<?php

class SimpleScraper {

    public static function scrape($url) {
        $html = self::fetchHtml($url);
        if (!$html) {
            return false;
        }
        return self::parseMetaTagsDom($html);
    }

    private static function fetchHtml($url) {
        $ch = curl_init($url);
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
        curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
        curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
        curl_setopt($ch, CURLOPT_TIMEOUT, 20);
        
        $facebook_agent = 'facebookexternalhit/1.1 (+www.facebook.com)';
        curl_setopt($ch, CURLOPT_USERAGENT, $facebook_agent);
        
        $html = curl_exec($ch);
        curl_close($ch);
        
        return $html;
    }

    // Function using DOMDocument for reliable parsing
    private static function parseMetaTagsDom($html) {
        $data = ['title' => null, 'og_tags' => []];

        // Create a new DOMDocument object
        $dom = new DOMDocument();
        
        // Suppress HTML5 parsing errors from messy real-world HTML
        @$dom->loadHTML($html); 
        
        // Use XPath to query the document
        $xpath = new DOMXPath($dom);

        // 1. Get the Title
        $titleNode = $xpath->query('//title');
        if ($titleNode->length > 0) {
            $data['title'] = trim($titleNode->item()->nodeValue);
        }

        // 2. Get Open Graph tags using XPath
        // Query all meta tags that have a 'property' attribute starting with 'og:'
        $ogMetaNodes = $xpath->query('//meta[starts-with(@property, "og:")]');

        foreach ($ogMetaNodes as $node) {
            $property = str_replace('og:', '', $node->getAttribute('property'));
            $content = $node->getAttribute('content');

            if ($property === 'image') {
                $data['og_tags']['image'][] = $content;
            } else {
                $data['og_tags'][$property] = $content;
            }
        }
        
        return $data;
    }
}