/**
* Media Extractor - Extracts images and videos from article content
*/
export interface MediaAsset {
id: string;
article_id: string;
media_type: 'image' | 'video';
url: string;
width?: number;
height?: number;
caption?: string;
}
export interface ExtractedMedia {
images: MediaAsset[];
videos: MediaAsset[];
totalCount: number;
}
export class MediaExtractor {
/**
* Extract all media (images and videos) from HTML content
*/
static extractAll(articleId: string, htmlContent: string): ExtractedMedia {
if (!htmlContent) {
return { images: [], videos: [], totalCount: 0 };
}
const images = this.extractImages(articleId, htmlContent);
const videos = this.extractVideos(articleId, htmlContent);
return {
images,
videos,
totalCount: images.length + videos.length
};
}
/**
* Extract images from HTML content
*/
static extractImages(articleId: string, htmlContent: string): MediaAsset[] {
const images: MediaAsset[] = [];
// Match img tags with src attribute
const imgRegex = /<img[^>]+src=["']([^"']+)["'][^>]*>/gi;
let match;
while ((match = imgRegex.exec(htmlContent)) !== null) {
const fullTag = match[0];
const url = match[1];
// Skip tiny images (likely tracking pixels)
if (this.isTinyImage(fullTag)) continue;
// Skip data URIs
if (url.startsWith('data:')) continue;
// Extract dimensions
const width = this.extractDimension(fullTag, 'width');
const height = this.extractDimension(fullTag, 'height');
// Extract alt text as caption
const altMatch = fullTag.match(/alt=["']([^"']+)["']/i);
const caption = altMatch ? altMatch[1] : undefined;
images.push({
id: this.generateId(),
article_id: articleId,
media_type: 'image',
url: this.normalizeUrl(url),
width,
height,
caption
});
}
// Also check for Open Graph images
const ogImageRegex = /<meta[^>]+property=["']og:image["'][^>]+content=["']([^"']+)["'][^>]*>/gi;
while ((match = ogImageRegex.exec(htmlContent)) !== null) {
const url = match[1];
if (!images.some(img => img.url === url)) {
images.push({
id: this.generateId(),
article_id: articleId,
media_type: 'image',
url: this.normalizeUrl(url),
caption: 'Open Graph Image'
});
}
}
return images;
}
/**
* Extract videos from HTML content
*/
static extractVideos(articleId: string, htmlContent: string): MediaAsset[] {
const videos: MediaAsset[] = [];
// Match video tags
const videoRegex = /<video[^>]*>[\s\S]*?<\/video>/gi;
let match;
while ((match = videoRegex.exec(htmlContent)) !== null) {
const videoTag = match[0];
// Extract source URL
const srcMatch = videoTag.match(/src=["']([^"']+)["']/i) ||
videoTag.match(/<source[^>]+src=["']([^"']+)["']/i);
if (srcMatch) {
const url = srcMatch[1];
const width = this.extractDimension(videoTag, 'width');
const height = this.extractDimension(videoTag, 'height');
videos.push({
id: this.generateId(),
article_id: articleId,
media_type: 'video',
url: this.normalizeUrl(url),
width,
height
});
}
}
// Extract YouTube embeds
const youtubeRegex = /(?:youtube\.com\/embed\/|youtu\.be\/)([a-zA-Z0-9_-]+)/gi;
while ((match = youtubeRegex.exec(htmlContent)) !== null) {
const videoId = match[1];
videos.push({
id: this.generateId(),
article_id: articleId,
media_type: 'video',
url: `https://www.youtube.com/watch?v=${videoId}`,
caption: 'YouTube Video'
});
}
// Extract Vimeo embeds
const vimeoRegex = /vimeo\.com\/(?:video\/)?(\d+)/gi;
while ((match = vimeoRegex.exec(htmlContent)) !== null) {
const videoId = match[1];
videos.push({
id: this.generateId(),
article_id: articleId,
media_type: 'video',
url: `https://vimeo.com/${videoId}`,
caption: 'Vimeo Video'
});
}
return videos;
}
/**
* Check if image is too small (likely a tracking pixel)
*/
private static isTinyImage(imgTag: string): boolean {
const width = this.extractDimension(imgTag, 'width');
const height = this.extractDimension(imgTag, 'height');
if (width && height && (width < 10 || height < 10)) {
return true;
}
return false;
}
/**
* Extract dimension from image/video tag
*/
private static extractDimension(tag: string, dimension: 'width' | 'height'): number | undefined {
const regex = new RegExp(`${dimension}=["']?(\\d+)["']?`, 'i');
const match = tag.match(regex);
return match ? parseInt(match[1]) : undefined;
}
/**
* Normalize URL (convert relative to absolute if possible)
*/
private static normalizeUrl(url: string): string {
// Remove leading/trailing whitespace
url = url.trim();
// If it's already absolute, return as is
if (url.startsWith('http://') || url.startsWith('https://')) {
return url;
}
// For protocol-relative URLs
if (url.startsWith('//')) {
return 'https:' + url;
}
return url;
}
/**
* Generate unique ID for media asset
*/
private static generateId(): string {
return `media_${Date.now()}_${Math.random().toString(36).substr(2, 9)}`;
}
}
