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 '/[ ]{2,}/', // Runs of spaces, pre-handling
149 '/<script[^>]*>.*?<\/script>/i', // <script>s -- which strip_tags supposedly has problems with
150 '/<style[^>]*>.*?<\/style>/i', // <style>s -- which strip_tags supposedly has problems with
151 '/<p[^>]*>/i', // <P>
152 '/<br[^>]*>/i', // <br>
153 '/<i[^>]*>(.*?)<\/i>/i', // <i>
154 '/<em[^>]*>(.*?)<\/em>/i', // <em>
155 '/(<ul[^>]*>|<\/ul>)/i', // <ul> and </ul>
156 '/(<ol[^>]*>|<\/ol>)/i', // <ol> and </ol>
157 '/<li[^>]*>(.*?)<\/li>/i', // <li> and </li>
158 '/<li[^>]*>/i', // <li>
159 '/<hr[^>]*>/i', // <hr>
160 '/<div[^>]*>/i', // <div>
161 '/(<table[^>]*>|<\/table>)/i', // <table> and </table>
162 '/(<tr[^>]*>|<\/tr>)/i', // <tr> and </tr>
163 '/<td[^>]*>(.*?)<\/td>/i', // <td> and </td>
164 '/&(nbsp|#160);/i', // Non-breaking space
165 '/&(quot|rdquo|ldquo|#8220|#8221|#147|#148);/i',
167 '/&(apos|rsquo|lsquo|#8216|#8217);/i', // Single quotes
168 '/>/i', // Greater-than
169 '/</i', // Less-than
170 '/&(copy|#169);/i', // Copyright
171 '/&(trade|#8482|#153);/i', // Trademark
172 '/&(reg|#174);/i', // Registered
173 '/&(mdash|#151|#8212);/i', // mdash
174 '/&(ndash|minus|#8211|#8722);/i', // ndash
175 '/&(bull|#149|#8226);/i', // Bullet
176 '/&(pound|#163);/i', // Pound sign
177 '/&(euro|#8364);/i', // Euro sign
178 '/&(amp|#38);/i', // Ampersand: see _converter()
179 '/[ ]{2,}/' // Runs of spaces, post-handling
183 * List of pattern replacements corresponding to patterns searched.
185 * @var array $replace
189 var $replace = array(
190 '', // Non-legal carriage return
191 ' ', // Newlines and tabs
192 ' ', // Runs of spaces, pre-handling
193 '', // <script>s -- which strip_tags supposedly has problems with
194 '', // <style>s -- which strip_tags supposedly has problems with
199 "\n\n", // <ul> and </ul>
200 "\n\n", // <ol> and </ol>
201 "\t* \\1\n", // <li> and </li>
203 "\n-------------------------\n", // <hr>
205 "\n\n", // <table> and </table>
206 "\n", // <tr> and </tr>
207 "\t\t\\1\n", // <td> and </td>
208 ' ', // Non-breaking space
209 '"', // Double quotes
210 "'", // Single quotes
220 'EUR', // Euro sign.
\80 ?
221 '|+|amp|+|', // Ampersand: see _converter()
222 ' ' // Runs of spaces, post-handling
226 * List of preg* regular expression patterns to search for
227 * and replace using callback function.
229 * @var array $callback_search
232 var $callback_search = array(
233 '/<(h)[123456][^>]*>(.*?)<\/h[123456]>/i', // H1 - H3
234 '/<(b)[^>]*>(.*?)<\/b>/i', // <b>
235 '/<(strong)[^>]*>(.*?)<\/strong>/i', // <strong>
236 '/<(a) [^>]*href=("|\')([^"\']+)\2[^>]*>(.*?)<\/a>/i',
238 '/<(th)[^>]*>(.*?)<\/th>/i', // <th> and </th>
242 * List of preg* regular expression patterns to search for in PRE body,
243 * used in conjunction with $pre_replace.
245 * @var array $pre_search
249 var $pre_search = array(
258 * List of pattern replacements corresponding to patterns searched for PRE body.
260 * @var array $pre_replace
264 var $pre_replace = array(
266 ' ',
273 * Contains a list of HTML tags to allow in the resulting text.
275 * @var string $allowed_tags
277 * @see set_allowed_tags()
279 var $allowed_tags = '';
282 * Contains the base URL that relative links should resolve to.
290 * Indicates whether content in the $html variable has been converted yet.
292 * @var boolean $_converted
296 var $_converted = false;
299 * Contains URL addresses from links to be rendered in plain text.
301 * @var string $_link_list
303 * @see _build_link_list()
305 var $_link_list = '';
308 * Number of valid links detected in the text, used for plain text
309 * display (rendered similar to footnotes).
311 * @var integer $_link_count
313 * @see _build_link_list()
315 var $_link_count = 0;
318 * Boolean flag, true if a table of link URLs should be listed after the text.
320 * @var boolean $_do_links
324 var $_do_links = true;
329 * If the HTML source string (or file) is supplied, the class
330 * will instantiate with that source propagated, all that has
331 * to be done it to call get_text().
333 * @param string $source HTML content
334 * @param boolean $from_file Indicates $source is a file to pull content from
335 * @param boolean $do_links Indicate whether a table of link URLs is desired
336 * @param integer $width Maximum width of the formatted text, 0 for no limit
340 function html2text( $source = '', $from_file = false, $do_links = true, $width = 75 )
342 if ( !empty($source) ) {
343 $this->set_html($source, $from_file);
346 $this->set_base_url();
347 $this->_do_links = $do_links;
348 $this->width = $width;
352 * Loads source HTML into memory, either from $source string or a file.
354 * @param string $source HTML content
355 * @param boolean $from_file Indicates $source is a file to pull content from
359 function set_html( $source, $from_file = false )
361 if ( $from_file && file_exists($source) ) {
362 $this->html = file_get_contents($source);
365 $this->html = $source;
367 $this->_converted = false;
371 * Returns the text, converted from HTML.
378 if ( !$this->_converted ) {
386 * Prints the text, converted from HTML.
391 function print_text()
393 print $this->get_text();
397 * Alias to print_text(), operates identically.
405 print $this->get_text();
409 * Sets the allowed HTML tags to pass through to the resulting text.
411 * Tags should be in the form "<p>", with no corresponding closing tag.
416 function set_allowed_tags( $allowed_tags = '' )
418 if ( !empty($allowed_tags) ) {
419 $this->allowed_tags = $allowed_tags;
424 * Sets a base URL to handle relative links.
429 function set_base_url( $url = '' )
432 if ( !empty($_SERVER['HTTP_HOST']) ) {
433 $this->url = 'http://' . $_SERVER['HTTP_HOST'];
438 // Strip any trailing slashes for consistency (relative
439 // URLs may already start with a slash like "/file.html")
440 if ( substr($url, -1) == '/' ) {
441 $url = substr($url, 0, -1);
448 * Workhorse function that does actual conversion (calls _converter() method).
455 // Variables used for building the link list
456 $this->_link_count = 0;
457 $this->_link_list = '';
459 $text = trim(stripslashes($this->html));
461 // Convert HTML to TXT
462 $this->_converter($text);
465 if ( !empty($this->_link_list) ) {
466 $text .= "\n\nLinks:\n------\n" . $this->_link_list;
471 $this->_converted = true;
475 * Workhorse function that does actual conversion.
477 * First performs custom tag replacement specified by $search and
478 * $replace arrays. Then strips any remaining HTML tags, reduces whitespace
479 * and newlines to a readable format, and word wraps the text to
482 * @param string Reference to HTML content string
487 function _converter(&$text)
489 // Convert <BLOCKQUOTE> (before PRE!)
490 $this->_convert_blockquotes($text);
493 $this->_convert_pre($text);
495 // Run our defined search-and-replace
496 $text = preg_replace($this->search, $this->replace, $text);
498 // Replace known html entities
499 $text = html_entity_decode($text, ENT_COMPAT, 'UTF-8');
501 // Run our defined search-and-replace with callback
502 $text = preg_replace_callback($this->callback_search, array('html2text', '_preg_callback'), $text);
504 // Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
505 $text = preg_replace('/&([a-zA-Z0-9]{2,6}|#[0-9]{2,4});/', '', $text);
507 // Convert "|+|amp|+|" into "&", need to be done after handling of unknown entities
508 // This properly handles situation of "&quot;" in input string
509 $text = str_replace('|+|amp|+|', '&', $text);
511 // Strip any other HTML tags
512 $text = strip_tags($text, $this->allowed_tags);
514 // Bring down number of empty lines to 2 max
515 $text = preg_replace("/\n\s+\n/", "\n\n", $text);
516 $text = preg_replace("/[\n]{3,}/", "\n\n", $text);
518 // remove leading empty lines (can be produced by eg. P tag on the beginning)
519 $text = preg_replace('/^\n+/', '', $text);
521 // Wrap the text to a readable format
522 // for PHP versions >= 4.0.2. Default width is 75
523 // If width is 0 or less, don't wrap the text.
524 if ( $this->width > 0 ) {
525 $text = wordwrap($text, $this->width);
530 * Helper function called by preg_replace() on link replacement.
532 * Maintains an internal list of links to be displayed at the end of the
533 * text, with numeric indices to the original point in the text they
534 * appeared. Also makes an effort at identifying and handling absolute
535 * and relative links.
537 * @param string $link URL of the link
538 * @param string $display Part of the text to associate number with
542 function _build_link_list( $link, $display )
544 if ( !$this->_do_links )
547 if ( substr($link, 0, 7) == 'http://' || substr($link, 0, 8) == 'https://' ||
548 substr($link, 0, 7) == 'mailto:'
550 $this->_link_count++;
551 $this->_link_list .= '[' . $this->_link_count . "] $link\n";
552 $additional = ' [' . $this->_link_count . ']';
553 } elseif ( substr($link, 0, 11) == 'javascript:' ) {
554 // Don't count the link; ignore it
556 // what about href="#anchor" ?
558 $this->_link_count++;
559 $this->_link_list .= '[' . $this->_link_count . '] ' . $this->url;
560 if ( substr($link, 0, 1) != '/' ) {
561 $this->_link_list .= '/';
563 $this->_link_list .= "$link\n";
564 $additional = ' [' . $this->_link_count . ']';
567 return $display . $additional;
571 * Helper function for PRE body conversion.
573 * @param string HTML content
576 function _convert_pre(&$text)
578 // get the content of PRE element
579 while (preg_match('/<pre[^>]*>(.*)<\/pre>/ismU', $text, $matches)) {
580 // convert the content
581 $this->pre_content = sprintf('<div><br>%s<br></div>',
582 preg_replace($this->pre_search, $this->pre_replace, $matches[1]));
583 // replace the content (use callback because content can contain $0 variable)
584 $text = preg_replace_callback('/<pre[^>]*>.*<\/pre>/ismU',
585 array('html2text', '_preg_pre_callback'), $text, 1);
587 $this->pre_content = '';
592 * Helper function for BLOCKQUOTE body conversion.
594 * @param string HTML content
597 function _convert_blockquotes(&$text)
599 if (preg_match_all('/<\/*blockquote[^>]*>/i', $text, $matches, PREG_OFFSET_CAPTURE)) {
602 foreach ($matches[0] as $m) {
603 if ($m[0][0] == '<' && $m[0][1] == '/') {
606 $level = 0; // malformed HTML: go to next blockquote
608 else if ($level > 0) {
609 // skip inner blockquote
613 $len = $end - $taglen - $start;
614 // Get blockquote content
615 $body = substr($text, $start + $taglen - $diff, $len);
618 $p_width = $this->width;
619 if ($this->width > 0) $this->width -= 2;
620 // Convert blockquote content
622 $this->_converter($body);
623 // Add citation markers and create PRE block
624 $body = preg_replace('/((^|\n)>*)/', '\\1> ', trim($body));
625 $body = '<pre>' . htmlspecialchars($body) . '</pre>';
627 $this->width = $p_width;
629 $text = substr($text, 0, $start - $diff)
630 . $body . substr($text, $end + strlen($m[0]) - $diff);
632 $diff = $len + $taglen + strlen($m[0]) - strlen($body);
639 $taglen = strlen($m[0]);
648 * Callback function for preg_replace_callback use.
650 * @param array PREG matches
653 private function _preg_callback($matches)
655 switch($matches[1]) {
658 return $this->_strtoupper($matches[2]);
660 return $this->_strtoupper("\t\t". $matches[2] ."\n");
662 return $this->_strtoupper("\n\n". $matches[2] ."\n\n");
664 // Remove spaces in URL (#1487805)
665 $url = str_replace(' ', '', $matches[3]);
666 return $this->_build_link_list($url, $matches[4]);
671 * Callback function for preg_replace_callback use in PRE content handler.
673 * @param array PREG matches
676 private function _preg_pre_callback($matches)
678 return $this->pre_content;
682 * Strtoupper multibyte wrapper function
687 private function _strtoupper($str)
689 if (function_exists('mb_strtoupper'))
690 return mb_strtoupper($str);
692 return strtoupper($str);