githubFork/roundcubemail.git

alecpl

2008-04-29 6ee5ed253b92593ac5784300ac425f34f7fd1728

commit \| author \| age
4e17e6	1	<?php
T	2
	3	/*************************************************************************
	4	* *
	5	* class.html2text.inc *
	6	* *
	7	*************************************************************************
	8	* *
	9	* Converts HTML to formatted plain text *
	10	* *
	11	* Copyright (c) 2005 Jon Abernathy <jon@chuggnutt.com> *
	12	* All rights reserved. *
	13	* *
	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. *
	18	* *
	19	* The GNU General Public License can be found at *
	20	* http://www.gnu.org/copyleft/gpl.html. *
	21	* *
	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. *
	26	* *
	27	* Author(s): Jon Abernathy <jon@chuggnutt.com> *
	28	* *
	29	* Last modified: 04/06/05 *
	30	* Modified: 2004/05/19 (tbr) *
	31	* *
	32	*************************************************************************/
	33
	34
	35	/**
	36	* Takes HTML and converts it to formatted, plain text.
	37	*
	38	* Thanks to Alexander Krug (http://www.krugar.de/) to pointing out and
	39	* correcting an error in the regexp search array. Fixed 7/30/03.
	40	*
	41	* Updated set_html() function's file reading mechanism, 9/25/03.
	42	*
	43	* Thanks to Joss Sanglier (http://www.dancingbear.co.uk/) for adding
	44	* several more HTML entity codes to the $search and $replace arrays.
	45	* Updated 11/7/03.
	46	*
	47	* Thanks to Darius Kasperavicius (http://www.dar.dar.lt/) for
	48	* suggesting the addition of $allowed_tags and its supporting function
	49	* (which I slightly modified). Updated 3/12/04.
	50	*
	51	* Thanks to Justin Dearing for pointing out that a replacement for the
	52	* <TH> tag was missing, and suggesting an appropriate fix.
	53	* Updated 8/25/04.
	54	*
	55	* Thanks to Mathieu Collas (http://www.myefarm.com/) for finding a
	56	* display/formatting bug in the _build_link_list() function: email
	57	* readers would show the left bracket and number ("[1") as part of the
	58	* rendered email address.
	59	* Updated 12/16/04.
	60	*
	61	* Thanks to Wojciech Bajon (http://histeria.pl/) for submitting code
	62	* to handle relative links, which I hadn't considered. I modified his
	63	* code a bit to handle normal HTTP links and MAILTO links. Also for
	64	* suggesting three additional HTML entity codes to search for.
	65	* Updated 03/02/05.
	66	*
	67	* Thanks to Jacob Chandler for pointing out another link condition
	68	* for the _build_link_list() function: "https".
	69	* Updated 04/06/05.
	70	*
	71	* @author Jon Abernathy <jon@chuggnutt.com>
	72	* @version 0.6.1
	73	* @since PHP 4.0.2
	74	*/
	75	class html2text
	76	{
	77
	78	/**
	79	* Contains the HTML content to convert.
	80	*
	81	* @var string $html
	82	* @access public
	83	*/
	84	var $html;
	85
	86	/**
	87	* Contains the converted, formatted text.
	88	*
	89	* @var string $text
	90	* @access public
	91	*/
	92	var $text;
	93
	94	/**
	95	* Maximum width of the formatted text, in columns.
	96	*
	97	* @var integer $width
	98	* @access public
	99	*/
	100	var $width = 70;
	101
	102	/**
	103	* List of preg* regular expression patterns to search for,
	104	* used in conjunction with $replace.
	105	*
	106	* @var array $search
	107	* @access public
	108	* @see $replace
	109	*/
	110	var $search = array(
	111	"/\r/", // Non-legal carriage return
	112	"/[\n\t]+/", // Newlines and tabs
	113	'/<script[^>]>.?<\/script>/i', // <script>s -- which strip_tags supposedly has problems with
	114	//'/<!-- .* -->/', // Comments -- which strip_tags might have problem a with
ab6f80	115	'/<a [^>]href=("\|\')([^"\']+)\1[^>]>(.+?)<\/a>/ie', // <a href="">
4e17e6	116	'/<h[123][^>]*>(.+?)<\/h[123]>/ie', // H1 - H3
T	117	'/<h[456][^>]*>(.+?)<\/h[456]>/ie', // H4 - H6
	118	'/<p[^>]*>/i', // <P>
	119	'/<br[^>]*>/i', // <br>
	120	'/<b[^>]*>(.+?)<\/b>/ie', // <b>
	121	'/<i[^>]*>(.+?)<\/i>/i', // <i>
	122	'/(<ul[^>]*>\|<\/ul>)/i', // <ul> and </ul>
	123	'/(<ol[^>]*>\|<\/ol>)/i', // <ol> and </ol>
	124	'/<li[^>]*>/i', // <li>
	125	'/<hr[^>]*>/i', // <hr>
	126	'/(<table[^>]*>\|<\/table>)/i', // <table> and </table>
	127	'/(<tr[^>]*>\|<\/tr>)/i', // <tr> and </tr>
	128	'/<td[^>]*>(.+?)<\/td>/i', // <td> and </td>
40ed9b	129	'/<th[^>]*>(.+?)<\/th>/ie', // <th> and </th>
4e17e6	130	'/ /i',
T	131	'/"/i',
	132	'/>/i',
	133	'/</i',
40ed9b	134	'/&(amp\|#38);/i',
4e17e6	135	'/©/i',
T	136	'/™/i',
	137	'/“/',
	138	'/”/',
	139	'/–/',
40ed9b	140	'/&#(8217\|39);/',
4e17e6	141	'/©/',
T	142	'/™/',
	143	'//',
	144	'//',
	145	'//',
	146	'//',
	147	'/®/i',
	148	'/•/i',
	149	'/&[&;]+;/i'
	150	);
	151
	152	/**
	153	* List of pattern replacements corresponding to patterns searched.
	154	*
	155	* @var array $replace
	156	* @access public
	157	* @see $search
	158	*/
	159	var $replace = array(
	160	'', // Non-legal carriage return
	161	' ', // Newlines and tabs
	162	'', // <script>s -- which strip_tags supposedly has problems with
5cc4b1	163	//'', // Comments -- which strip_tags might have problem a with
ab6f80	164	'$this->_build_link_list("\\2", "\\3")', // <a href="">
4e17e6	165	"strtoupper(\"\n\n\\1\n\n\")", // H1 - H3
5cc4b1	166	"ucwords(\"\n\n\\1\n\")", // H4 - H6
T	167	"\n\n", // <P>
4e17e6	168	"\n", // <br>
T	169	'strtoupper("\\1")', // <b>
	170	'_\\1_', // <i>
	171	"\n\n", // <ul> and </ul>
	172	"\n\n", // <ol> and </ol>
	173	"\t*", // <li>
	174	"\n-------------------------\n", // <hr>
	175	"\n\n", // <table> and </table>
	176	"\n", // <tr> and </tr>
	177	"\t\t\\1\n", // <td> and </td>
	178	"strtoupper(\"\t\t\\1\n\")", // <th> and </th>
	179	' ',
	180	'"',
	181	'>',
	182	'<',
	183	'&',
	184	'(c)',
	185	'(tm)',
	186	'"',
	187	'"',
	188	'-',
	189	"'",
	190	'(c)',
	191	'(tm)',
	192	'--',
	193	'"',
	194	'"',
	195	'*',
	196	'(R)',
	197	'*',
	198	''
	199	);
	200
	201	/**
	202	* Contains a list of HTML tags to allow in the resulting text.
	203	*
	204	* @var string $allowed_tags
	205	* @access public
	206	* @see set_allowed_tags()
	207	*/
	208	var $allowed_tags = '';
	209
	210	/**
	211	* Contains the base URL that relative links should resolve to.
	212	*
	213	* @var string $url
	214	* @access public
	215	*/
	216	var $url;
	217
	218	/**
	219	* Indicates whether content in the $html variable has been converted yet.
	220	*
	221	* @var boolean $converted
	222	* @access private
	223	* @see $html, $text
	224	*/
	225	var $_converted = false;
	226
	227	/**
	228	* Contains URL addresses from links to be rendered in plain text.
	229	*
	230	* @var string $link_list
	231	* @access private
	232	* @see _build_link_list()
	233	*/
	234	var $_link_list = array();
dd792e	235
S	236	/**
	237	* Boolean flag, true if a table of link URLs should be listed after the text.
	238	*
	239	* @var boolean $_do_links
	240	* @access private
	241	* @see html2text()
	242	*/
	243	var $_do_links = true;
4e17e6	244
T	245	/**
	246	* Constructor.
	247	*
	248	* If the HTML source string (or file) is supplied, the class
	249	* will instantiate with that source propagated, all that has
	250	* to be done it to call get_text().
	251	*
	252	* @param string $source HTML content
	253	* @param boolean $from_file Indicates $source is a file to pull content from
dd792e	254	* @param boolean $do_link_table indicate whether a table of link URLs is desired
4e17e6	255	* @access public
T	256	* @return void
	257	*/
5cc4b1	258	function html2text( $source = '', $from_file = false, $produce_link_table = true )
4e17e6	259	{
T	260	if ( !empty($source) ) {
	261	$this->set_html($source, $from_file);
	262	}
	263	$this->set_base_url();
dd792e	264	$this->_do_links = $produce_link_table;
4e17e6	265	}
T	266
	267	/**
	268	* Loads source HTML into memory, either from $source string or a file.
	269	*
	270	* @param string $source HTML content
	271	* @param boolean $from_file Indicates $source is a file to pull content from
	272	* @access public
	273	* @return void
	274	*/
	275	function set_html( $source, $from_file = false )
	276	{
	277	$this->html = $source;
	278
	279	if ( $from_file && file_exists($source) ) {
	280	$fp = fopen($source, 'r');
	281	$this->html = fread($fp, filesize($source));
	282	fclose($fp);
	283	}
	284
	285	$this->_converted = false;
	286	}
	287
	288	/**
	289	* Returns the text, converted from HTML.
	290	*
	291	* @access public
	292	* @return string
	293	*/
	294	function get_text()
	295	{
	296	if ( !$this->_converted ) {
	297	$this->_convert();
	298	}
	299
	300	return $this->text;
	301	}
	302
	303	/**
	304	* Prints the text, converted from HTML.
	305	*
	306	* @access public
	307	* @return void
	308	*/
	309	function print_text()
	310	{
	311	print $this->get_text();
	312	}
	313
	314	/**
	315	* Alias to print_text(), operates identically.
	316	*
	317	* @access public
	318	* @return void
	319	* @see print_text()
	320	*/
	321	function p()
	322	{
	323	print $this->get_text();
	324	}
	325
	326	/**
	327	* Sets the allowed HTML tags to pass through to the resulting text.
	328	*
	329	* Tags should be in the form "<p>", with no corresponding closing tag.
	330	*
	331	* @access public
	332	* @return void
	333	*/
	334	function set_allowed_tags( $allowed_tags = '' )
	335	{
	336	if ( !empty($allowed_tags) ) {
	337	$this->allowed_tags = $allowed_tags;
	338	}
	339	}
	340
	341	/**
	342	* Sets a base URL to handle relative links.
	343	*
	344	* @access public
	345	* @return void
	346	*/
	347	function set_base_url( $url = '' )
	348	{
	349	if ( empty($url) ) {
	350	$this->url = 'http://' . $_SERVER['HTTP_HOST'];
	351	} else {
	352	// Strip any trailing slashes for consistency (relative
	353	// URLs may already start with a slash like "/file.html")
	354	if ( substr($url, -1) == '/' ) {
	355	$url = substr($url, 0, -1);
	356	}
	357	$this->url = $url;
	358	}
	359	}
	360
	361	/**
	362	* Workhorse function that does actual conversion.
	363	*
	364	* First performs custom tag replacement specified by $search and
	365	* $replace arrays. Then strips any remaining HTML tags, reduces whitespace
	366	* and newlines to a readable format, and word wraps the text to
	367	* $width characters.
	368	*
	369	* @access private
	370	* @return void
	371	*/
	372	function _convert()
	373	{
	374	// Variables used for building the link list
	375	//$link_count = 1;
	376	//$this->_link_list = '';
	377
	378	$text = trim(stripslashes($this->html));
	379
	380	// Run our defined search-and-replace
	381	$text = preg_replace($this->search, $this->replace, $text);
	382
	383	// Strip any other HTML tags
	384	$text = strip_tags($text, $this->allowed_tags);
	385
	386	// Bring down number of empty lines to 2 max
	387	$text = preg_replace("/\n\s+\n/", "\n", $text);
	388	$text = preg_replace("/[\n]{3,}/", "\n\n", $text);
	389
	390	// Add link list
	391	if ( sizeof($this->_link_list) ) {
	392	$text .= "\n\nLinks:\n------\n";
	393	foreach ($this->_link_list as $id => $link) {
	394	$text .= '[' . ($id+1) . '] ' . $link . "\n";
	395	}
	396	}
	397
	398	// Wrap the text to a readable format
	399	// for PHP versions >= 4.0.2. Default width is 75
	400	$text = wordwrap($text, $this->width);
	401
	402	$this->text = $text;
	403
	404	$this->_converted = true;
	405	}
	406
	407	/**
	408	* Helper function called by preg_replace() on link replacement.
	409	*
	410	* Maintains an internal list of links to be displayed at the end of the
	411	* text, with numeric indices to the original point in the text they
	412	* appeared. Also makes an effort at identifying and handling absolute
	413	* and relative links.
	414	*
	415	* @param integer $link_count Counter tracking current link number
	416	* @param string $link URL of the link
	417	* @param string $display Part of the text to associate number with
	418	* @access private
	419	* @return string
	420	*/
	421	function _build_link_list($link, $display)
	422	{
dd792e	423	if (! $this->_do_links) return $display;
S	424
4e17e6	425	$link_lc = strtolower($link);
T	426
	427	if (substr($link_lc, 0, 7) == 'http://' \|\| substr($link_lc, 0, 8) == 'https://' \|\| substr($link_lc, 0, 7) == 'mailto:')
	428	{
	429	$url = $link;
	430	}
	431	else
	432	{
	433	$url = $this->url;
	434	if ($link{0} != '/') {
	435	$url .= '/';
	436	}
	437	$url .= $link;
	438	}
	439
	440	$index = array_search($url, $this->_link_list);
	441	if ($index===FALSE)
	442	{
	443	$index = sizeof($this->_link_list);
	444	$this->_link_list[$index] = $url;
	445	}
	446
	447	return $display . ' [' . ($index+1) . ']';
	448	}
	449	}
	450
	451	?>