trying another way to simplify

pull/73/head
Huda Joad 2023-11-25 00:59:25 +03:00
rodzic e6b9cfc6dd
commit 5145b0fe3e
1 zmienionych plików z 13 dodań i 7 usunięć

Wyświetl plik

@ -25,13 +25,19 @@ function simplifyContent(content) {
simplifiedContent = simplifiedContent.replace(/<style.*?>.*?<\/style>/gms, ''); simplifiedContent = simplifiedContent.replace(/<style.*?>.*?<\/style>/gms, '');
// Remove all remaining HTML tags, leaving the inner text // Remove all remaining HTML tags, leaving the inner text
simplifiedContent = simplifiedContent.replace(/<[^>]+>/g, ''); simplifiedContent = simplifiedContent.replace(/<[^>]+>/g, '');
// Decode HTML entities - for a Node.js environment, consider using a library like 'he' // // Decode HTML entities - for a Node.js environment, consider using a library like 'he'
simplifiedContent = simplifiedContent.replace(/&[a-z]+;/gi, match => { // simplifiedContent = simplifiedContent.replace(/&[a-z]+;/gi, match => {
// This part is for browser environments, adjust for Node.js if necessary // // This part is for browser environments, adjust for Node.js if necessary
const span = document.createElement('span'); // const span = document.createElement('span');
span.innerHTML = match; // span.innerHTML = match;
return span.textContent || span.innerText; // return span.textContent || span.innerText;
}); // });
simplifiedContent = simplifiedContent
.replace(/&amp;/g, '&')
.replace(/&lt;/g, '<')
.replace(/&gt;/g, '>')
.replace(/&quot;/g, '"')
.replace(/&#39;/g, "'");
// Remove inline CSS and JavaScript event handlers // Remove inline CSS and JavaScript event handlers
simplifiedContent = simplifiedContent.replace(/style\s*=\s*'.*?'/gi, ''); simplifiedContent = simplifiedContent.replace(/style\s*=\s*'.*?'/gi, '');
simplifiedContent = simplifiedContent.replace(/on\w+\s*=\s*".*?"/gi, ''); simplifiedContent = simplifiedContent.replace(/on\w+\s*=\s*".*?"/gi, '');