Refactor

src/Cision.php

@@ -5,11 +5,16 @@ namespace Mattitja\Cision;
 use Carbon\Carbon;
 use Illuminate\Support\Facades\Http;
 use Illuminate\Http\Client\Response;
+use Symfony\Component\DomCrawler\Crawler;
+use Mattitja\Cision\Support\HtmlCleaner;
 
 class Cision
 {
     protected string $slug;
 
+    /**
+     * Constructor that accepts a slug identifier for the feed.
+     */
     public function __construct(string $slug = null)
     {
         $this->slug = $slug ?? env('LARAVEL_CISION_FEED_SLUG', '');
@@ -51,6 +56,39 @@ class Cision
         return $this->parseResponse($response->body(), $page);
     }
 
+    /**
+     * Fetch and parse the HTML content of a Cision article URL.
+     */
+    public function fetchContent(string $url): array
+    {
+        $html = Http::get($url)->body();
+        $crawler = new Crawler($html);
+
+        $article = $crawler->filter('article')->first();
+
+        $header = trim($article->filter('h1')->text());
+        $publishedAt = $article->filter('time')->attr('datetime');
+
+        $allowedTags = ['p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'ul', 'ol', 'li'];
+        $allowedInline = ['strong', 'b', 'em', 'i', 'br'];
+
+        $content = [];
+
+        foreach ($article as $node) {
+            foreach ($node->childNodes as $child) {
+                $content[] = HtmlCleaner::renderNode($child, $allowedTags, $allowedInline);
+            }
+        }
+
+        $cleaned = HtmlCleaner::clean(implode("\n", array_filter($content)));
+
+        return [
+            'header' => $header,
+            'published_at' => $publishedAt,
+            'content' => $cleaned,
+        ];
+    }
+
     protected function makeRequest(int $page, ?string $type = null): Response|false
     {
         if (empty($this->slug)) {
@@ -102,4 +140,4 @@ class Cision
             'next_page' => count($items) === 24 ? $page + 1 : false,
         ];
     }
-}
+}

src/Support/HtmlCleaner.php

@@ -0,0 +1,107 @@
+<?php
+
+namespace Mattitja\Cision\Support;
+
+use DOMNode;
+use DOMText;
+use DOMElement;
+
+class HtmlCleaner
+{
+    /**
+     * Render a DOM node into safe HTML, preserving allowed block and inline tags.
+     */
+    public static function renderNode(DOMNode $node, array $allowedTags, array $allowedInline, string $parentTag = ''): ?string
+    {
+        if ($node instanceof DOMText) {
+            return $node->wholeText;
+        }
+
+        if (!($node instanceof DOMElement)) {
+            return null;
+        }
+
+        $tag = $node->nodeName;
+
+        // Special case: handle <br> within <p>
+        if ($tag === 'br') {
+            if ($parentTag === 'p') {
+                $prev = $node->previousSibling;
+                $next = $node->nextSibling;
+
+                $hasTextBefore = $prev instanceof DOMText && trim($prev->textContent) !== '';
+                $hasTextAfter = $next instanceof DOMText && trim($next->textContent) !== '';
+
+                if ($hasTextBefore && $hasTextAfter) {
+                    return '<br>';
+                }
+            }
+
+            // Skip all other <br>
+            return null;
+        }
+
+        // Recursively render children
+        $inner = '';
+        foreach ($node->childNodes as $child) {
+            $rendered = self::renderNode($child, $allowedTags, $allowedInline, $tag);
+            if ($rendered !== null) {
+                $inner .= $rendered;
+            }
+        }
+
+        // Handle block tags
+        if (in_array($tag, $allowedTags)) {
+            if ($tag === 'p') {
+                $html = $node->C14N();
+                $isItalic = str_starts_with(trim($html), '<em') || str_starts_with(trim($html), '<i');
+                $class = $isItalic ? ' class="italic"' : '';
+                return "<p{$class}>" . trim($inner) . "</p>";
+            }
+
+            return "<{$tag}>" . trim($inner) . "</{$tag}>";
+        }
+
+        // Handle inline tags
+        if (in_array($tag, $allowedInline)) {
+            return "<{$tag}>{$inner}</{$tag}>";
+        }
+
+        // Unknown tags are ignored but their content is preserved
+        return $inner;
+    }
+
+    /**
+     * Clean a raw HTML string: remove unwanted tags, empty elements, garbage phrases, and excessive whitespace.
+     */
+    public static function clean(string $html): string
+    {
+        // Remove first <h1>
+        $html = preg_replace('/<h1\b[^>]*>.*?<\/h1>/i', '', $html, 1);
+
+        // Remove empty tags like <i></i>, <span></span>
+        $html = preg_replace('/<(\w+)[^>]*>\s*<\/\1>/', '', $html);
+
+        // Remove specific garbage strings
+        $unwantedStrings = [
+            'Report this content',
+            '<p><strong>Taggar:</strong></p>',
+        ];
+
+        foreach ($unwantedStrings as $str) {
+            $html = str_replace($str, '', $html);
+        }
+
+        // Remove localized date lines like: fre, jul 04, 2025 09:55 CET
+        $html = preg_replace('/\b[a-z]{3},\s[a-z]{3}\s\d{2},\s\d{4}\s\d{2}:\d{2}\sCET\b/i', '', $html);
+
+        // Normalize whitespace
+        $html = str_replace(["\u{A0}", "\t", "\r"], ' ', $html);          // Remove non-breaking spaces and tabs
+        $html = preg_replace('/[ ]{2,}/', ' ', $html);                    // Collapse double spaces
+        $html = preg_replace("/(\n\s*){2,}/", "\n", $html);               // Collapse multiple newlines
+        $html = str_replace(['  ', ' \n'], '', $html);                    // Additional spacing clean
+        $html = trim($html);
+
+        return $html;
+    }
+}