Element_Array object and add it to the font's
  213. * object factory and resource dictionary.
  214. */
  215. $widthsArrayElement = new Zend_Pdf_Element_Array($pdfCharsWidths);
  216. $widthsObject = $this->_objectFactory->newObject($widthsArrayElement);
  217. $this->_resource->W = $widthsObject;
  218. /* CIDSystemInfo dictionary */
  219. $cidSystemInfo = new Zend_Pdf_Element_Dictionary();
  220. $cidSystemInfo->Registry = new Zend_Pdf_Element_String('Adobe');
  221. $cidSystemInfo->Ordering = new Zend_Pdf_Element_String('UCS');
  222. $cidSystemInfo->Supplement = new Zend_Pdf_Element_Numeric(0);
  223. $cidSystemInfoObject = $this->_objectFactory->newObject($cidSystemInfo);
  224. $this->_resource->CIDSystemInfo = $cidSystemInfoObject;
  225. }
  226. /**
  227. * Returns an array of glyph numbers corresponding to the Unicode characters.
  228. *
  229. * If a particular character doesn't exist in this font, the special 'missing
  230. * character glyph' will be substituted.
  231. *
  232. * See also {@link glyphNumberForCharacter()}.
  233. *
  234. * @param array $characterCodes Array of Unicode character codes (code points).
  235. * @return array Array of glyph numbers.
  236. */
  237. public function glyphNumbersForCharacters($characterCodes)
  238. {
  239. /**
  240. * CIDFont object is not actually a font. It does not have an Encoding entry,
  241. * it cannot be listed in the Font subdictionary of a resource dictionary, and
  242. * it cannot be used as the operand of the Tf operator.
  243. *
  244. * Throw an exception.
  245. */
  246. require_once 'Zend/Pdf/Exception.php';
  247. throw new Zend_Pdf_Exception('CIDFont PDF objects could not be used as the operand of the text drawing operators');
  248. }
  249. /**
  250. * Returns the glyph number corresponding to the Unicode character.
  251. *
  252. * If a particular character doesn't exist in this font, the special 'missing
  253. * character glyph' will be substituted.
  254. *
  255. * See also {@link glyphNumbersForCharacters()} which is optimized for bulk
  256. * operations.
  257. *
  258. * @param integer $characterCode Unicode character code (code point).
  259. * @return integer Glyph number.
  260. */
  261. public function glyphNumberForCharacter($characterCode)
  262. {
  263. /**
  264. * CIDFont object is not actually a font. It does not have an Encoding entry,
  265. * it cannot be listed in the Font subdictionary of a resource dictionary, and
  266. * it cannot be used as the operand of the Tf operator.
  267. *
  268. * Throw an exception.
  269. */
  270. require_once 'Zend/Pdf/Exception.php';
  271. throw new Zend_Pdf_Exception('CIDFont PDF objects could not be used as the operand of the text drawing operators');
  272. }
  273. /**
  274. * Returns a number between 0 and 1 inclusive that indicates the percentage
  275. * of characters in the string which are covered by glyphs in this font.
  276. *
  277. * Since no one font will contain glyphs for the entire Unicode character
  278. * range, this method can be used to help locate a suitable font when the
  279. * actual contents of the string are not known.
  280. *
  281. * Note that some fonts lie about the characters they support. Additionally,
  282. * fonts don't usually contain glyphs for control characters such as tabs
  283. * and line breaks, so it is rare that you will get back a full 1.0 score.
  284. * The resulting value should be considered informational only.
  285. *
  286. * @param string $string
  287. * @param string $charEncoding (optional) Character encoding of source text.
  288. * If omitted, uses 'current locale'.
  289. * @return float
  290. */
  291. public function getCoveredPercentage($string, $charEncoding = '')
  292. {
  293. /* Convert the string to UTF-16BE encoding so we can match the string's
  294. * character codes to those found in the cmap.
  295. */
  296. if ($charEncoding != 'UTF-16BE') {
  297. $string = iconv($charEncoding, 'UTF-16BE', $string);
  298. }
  299. $charCount = iconv_strlen($string, 'UTF-16BE');
  300. if ($charCount == 0) {
  301. return 0;
  302. }
  303. /* Calculate the score by doing a lookup for each character.
  304. */
  305. $score = 0;
  306. $maxIndex = strlen($string);
  307. for ($i = 0; $i < $maxIndex; $i++) {
  308. /**
  309. * @todo Properly handle characters encoded as surrogate pairs.
  310. */
  311. $charCode = (ord($string[$i]) << 8) | ord($string[++$i]);
  312. /* This could probably be optimized a bit with a binary search...
  313. */
  314. if (isset($this->_charWidths[$charCode])) {
  315. $score++;
  316. }
  317. }
  318. return $score / $charCount;
  319. }
  320. /**
  321. * Returns the widths of the Chars.
  322. *
  323. * The widths are expressed in the font's glyph space. You are responsible
  324. * for converting to user space as necessary. See {@link unitsPerEm()}.
  325. *
  326. * See also {@link widthForChar()}.
  327. *
  328. * @param array &$glyphNumbers Array of glyph numbers.
  329. * @return array Array of glyph widths (integers).
  330. */
  331. public function widthsForChars($charCodes)
  332. {
  333. $widths = array();
  334. foreach ($charCodes as $key => $charCode) {
  335. if (!isset($this->_charWidths[$charCode])) {
  336. $widths[$key] = $this->_missingCharWidth;
  337. } else {
  338. $widths[$key] = $this->_charWidths[$charCode];
  339. }
  340. }
  341. return $widths;
  342. }
  343. /**
  344. * Returns the width of the character.
  345. *
  346. * Like {@link widthsForChars()} but used for one char at a time.
  347. *
  348. * @param integer $charCode
  349. * @return integer
  350. */
  351. public function widthForChar($charCode)
  352. {
  353. if (!isset($this->_charWidths[$charCode])) {
  354. return $this->_missingCharWidth;
  355. }
  356. return $this->_charWidths[$charCode];
  357. }
  358. /**
  359. * Returns the widths of the glyphs.
  360. *
  361. * @param array &$glyphNumbers Array of glyph numbers.
  362. * @return array Array of glyph widths (integers).
  363. * @throws Zend_Pdf_Exception
  364. */
  365. public function widthsForGlyphs($glyphNumbers)
  366. {
  367. /**
  368. * CIDFont object is not actually a font. It does not have an Encoding entry,
  369. * it cannot be listed in the Font subdictionary of a resource dictionary, and
  370. * it cannot be used as the operand of the Tf operator.
  371. *
  372. * Throw an exception.
  373. */
  374. require_once 'Zend/Pdf/Exception.php';
  375. throw new Zend_Pdf_Exception('CIDFont PDF objects could not be used as the operand of the text drawing operators');
  376. }
  377. /**
  378. * Returns the width of the glyph.
  379. *
  380. * Like {@link widthsForGlyphs()} but used for one glyph at a time.
  381. *
  382. * @param integer $glyphNumber
  383. * @return integer
  384. * @throws Zend_Pdf_Exception
  385. */
  386. public function widthForGlyph($glyphNumber)
  387. {
  388. /**
  389. * CIDFont object is not actually a font. It does not have an Encoding entry,
  390. * it cannot be listed in the Font subdictionary of a resource dictionary, and
  391. * it cannot be used as the operand of the Tf operator.
  392. *
  393. * Throw an exception.
  394. */
  395. require_once 'Zend/Pdf/Exception.php';
  396. throw new Zend_Pdf_Exception('CIDFont PDF objects could not be used as the operand of the text drawing operators');
  397. }
  398. /**
  399. * Convert string to the font encoding.
  400. *
  401. * @param string $string
  402. * @param string $charEncoding Character encoding of source text.
  403. * @return string
  404. * @throws Zend_Pdf_Exception
  405. * */
  406. public function encodeString($string, $charEncoding)
  407. {
  408. /**
  409. * CIDFont object is not actually a font. It does not have an Encoding entry,
  410. * it cannot be listed in the Font subdictionary of a resource dictionary, and
  411. * it cannot be used as the operand of the Tf operator.
  412. *
  413. * Throw an exception.
  414. */
  415. require_once 'Zend/Pdf/Exception.php';
  416. throw new Zend_Pdf_Exception('CIDFont PDF objects could not be used as the operand of the text drawing operators');
  417. }
  418. /**
  419. * Convert string from the font encoding.
  420. *
  421. * @param string $string
  422. * @param string $charEncoding Character encoding of resulting text.
  423. * @return string
  424. * @throws Zend_Pdf_Exception
  425. */
  426. public function decodeString($string, $charEncoding)
  427. {
  428. /**
  429. * CIDFont object is not actually a font. It does not have an Encoding entry,
  430. * it cannot be listed in the Font subdictionary of a resource dictionary, and
  431. * it cannot be used as the operand of the Tf operator.
  432. *
  433. * Throw an exception.
  434. */
  435. require_once 'Zend/Pdf/Exception.php';
  436. throw new Zend_Pdf_Exception('CIDFont PDF objects could not be used as the operand of the text drawing operators');
  437. }
  438. }