3 /*************************************************************************
5 * class.html2text.inc *
7 *************************************************************************
9 * Converts HTML to formatted plain text *
11 * Copyright (c) 2005-2007 Jon Abernathy <jon@chuggnutt.com> *
12 * All rights reserved. *
14 * This script is free software; you can redistribute it and/or modify *
15 * it under the terms of the GNU General Public License as published by *
16 * the Free Software Foundation; either version 2 of the License, or *
17 * (at your option) any later version. *
19 * The GNU General Public License can be found at *
20 * http://www.gnu.org/copyleft/gpl.html. *
22 * This script is distributed in the hope that it will be useful, *
23 * but WITHOUT ANY WARRANTY; without even the implied warranty of *
24 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the *
25 * GNU General Public License for more details. *
27 * Author(s): Jon Abernathy <jon@chuggnutt.com> *
29 * Last modified: 08/08/07 *
31 *************************************************************************/
35 * Takes HTML and converts it to formatted, plain text.
37 * Thanks to Alexander Krug (http://www.krugar.de/) to pointing out and
38 * correcting an error in the regexp search array. Fixed 7/30/03.
40 * Updated set_html() function's file reading mechanism, 9/25/03.
42 * Thanks to Joss Sanglier (http://www.dancingbear.co.uk/) for adding
43 * several more HTML entity codes to the $search and $replace arrays.
46 * Thanks to Darius Kasperavicius (http://www.dar.dar.lt/) for
47 * suggesting the addition of $allowed_tags and its supporting function
48 * (which I slightly modified). Updated 3/12/04.
50 * Thanks to Justin Dearing for pointing out that a replacement for the
51 * <TH> tag was missing, and suggesting an appropriate fix.
54 * Thanks to Mathieu Collas (http://www.myefarm.com/) for finding a
55 * display/formatting bug in the _build_link_list() function: email
56 * readers would show the left bracket and number ("[1") as part of the
57 * rendered email address.
60 * Thanks to Wojciech Bajon (http://histeria.pl/) for submitting code
61 * to handle relative links, which I hadn't considered. I modified his
62 * code a bit to handle normal HTTP links and MAILTO links. Also for
63 * suggesting three additional HTML entity codes to search for.
66 * Thanks to Jacob Chandler for pointing out another link condition
67 * for the _build_link_list() function: "https".
70 * Thanks to Marc Bertrand (http://www.dresdensky.com/) for
71 * suggesting a revision to the word wrapping functionality; if you
72 * specify a $width of 0 or less, word wrapping will be ignored.
75 * *** Big housecleaning updates below:
77 * Thanks to Colin Brown (http://www.sparkdriver.co.uk/) for
78 * suggesting the fix to handle </li> and blank lines (whitespace).
79 * Christian Basedau (http://www.movetheweb.de/) also suggested the
82 * Special thanks to Marcus Bointon (http://www.synchromedia.co.uk/),
83 * Christian Basedau, Norbert Laposa (http://ln5.co.uk/),
84 * Bas van de Weijer, and Marijn van Butselaar
85 * for pointing out my glaring error in the <th> handling. Marcus also
86 * supplied a host of fixes.
88 * Thanks to Jeffrey Silverman (http://www.newtnotes.com/) for pointing
89 * out that extra spaces should be compressed--a problem addressed with
90 * Marcus Bointon's fixes but that I had not yet incorporated.
92 * Thanks to Daniel Schledermann (http://www.typoconsult.dk/) for
93 * suggesting a valuable fix with <a> tag handling.
95 * Thanks to Wojciech Bajon (again!) for suggesting fixes and additions,
96 * including the <a> tag handling that Daniel Schledermann pointed
97 * out but that I had not yet incorporated. I haven't (yet)
98 * incorporated all of Wojciech's changes, though I may at some
101 * *** End of the housecleaning updates. Updated 08/08/07.
103 * @author Jon Abernathy <jon@chuggnutt.com>
111 * Contains the HTML content to convert.
119 * Contains the converted, formatted text.
127 * Maximum width of the formatted text, in columns.
129 * Set this value to 0 (or less) to ignore word wrapping
130 * and not constrain text to a fixed-width column.
132 * @var integer $width
138 * List of preg* regular expression patterns to search for,
139 * used in conjunction with $replace.
146 "/\r/", // Non-legal carriage return
147 "/[\n\t]+/", // Newlines and tabs
148 '/<script[^>]*>.*?<\/script>/i', // <script>s -- which strip_tags supposedly has problems with
149 '/<style[^>]*>.*?<\/style>/i', // <style>s -- which strip_tags supposedly has problems with
150 '/<p[^>]*>/i', // <P>
151 '/<br[^>]*>/i', // <br>
152 '/<i[^>]*>(.*?)<\/i>/i', // <i>
153 '/<em[^>]*>(.*?)<\/em>/i', // <em>
154 '/(<ul[^>]*>|<\/ul>)/i', // <ul> and </ul>
155 '/(<ol[^>]*>|<\/ol>)/i', // <ol> and </ol>
156 '/<li[^>]*>(.*?)<\/li>/i', // <li> and </li>
157 '/<li[^>]*>/i', // <li>
158 '/<hr[^>]*>/i', // <hr>
159 '/<div[^>]*>/i', // <div>
160 '/(<table[^>]*>|<\/table>)/i', // <table> and </table>
161 '/(<tr[^>]*>|<\/tr>)/i', // <tr> and </tr>
162 '/<td[^>]*>(.*?)<\/td>/i', // <td> and </td>
166 * List of pattern replacements corresponding to patterns searched.
168 * @var array $replace
172 var $replace = array(
173 '', // Non-legal carriage return
174 ' ', // Newlines and tabs
175 '', // <script>s -- which strip_tags supposedly has problems with
176 '', // <style>s -- which strip_tags supposedly has problems with
181 "\n\n", // <ul> and </ul>
182 "\n\n", // <ol> and </ol>
183 "\t* \\1\n", // <li> and </li>
185 "\n-------------------------\n", // <hr>
187 "\n\n", // <table> and </table>
188 "\n", // <tr> and </tr>
189 "\t\t\\1\n", // <td> and </td>
193 * List of preg* regular expression patterns to search for,
194 * used in conjunction with $ent_replace.
196 * @var array $ent_search
200 var $ent_search = array(
201 '/&(nbsp|#160);/i', // Non-breaking space
202 '/&(quot|rdquo|ldquo|#8220|#8221|#147|#148);/i',
204 '/&(apos|rsquo|lsquo|#8216|#8217);/i', // Single quotes
205 '/>/i', // Greater-than
206 '/</i', // Less-than
207 '/&(copy|#169);/i', // Copyright
208 '/&(trade|#8482|#153);/i', // Trademark
209 '/&(reg|#174);/i', // Registered
210 '/&(mdash|#151|#8212);/i', // mdash
211 '/&(ndash|minus|#8211|#8722);/i', // ndash
212 '/&(bull|#149|#8226);/i', // Bullet
213 '/&(pound|#163);/i', // Pound sign
214 '/&(euro|#8364);/i', // Euro sign
215 '/&(amp|#38);/i', // Ampersand: see _converter()
216 '/[ ]{2,}/', // Runs of spaces, post-handling
220 * List of pattern replacements corresponding to patterns searched.
222 * @var array $ent_replace
226 var $ent_replace = array(
227 ' ', // Non-breaking space
228 '"', // Double quotes
229 "'", // Single quotes
239 'EUR', // Euro sign.
\80 ?
240 '|+|amp|+|', // Ampersand: see _converter()
241 ' ', // Runs of spaces, post-handling
245 * List of preg* regular expression patterns to search for
246 * and replace using callback function.
248 * @var array $callback_search
251 var $callback_search = array(
252 '/<(h)[123456][^>]*>(.*?)<\/h[123456]>/i', // H1 - H3
253 '/<(b)[^>]*>(.*?)<\/b>/i', // <b>
254 '/<(strong)[^>]*>(.*?)<\/strong>/i', // <strong>
255 '/<(a) [^>]*href=("|\')([^"\']+)\2[^>]*>(.*?)<\/a>/i',
257 '/<(th)[^>]*>(.*?)<\/th>/i', // <th> and </th>
261 * List of preg* regular expression patterns to search for in PRE body,
262 * used in conjunction with $pre_replace.
264 * @var array $pre_search
268 var $pre_search = array(
277 * List of pattern replacements corresponding to patterns searched for PRE body.
279 * @var array $pre_replace
283 var $pre_replace = array(
285 ' ',
292 * Contains a list of HTML tags to allow in the resulting text.
294 * @var string $allowed_tags
296 * @see set_allowed_tags()
298 var $allowed_tags = '';
301 * Contains the base URL that relative links should resolve to.
309 * Indicates whether content in the $html variable has been converted yet.
311 * @var boolean $_converted
315 var $_converted = false;
318 * Contains URL addresses from links to be rendered in plain text.
320 * @var string $_link_list
322 * @see _build_link_list()
324 var $_link_list = '';
327 * Number of valid links detected in the text, used for plain text
328 * display (rendered similar to footnotes).
330 * @var integer $_link_count
332 * @see _build_link_list()
334 var $_link_count = 0;
337 * Boolean flag, true if a table of link URLs should be listed after the text.
339 * @var boolean $_do_links
343 var $_do_links = true;
348 * If the HTML source string (or file) is supplied, the class
349 * will instantiate with that source propagated, all that has
350 * to be done it to call get_text().
352 * @param string $source HTML content
353 * @param boolean $from_file Indicates $source is a file to pull content from
354 * @param boolean $do_links Indicate whether a table of link URLs is desired
355 * @param integer $width Maximum width of the formatted text, 0 for no limit
359 function html2text( $source = '', $from_file = false, $do_links = true, $width = 75 )
361 if ( !empty($source) ) {
362 $this->set_html($source, $from_file);
365 $this->set_base_url();
366 $this->_do_links = $do_links;
367 $this->width = $width;
371 * Loads source HTML into memory, either from $source string or a file.
373 * @param string $source HTML content
374 * @param boolean $from_file Indicates $source is a file to pull content from
378 function set_html( $source, $from_file = false )
380 if ( $from_file && file_exists($source) ) {
381 $this->html = file_get_contents($source);
384 $this->html = $source;
386 $this->_converted = false;
390 * Returns the text, converted from HTML.
397 if ( !$this->_converted ) {
405 * Prints the text, converted from HTML.
410 function print_text()
412 print $this->get_text();
416 * Alias to print_text(), operates identically.
424 print $this->get_text();
428 * Sets the allowed HTML tags to pass through to the resulting text.
430 * Tags should be in the form "<p>", with no corresponding closing tag.
435 function set_allowed_tags( $allowed_tags = '' )
437 if ( !empty($allowed_tags) ) {
438 $this->allowed_tags = $allowed_tags;
443 * Sets a base URL to handle relative links.
448 function set_base_url( $url = '' )
451 if ( !empty($_SERVER['HTTP_HOST']) ) {
452 $this->url = 'http://' . $_SERVER['HTTP_HOST'];
457 // Strip any trailing slashes for consistency (relative
458 // URLs may already start with a slash like "/file.html")
459 if ( substr($url, -1) == '/' ) {
460 $url = substr($url, 0, -1);
467 * Workhorse function that does actual conversion (calls _converter() method).
474 // Variables used for building the link list
475 $this->_link_count = 0;
476 $this->_link_list = '';
478 $text = trim(stripslashes($this->html));
480 // Convert HTML to TXT
481 $this->_converter($text);
484 if ( !empty($this->_link_list) ) {
485 $text .= "\n\nLinks:\n------\n" . $this->_link_list;
490 $this->_converted = true;
494 * Workhorse function that does actual conversion.
496 * First performs custom tag replacement specified by $search and
497 * $replace arrays. Then strips any remaining HTML tags, reduces whitespace
498 * and newlines to a readable format, and word wraps the text to
501 * @param string Reference to HTML content string
506 function _converter(&$text)
508 // Convert <BLOCKQUOTE> (before PRE!)
509 $this->_convert_blockquotes($text);
512 $this->_convert_pre($text);
514 // Run our defined tags search-and-replace
515 $text = preg_replace($this->search, $this->replace, $text);
517 // Run our defined tags search-and-replace with callback
518 $text = preg_replace_callback($this->callback_search, array('html2text', '_preg_callback'), $text);
520 // Strip any other HTML tags
521 $text = strip_tags($text, $this->allowed_tags);
523 // Run our defined entities/characters search-and-replace
524 $text = preg_replace($this->ent_search, $this->ent_replace, $text);
526 // Replace known html entities
527 $text = html_entity_decode($text, ENT_COMPAT, 'UTF-8');
529 // Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
530 $text = preg_replace('/&([a-zA-Z0-9]{2,6}|#[0-9]{2,4});/', '', $text);
532 // Convert "|+|amp|+|" into "&", need to be done after handling of unknown entities
533 // This properly handles situation of "&quot;" in input string
534 $text = str_replace('|+|amp|+|', '&', $text);
536 // Bring down number of empty lines to 2 max
537 $text = preg_replace("/\n\s+\n/", "\n\n", $text);
538 $text = preg_replace("/[\n]{3,}/", "\n\n", $text);
540 // remove leading empty lines (can be produced by eg. P tag on the beginning)
541 $text = ltrim($text, "\n");
543 // Wrap the text to a readable format
544 // for PHP versions >= 4.0.2. Default width is 75
545 // If width is 0 or less, don't wrap the text.
546 if ( $this->width > 0 ) {
547 $text = wordwrap($text, $this->width);
552 * Helper function called by preg_replace() on link replacement.
554 * Maintains an internal list of links to be displayed at the end of the
555 * text, with numeric indices to the original point in the text they
556 * appeared. Also makes an effort at identifying and handling absolute
557 * and relative links.
559 * @param string $link URL of the link
560 * @param string $display Part of the text to associate number with
564 function _build_link_list( $link, $display )
566 if ( !$this->_do_links )
569 if ( preg_match('!^(https?://|mailto:)!', $link) ) {
570 $this->_link_count++;
571 $this->_link_list .= '[' . $this->_link_count . "] $link\n";
572 $additional = ' [' . $this->_link_count . ']';
573 } elseif ( substr($link, 0, 11) == 'javascript:' ) {
574 // Don't count the link; ignore it
576 // what about href="#anchor" ?
578 $this->_link_count++;
579 $this->_link_list .= '[' . $this->_link_count . '] ' . $this->url;
580 if ( substr($link, 0, 1) != '/' ) {
581 $this->_link_list .= '/';
583 $this->_link_list .= "$link\n";
584 $additional = ' [' . $this->_link_count . ']';
587 return $display . $additional;
591 * Helper function for PRE body conversion.
593 * @param string HTML content
596 function _convert_pre(&$text)
598 // get the content of PRE element
599 while (preg_match('/<pre[^>]*>(.*)<\/pre>/ismU', $text, $matches)) {
600 // convert the content
601 $this->pre_content = sprintf('<div><br>%s<br></div>',
602 preg_replace($this->pre_search, $this->pre_replace, $matches[1]));
603 // replace the content (use callback because content can contain $0 variable)
604 $text = preg_replace_callback('/<pre[^>]*>.*<\/pre>/ismU',
605 array('html2text', '_preg_pre_callback'), $text, 1);
607 $this->pre_content = '';
612 * Helper function for BLOCKQUOTE body conversion.
614 * @param string HTML content
617 function _convert_blockquotes(&$text)
619 if (preg_match_all('/<\/*blockquote[^>]*>/i', $text, $matches, PREG_OFFSET_CAPTURE)) {
622 foreach ($matches[0] as $m) {
623 if ($m[0][0] == '<' && $m[0][1] == '/') {
626 $level = 0; // malformed HTML: go to next blockquote
628 else if ($level > 0) {
629 // skip inner blockquote
633 $len = $end - $taglen - $start;
634 // Get blockquote content
635 $body = substr($text, $start + $taglen - $diff, $len);
638 $p_width = $this->width;
639 if ($this->width > 0) $this->width -= 2;
640 // Convert blockquote content
642 $this->_converter($body);
643 // Add citation markers and create PRE block
644 $body = preg_replace('/((^|\n)>*)/', '\\1> ', trim($body));
645 $body = '<pre>' . htmlspecialchars($body) . '</pre>';
647 $this->width = $p_width;
649 $text = substr($text, 0, $start - $diff)
650 . $body . substr($text, $end + strlen($m[0]) - $diff);
652 $diff = $len + $taglen + strlen($m[0]) - strlen($body);
659 $taglen = strlen($m[0]);
668 * Callback function for preg_replace_callback use.
670 * @param array PREG matches
673 private function _preg_callback($matches)
675 switch($matches[1]) {
678 return $this->_strtoupper($matches[2]);
680 return $this->_strtoupper("\t\t". $matches[2] ."\n");
682 return $this->_strtoupper("\n\n". $matches[2] ."\n\n");
684 // Remove spaces in URL (#1487805)
685 $url = str_replace(' ', '', $matches[3]);
686 return $this->_build_link_list($url, $matches[4]);
691 * Callback function for preg_replace_callback use in PRE content handler.
693 * @param array PREG matches
696 private function _preg_pre_callback($matches)
698 return $this->pre_content;
702 * Strtoupper multibyte wrapper function
707 private function _strtoupper($str)
709 if (function_exists('mb_strtoupper'))
710 return mb_strtoupper($str);
712 return strtoupper($str);