Removing tags from HTML

I'm not that great with php but to me this looks like you're blacklisting tags. This is not an approach I would recommend as it is easy to miss some tags or combinations of attributes.

Instead I would recommend to use a white list of tags that you allow and block everything else. This is more secure as if you miss to white list something, your users will complain and you can fix it. Xss attackers typically won't inform you if you forget to blacklist a tag and have a vulnerability ;) And you are also automatically protected from new tags and attribute combinations.

Or you could simply strip ALL tags or html escape the entire body.

I may have flaws in my suggested code (because I didn't test it), but it seems to me that you could afford to implement the DRY principle. Writing a loop, to generate an array of items, then writing another loop to traverse the array of items just doesn't make sense to me -- just perform all necessary processes in the first loop (no temporary arrays).

I recommend curly brackets for your conditionals and loops; not because they are essential, but because they improve readability for most people (never assume that you are the only person to see your code).

Untested Suggested Code:

class html 
{
    public static function parse(String $html) : String
    {
        $tidy = new \tidy();
        $html = $tidy->repairString($html);
        $dom = new \DOMDocument();
        $dom->loadHTML('<?xml encoding="UTF-8">' . $html);
        // dirty fix
        foreach ($dom->childNodes as $item) {
            if ($item->nodeType == XML_PI_NODE) {
                $dom->removeChild($item); // remove hack
            }
        }
        $dom->encoding = 'UTF-8'; // insert proper
        $tagnames = ['script','style','iframe','applet','video','audio','link','meta','head','form','input','textarea'];  // use array to make DRY
        foreach ($tagnames as $tagname) {
            $dom->getElementsByTagName($tagname)->parentNode->removeChild($item);
        }

        $imgs = $dom->getElementsByTagName('img');
        foreach ($imgs as $img) {
            foreach ($img->attributes as $attr) {
                if ($attr->nodeName == 'src' && !in_array(substr($attr->nodeValue,0,7),['base64:','https:/','http://'])) {  // condensed this line
                    $img->parentNode->removeChild($item);
                }
            }
        }

        $xpath = new \DOMXPath($dom);
        $nodes = $xpath->query('//@*');
        foreach ($nodes as $node) {
            if (!in_array(str_replace('data-mce-','',$node->nodeName),['href','src','class','style','width','height','data-href','title','target','rel'])) { // i use tinyMCE and sometimes it places data-mce-* attributes
                $node->parentNode->removeAttribute($node->nodeName);
            }
        }
        $html = $dom->saveHTML();
        $buffer = strip_tags($html, '<figure><section><p><strong><em><u><h1><h2><h3><h4><h5><h6><img><li><ol><ul><span><div><br><ins><del><a>');
        $clean = $tidy->repairString($buffer);
        return $clean;
    }
}

These changes not only shorten your code block, they improve readability with curly brackets and tabbing, and improve efficiency by reducing the number of loops.