fix: decode KDCA press release entities

Author: neuwcodebox

src/modules/ingest/app/sources/_shared/decode-html-entities.test.ts

@@ -0,0 +1,17 @@
+import { describe, expect, it } from 'vitest';
+import { decodeHtmlEntities } from './decode-html-entities';
+
+describe('decodeHtmlEntities', () => {
+  it('should decode named, decimal, and hexadecimal entities', () => {
+    expect(decodeHtmlEntities('&lt;경보&nbsp;&#9312;&#x2461;&gt;')).toBe('<경보 ①②>');
+  });
+
+  it('should decode escaped entities up to max passes', () => {
+    expect(decodeHtmlEntities('&amp;#9312;')).toBe('&#9312;');
+    expect(decodeHtmlEntities('&amp;#9312;', { maxPasses: 2 })).toBe('①');
+  });
+
+  it('should keep unknown and invalid entities unchanged', () => {
+    expect(decodeHtmlEntities('&unknown; &#9999999999;')).toBe('&unknown; &#9999999999;');
+  });
+});

src/modules/ingest/app/sources/_shared/decode-html-entities.ts

@@ -0,0 +1,44 @@
+const NAMED_ENTITIES = new Map<string, string>([
+  ['nbsp', ' '],
+  ['lt', '<'],
+  ['gt', '>'],
+  ['amp', '&'],
+  ['quot', '"'],
+  ['apos', "'"],
+]);
+
+type DecodeHtmlEntitiesOptions = {
+  maxPasses?: number;
+};
+
+export function decodeHtmlEntities(value: string, options: DecodeHtmlEntitiesOptions = {}): string {
+  const maxPasses = Math.max(1, options.maxPasses ?? 1);
+  let decoded = value;
+
+  for (let index = 0; index < maxPasses; index += 1) {
+    const next = decodeHtmlEntitiesOnce(decoded);
+    if (next === decoded) {
+      return next;
+    }
+    decoded = next;
+  }
+
+  return decoded;
+}
+
+function decodeHtmlEntitiesOnce(value: string): string {
+  return value
+    .replace(/&#x([0-9a-fA-F]+);/g, (full, hex: string) => {
+      const codePoint = Number.parseInt(hex, 16);
+      return isValidCodePoint(codePoint) ? String.fromCodePoint(codePoint) : full;
+    })
+    .replace(/&#(\d+);/g, (full, num: string) => {
+      const codePoint = Number.parseInt(num, 10);
+      return isValidCodePoint(codePoint) ? String.fromCodePoint(codePoint) : full;
+    })
+    .replace(/&([a-zA-Z]+);/g, (full, name: string) => NAMED_ENTITIES.get(name) ?? full);
+}
+
+function isValidCodePoint(value: number): boolean {
+  return Number.isInteger(value) && value >= 0 && value <= 0x10ffff;
+}

src/modules/ingest/app/sources/kdca-press-release.source.test.ts

@@ -100,4 +100,32 @@ describe('KdcaPressReleaseSource', () => {
     const payload = result.events[0].payload as { id?: string };
     expect(payload.id).toBe('전염병 경보 상향');
   });
+
+  it('should decode escaped html entities and remove trailing title artifact', async () => {
+    vi.useFakeTimers();
+    vi.setSystemTime(new Date('2026-06-17T06:00:00.000Z'));
+
+    const xml = buildRss(
+      buildItem({
+        title: '일본뇌염 바이러스 검출, 전국 경보 발령(6.17.수)}',
+        link: '/bbs/kdca/41/311453/artclView.do?layout=unknown',
+        pubDate: '2026-06-17 14:00:00.0',
+        author: '대변인',
+        description:
+          '일본뇌염 경보 발령 [ 일본뇌염 주의보 및 경보 기준 ] (일본뇌염 주의보)① 일본뇌염 매개모기 최초 채집 시',
+      }),
+    );
+
+    const fetchMock = vi.fn().mockImplementation(() => Promise.resolve(new Response(xml, { status: 200 })));
+    vi.stubGlobal('fetch', fetchMock);
+
+    const source = new KdcaPressReleaseSource();
+    const result = await source.run(null);
+
+    expect(result.events).toHaveLength(1);
+    expect(result.events[0].title).toBe('일본뇌염 바이러스 검출, 전국 경보 발령(6.17.수)');
+    expect(result.events[0].body).toBe(
+      '일본뇌염 경보 발령 [ 일본뇌염 주의보 및 경보 기준 ] (일본뇌염 주의보)① 일본뇌염 매개모기 최초 채집 시',
+    );
+  });
 });

src/modules/ingest/app/sources/kdca-press-release.source.ts

@@ -2,6 +2,7 @@ import { load } from 'cheerio';
 import { logger } from '@/core/logger';
 import { EventKinds, EventLevels, EventSources } from '@/modules/events/domain/event.enums';
 import type { Source, SourceEvent, SourceRunResult } from '../../domain/port/source.interface';
+import { decodeHtmlEntities } from './_shared/decode-html-entities';
 import { fetchWithTimeout } from './_shared/fetch-with-timeout';
 import { isTooOld } from './_shared/is-too-old';
 import { normalizeText } from './_shared/normalize';
@@ -119,7 +120,7 @@ function parseRssItems(xml: string, now: Date): KdcaPressItem[] {
 
   const items: KdcaPressItem[] = [];
   for (const node of nodes) {
-    const title = normalizeText($(node).find('title').first().text());
+    const title = normalizeRssTitle($(node).find('title').first().text());
     const rawLink = normalizeText($(node).find('link').first().text());
     if (!title || !rawLink) {
       continue;
@@ -135,8 +136,8 @@ function parseRssItems(xml: string, now: Date): KdcaPressItem[] {
       logger.warn({ link, title }, 'Using title as KDCA press release id fallback');
     }
 
-    const author = normalizeText($(node).find('author').first().text());
-    const description = normalizeText($(node).find('description').first().text());
+    const author = normalizeRssText($(node).find('author').first().text());
+    const description = normalizeRssText($(node).find('description').first().text());
     const rawDate = normalizeText($(node).find('pubDate').first().text());
     const occurredAt = parseKdcaDate(rawDate, now);
     if (rawDate && !occurredAt) {
@@ -157,6 +158,28 @@ function parseRssItems(xml: string, now: Date): KdcaPressItem[] {
   return items;
 }
 
+function normalizeRssTitle(value: string | null | undefined): string | null {
+  const normalized = normalizeRssText(value);
+  if (!normalized) {
+    return null;
+  }
+
+  if (!normalized.includes('{')) {
+    return normalizeText(normalized.replace(/\}+$/, ''));
+  }
+
+  return normalized;
+}
+
+function normalizeRssText(value: string | null | undefined): string | null {
+  const normalized = normalizeText(value);
+  if (!normalized) {
+    return null;
+  }
+
+  return decodeHtmlEntities(normalized, { maxPasses: 3 });
+}
+
 function mapCrisisLevel(title: string): EventLevels {
   const candidates: Array<{ keyword: string; level: EventLevels }> = [
     { keyword: '심각', level: EventLevels.Severe },

src/modules/ingest/app/sources/kma-micro-earthquake.source.ts

@@ -2,19 +2,11 @@ import { logger } from '@/core/logger';
 import type { EventPayload } from '@/modules/events/domain/entity/event.entity';
 import { EventKinds, EventLevels, EventSources } from '@/modules/events/domain/event.enums';
 import type { Source, SourceEvent, SourceRunResult } from '../../domain/port/source.interface';
+import { decodeHtmlEntities } from './_shared/decode-html-entities';
 import { fetchWithTimeout } from './_shared/fetch-with-timeout';
 
 const KMA_MICRO_EARTHQUAKE_ENDPOINT = 'https://www.weather.go.kr/w/wnuri-eqk-vol/eqk/eqk-micro.do';
 
-const NAMED_ENTITIES = new Map<string, string>([
-  ['nbsp', ' '],
-  ['lt', '<'],
-  ['gt', '>'],
-  ['amp', '&'],
-  ['quot', '"'],
-  ['apos', "'"],
-]);
-
 type MicroEarthquakeDetail = {
   occurredAt: string | null;
   regionText: string | null;
@@ -196,19 +188,6 @@ const sanitizeHtmlFragment = (fragment: string): string => {
   return text.replace(/\r\n/g, '\n').replace(/\r/g, '\n').trim();
 };
 
-const decodeHtmlEntities = (text: string): string => {
-  return text
-    .replace(/&#x([0-9a-fA-F]+);/g, (_, hex: string) => {
-      const codePoint = Number.parseInt(hex, 16);
-      return Number.isFinite(codePoint) ? String.fromCodePoint(codePoint) : _;
-    })
-    .replace(/&#(\d+);/g, (_, num: string) => {
-      const codePoint = Number.parseInt(num, 10);
-      return Number.isFinite(codePoint) ? String.fromCodePoint(codePoint) : _;
-    })
-    .replace(/&([a-zA-Z]+);/g, (full, name: string) => NAMED_ENTITIES.get(name) ?? full);
-};
-
 const parseKstDateTime = (value: string): string | null => {
   const matched = value.match(/(\d{4})[./-](\d{2})[./-](\d{2})\s+(\d{2}):(\d{2}):(\d{2})/);
   if (!matched) {