/**
* \brief Parsed HTML tree walker( used by HTML sucker )
*
* This is initial implementation( stupid... w/o any intellegence( almost : ) )
* It is rapidly designed version... just for test: 'can this feature be useful'.
* Later it should be replaced by well designed one :) don't bash me now : )
*
* \param &$c array -- parsed HTML
* \param &$src string -- output string
* \param &$p array -- ['stack'] = closing strings stack,
['listack'] = stack of list types currently opened
['first_td'] = flag: 'is <tr> was just before this <td>'
*/
function walk_and_parse(&$c, &$src, &$p)
{
for ($i = 0; $i <= $c["contentpos"]; $i++) {
// If content type 'text' output it to destination...
if ($c[$i]["type"] == "text") {
$src .= $c[$i]["data"];
} elseif ($c[$i]["type"] == "tag") {
if ($c[$i]["data"]["type"] == "open") {
// Open tag type
switch ($c[$i]["data"]["name"]) {
case "br":
$src .= "\n";
break;
case "title":
$src .= "\n!";
$p['stack'][] = array('tag' => 'title', 'string' => "\n");
break;
case "p":
$src .= "\n";
$p['stack'][] = array('tag' => 'p', 'string' => "\n");
break;
case "b":
$src .= '__';
$p['stack'][] = array('tag' => 'b', 'string' => '__');
break;
case "i":
$src .= "''";
$p['stack'][] = array('tag' => 'i', 'string' => "''");
break;
case "u":
$src .= "==";
$p['stack'][] = array('tag' => 'u', 'string' => "==");
break;
case "center":
$src .= '::';
$p['stack'][] = array('tag' => 'center', 'string' => '::');
break;
case "code":
$src .= '-+';
$p['stack'][] = array('tag' => 'code', 'string' => '+-');
break;
// headers detection looks like real suxx code...
// but possible it run faster :) I don't know where is profiler in PHP...
// headers detection looks like real suxx code...
// but possible it run faster :) I don't know where is profiler in PHP...
case "h1":
$src .= "\n!";
$p['stack'][] = array('tag' => 'h1', 'string' => "\n");
break;
case "h2":
$src .= "\n!!";
$p['stack'][] = array('tag' => 'h2', 'string' => "\n");
break;
case "h3":
$src .= "\n!!!";
$p['stack'][] = array('tag' => 'h3', 'string' => "\n");
break;
case "h4":
$src .= "\n!!!!";
$p['stack'][] = array('tag' => 'h4', 'string' => "\n");
break;
case "h5":
$src .= "\n!!!!!";
$p['stack'][] = array('tag' => 'h5', 'string' => "\n");
break;
case "h6":
$src .= "\n!!!!!!";
$p['stack'][] = array('tag' => 'h6', 'string' => "\n");
break;
case "pre":
$src .= '~pp~';
$p['stack'][] = array('tag' => 'pre', 'string' => '~/pp~');
break;
// Table parser
// Table parser
case "table":
$src .= '||';
$p['stack'][] = array('tag' => 'table', 'string' => '||');
break;
case "tr":
$p['first_td'] = true;
break;
case "td":
$src .= $p['first_td'] ? '' : '|';
$p['first_td'] = false;
break;
// Lists parser
// Lists parser
case "ul":
$p['listack'][] = '*';
break;
case "ol":
$p['listack'][] = '#';
break;
case "li":
// Generate wiki list item according to current list depth.
//( ensure '*/#' starts from begining of line )
for ($l = ''; strlen($l) < count($p['listack']); $l .= end($p['listack'])) {
}
$src .= "\n{$l} ";
break;
case "font":
// If color attribute present in <font> tag
if (isset($c[$i]["pars"]["color"]["value"])) {
$src .= '~~' . $c[$i]["pars"]["color"]["value"] . ':';
$p['stack'][] = array('tag' => 'font', 'string' => '~~');
}
break;
case "img":
// If src attribute present in <img> tag
if (isset($c[$i]["pars"]["src"]["value"])) {
// Note what it produce( img ) not {img}! Will fix this below...
$src .= '( img src=' . $c[$i]["pars"]["src"]["value"] . ' )';
}
break;
case "a":
// If href attribute present in <a> tag
if (isset($c[$i]["pars"]["href"]["value"])) {
$src .= '[' . $c[$i]["pars"]["href"]["value"] . '|';
$p['stack'][] = array('tag' => 'a', 'string' => ']');
}
break;
}
} else {
// This is close tag type. Is that smth we r waiting for?
switch ($c[$i]["data"]["name"]) {
case "ul":
if (end($p['listack']) == '*') {
array_pop($p['listack']);
}
break;
case "ol":
if (end($p['listack']) == '#') {
array_pop($p['listack']);
}
break;
default:
$e = end($p['stack']);
if ($c[$i]["data"]["name"] == $e['tag']) {
$src .= $e['string'];
array_pop($p['stack']);
}
break;
}
}
}
// Recursive call on tags with content...
if (isset($c[$i]["content"])) {
// if( substr( $src, -1 )!= " " )$src .= " ";
walk_and_parse($c[$i]["content"], $src, $p);
}
}
}
/**
* wrapper around zaufi's HTML sucker code just to use the html to wiki bit
*
* \param &$c string -- HTML in
* \param &$src string -- output string
*/
function parse_html(&$inHtml)
{
//error_reporting(6143);
// Read compiled (serialized) grammar
$grammarfile = 'lib/htmlparser/htmlgrammar.cmp';
if (!($fp = @fopen($grammarfile, 'r'))) {
$smarty->assign('msg', tra("Can't parse HTML data - no grammar file"));
$smarty->display("error.tpl");
die;
}
$grammar = unserialize(fread($fp, filesize($grammarfile)));
fclose($fp);
// create parser object, insert html code and parse it
$htmlparser = new HtmlParser($inHtml, $grammar, '', 0);
$htmlparser->Parse();
// Should I try to convert HTML to wiki?
$out_data = '';
$p = array('stack' => array(), 'listack' => array(), 'first_td' => false, 'first_tr' => false);
walk_and_parse($htmlparser->content, $out_data, $p, '');
// Is some tags still opened? (It can be if HTML not valid, but this is not reason
// to produce invalid wiki :)
while (count($p['stack'])) {
$e = end($p['stack']);
$out_data .= $e['string'];
array_pop($p['stack']);
}
// Unclosed lists r ignored... wiki have no special start/end lists syntax....
// OK. Things remains to do:
// 1) fix linked images
$out_data = preg_replace(',\\[(.*)\\|\\(img src=(.*)\\)\\],mU', '{img src=$2 link=$1}', $out_data);
// 2) fix remains images (not in links)
$out_data = preg_replace(',\\(img src=(.*)\\),mU', '{img src=$1}', $out_data);
// 3) remove empty lines
$out_data = preg_replace(",[\n]+,mU", "\n", $out_data);
// 4) remove nbsp's
$out_data = preg_replace(", ,mU", " ", $out_data);
return $out_data;
}
