import { ContentExtractor, ExtractedContent } from './content-extractor-types.js';
export class GitHubPagesExtractor implements ContentExtractor {
async extractContent(document: Document): Promise<ExtractedContent> {
// Remove navigation and footer
document.querySelectorAll('nav, header, footer').forEach((el) => el.remove());
// Get main content
const main = document.querySelector('main, article, .markdown-body');
if (!main) {
return {
content: '',
metadata: { type: 'overview' },
};
}
const clone = main.cloneNode(true) as Element;
clone.querySelectorAll('script, style').forEach((el) => el.remove());
// Extract title and description
const title = clone.querySelector('h1')?.textContent?.trim();
const firstParagraph = clone.querySelector('h1 + p')?.textContent?.trim();
return {
content: clone.textContent?.trim() || '',
metadata: {
type: 'overview',
pattern: {
name: title || '',
type: 'component',
description: firstParagraph || '',
usageContexts: [],
relatedPatterns: [],
},
},
};
}
}
