feat(scoring): implemented semantic context padding and optimized excerpting v2.76.0

tests/scoring.test.ts

@@ -94,6 +94,21 @@ describe('Scoring Engine Unit Tests (v2.72.0)', () => {
         expect(duration).toBeLessThan(100); // Tokenizer should be efficient even for long text
     });
 
+    test('Contextual Completeness: should include adjacent sentences for semantic padding', () => {
+        const content = `
+            도입부 문장입니다.
+            핵심 키워드 성능 최적화가 포함된 문장입니다.
+            마무리 문장입니다.
+        `;
+        const query = ['성능', '최적화'];
+        const excerpt = extractBestExcerpt(content, query, 200);
+        
+        // Should include introduction and conclusion due to padding
+        expect(excerpt).toContain('도입부 문장');
+        expect(excerpt).toContain('핵심 키워드');
+        expect(excerpt).toContain('마무리 문장');
+    });
+
     test('Performance Benchmark: should process 100 documents within threshold', () => {
         const query = tokenize('performance optimization');
         const largeDocs = Array.from({ length: 100 }, (_, i) => ({