feat: Medium 서브도메인 RSS 지원 + 썸네일 fallback

packages/bot/src/scheduler-registry.ts

@@ -21,7 +21,7 @@ import { getScoreService } from './services/score.service';
 import { getAttendanceService, getFineService } from './services';
 
 import { ActivityScoreType, curationSources, getDb, members } from '@blog-study/shared/db';
-import { extractOgImage } from '@blog-study/shared/utils';
+import { extractFirstImage, extractOgImage } from '@blog-study/shared/utils';
 import { getCurrentRound } from './services/round.service';
 import { eq } from 'drizzle-orm';
 import logger from './lib/logger';
@@ -80,8 +80,9 @@ export async function registerAllJobs(boss: PgBoss, client: Client): Promise<voi
     for (const item of items) {
       if (item.pubDate < POST_CUTOFF_DATE) continue;
 
-      // OG 이미지 추출 (실패해도 글 등록은 진행)
-      const thumbnailUrl = await extractOgImage(item.link).catch(() => null);
+      // OG 이미지 추출 (실패 시 RSS content 첫 이미지 fallback)
+      const thumbnailUrl = await extractOgImage(item.link).catch(() => null)
+        ?? extractFirstImage(item.description);
 
       const result = await postService.create({
         memberId: member.id,
@@ -262,7 +263,8 @@ export async function registerAllJobs(boss: PgBoss, client: Client): Promise<voi
         category: '',
         tags: item.categories ?? [],
         description: item.description,
-        thumbnailUrl: result?.status === 'fulfilled' ? result.value : null,
+        thumbnailUrl: (result?.status === 'fulfilled' ? result.value : null)
+          ?? extractFirstImage(item.description),
       };
     });
 

packages/shared/src/utils/feed-parser.ts

@@ -25,9 +25,13 @@ export interface NormalizedFeedItem {
 export function extractFeedItems(result: ReturnType<typeof parseFeed>): NormalizedFeedItem[] {
   const { format, feed } = result;
 
+  // HTML 엔티티 디코딩 (&quot; &lsquo; 등)
+  const decodeField = (val: string | undefined): string | undefined =>
+    val ? decode(val) : undefined;
+
   if (format === 'atom') {
     return (feed.entries ?? []).map((entry) => ({
-      title: entry.title,
+      title: decodeField(entry.title),
       link: entry.links?.[0]?.href,
       pubDate: entry.published ?? entry.updated,
       description: entry.summary ?? entry.content,
@@ -37,7 +41,7 @@ export function extractFeedItems(result: ReturnType<typeof parseFeed>): Normaliz
 
   if (format === 'rss') {
     return (feed.items ?? []).map((item) => ({
-      title: item.title,
+      title: decodeField(item.title),
       link: item.link,
       pubDate: item.pubDate ? String(item.pubDate) : undefined,
       description: item.description,
@@ -49,7 +53,7 @@ export function extractFeedItems(result: ReturnType<typeof parseFeed>): Normaliz
 
   if (format === 'json') {
     return (feed.items ?? []).map((item) => ({
-      title: item.title,
+      title: decodeField(item.title),
       link: item.url ?? item.external_url,
       pubDate: item.date_published ?? item.date_modified,
       description: item.summary ?? item.content_text,
@@ -59,7 +63,7 @@ export function extractFeedItems(result: ReturnType<typeof parseFeed>): Normaliz
 
   // RDF
   return (feed.items ?? []).map((item) => ({
-    title: item.title,
+    title: decodeField(item.title),
     link: item.link,
     pubDate: item.dc?.date,
     description: item.description,
@@ -100,6 +104,17 @@ export function sanitizeDescription(html: string | undefined): string | null {
   return sanitized.length > 300 ? sanitized.slice(0, 300) + '...' : sanitized;
 }
 
+/**
+ * HTML content에서 첫 번째 이미지 URL 추출 (RSS content:encoded fallback용)
+ */
+export function extractFirstImage(html: string | null | undefined): string | null {
+  if (!html) return null;
+  const match = html.match(/<img[^>]+src=["']([^"']+)["']/i);
+  const imgUrl = match?.[1] ?? null;
+  if (imgUrl && !isSafeUrl(imgUrl)) return null;
+  return imgUrl;
+}
+
 /**
  * URL에서 og:image 메타태그 추출 (5초 타임아웃)
  * SSRF 보호: 내부 URL 차단 + OG 이미지 URL 검증
@@ -124,7 +139,19 @@ export async function extractOgImage(url: string): Promise<string | null> {
       html.match(/<meta[^>]+property=["']og:image["'][^>]+content=["']([^"']+)["']/i) ||
       html.match(/<meta[^>]+content=["']([^"']+)["'][^>]+property=["']og:image["']/i);
 
-    const ogImageUrl = match?.[1] ?? null;
+    let ogImageUrl = match?.[1] ?? null;
+
+    // og:image 없으면 JSON-LD Schema.org image fallback (Medium 등)
+    if (!ogImageUrl) {
+      const jsonLdMatch = html.match(/<script[^>]*type=["']application\/ld\+json["'][^>]*>([\s\S]*?)<\/script>/i);
+      if (jsonLdMatch?.[1] && jsonLdMatch[1].length < 100_000) {
+        try {
+          const ld = JSON.parse(jsonLdMatch[1]);
+          const ldImage = ld.image?.url || ld.image?.contentUrl || (typeof ld.image === 'string' ? ld.image : null);
+          if (ldImage && isSafeUrl(ldImage)) ogImageUrl = ldImage;
+        } catch { /* invalid JSON-LD */ }
+      }
+    }
 
     // SSRF 방지: OG 이미지 URL 자체도 안전한지 검증
     if (ogImageUrl && !isSafeUrl(ogImageUrl)) {

packages/shared/src/utils/url-validator.ts

@@ -61,7 +61,7 @@ export interface UrlValidationResult {
 const PLATFORM_PATTERNS: Record<BlogPlatform, RegExp> = {
   velog: /^https?:\/\/velog\.io\/@[\w-]+(\/posts)?\/?$/,
   tistory: /^https?:\/\/[\w-]+\.tistory\.com\/?$/,
-  medium: /^https?:\/\/medium\.com\/@[\w-]+\/?$/,
+  medium: /^https?:\/\/(medium\.com\/@[\w-]+|(?!www\.)[\w-]+\.medium\.com)\/?$/,
   unknown: /^https?:\/\/.+/,
 };
 

packages/web/src/app/api/posts/preview/route.ts

@@ -68,6 +68,24 @@ export async function POST(request: Request) {
     if (thumbnailUrl) thumbnailUrl = decodeHtmlEntities(thumbnailUrl);
     if (thumbnailUrl && !isSafeUrl(thumbnailUrl)) thumbnailUrl = null;
 
+    // og:image 없으면 JSON-LD → 본문 첫 이미지 순서로 fallback
+    if (!thumbnailUrl) {
+      // JSON-LD Schema.org image (Medium 등 JS 렌더링 플랫폼)
+      const jsonLdMatch = html.match(/<script[^>]*type=["']application\/ld\+json["'][^>]*>([\s\S]*?)<\/script>/i);
+      if (jsonLdMatch?.[1] && jsonLdMatch[1].length < 100_000) {
+        try {
+          const ld = JSON.parse(jsonLdMatch[1]);
+          const ldImage = ld.image?.url || ld.image?.contentUrl || (typeof ld.image === 'string' ? ld.image : null);
+          if (ldImage && isSafeUrl(ldImage)) thumbnailUrl = ldImage;
+        } catch { /* invalid JSON-LD */ }
+      }
+    }
+    if (!thumbnailUrl) {
+      const imgMatch = html.match(/<img[^>]+src=["']([^"']+)["']/i);
+      const fallback = imgMatch?.[1] || null;
+      if (fallback && isSafeUrl(fallback)) thumbnailUrl = fallback;
+    }
+
     // og:description > meta description
     const ogDescMatch =
       html.match(/<meta[^>]*property=["']og:description["'][^>]*content=["']([^"']+)["']/i) ||

packages/web/src/lib/rss-detect.ts

@@ -28,8 +28,13 @@ function constructRssUrl(blogUrl: string, platform: string): string | null {
       case 'tistory':
         return `${url.protocol}//${url.hostname}/rss`;
       case 'medium': {
-        const match = url.pathname.match(/^\/@([\w-]+)\/?$/);
-        return match ? `https://medium.com/feed/@${match[1]}` : null;
+        // medium.com/@username 형식
+        const pathMatch = url.pathname.match(/^\/@([\w-]+)\/?$/);
+        if (pathMatch) return `https://medium.com/feed/@${pathMatch[1]}`;
+        // username.medium.com 서브도메인 형식
+        const subdomainMatch = url.hostname.match(/^([\w-]+)\.medium\.com$/);
+        if (subdomainMatch) return `https://${subdomainMatch[1]}.medium.com/feed`;
+        return null;
       }
       default:
         return null;