@sparkelf/dsh-knowledge-base 0.1.29 → 0.1.30

package/lib/index.js

@@ -9053,6 +9053,31 @@ async function syncIndexUnlocked(options) {
 * few unreadable descendants cannot starve the result; the returned list is still
 * capped at `topK`.
 */
+/**
+* 把一段连续中文切成长度 3 的滑动窗口（去重，最多 `limit` 个，默认 12）。
+*
+* trigram FTS 只能匹配长度 ≥3 的片段：整句中文当 phrase 查通常 0 命中，
+* 而窗口化之后「加油站」这样的 trigram 就能命中索引，再按 chunk 合并回来。
+*
+* 成本上限（有意为之，别去掉）：只有在「整句 phrase 查询返回 0 行」时才会走到这里，
+* 且最多展开 `limit` 个窗口 → 最多多跑 `limit` 次 FTS 查询；每个窗口仍受原来的
+* `candidateWindow`（topK*4）限制，所以单次检索的 SQL 次数与返回量都是常数级的。
+* 12 个窗口足够覆盖一句 14 字左右的中文提问里最有信息量的前缀片段；再长也不会增加开销。
+*/
+function trigramWindows(text, limit = 12) {
+	const chars = [...text].filter((char) => /[\p{Script=Han}A-Za-z0-9]/u.test(char));
+	if (chars.length < 3) return [];
+	const windows = [];
+	const seen = /* @__PURE__ */ new Set();
+	for (let i = 0; i + 3 <= chars.length; i += 1) {
+		const window = chars.slice(i, i + 3).join("");
+		if (seen.has(window)) continue;
+		seen.add(window);
+		windows.push(window);
+		if (windows.length >= limit) break;
+	}
+	return windows;
+}
 async function searchIndex(options) {
 	return withIndexLock(() => withBusyRetry(() => searchIndexUnlocked(options)));
 }
@@ -9089,13 +9114,11 @@ async function searchIndexUnlocked(options) {
 		ensureChunkColumns(db);
 		const clauses = ["d.scope = 'global'"];
 		const params = [];
-		if (!options.caller.global) {
-			clauses.push("(d.scope = 'personal' AND d.owner_id = ?)");
-			params.push(options.caller.callerId);
-			if (options.caller.departmentCodes.length > 0) {
-				clauses.push(`(d.scope = 'department' AND d.department_code IN (${options.caller.departmentCodes.map(() => "?").join(",")}))`);
-				params.push(...options.caller.departmentCodes);
-			}
+		clauses.push("(d.scope = 'personal' AND d.owner_id = ?)");
+		params.push(options.caller.callerId);
+		if (options.caller.departmentCodes.length > 0) {
+			clauses.push(`(d.scope = 'department' AND d.department_code IN (${options.caller.departmentCodes.map(() => "?").join(",")}))`);
+			params.push(...options.caller.departmentCodes);
 		}
 		const narrowed = options.scope === "personal" || options.scope === "department" || options.scope === "global" ? options.scope : void 0;
 		const scopeClause = narrowed === void 0 ? "" : " AND d.scope = ?";
@@ -9133,7 +9156,8 @@ async function searchIndexUnlocked(options) {
 		};
 		const rows = runTerm(query);
 		const tokens = query.split(/\s+/u).map((token) => token.trim()).filter((token) => token !== "");
-		const relaxed = tokens.length > 1 && rows.length < 2;
+		const cjkWindows = tokens.length === 1 && rows.length === 0 ? trigramWindows(query) : [];
+		const relaxed = tokens.length > 1 && rows.length < 2 || cjkWindows.length > 0;
 		const merged = /* @__PURE__ */ new Map();
 		const addRows = (candidates) => {
 			for (const row of candidates) {
@@ -9142,7 +9166,7 @@ async function searchIndexUnlocked(options) {
 				if (existing === void 0 || row.score < existing.score) merged.set(key, row);
 			}
 		};
-		const queries = relaxed ? tokens.slice(0, 8) : [query];
+		const queries = relaxed ? cjkWindows.length > 0 ? cjkWindows : tokens.slice(0, 8) : [query];
 		if (relaxed) for (const token of queries) addRows(runTerm(token));
 		else addRows(rows);
 		const ordered = [...merged.values()].sort((a, b) => {

package/lib/types/ftsIndex.d.ts

@@ -287,6 +287,18 @@ export declare function indexStats(indexPath: string): Promise<{
  * few unreadable descendants cannot starve the result; the returned list is still
  * capped at `topK`.
  */
+/**
+ * 把一段连续中文切成长度 3 的滑动窗口（去重，最多 `limit` 个，默认 12）。
+ *
+ * trigram FTS 只能匹配长度 ≥3 的片段：整句中文当 phrase 查通常 0 命中，
+ * 而窗口化之后「加油站」这样的 trigram 就能命中索引，再按 chunk 合并回来。
+ *
+ * 成本上限（有意为之，别去掉）：只有在「整句 phrase 查询返回 0 行」时才会走到这里，
+ * 且最多展开 `limit` 个窗口 → 最多多跑 `limit` 次 FTS 查询；每个窗口仍受原来的
+ * `candidateWindow`（topK*4）限制，所以单次检索的 SQL 次数与返回量都是常数级的。
+ * 12 个窗口足够覆盖一句 14 字左右的中文提问里最有信息量的前缀片段；再长也不会增加开销。
+ */
+export declare function trigramWindows(text: string, limit?: number): string[];
 export declare function searchIndex(options: {
     indexPath: string;
     query: string;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sparkelf/dsh-knowledge-base",
-  "version": "0.1.29",
+  "version": "0.1.30",
   "description": "DataOps-backed knowledge base for DeepSeek Harness: a left-sidebar destination, Settings configuration, and index tooling over a mounted NAS view.",
   "type": "module",
   "main": "./lib/index.js",