1 | <?php |
---|
2 | /** |
---|
3 | * DOMPDF - PHP5 HTML to PDF renderer |
---|
4 | * |
---|
5 | * File: $RCSfile: dompdf.cls.php,v $ |
---|
6 | * Created on: 2004-06-09 |
---|
7 | * |
---|
8 | * Copyright (c) 2004 - Benj Carson <benjcarson@digitaljunkies.ca> |
---|
9 | * |
---|
10 | * This library is free software; you can redistribute it and/or |
---|
11 | * modify it under the terms of the GNU Lesser General Public |
---|
12 | * License as published by the Free Software Foundation; either |
---|
13 | * version 2.1 of the License, or (at your option) any later version. |
---|
14 | * |
---|
15 | * This library is distributed in the hope that it will be useful, |
---|
16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
---|
18 | * Lesser General Public License for more details. |
---|
19 | * |
---|
20 | * You should have received a copy of the GNU Lesser General Public License |
---|
21 | * along with this library in the file LICENSE.LGPL; if not, write to the |
---|
22 | * Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA |
---|
23 | * 02111-1307 USA |
---|
24 | * |
---|
25 | * Alternatively, you may distribute this software under the terms of the |
---|
26 | * PHP License, version 3.0 or later. A copy of this license should have |
---|
27 | * been distributed with this file in the file LICENSE.PHP . If this is not |
---|
28 | * the case, you can obtain a copy at http://www.php.net/license/3_0.txt. |
---|
29 | * |
---|
30 | * The latest version of DOMPDF might be available at: |
---|
31 | * http://www.digitaljunkies.ca/dompdf |
---|
32 | * |
---|
33 | * @link http://www.digitaljunkies.ca/dompdf |
---|
34 | * @copyright 2004 Benj Carson |
---|
35 | * @author Benj Carson <benjcarson@digitaljunkies.ca> |
---|
36 | * @package dompdf |
---|
37 | * @version 0.5.1 |
---|
38 | */ |
---|
39 | |
---|
40 | /* $Id: dompdf.cls.php,v 1.18 2006/07/07 21:31:03 benjcarson Exp $ */ |
---|
41 | |
---|
42 | /** |
---|
43 | * DOMPDF - PHP5 HTML to PDF renderer |
---|
44 | * |
---|
45 | * DOMPDF loads HTML and does its best to render it as a PDF. It gets its |
---|
46 | * name from the new DomDocument PHP5 extension. Source HTML is first |
---|
47 | * parsed by a DomDocument object. DOMPDF takes the resulting DOM tree and |
---|
48 | * attaches a {@link Frame} object to each node. {@link Frame} objects store |
---|
49 | * positioning and layout information and each has a reference to a {@link |
---|
50 | * Style} object. |
---|
51 | * |
---|
52 | * Style information is loaded and parsed (see {@link Stylesheet}) and is |
---|
53 | * applied to the frames in the tree by using XPath. CSS selectors are |
---|
54 | * converted into XPath queries, and the computed {@link Style} objects are |
---|
55 | * applied to the {@link Frame}s. |
---|
56 | * |
---|
57 | * {@link Frame}s are then decorated (in the design pattern sense of the |
---|
58 | * word) based on their CSS display property ({@link |
---|
59 | * http://www.w3.org/TR/CSS21/visuren.html#propdef-display}). |
---|
60 | * Frame_Decorators augment the basic {@link Frame} class by adding |
---|
61 | * additional properties and methods specific to the particular type of |
---|
62 | * {@link Frame}. For example, in the CSS layout model, block frames |
---|
63 | * (display: block;) contain line boxes that are usually filled with text or |
---|
64 | * other inline frames. The Block_Frame_Decorator therefore adds a $lines |
---|
65 | * property as well as methods to add {@link Frame}s to lines and to add |
---|
66 | * additional lines. {@link Frame}s also are attached to specific |
---|
67 | * Positioner and {@link Frame_Reflower} objects that contain the |
---|
68 | * positioining and layout algorithm for a specific type of frame, |
---|
69 | * respectively. This is an application of the Strategy pattern. |
---|
70 | * |
---|
71 | * Layout, or reflow, proceeds recursively (post-order) starting at the root |
---|
72 | * of the document. Space constraints (containing block width & height) are |
---|
73 | * pushed down, and resolved positions and sizes bubble up. Thus, every |
---|
74 | * {@link Frame} in the document tree is traversed once (except for tables |
---|
75 | * which use a two-pass layout algorithm). If you are interested in the |
---|
76 | * details, see the reflow() method of the Reflower classes. |
---|
77 | * |
---|
78 | * Rendering is relatively straightforward once layout is complete. {@link |
---|
79 | * Frame}s are rendered using an adapted {@link Cpdf} class, originally |
---|
80 | * written by Wayne Munro, http://www.ros.co.nz/pdf/. (Some performance |
---|
81 | * related changes have been made to the original {@link Cpdf} class, and |
---|
82 | * the {@link CPDF_Adapter} class provides a simple, stateless interface to |
---|
83 | * PDF generation.) PDFLib support has now also been added, via the {@link |
---|
84 | * PDFLib_Adapter}. |
---|
85 | * |
---|
86 | * |
---|
87 | * @package dompdf |
---|
88 | */ |
---|
89 | class DOMPDF { |
---|
90 | |
---|
91 | |
---|
92 | /** |
---|
93 | * DomDocument representing the HTML document |
---|
94 | * |
---|
95 | * @var DomDocument |
---|
96 | */ |
---|
97 | protected $_xml; |
---|
98 | |
---|
99 | /** |
---|
100 | * Frame_Tree derived from the DOM tree |
---|
101 | * |
---|
102 | * @var Frame_Tree |
---|
103 | */ |
---|
104 | protected $_tree; |
---|
105 | |
---|
106 | /** |
---|
107 | * Stylesheet for the document |
---|
108 | * |
---|
109 | * @var Stylesheet |
---|
110 | */ |
---|
111 | protected $_css; |
---|
112 | |
---|
113 | /** |
---|
114 | * Actual PDF renderer |
---|
115 | * |
---|
116 | * @var Canvas |
---|
117 | */ |
---|
118 | protected $_pdf; |
---|
119 | |
---|
120 | /** |
---|
121 | * Desired paper size ('letter', 'legal', 'A4', etc.) |
---|
122 | * |
---|
123 | * @var string |
---|
124 | */ |
---|
125 | protected $_paper_size; |
---|
126 | |
---|
127 | /** |
---|
128 | * Paper orientation ('portrait' or 'landscape') |
---|
129 | * |
---|
130 | * @var string |
---|
131 | */ |
---|
132 | protected $_paper_orientation; |
---|
133 | |
---|
134 | private $_cache_id; |
---|
135 | |
---|
136 | /** |
---|
137 | * Base hostname |
---|
138 | * |
---|
139 | * Used for relative paths/urls |
---|
140 | * @var string |
---|
141 | */ |
---|
142 | protected $_base_host; |
---|
143 | |
---|
144 | /** |
---|
145 | * Absolute base path |
---|
146 | * |
---|
147 | * Used for relative paths/urls |
---|
148 | * @var string |
---|
149 | */ |
---|
150 | protected $_base_path; |
---|
151 | |
---|
152 | /** |
---|
153 | * Protcol used to request file (file://, http://, etc) |
---|
154 | * |
---|
155 | * @var string |
---|
156 | */ |
---|
157 | protected $_protocol; |
---|
158 | |
---|
159 | |
---|
160 | /** |
---|
161 | * Class constructor |
---|
162 | */ |
---|
163 | function __construct() { |
---|
164 | $this->_messages = array(); |
---|
165 | $this->_xml = new DomDocument(); |
---|
166 | $this->_xml->preserveWhiteSpace = true; |
---|
167 | $this->_tree = new Frame_Tree($this->_xml); |
---|
168 | $this->_css = new Stylesheet(); |
---|
169 | $this->_pdf = null; |
---|
170 | $this->_paper_size = "letter"; |
---|
171 | $this->_paper_orientation = "portrait"; |
---|
172 | $this->_base_host = ""; |
---|
173 | $this->_base_path = ""; |
---|
174 | $this->_cache_id = null; |
---|
175 | } |
---|
176 | |
---|
177 | /** |
---|
178 | * Returns the underlying {@link Frame_Tree} object |
---|
179 | * |
---|
180 | * @return Frame_Tree |
---|
181 | */ |
---|
182 | function get_tree() { return $this->_tree; } |
---|
183 | |
---|
184 | //........................................................................ |
---|
185 | |
---|
186 | /** |
---|
187 | * Sets the protocol to use |
---|
188 | * |
---|
189 | * @param string $proto |
---|
190 | */ |
---|
191 | // FIXME: validate these |
---|
192 | function set_protocol($proto) { $this->_protocol = $proto; } |
---|
193 | |
---|
194 | /** |
---|
195 | * Sets the base hostname |
---|
196 | * |
---|
197 | * @param string $host |
---|
198 | */ |
---|
199 | function set_host($host) { $this->_base_host = $host; } |
---|
200 | |
---|
201 | /** |
---|
202 | * Sets the base path |
---|
203 | * |
---|
204 | * @param string $path |
---|
205 | */ |
---|
206 | function set_base_path($path) { $this->_base_path = $path; } |
---|
207 | |
---|
208 | /** |
---|
209 | * Returns the protocol in use |
---|
210 | * |
---|
211 | * @return string |
---|
212 | */ |
---|
213 | function get_protocol() { return $this->_protocol; } |
---|
214 | |
---|
215 | /** |
---|
216 | * Returns the base hostname |
---|
217 | * |
---|
218 | * @return string |
---|
219 | */ |
---|
220 | function get_host() { return $this->_base_host; } |
---|
221 | |
---|
222 | /** |
---|
223 | * Returns the base path |
---|
224 | * |
---|
225 | * @return string |
---|
226 | */ |
---|
227 | function get_base_path() { return $this->_base_path; } |
---|
228 | |
---|
229 | /** |
---|
230 | * Return the underlying Canvas instance (e.g. CPDF_Adapter, GD_Adapter) |
---|
231 | * |
---|
232 | * @return Canvas |
---|
233 | */ |
---|
234 | function get_canvas() { return $this->_pdf; } |
---|
235 | |
---|
236 | //........................................................................ |
---|
237 | |
---|
238 | /** |
---|
239 | * Loads an HTML file |
---|
240 | * |
---|
241 | * Parse errors are stored in the global array _dompdf_warnings. |
---|
242 | * |
---|
243 | * @param string $file a filename or url to load |
---|
244 | */ |
---|
245 | function load_html_file($file) { |
---|
246 | // Store parsing warnings as messages (this is to prevent output to the |
---|
247 | // browser if the html is ugly and the dom extension complains, |
---|
248 | // preventing the pdf from being streamed.) |
---|
249 | list($this->_protocol, $this->_base_host, $this->_base_path) = explode_url($file); |
---|
250 | |
---|
251 | if ( !DOMPDF_ENABLE_REMOTE && |
---|
252 | ($this->_protocol != "" && $this->_protocol != "file://" ) ) |
---|
253 | throw new DOMPDF_Exception("Remote file requested, but DOMPDF_ENABLE_REMOTE is false."); |
---|
254 | |
---|
255 | if ( !DOMPDF_ENABLE_PHP ) { |
---|
256 | set_error_handler("record_warnings"); |
---|
257 | $this->_xml->loadHTMLFile($file); |
---|
258 | restore_error_handler(); |
---|
259 | |
---|
260 | } else |
---|
261 | $this->load_html(file_get_contents($file)); |
---|
262 | |
---|
263 | } |
---|
264 | |
---|
265 | /** |
---|
266 | * Loads an HTML string |
---|
267 | * |
---|
268 | * Parse errors are stored in the global array _dompdf_warnings. |
---|
269 | * |
---|
270 | * @param string $str HTML text to load |
---|
271 | */ |
---|
272 | function load_html($str) { |
---|
273 | |
---|
274 | // Parse embedded php, first-pass |
---|
275 | if ( DOMPDF_ENABLE_PHP ) { |
---|
276 | ob_start(); |
---|
277 | eval("?" . ">$str"); |
---|
278 | $str = ob_get_contents(); |
---|
279 | ob_end_clean(); |
---|
280 | } |
---|
281 | |
---|
282 | // Store parsing warnings as messages |
---|
283 | set_error_handler("record_warnings"); |
---|
284 | $this->_xml->loadHTML($str); |
---|
285 | restore_error_handler(); |
---|
286 | } |
---|
287 | |
---|
288 | /** |
---|
289 | * Builds the {@link Frame_Tree}, loads any CSS and applies the styles to |
---|
290 | * the {@link Frame_Tree} |
---|
291 | */ |
---|
292 | protected function _process_html() { |
---|
293 | $this->_tree->build_tree(); |
---|
294 | |
---|
295 | $this->_css->load_css_file(Stylesheet::DEFAULT_STYLESHEET); |
---|
296 | |
---|
297 | // load <link rel="STYLESHEET" ... /> tags |
---|
298 | $links = $this->_xml->getElementsByTagName("link"); |
---|
299 | foreach ($links as $link) { |
---|
300 | if ( mb_strtolower($link->getAttribute("rel")) == "stylesheet" || |
---|
301 | mb_strtolower($link->getAttribute("type")) == "text/css" ) { |
---|
302 | $url = $link->getAttribute("href"); |
---|
303 | $url = build_url($this->_protocol, $this->_base_host, $this->_base_path, $url); |
---|
304 | |
---|
305 | $this->_css->load_css_file($url); |
---|
306 | } |
---|
307 | |
---|
308 | } |
---|
309 | |
---|
310 | // load <style> tags |
---|
311 | $styles = $this->_xml->getElementsByTagName("style"); |
---|
312 | foreach ($styles as $style) { |
---|
313 | |
---|
314 | // Accept all <style> tags by default (note this is contrary to W3C |
---|
315 | // HTML 4.0 spec: |
---|
316 | // http://www.w3.org/TR/REC-html40/present/styles.html#adef-media |
---|
317 | // which states that the default media type is 'screen' |
---|
318 | if ( $style->hasAttributes() && |
---|
319 | ($media = $style->getAttribute("media")) && |
---|
320 | !in_array($media, Stylesheet::$ACCEPTED_MEDIA_TYPES) ) |
---|
321 | continue; |
---|
322 | |
---|
323 | $css = ""; |
---|
324 | if ( $style->hasChildNodes() ) { |
---|
325 | |
---|
326 | $child = $style->firstChild; |
---|
327 | while ( $child ) { |
---|
328 | $css .= $child->nodeValue; // Handle <style><!-- blah --></style> |
---|
329 | $child = $child->nextSibling; |
---|
330 | } |
---|
331 | |
---|
332 | } else |
---|
333 | $css = $style->nodeValue; |
---|
334 | |
---|
335 | // Set the base path of the Stylesheet to that of the file being processed |
---|
336 | $this->_css->set_protocol($this->_protocol); |
---|
337 | $this->_css->set_host($this->_base_host); |
---|
338 | $this->_css->set_base_path($this->_base_path); |
---|
339 | |
---|
340 | $this->_css->load_css($css); |
---|
341 | } |
---|
342 | |
---|
343 | } |
---|
344 | |
---|
345 | //........................................................................ |
---|
346 | |
---|
347 | /** |
---|
348 | * Sets the paper size & orientation |
---|
349 | * |
---|
350 | * @param string $size 'letter', 'legal', 'A4', etc. {@link CPDF_Adapter::$PAPER_SIZES} |
---|
351 | * @param string $orientation 'portrait' or 'landscape' |
---|
352 | */ |
---|
353 | function set_paper($size, $orientation = "portrait") { |
---|
354 | $this->_paper_size = $size; |
---|
355 | $this->_paper_orientation = $orientation; |
---|
356 | } |
---|
357 | |
---|
358 | //........................................................................ |
---|
359 | |
---|
360 | /** |
---|
361 | * Enable experimental caching capability |
---|
362 | * @access private |
---|
363 | */ |
---|
364 | function enable_caching($cache_id) { |
---|
365 | $this->_cache_id = $cache_id; |
---|
366 | } |
---|
367 | |
---|
368 | //........................................................................ |
---|
369 | |
---|
370 | /** |
---|
371 | * Renders the HTML to PDF |
---|
372 | */ |
---|
373 | function render() { |
---|
374 | |
---|
375 | //enable_mem_profile(); |
---|
376 | |
---|
377 | $this->_process_html(); |
---|
378 | |
---|
379 | $this->_css->apply_styles($this->_tree); |
---|
380 | |
---|
381 | $root = null; |
---|
382 | |
---|
383 | foreach ($this->_tree->get_frames() as $frame) { |
---|
384 | |
---|
385 | // Set up the root frame |
---|
386 | if ( is_null($root) ) { |
---|
387 | $root = Frame_Factory::decorate_root( $this->_tree->get_root(), $this ); |
---|
388 | continue; |
---|
389 | } |
---|
390 | |
---|
391 | // Create the appropriate decorators, reflowers & positioners. |
---|
392 | $deco = Frame_Factory::decorate_frame($frame, $this); |
---|
393 | $deco->set_root($root); |
---|
394 | |
---|
395 | // FIXME: handle generated content |
---|
396 | if ( $frame->get_style()->display == "list-item" ) { |
---|
397 | |
---|
398 | // Insert a list-bullet frame |
---|
399 | $node = $this->_xml->createElement("bullet"); // arbitrary choice |
---|
400 | $b_f = new Frame($node); |
---|
401 | |
---|
402 | $style = $this->_css->create_style(); |
---|
403 | $style->display = "-dompdf-list-bullet"; |
---|
404 | $style->inherit($frame->get_style()); |
---|
405 | $b_f->set_style($style); |
---|
406 | |
---|
407 | $deco->prepend_child( Frame_Factory::decorate_frame($b_f, $this) ); |
---|
408 | } |
---|
409 | } |
---|
410 | |
---|
411 | $this->_pdf = Canvas_Factory::get_instance($this->_paper_size, $this->_paper_orientation); |
---|
412 | |
---|
413 | $root->set_containing_block(0, 0, $this->_pdf->get_width(), $this->_pdf->get_height()); |
---|
414 | $root->set_renderer(new Renderer($this)); |
---|
415 | |
---|
416 | // This is where the magic happens: |
---|
417 | $root->reflow(); |
---|
418 | |
---|
419 | // Clean up cached images |
---|
420 | Image_Cache::clear(); |
---|
421 | } |
---|
422 | |
---|
423 | //........................................................................ |
---|
424 | |
---|
425 | /** |
---|
426 | * Streams the PDF to the client |
---|
427 | * |
---|
428 | * The file will open a download dialog by default. The options |
---|
429 | * parameter controls the output headers. Accepted headers are: |
---|
430 | * |
---|
431 | * 'Accept-Ranges' => 1 or 0 - if this is not set to 1, then this |
---|
432 | * header is not included, off by default this header seems to |
---|
433 | * have caused some problems despite tha fact that it is supposed |
---|
434 | * to solve them, so I am leaving it off by default. |
---|
435 | * |
---|
436 | * 'compress' = > 1 or 0 - apply content stream compression, this is |
---|
437 | * on (1) by default |
---|
438 | * |
---|
439 | * 'Attachment' => 1 or 0 - if 1, force the browser to open a |
---|
440 | * download dialog, on (1) by default |
---|
441 | * |
---|
442 | * @param string $filename the name of the streamed file |
---|
443 | * @param array $options header options (see above) |
---|
444 | */ |
---|
445 | function stream($filename, $options = null) { |
---|
446 | if (!is_null($this->_pdf)) |
---|
447 | $this->_pdf->stream($filename, $options); |
---|
448 | } |
---|
449 | |
---|
450 | /** |
---|
451 | * Returns the PDF as a string |
---|
452 | * |
---|
453 | * @return string |
---|
454 | */ |
---|
455 | function output() { |
---|
456 | global $_dompdf_debug; |
---|
457 | if ( is_null($this->_pdf) ) |
---|
458 | return null; |
---|
459 | |
---|
460 | return $this->_pdf->output( $_dompdf_debug ); |
---|
461 | } |
---|
462 | |
---|
463 | //........................................................................ |
---|
464 | |
---|
465 | } |
---|
466 | ?> |
---|