You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
roundcubemail/program/lib/Roundcube/rcube_html2text.php

695 lines
23 KiB
PHTML

19 years ago
<?php
/**
+-----------------------------------------------------------------------+
| This file is part of the Roundcube Webmail client |
| Copyright (C) 2008-2012, The Roundcube Dev Team |
| Copyright (c) 2005-2007, Jon Abernathy <jon@chuggnutt.com> |
| |
| Licensed under the GNU General Public License version 3 or |
| any later version with exceptions for skins & plugins. |
| See the README file for a full license statement. |
| |
| PURPOSE: |
| Converts HTML to formatted plain text (based on html2text class) |
+-----------------------------------------------------------------------+
| Author: Thomas Bruederli <roundcube@gmail.com> |
| Author: Aleksander Machniak <alec@alec.pl> |
| Author: Jon Abernathy <jon@chuggnutt.com> |
+-----------------------------------------------------------------------+
*/
19 years ago
/**
* Takes HTML and converts it to formatted, plain text.
*
* Thanks to Alexander Krug (http://www.krugar.de/) to pointing out and
* correcting an error in the regexp search array. Fixed 7/30/03.
*
* Updated set_html() function's file reading mechanism, 9/25/03.
*
* Thanks to Joss Sanglier (http://www.dancingbear.co.uk/) for adding
* several more HTML entity codes to the $search and $replace arrays.
* Updated 11/7/03.
*
* Thanks to Darius Kasperavicius (http://www.dar.dar.lt/) for
* suggesting the addition of $allowed_tags and its supporting function
* (which I slightly modified). Updated 3/12/04.
*
* Thanks to Justin Dearing for pointing out that a replacement for the
* <TH> tag was missing, and suggesting an appropriate fix.
* Updated 8/25/04.
*
* Thanks to Mathieu Collas (http://www.myefarm.com/) for finding a
* display/formatting bug in the _build_link_list() function: email
* readers would show the left bracket and number ("[1") as part of the
* rendered email address.
* Updated 12/16/04.
*
* Thanks to Wojciech Bajon (http://histeria.pl/) for submitting code
* to handle relative links, which I hadn't considered. I modified his
* code a bit to handle normal HTTP links and MAILTO links. Also for
* suggesting three additional HTML entity codes to search for.
* Updated 03/02/05.
*
* Thanks to Jacob Chandler for pointing out another link condition
* for the _build_link_list() function: "https".
* Updated 04/06/05.
*
* Thanks to Marc Bertrand (http://www.dresdensky.com/) for
* suggesting a revision to the word wrapping functionality; if you
* specify a $width of 0 or less, word wrapping will be ignored.
* Updated 11/02/06.
*
* *** Big housecleaning updates below:
*
* Thanks to Colin Brown (http://www.sparkdriver.co.uk/) for
* suggesting the fix to handle </li> and blank lines (whitespace).
* Christian Basedau (http://www.movetheweb.de/) also suggested the
* blank lines fix.
*
* Special thanks to Marcus Bointon (http://www.synchromedia.co.uk/),
* Christian Basedau, Norbert Laposa (http://ln5.co.uk/),
* Bas van de Weijer, and Marijn van Butselaar
* for pointing out my glaring error in the <th> handling. Marcus also
* supplied a host of fixes.
*
* Thanks to Jeffrey Silverman (http://www.newtnotes.com/) for pointing
* out that extra spaces should be compressed--a problem addressed with
* Marcus Bointon's fixes but that I had not yet incorporated.
*
* Thanks to Daniel Schledermann (http://www.typoconsult.dk/) for
* suggesting a valuable fix with <a> tag handling.
*
* Thanks to Wojciech Bajon (again!) for suggesting fixes and additions,
* including the <a> tag handling that Daniel Schledermann pointed
* out but that I had not yet incorporated. I haven't (yet)
* incorporated all of Wojciech's changes, though I may at some
* future time.
*
* *** End of the housecleaning updates. Updated 08/08/07.
*/
/**
* Converts HTML to formatted plain text
*
* @package Framework
* @subpackage Utils
*/
class rcube_html2text
19 years ago
{
/**
* Contains the HTML content to convert.
19 years ago
*
* @var string $html
19 years ago
*/
protected $html;
19 years ago
/**
* Contains the converted, formatted text.
19 years ago
*
* @var string $text
19 years ago
*/
protected $text;
19 years ago
/**
* Maximum width of the formatted text, in columns.
19 years ago
*
* Set this value to 0 (or less) to ignore word wrapping
* and not constrain text to a fixed-width column.
*
* @var integer $width
19 years ago
*/
protected $width = 70;
19 years ago
/**
* Target character encoding for output text
*
* @var string $charset
*/
protected $charset = 'UTF-8';
19 years ago
/**
* List of preg* regular expression patterns to search for,
* used in conjunction with $replace.
19 years ago
*
* @var array $search
* @see $replace
19 years ago
*/
protected $search = array(
19 years ago
"/\r/", // Non-legal carriage return
"/[\n\t]+/", // Newlines and tabs
'/<head[^>]*>.*?<\/head>/i', // <head>
19 years ago
'/<script[^>]*>.*?<\/script>/i', // <script>s -- which strip_tags supposedly has problems with
'/<style[^>]*>.*?<\/style>/i', // <style>s -- which strip_tags supposedly has problems with
19 years ago
'/<p[^>]*>/i', // <P>
'/<br[^>]*>/i', // <br>
'/<i[^>]*>(.*?)<\/i>/i', // <i>
'/<em[^>]*>(.*?)<\/em>/i', // <em>
19 years ago
'/(<ul[^>]*>|<\/ul>)/i', // <ul> and </ul>
'/(<ol[^>]*>|<\/ol>)/i', // <ol> and </ol>
'/<li[^>]*>(.*?)<\/li>/i', // <li> and </li>
19 years ago
'/<li[^>]*>/i', // <li>
'/<hr[^>]*>/i', // <hr>
'/<div[^>]*>/i', // <div>
19 years ago
'/(<table[^>]*>|<\/table>)/i', // <table> and </table>
'/(<tr[^>]*>|<\/tr>)/i', // <tr> and </tr>
'/<td[^>]*>(.*?)<\/td>/i', // <td> and </td>
19 years ago
);
/**
* List of pattern replacements corresponding to patterns searched.
19 years ago
*
* @var array $replace
* @see $search
19 years ago
*/
protected $replace = array(
19 years ago
'', // Non-legal carriage return
' ', // Newlines and tabs
'', // <head>
19 years ago
'', // <script>s -- which strip_tags supposedly has problems with
'', // <style>s -- which strip_tags supposedly has problems with
"\n\n", // <P>
19 years ago
"\n", // <br>
'_\\1_', // <i>
'_\\1_', // <em>
19 years ago
"\n\n", // <ul> and </ul>
"\n\n", // <ol> and </ol>
"\t* \\1\n", // <li> and </li>
"\n\t* ", // <li>
"\n-------------------------\n", // <hr>
"<div>\n", // <div>
"\n\n", // <table> and </table>
19 years ago
"\n", // <tr> and </tr>
"\t\t\\1\n", // <td> and </td>
);
/**
* List of preg* regular expression patterns to search for,
* used in conjunction with $ent_replace.
*
* @var array $ent_search
* @see $ent_replace
*/
protected $ent_search = array(
'/&(nbsp|#160);/i', // Non-breaking space
'/&(quot|rdquo|ldquo|#8220|#8221|#147|#148);/i',
// Double quotes
'/&(apos|rsquo|lsquo|#8216|#8217);/i', // Single quotes
'/&gt;/i', // Greater-than
'/&lt;/i', // Less-than
'/&(copy|#169);/i', // Copyright
'/&(trade|#8482|#153);/i', // Trademark
'/&(reg|#174);/i', // Registered
'/&(mdash|#151|#8212);/i', // mdash
'/&(ndash|minus|#8211|#8722);/i', // ndash
'/&(bull|#149|#8226);/i', // Bullet
'/&(pound|#163);/i', // Pound sign
'/&(euro|#8364);/i', // Euro sign
'/&(amp|#38);/i', // Ampersand: see _converter()
'/[ ]{2,}/', // Runs of spaces, post-handling
);
/**
* List of pattern replacements corresponding to patterns searched.
*
* @var array $ent_replace
* @see $ent_search
*/
protected $ent_replace = array(
' ', // Non-breaking space
'"', // Double quotes
"'", // Single quotes
19 years ago
'>',
'<',
'(c)',
'(tm)',
'(R)',
'--',
'-',
19 years ago
'*',
'£',
'EUR', // Euro sign. <20> ?
'|+|amp|+|', // Ampersand: see _converter()
' ', // Runs of spaces, post-handling
19 years ago
);
/**
* List of preg* regular expression patterns to search for
* and replace using callback function.
*
* @var array $callback_search
*/
protected $callback_search = array(
'/<(a) [^>]*href=("|\')([^"\']+)\2[^>]*>(.*?)<\/a>/i', // <a href="">
'/<(h)[123456]( [^>]*)?>(.*?)<\/h[123456]>/i', // h1 - h6
'/<(b)( [^>]*)?>(.*?)<\/b>/i', // <b>
'/<(strong)( [^>]*)?>(.*?)<\/strong>/i', // <strong>
'/<(th)( [^>]*)?>(.*?)<\/th>/i', // <th> and </th>
);
/**
* List of preg* regular expression patterns to search for in PRE body,
* used in conjunction with $pre_replace.
*
* @var array $pre_search
* @see $pre_replace
*/
protected $pre_search = array(
"/\n/",
"/\t/",
'/ /',
'/<pre[^>]*>/',
'/<\/pre>/'
);
/**
* List of pattern replacements corresponding to patterns searched for PRE body.
*
* @var array $pre_replace
* @see $pre_search
*/
protected $pre_replace = array(
'<br>',
'&nbsp;&nbsp;&nbsp;&nbsp;',
'&nbsp;',
'',
''
);
19 years ago
/**
* Contains a list of HTML tags to allow in the resulting text.
19 years ago
*
* @var string $allowed_tags
* @see set_allowed_tags()
19 years ago
*/
protected $allowed_tags = '';
19 years ago
/**
* Contains the base URL that relative links should resolve to.
19 years ago
*
* @var string $url
19 years ago
*/
protected $url;
19 years ago
/**
* Indicates whether content in the $html variable has been converted yet.
19 years ago
*
* @var boolean $_converted
* @see $html, $text
19 years ago
*/
protected $_converted = false;
19 years ago
/**
* Contains URL addresses from links to be rendered in plain text.
19 years ago
*
* @var array $_link_list
* @see _build_link_list()
19 years ago
*/
protected $_link_list = array();
19 years ago
/**
* Boolean flag, true if a table of link URLs should be listed after the text.
*
* @var boolean $_do_links
* @see __construct()
*/
protected $_do_links = true;
19 years ago
/**
* Constructor.
19 years ago
*
* If the HTML source string (or file) is supplied, the class
* will instantiate with that source propagated, all that has
* to be done it to call get_text().
19 years ago
*
* @param string $source HTML content
* @param boolean $from_file Indicates $source is a file to pull content from
* @param boolean $do_links Indicate whether a table of link URLs is desired
* @param integer $width Maximum width of the formatted text, 0 for no limit
19 years ago
*/
function __construct($source = '', $from_file = false, $do_links = true, $width = 75, $charset = 'UTF-8')
19 years ago
{
if (!empty($source)) {
19 years ago
$this->set_html($source, $from_file);
}
19 years ago
$this->set_base_url();
$this->_do_links = $do_links;
$this->width = $width;
$this->charset = $charset;
19 years ago
}
/**
* Loads source HTML into memory, either from $source string or a file.
19 years ago
*
* @param string $source HTML content
* @param boolean $from_file Indicates $source is a file to pull content from
19 years ago
*/
function set_html($source, $from_file = false)
19 years ago
{
if ($from_file && file_exists($source)) {
$this->html = file_get_contents($source);
19 years ago
}
else {
$this->html = $source;
}
19 years ago
$this->_converted = false;
}
/**
* Returns the text, converted from HTML.
19 years ago
*
* @return string Plain text
19 years ago
*/
function get_text()
{
if (!$this->_converted) {
19 years ago
$this->_convert();
}
return $this->text;
}
/**
* Prints the text, converted from HTML.
19 years ago
*/
function print_text()
{
print $this->get_text();
}
/**
* Sets the allowed HTML tags to pass through to the resulting text.
19 years ago
*
* Tags should be in the form "<p>", with no corresponding closing tag.
19 years ago
*/
function set_allowed_tags($allowed_tags = '')
19 years ago
{
if (!empty($allowed_tags)) {
19 years ago
$this->allowed_tags = $allowed_tags;
}
}
/**
* Sets a base URL to handle relative links.
19 years ago
*/
function set_base_url($url = '')
19 years ago
{
if (empty($url)) {
if (!empty($_SERVER['HTTP_HOST'])) {
$this->url = 'http://' . $_SERVER['HTTP_HOST'];
}
else {
$this->url = '';
}
}
else {
19 years ago
// Strip any trailing slashes for consistency (relative
// URLs may already start with a slash like "/file.html")
if (substr($url, -1) == '/') {
19 years ago
$url = substr($url, 0, -1);
}
$this->url = $url;
}
}
/**
* Workhorse function that does actual conversion (calls _converter() method).
19 years ago
*/
protected function _convert()
19 years ago
{
// Variables used for building the link list
$this->_link_list = array();
19 years ago
$text = trim(stripslashes($this->html));
// Convert HTML to TXT
$this->_converter($text);
// Add link list
if (!empty($this->_link_list)) {
$text .= "\n\nLinks:\n------\n";
foreach ($this->_link_list as $idx => $url) {
$text .= '[' . ($idx+1) . '] ' . $url . "\n";
}
}
$this->text = $text;
$this->_converted = true;
}
/**
* Workhorse function that does actual conversion.
*
* First performs custom tag replacement specified by $search and
* $replace arrays. Then strips any remaining HTML tags, reduces whitespace
* and newlines to a readable format, and word wraps the text to
* $width characters.
*
* @param string Reference to HTML content string
*/
protected function _converter(&$text)
{
// Convert <BLOCKQUOTE> (before PRE!)
$this->_convert_blockquotes($text);
// Convert <PRE>
$this->_convert_pre($text);
// Run our defined tags search-and-replace
19 years ago
$text = preg_replace($this->search, $this->replace, $text);
// Run our defined tags search-and-replace with callback
$text = preg_replace_callback($this->callback_search, array($this, 'tags_preg_callback'), $text);
// Strip any other HTML tags
$text = strip_tags($text, $this->allowed_tags);
// Run our defined entities/characters search-and-replace
$text = preg_replace($this->ent_search, $this->ent_replace, $text);
// Replace known html entities
$text = html_entity_decode($text, ENT_QUOTES, $this->charset);
// Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
$text = preg_replace('/&([a-zA-Z0-9]{2,6}|#[0-9]{2,4});/', '', $text);
// Convert "|+|amp|+|" into "&", need to be done after handling of unknown entities
// This properly handles situation of "&amp;quot;" in input string
$text = str_replace('|+|amp|+|', '&', $text);
19 years ago
// Bring down number of empty lines to 2 max
$text = preg_replace("/\n\s+\n/", "\n\n", $text);
19 years ago
$text = preg_replace("/[\n]{3,}/", "\n\n", $text);
// remove leading empty lines (can be produced by eg. P tag on the beginning)
$text = ltrim($text, "\n");
19 years ago
// Wrap the text to a readable format
// for PHP versions >= 4.0.2. Default width is 75
// If width is 0 or less, don't wrap the text.
if ( $this->width > 0 ) {
$text = wordwrap($text, $this->width);
}
19 years ago
}
/**
* Helper function called by preg_replace() on link replacement.
19 years ago
*
* Maintains an internal list of links to be displayed at the end of the
* text, with numeric indices to the original point in the text they
* appeared. Also makes an effort at identifying and handling absolute
* and relative links.
19 years ago
*
* @param string $link URL of the link
* @param string $display Part of the text to associate number with
*/
protected function _build_link_list( $link, $display )
{
if (!$this->_do_links || empty($link)) {
return $display;
}
// Ignored link types
if (preg_match('!^(javascript:|mailto:|#)!i', $link)) {
return $display;
}
if (preg_match('!^([a-z][a-z0-9.+-]+:)!i', $link)) {
$url = $link;
}
else {
$url = $this->url;
if (substr($link, 0, 1) != '/') {
$url .= '/';
19 years ago
}
$url .= "$link";
}
if (($index = array_search($url, $this->_link_list)) === false) {
$index = count($this->_link_list);
$this->_link_list[] = $url;
19 years ago
}
return $display . ' [' . ($index+1) . ']';
}
/**
* Helper function for PRE body conversion.
*
* @param string HTML content
*/
protected function _convert_pre(&$text)
{
// get the content of PRE element
while (preg_match('/<pre[^>]*>(.*)<\/pre>/ismU', $text, $matches)) {
$this->pre_content = $matches[1];
// Run our defined tags search-and-replace with callback
$this->pre_content = preg_replace_callback($this->callback_search,
array($this, 'tags_preg_callback'), $this->pre_content);
// convert the content
$this->pre_content = sprintf('<div><br>%s<br></div>',
preg_replace($this->pre_search, $this->pre_replace, $this->pre_content));
// replace the content (use callback because content can contain $0 variable)
$text = preg_replace_callback('/<pre[^>]*>.*<\/pre>/ismU',
array($this, 'pre_preg_callback'), $text, 1);
// free memory
$this->pre_content = '';
}
}
/**
* Helper function for BLOCKQUOTE body conversion.
*
* @param string HTML content
*/
protected function _convert_blockquotes(&$text)
{
$level = 0;
$offset = 0;
while (($start = strpos($text, '<blockquote', $offset)) !== false) {
$offset = $start + 12;
do {
$end = strpos($text, '</blockquote>', $offset);
$next = strpos($text, '<blockquote', $offset);
// nested <blockquote>, skip
if ($next !== false && $next < $end) {
$offset = $next + 12;
$level++;
}
// nested </blockquote> tag
if ($end !== false && $level > 0) {
$offset = $end + 12;
$level--;
}
// found matching end tag
else if ($end !== false && $level == 0) {
$taglen = strpos($text, '>', $start) - $start;
$startpos = $start + $taglen + 1;
// get blockquote content
$body = trim(substr($text, $startpos, $end - $startpos));
// replace content with inner blockquotes
$this->_converter($body);
// Add citation markers and create <pre> block
$body = preg_replace_callback('/((?:^|\n)>*)([^\n]*)/', array($this, 'blockquote_citation_ballback'), trim($body));
$body = '<pre>' . htmlspecialchars($body) . '</pre>';
$text = substr($text, 0, $start) . $body . "\n" . substr($text, $end + 13);
$offset = 0;
break;
}
} while ($end || $next);
}
}
/**
* Callback function to correctly add citation markers for blockquote contents
*/
public function blockquote_citation_ballback($m)
{
$line = ltrim($m[2]);
$space = $line[0] == '>' ? '' : ' ';
return $m[1] . '>' . $space . $line;
}
/**
* Callback function for preg_replace_callback use.
*
* @param array PREG matches
* @return string
*/
public function tags_preg_callback($matches)
{
switch (strtolower($matches[1])) {
case 'b':
case 'strong':
return $this->_toupper($matches[3]);
case 'th':
return $this->_toupper("\t\t". $matches[3] ."\n");
case 'h':
return $this->_toupper("\n\n". $matches[3] ."\n\n");
case 'a':
// Remove spaces in URL (#1487805)
$url = str_replace(' ', '', $matches[3]);
return $this->_build_link_list($url, $matches[4]);
}
}
/**
* Callback function for preg_replace_callback use in PRE content handler.
*
* @param array PREG matches
* @return string
*/
public function pre_preg_callback($matches)
{
return $this->pre_content;
}
/**
* Strtoupper function with HTML tags and entities handling.
*
* @param string $str Text to convert
* @return string Converted text
*/
private function _toupper($str)
{
// string can containg HTML tags
$chunks = preg_split('/(<[^>]*>)/', $str, null, PREG_SPLIT_NO_EMPTY | PREG_SPLIT_DELIM_CAPTURE);
// convert toupper only the text between HTML tags
foreach ($chunks as $idx => $chunk) {
if ($chunk[0] != '<') {
$chunks[$idx] = $this->_strtoupper($chunk);
}
}
return implode($chunks);
}
/**
* Strtoupper multibyte wrapper function with HTML entities handling.
*
* @param string $str Text to convert
* @return string Converted text
*/
private function _strtoupper($str)
{
$str = html_entity_decode($str, ENT_COMPAT, $this->charset);
$str = mb_strtoupper($str);
$str = htmlspecialchars($str, ENT_COMPAT, $this->charset);
return $str;
}
19 years ago
}