LesLibres.org
/
Rss-Bridge
2
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

[EconomistBridge] Fixes for fetching new page structure (#1836)

This commit is contained in:
Tobias Alexander Franke 2020-11-29 10:31:20 +00:00 committed by GitHub
parent 7705d097e3
commit 56eb829a66
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 21 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
bridges/EconomistBridge.php
View File

@ -14,17 +14,28 @@ class EconomistBridge extends BridgeAbstract {
$html = getSimpleHTMLDOM(self::URI . '/latest/')
or returnServerError('Could not fetch latest updates form The Economist.');
foreach($html->find('article') as $element) {
foreach($html->find('div.teaser') as $element) {
$a = $element->find('a.headline-link', 0);
$href = $a->href;
if (substr($href, 0, 4) != 'http')
$href = self::URI . $a->href;
$a = $element->find('a', 0);
$href = self::URI . $a->href;
$full = getSimpleHTMLDOMCached($href);
$article = $full->find('article', 0);
$header = $article->find('span[itemprop="headline"]', 0);
$headerimg = $article->find('div[itemprop="image"]', 0)->find('img', 0);
$author = $article->find('p[itemprop="byline"]', 0);
$time = $article->find('time', 0);
$content = $article->find('div[itemprop="text"]', 0);
$section = array( $article->find('strong[itemprop="articleSection"]', 0)->plaintext );
$header = $article->find('h1', 0);
$author = $article->find('span[itemprop="author"]', 0);
$time = $article->find('time[itemprop="dateCreated"]', 0);
$content = $article->find('div[itemprop="description"]', 0);
// Author
if ($author)
$author = substr($author->innertext, 3, strlen($author));
else
$author = 'The Economist';
// Remove newsletter subscription box
$newsletter = $content->find('div[class="newsletter-form__message"]', 0);
@ -40,19 +51,15 @@ class EconomistBridge extends BridgeAbstract {
if ($nextprev)
$nextprev->outertext = '';
$section = array( $article->find('h3[itemprop="articleSection"]', 0)->plaintext );
$item = array();
$item['title'] = $header->find('span', 0)->innertext . ': '
. $header->find('span', 1)->innertext;
$item['title'] = $header->innertext;
$item['uri'] = $href;
$item['timestamp'] = strtotime($time->datetime);
$item['author'] = $author->innertext;
$item['author'] = $author;
$item['categories'] = $section;
$item['content'] = '<img style="max-width: 100%" src="'
. $a->find('img', 0)->src . '">' . $content->innertext;
. $headerimg->src . '">' . $content->innertext;
$this->items[] = $item;