<?php
class EconomistBridge extends FeedExpander {

	const MAINTAINER = 'bockiii';
	const NAME = 'Economist Bridge';
	const URI = 'https://www.economist.com/';
	const CACHE_TIMEOUT = 3600; //1hour
	const DESCRIPTION = 'Returns the latest articles for the selected category';

	const PARAMETERS = array(
		'global' => array(
			'limit' => array(
				'name' => 'Feed Item Limit',
				'required' => true,
				'type' => 'number',
				'defaultValue' => 10,
				'title' => 'Maximum number of returned feed items. Maximum 30, default 10'
			)
		),
		'Topics' => array(
			'topic' => array(
				'name' => 'Topics',
				'type' => 'list',
				'title' => 'Select a Topic',
				'defaultValue' => 'latest',
				'values' => array(
					'Latest' => 'latest',
					'The world this week' => 'the-world-this-week',
					'Letters' => 'letters',
					'Leaders' => 'leaders',
					'Briefings' => 'briefing',
					'Special reports' => 'special-report',
					'Britain' => 'britain',
					'Europe' => 'europe',
					'United States' => 'united-states',
					'The Americas' => 'the-americas',
					'Middle East and Africa' => 'middle-east-and-africa',
					'Asia' => 'asia',
					'China' => 'china',
					'International' => 'international',
					'Business' => 'business',
					'Finance and economics' => 'finance-and-economics',
					'Science and technology' => 'science-and-technology',
					'Books and arts' => 'books-and-arts',
					'Obituaries' => 'obituary',
					'Graphic detail' => 'graphic-detail',
					'Indicators' => 'economic-and-financial-indicators',
				)
			)
		),
		'Blogs' => array(
			'blog' => array(
				'name' => 'Blogs',
				'type' => 'list',
				'title' => 'Select a Blog',
				'values' => array(
					'Bagehots notebook' => 'bagehots-notebook',
					'Bartleby' => 'bartleby',
					'Buttonwoods notebook' => 'buttonwoods-notebook',
					'Charlemagnes notebook' => 'charlemagnes-notebook',
					'Democracy in America' => 'democracy-in-america',
					'Erasmus' => 'erasmus',
					'Free exchange' => 'free-exchange',
					'Game theory' => 'game-theory',
					'Gulliver' => 'gulliver',
					'Kaffeeklatsch' => 'kaffeeklatsch',
					'Prospero' => 'prospero',
					'The Economist Explains' => 'the-economist-explains',
				)
			)
		)
	);

	public function collectData(){
		// get if topics or blogs were selected and store the selected category
		switch ($this->queriedContext) {
			case 'Topics':
				$category = $this->getInput('topic');
				break;
			case 'Blogs':
				$category = $this->getInput('blog');
				break;
			default:
				$category = 'latest';
		}
		// limit the returned articles to 30 at max
		if ((int)$this->getInput('limit') <= 30) {
			$limit = (int)$this->getInput('limit');
		} else {
			$limit = 30;
		}

		$this->collectExpandableDatas('https://www.economist.com/' . $category . '/rss.xml', $limit);
	}

	protected function parseItem($feedItem){
		$item = parent::parseItem($feedItem);
		$article = getSimpleHTMLDOM($item['uri']);
		// before the article can be added, it needs to be cleaned up, thus, the extra function
		// We also need to distinguish between old style and new style articles
		if ($article->find('article', 0)->getAttribute('data-test-id') == 'Article') {
			$contentNode = 'div.layout-article-body';
			$imgNode = 'div.article__lead-image';
			$categoryNode = 'span.article__subheadline';
		} elseif ($article->find('article', 0)->getAttribute('data-test-id') === 'NewArticle') {
			$contentNode = 'section';
			$imgNode = 'figure.css-12eysrk.e3y6nua0';
			$categoryNode = 'span.ern1uyf0';
		} else {
			return;
		}

		$item['content'] = $this->cleanContent($article, $contentNode);
		// only the article lead image is retained if it's there
		if (!is_null($article->find($imgNode, 0))) {
			$item['enclosures'][] = $article->find($imgNode, 0)->find('img', 0)->getAttribute('src');
		} else {
			$item['enclosures'][] = '';
		}
		// add the subheadline as category. This will create a link in new articles
		// and a text in old articles
		$item['categories'][] = $article->find($categoryNode, 0)->innertext;

		return $item;
	}

	private function cleanContent($article, $contentNode){
		// the actual article is in this div
		$content = $article->find($contentNode, 0)->innertext;
		// clean the article content. Remove all div's since the text is in paragraph elements
		foreach (array(
			'<div '
		) as $tag_start) {
			$content = stripRecursiveHTMLSection($content, 'div', $tag_start);
		}
		// now remove embedded iframes. The podcast postings contain these for example
		$content = preg_replace('/<iframe.*?\/iframe>/i', '', $content);
		// fix the relative links
		$content = defaultLinkTo($content, $this->getURI());

		return $content;
	}
}