fix: recognize statute sections with trailing letters (§ 1028A) (#27)

Tokenization patterns for USC and state code statutes required purely
numeric section numbers. Updated capture groups from (\d+) to
(\d+[A-Za-z]*) so sections like § 1028A and § 2339B are matched.

Closes #2

Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Author: medelman17

.changeset/statute-trailing-letters.md

@@ -0,0 +1,5 @@
+---
+"eyecite-ts": patch
+---
+
+Fix statute sections with trailing letters (e.g., "18 U.S.C. § 1028A") not being recognized. Updated tokenization patterns for both USC and state code statutes to allow alphanumeric section suffixes.

src/patterns/statutePatterns.ts

@@ -16,13 +16,13 @@ import type { Pattern } from './casePatterns'
 export const statutePatterns: Pattern[] = [
   {
     id: 'usc',
-    regex: /\b(\d+)\s+U\.S\.C\.?\s+§+\s*(\d+)\b/g,
+    regex: /\b(\d+)\s+U\.S\.C\.?\s+§+\s*(\d+[A-Za-z]*)\b/g,
     description: 'U.S. Code citations (e.g., "42 U.S.C. § 1983")',
     type: 'statute',
   },
   {
     id: 'state-code',
-    regex: /\b([A-Z][a-z]+\.?\s+[A-Za-z.]+\s+Code)\s+§\s*(\d+)\b/g,
+    regex: /\b([A-Z][a-z]+\.?\s+[A-Za-z.]+\s+Code)\s+§\s*(\d+[A-Za-z]*)\b/g,
     description: 'State code citations (broad pattern, e.g., "Cal. Penal Code § 187")',
     type: 'statute',
   },

tests/extract/extractStatute.test.ts

@@ -1,5 +1,5 @@
 import { describe, it, expect } from 'vitest'
-import { extractStatute } from '@/extract'
+import { extractStatute, extractCitations } from '@/extract'
 import type { Token } from '@/tokenize'
 import type { TransformationMap } from '@/types/span'
 
@@ -197,6 +197,25 @@ describe('extractStatute', () => {
 		})
 	})
 
+	describe('trailing letters via full pipeline', () => {
+		it('should extract section with trailing uppercase letter', () => {
+			const citations = extractCitations('18 U.S.C. § 1028A')
+			expect(citations).toHaveLength(1)
+			expect(citations[0].type).toBe('statute')
+			if (citations[0].type === 'statute') {
+				expect(citations[0].section).toBe('1028A')
+			}
+		})
+
+		it('should extract section with trailing lowercase letter', () => {
+			const citations = extractCitations('18 U.S.C. § 2339B')
+			expect(citations).toHaveLength(1)
+			if (citations[0].type === 'statute') {
+				expect(citations[0].section).toBe('2339B')
+			}
+		})
+	})
+
 	describe('metadata fields', () => {
 		it('should include all required CitationBase fields', () => {
 			const token: Token = {