<?php
// https://github.com/mihaeu/html-formatter

namespace Mihaeu;

class HtmlFormatter
{
    /**
     * Formats HTML by re-indenting the tags and removing unnecessary whitespace.
     *
     * @param string $html HTML string.
     * @param string $indentWith Characters that are being used for indentation (default = 4 spaces).
     * @param string $tagsWithoutIndentation Comma-separated list of HTML tags that should not be indented (default = html,link,img,meta)
     * @return string Re-indented HTML.
     */
    public static function format($html, $indentWith = '    ', $tagsWithoutIndentation = 'html,link,img,meta')
    {
        // replace newlines (CRLF and LF), followed by a non-whitespace character, with a space
        $html = preg_replace('/\\r?\\n([^\s])/', ' $1', $html);

        // remove all remaining line feeds and replace tabs with spaces
        $html = str_replace(["\n", "\r", "\t"], ['', '', ' '], $html);
        $elements = preg_split('/(<.+>)/U', $html, -1, PREG_SPLIT_NO_EMPTY | PREG_SPLIT_DELIM_CAPTURE);
        $dom = self::parseDom($elements);

        $indent = 0;
        $output = array();
        foreach ($dom as $index => $element)
        {
            if ($element['opening'])
            {
                $output[] = "\n".str_repeat($indentWith, $indent).trim($element['content']);

                // make sure that only the elements who have not been blacklisted are being indented
                if ( ! in_array($element['type'], explode(',', $tagsWithoutIndentation)))
                {
                    ++$indent;
                }
            }
            else if ($element['standalone'])
            {
                $output[] = "\n".str_repeat($indentWith, $indent).trim($element['content']);
            }
            else if ($element['closing'])
            {
                --$indent;
                $lf = "\n".str_repeat($indentWith, abs($indent));
                if (isset($dom[$index - 1]) && $dom[$index - 1]['opening'])
                {
                    $lf = '';
                }
                $output[] = $lf.trim($element['content']);
            }
            else if ($element['text'])
            {
                // $output[] = "\n".str_repeat($indentWith, $indent).trim($element['content']);
                $output[] = "\n".str_repeat($indentWith, $indent).preg_replace('/ [ \t]*/', ' ', $element['content']);
            }
            else if ($element['comment'])
            {
                $output[] = "\n".str_repeat($indentWith, $indent).trim($element['content']);
            }
        }

        return trim(implode('', $output));
    }

    /**
     * Parses an array of HTML tokens and adds basic information about about the type of 
     * tag the token represents.
     *
     * @param Array $elements Array of HTML tokens (tags and text tokens).
     * @return Array HTML elements with extra information.
     */
    public static function parseDom(Array $elements)
    {
        $dom = array();
        foreach ($elements as $element)
        {
            $isText = false;
            $isComment = false;
            $isClosing = false;
            $isOpening = false;
            $isStandalone = false;

            $currentElement = trim($element);

            // comment
            if (strpos($currentElement, '<!') === 0)
            {
                $isComment = true;
            }
            // closing tag
            else if (strpos($currentElement, '</') === 0)
            {
                $isClosing = true;
            }
            // stand-alone tag
            else if (preg_match('/\/>$/', $currentElement))
            {
                $isStandalone = true;
            }
            // normal opening tag
            else if (strpos($currentElement, '<') === 0)
            {
                $isOpening = true;
            }
            // text
            else
            {
                $isText = true;
            }

            $dom[] = array(
                'text'              => $isText,
                'comment'           => $isComment,
                'closing'           => $isClosing,
                'opening'           => $isOpening,
                'standalone'        => $isStandalone,
                'content'           => $element,
                'type'              => preg_replace('/^<\/?(\w+)[ >].*$/U', '$1', $element)
            );
        }
        return $dom;
    }
}