feat(import): add Confluence emoji-id parser with tests

2026-05-20 00:14:10 +08:00 · 2026-05-16 01:28:10 +01:00
parent 3e66aff1e2
commit bbb1b5eb26
2 changed files with 74 additions and 0 deletions
@@ -0,0 +1,46 @@
+import { parseConfluenceEmojiId } from './confluence-emoji';
+
+describe('parseConfluenceEmojiId', () => {
+  it('parses a single code point id', () => {
+    expect(parseConfluenceEmojiId('1f600')).toBe('😀');
+    expect(parseConfluenceEmojiId('1F600')).toBe('😀');
+  });
+
+  it('parses a country flag (two regional indicator code points)', () => {
+    expect(parseConfluenceEmojiId('1f1f3-1f1ec')).toBe('🇳🇬');
+    expect(parseConfluenceEmojiId('1f1fa-1f1f8')).toBe('🇺🇸');
+  });
+
+  it('parses a ZWJ sequence (three code points)', () => {
+    expect(parseConfluenceEmojiId('1f468-200d-1f4bb')).toBe('👨‍💻');
+  });
+
+  it('parses a five-component family ZWJ sequence', () => {
+    // 👨‍👩‍👧‍👦 = man, ZWJ, woman, ZWJ, girl, ZWJ, boy
+    expect(parseConfluenceEmojiId('1f468-200d-1f469-200d-1f467-200d-1f466')).toBe(
+      '👨‍👩‍👧‍👦',
+    );
+  });
+
+  it('returns null for missing input', () => {
+    expect(parseConfluenceEmojiId(undefined)).toBeNull();
+    expect(parseConfluenceEmojiId(null)).toBeNull();
+    expect(parseConfluenceEmojiId('')).toBeNull();
+  });
+
+  it('returns null when any segment is not pure hex', () => {
+    expect(parseConfluenceEmojiId('1f600-NG')).toBeNull();
+    expect(parseConfluenceEmojiId('not-hex')).toBeNull();
+    expect(parseConfluenceEmojiId('1f600--1f1ec')).toBeNull();
+    expect(parseConfluenceEmojiId('1f600 1f1ec')).toBeNull();
+  });
+
+  it('returns null when a segment parses to a non-positive value', () => {
+    expect(parseConfluenceEmojiId('0')).toBeNull();
+  });
+
+  it('returns null for code points outside the valid Unicode range', () => {
+    // 0x110000 is one past the highest valid code point.
+    expect(parseConfluenceEmojiId('110000')).toBeNull();
+  });
+});
@@ -0,0 +1,28 @@
+/**
+ * Parse a Confluence emoji id (hex code points joined by hyphens) into a
+ * Unicode string. Confluence emits ids in both single- and multi-code-point
+ * forms:
+ *
+ *   "1f600"             → "😀"
+ *   "1f1f3-1f1ec"       → "🇳🇬" (flag: Nigeria)
+ *   "1f468-200d-1f4bb"  → "👨‍💻" (man technologist, ZWJ sequence)
+ *
+ * Returns null when the input is missing, empty, or doesn't parse cleanly as
+ * hyphen-separated hex code points.
+ */
+export function parseConfluenceEmojiId(
+  raw: string | undefined | null,
+): string | null {
+  if (!raw) return null;
+  const parts = raw.split('-');
+  if (parts.length === 0) return null;
+  if (!parts.every((p) => /^[0-9a-fA-F]+$/.test(p))) return null;
+  const codePoints = parts.map((p) => parseInt(p, 16));
+  if (codePoints.some((cp) => !Number.isFinite(cp) || cp <= 0)) return null;
+  try {
+    return String.fromCodePoint(...codePoints);
+  } catch {
+    // Out-of-range code points throw RangeError on String.fromCodePoint.
+    return null;
+  }
+}