feat: Add exact phrase search support to FTS

src/store.test.ts

@@ -35,6 +35,7 @@ import {
   parseVirtualPath,
   normalizeDocid,
   isDocid,
+  buildFTS5Query,
   type Store,
   type DocumentResult,
   type SearchResult,
@@ -935,6 +936,123 @@ describe("FTS Search", () => {
   });
 });
 
+// =============================================================================
+// buildFTS5Query Unit Tests
+// =============================================================================
+
+describe("buildFTS5Query", () => {
+  test("returns null for empty query", () => {
+    expect(buildFTS5Query("")).toBeNull();
+    expect(buildFTS5Query("   ")).toBeNull();
+  });
+
+  test("single term becomes prefix match", () => {
+    expect(buildFTS5Query("hello")).toBe('"hello"*');
+  });
+
+  test("multiple terms become prefix matches with AND", () => {
+    expect(buildFTS5Query("hello world")).toBe('"hello"* AND "world"*');
+  });
+
+  test("quoted phrase becomes exact phrase match", () => {
+    expect(buildFTS5Query('"hello world"')).toBe('"hello world"');
+  });
+
+  test("mixed quoted and unquoted terms", () => {
+    expect(buildFTS5Query('meeting "Q1 planning"')).toBe('"q1 planning" AND "meeting"*');
+  });
+
+  test("multiple quoted phrases", () => {
+    expect(buildFTS5Query('"hello world" "foo bar"')).toBe('"hello world" AND "foo bar"');
+  });
+
+  test("quoted phrase with unquoted terms on both sides", () => {
+    expect(buildFTS5Query('before "exact phrase" after')).toBe('"exact phrase" AND "before"* AND "after"*');
+  });
+
+  test("sanitizes special characters in terms", () => {
+    expect(buildFTS5Query("foo(bar)")).toBe('"foobar"*');
+  });
+
+  test("sanitizes special characters in phrases", () => {
+    expect(buildFTS5Query('"foo(bar) baz"')).toBe('"foo bar baz"');
+  });
+
+  test("handles unicode in terms", () => {
+    expect(buildFTS5Query("日本語")).toBe('"日本語"*');
+  });
+
+  test("handles unicode in phrases", () => {
+    expect(buildFTS5Query('"日本語 テスト"')).toBe('"日本語 テスト"');
+  });
+
+  test("normalizes case in terms", () => {
+    expect(buildFTS5Query("Hello WORLD")).toBe('"hello"* AND "world"*');
+  });
+
+  test("normalizes case in phrases", () => {
+    expect(buildFTS5Query('"Hello WORLD"')).toBe('"hello world"');
+  });
+
+  test("empty quoted phrase is ignored", () => {
+    expect(buildFTS5Query('"" hello')).toBe('"hello"*');
+  });
+
+  test("phrase with only special chars is ignored", () => {
+    expect(buildFTS5Query('"()" hello')).toBe('"hello"*');
+  });
+});
+
+describe("FTS Search with exact phrases", () => {
+  test("searchFTS finds exact phrase match", async () => {
+    const store = await createTestStore();
+    const collectionName = await createTestCollection();
+
+    await insertTestDocument(store.db, collectionName, {
+      name: "doc1",
+      body: "The quick brown fox jumps over the lazy dog",
+      displayPath: "test/doc1.md",
+    });
+
+    await insertTestDocument(store.db, collectionName, {
+      name: "doc2",
+      body: "The brown quick fox is here",
+      displayPath: "test/doc2.md",
+    });
+
+    // Exact phrase "quick brown" should only match doc1
+    const results = store.searchFTS('"quick brown"', 10);
+    expect(results).toHaveLength(1);
+    expect(results[0]!.displayPath).toBe(`${collectionName}/test/doc1.md`);
+
+    await cleanupTestDb(store);
+  });
+
+  test("searchFTS combines exact phrase with prefix term", async () => {
+    const store = await createTestStore();
+    const collectionName = await createTestCollection();
+
+    await insertTestDocument(store.db, collectionName, {
+      name: "doc1",
+      body: "Meeting notes about project timeline discussion",
+      displayPath: "test/doc1.md",
+    });
+
+    await insertTestDocument(store.db, collectionName, {
+      name: "doc2",
+      body: "The project timeline is important",
+      displayPath: "test/doc2.md",
+    });
+
+    // Should match doc1 which has both "meeting" and "project timeline"
+    const results = store.searchFTS('meeting "project timeline"', 10);
+    expect(results).toHaveLength(1);
+    expect(results[0]!.displayPath).toBe(`${collectionName}/test/doc1.md`);
+
+    await cleanupTestDb(store);
+  });
+});
+
 // =============================================================================
 // Document Retrieval Tests
 // =============================================================================

src/store.ts

@@ -1834,13 +1834,51 @@ function sanitizeFTS5Term(term: string): string {
   return term.replace(/[^\p{L}\p{N}']/gu, '').toLowerCase();
 }
 
-function buildFTS5Query(query: string): string | null {
-  const terms = query.split(/\s+/)
+/**
+ * Sanitize a phrase for FTS5 exact phrase matching.
+ * Removes characters that could break FTS5 syntax but preserves spaces.
+ */
+function sanitizeFTS5Phrase(phrase: string): string {
+  // Remove quotes and special FTS5 operators, but keep spaces and alphanumerics
+  return phrase.replace(/[^\p{L}\p{N}\s']/gu, ' ').replace(/\s+/g, ' ').trim().toLowerCase();
+}
+
+/**
+ * Build an FTS5 query string that supports:
+ * - Quoted phrases: "hello world" → exact phrase match
+ * - Unquoted terms: hello world → prefix matches with AND
+ * - Mixed: meeting "Q1 planning" → "meeting"* AND "Q1 planning"
+ */
+export function buildFTS5Query(query: string): string | null {
+  const parts: string[] = [];
+  let remaining = query;
+
+  // Extract quoted phrases first
+  const phraseRegex = /"([^"]+)"/g;
+  let match;
+  while ((match = phraseRegex.exec(query)) !== null) {
+    const phrase = sanitizeFTS5Phrase(match[1]);
+    if (phrase.length > 0) {
+      // Exact phrase match (no asterisk)
+      parts.push(`"${phrase}"`);
+    }
+  }
+  // Remove quoted phrases from remaining text
+  remaining = remaining.replace(phraseRegex, ' ');
+
+  // Process remaining unquoted terms as prefix matches
+  const terms = remaining.split(/\s+/)
     .map(t => sanitizeFTS5Term(t))
     .filter(t => t.length > 0);
-  if (terms.length === 0) return null;
-  if (terms.length === 1) return `"${terms[0]}"*`;
-  return terms.map(t => `"${t}"*`).join(' AND ');
+
+  for (const term of terms) {
+    // Prefix match with asterisk
+    parts.push(`"${term}"*`);
+  }
+
+  if (parts.length === 0) return null;
+  if (parts.length === 1) return parts[0];
+  return parts.join(' AND ');
 }
 
 export function searchFTS(db: Database, query: string, limit: number = 20, collectionId?: number): SearchResult[] {