spectra-cli 4.2.0 → 4.3.0

package/dist/scaffold-kb/entity-heuristic.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"entity-heuristic.js","sourceRoot":"","sources":["../../src/scaffold-kb/entity-heuristic.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAGH,OAAO,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAEhD,MAAM,oBAAoB,GAAG,GAAG,CAAC;AAEjC,wEAAwE;AACxE,MAAM,MAAM,GACV,+GAA+G,CAAC;AAClH,oDAAoD;AACpD,MAAM,UAAU,GAAG,uEAAuE,CAAC;AAC3F,8BAA8B;AAC9B,MAAM,aAAa,GAAG,mEAAmE,CAAC;AAE1F,6BAA6B;AAC7B,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC;IAC1B,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,UAAU,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO;IAC/E,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK;IAC9E,QAAQ,EAAE,YAAY,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO;CAClF,CAAC,CAAC;AAEH,SAAS,WAAW,CAAC,IAAY;IAC/B,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACjC,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC;IACzE,OAAO,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC;AAC7C,CAAC;AAED,wDAAwD;AACxD,SAAS,WAAW,CAAC,QAAgB;IACnC,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,EAAE,CAAC;IAChC,IAAI,CAAC,OAAO;QAAE,OAAO,EAAE,CAAC;IACxB,OAAO,OAAO;SACX,KAAK,CAAC,GAAG,CAAC;SACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC;SAC3B,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;QACT,6DAA6D;QAC7D,MAAM,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAC3B,MAAM,OAAO,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;QAC3D,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QACtE,MAAM,IAAI,GAAG,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;QACvE,MAAM,KAAK,GAA6C,EAAE,IAAI,EAAE,CAAC;QACjE,IAAI,IAAI;YAAE,KAAK,CAAC,IAAI,GAAG,IAAI,CAAC;QAC5B,OAAO,KAAK,CAAC;IACf,CAAC,CAAC;SACD,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;AACpE,CAAC;AAED,SAAS,eAAe,CAAC,IAAY;IACnC,MAAM,CAAC,GAAG,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACnC,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,YAAY,EAAE,KAAK,EAAE,CAAC;IACvC,MAAM,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC;IACnC,OAAO,KAAK,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,IAAI,EAAE,CAAC;AACxE,CAAC;AAED,oCAAoC;AACpC,MAAM,UAAU,gBAAgB,CAAC,OAA0B;IACzD,MAAM,GAAG,GAAgB,EAAE,CAAC;IAC5B,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,YAAY,GAAG,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IAC/C,MAAM,GAAG,GAAG,eAAe,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAE1C,MAAM,IAAI,GAAG,CACX,IAAY,EACZ,IAAuB,EACvB,SAAwB,EACxB,MAA2B,EACrB,EAAE;QACR,MAAM,EAAE,GAAG,YAAY,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;QACpC,IAAI,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;YAAE,OAAO;QACzB,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;QACb,MAAM,CAAC,GAAc;YACnB,EAAE;YACF,IAAI;YACJ,aAAa,EAAE,IAAI;YACnB,IAAI;YACJ,SAAS;YACT,WAAW,EAAE,OAAO,CAAC,KAAK;YAC1B,aAAa,EAAE,YAAY;YAC3B,cAAc,EAAE,OAAO,CAAC,QAAQ;YAChC,YAAY,EAAE,OAAO,CAAC,MAAM;YAC5B,IAAI,EAAE,OAAO,CAAC,IAAI;YAClB,UAAU,EAAE,oBAAoB;YAChC,gBAAgB,EAAE,WAAW;SAC9B,CAAC;QACF,IAAI,MAAM,KAAK,SAAS;YAAE,CAAC,CAAC,MAAM,GAAG,MAAM,CAAC;QAC5C,IAAI,GAAG,CAAC,YAAY;YAAE,CAAC,CAAC,UAAU,GAAG,GAAG,CAAC;QACzC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACd,CAAC,CAAC;IAEF,SAAS;IACT,KAAK,MAAM,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,EAAE,CAAC;QAClD,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAClB,IAAI,IAAI;YAAE,IAAI,CAAC,IAAI,EAAE,YAAY,EAAE,IAAI,EAAE,SAAS,CAAC,CAAC;IACtD,CAAC;IACD,aAAa;IACb,KAAK,MAAM,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,EAAE,CAAC;QAC9C,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;QAClB,MAAM,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAC5B,IAAI,CAAC,IAAI,IAAI,WAAW,CAAC,IAAI,CAAC;YAAE,SAAS;QACzC,IAAI,CAAC,IAAI,EAAE,UAAU,EAAE,GAAG,IAAI,IAAI,QAAQ,CAAC,IAAI,EAAE,GAAG,EAAE,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC;IAC/E,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC"}

package/dist/scaffold-kb/entity-matcher.d.ts

@@ -0,0 +1,20 @@
+/**
+ * F192 T007 — API 实体匹配（kb_api_lookup 用，FR-004 + W-2 消歧）
+ *
+ * 精确（name/qualified_name/末段 相等）优先，模糊（子串 / token 重叠）兜底；
+ * 支持 kind / container 过滤；多命中返回 top-N（不静默取一）。纯函数。
+ */
+import type { ApiEntity } from './types.js';
+export interface MatchQuery {
+    apiName: string;
+    kind?: ApiEntity['kind'] | null;
+    /** 按所属 class/module 限定（消歧重载/同名） */
+    container?: string | null;
+    topN?: number;
+}
+export interface EntityMatch extends ApiEntity {
+    matchType: 'exact' | 'fuzzy';
+}
+/** 在实体集中匹配查询，返回按相关度排序的 top-N（精确在前） */
+export declare function matchEntities(entities: ApiEntity[], q: MatchQuery): EntityMatch[];
+//# sourceMappingURL=entity-matcher.d.ts.map

package/dist/scaffold-kb/entity-matcher.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"entity-matcher.d.ts","sourceRoot":"","sources":["../../src/scaffold-kb/entity-matcher.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,YAAY,CAAC;AAG5C,MAAM,WAAW,UAAU;IACzB,OAAO,EAAE,MAAM,CAAC;IAChB,IAAI,CAAC,EAAE,SAAS,CAAC,MAAM,CAAC,GAAG,IAAI,CAAC;IAChC,mCAAmC;IACnC,SAAS,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,IAAI,CAAC,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,WAAY,SAAQ,SAAS;IAC5C,SAAS,EAAE,OAAO,GAAG,OAAO,CAAC;CAC9B;AAUD,sCAAsC;AACtC,wBAAgB,aAAa,CAAC,QAAQ,EAAE,SAAS,EAAE,EAAE,CAAC,EAAE,UAAU,GAAG,WAAW,EAAE,CAoCjF"}

package/dist/scaffold-kb/entity-matcher.js

@@ -0,0 +1,53 @@
+/**
+ * F192 T007 — API 实体匹配（kb_api_lookup 用，FR-004 + W-2 消歧）
+ *
+ * 精确（name/qualified_name/末段 相等）优先，模糊（子串 / token 重叠）兜底；
+ * 支持 kind / container 过滤；多命中返回 top-N（不静默取一）。纯函数。
+ */
+import { tokenize } from './tokenizer.js';
+function norm(s) {
+    return s.trim().toLowerCase();
+}
+function lastSegment(qualified) {
+    return qualified.includes('.') ? (qualified.split('.').pop() ?? qualified) : qualified;
+}
+/** 在实体集中匹配查询，返回按相关度排序的 top-N（精确在前） */
+export function matchEntities(entities, q) {
+    const topN = q.topN ?? 10;
+    const qn = norm(q.apiName);
+    if (!qn)
+        return [];
+    const qTokens = new Set(tokenize(q.apiName).map(norm));
+    const candidates = entities.filter((e) => {
+        if (q.kind && e.kind !== q.kind)
+            return false;
+        if (q.container && norm(e.container ?? '') !== norm(q.container))
+            return false;
+        return true;
+    });
+    const scored = [];
+    for (const e of candidates) {
+        const name = norm(e.name);
+        const qualified = norm(e.qualifiedName);
+        const last = norm(lastSegment(e.qualifiedName));
+        if (name === qn || qualified === qn || last === qn) {
+            scored.push({ e, matchType: 'exact', rank: 0 });
+            continue;
+        }
+        // qn.includes(name) 仅在 name 足够长时启用，避免短名（on/api/get）成为乱串查询的误命中
+        const substr = name.includes(qn) || qualified.includes(qn) || (name.length >= 4 && qn.includes(name));
+        const eTokens = new Set([...tokenize(e.name), ...tokenize(e.qualifiedName)].map(norm));
+        let overlap = 0;
+        for (const t of qTokens)
+            if (t.length > 0 && eTokens.has(t))
+                overlap++;
+        if (substr || overlap > 0) {
+            // rank: 子串(1.0) < token 重叠(1 + 缺口比例)
+            const rank = substr ? 1.0 : 2.0 - overlap / Math.max(1, qTokens.size);
+            scored.push({ e, matchType: 'fuzzy', rank });
+        }
+    }
+    scored.sort((a, b) => a.rank - b.rank);
+    return scored.slice(0, topN).map(({ e, matchType }) => ({ ...e, matchType }));
+}
+//# sourceMappingURL=entity-matcher.js.map

package/dist/scaffold-kb/entity-matcher.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"entity-matcher.js","sourceRoot":"","sources":["../../src/scaffold-kb/entity-matcher.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAGH,OAAO,EAAE,QAAQ,EAAE,MAAM,gBAAgB,CAAC;AAc1C,SAAS,IAAI,CAAC,CAAS;IACrB,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;AAChC,CAAC;AAED,SAAS,WAAW,CAAC,SAAiB;IACpC,OAAO,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,SAAS,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AACzF,CAAC;AAED,sCAAsC;AACtC,MAAM,UAAU,aAAa,CAAC,QAAqB,EAAE,CAAa;IAChE,MAAM,IAAI,GAAG,CAAC,CAAC,IAAI,IAAI,EAAE,CAAC;IAC1B,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC;IAC3B,IAAI,CAAC,EAAE;QAAE,OAAO,EAAE,CAAC;IACnB,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC;IAEvD,MAAM,UAAU,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE;QACvC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,IAAI;YAAE,OAAO,KAAK,CAAC;QAC9C,IAAI,CAAC,CAAC,SAAS,IAAI,IAAI,CAAC,CAAC,CAAC,SAAS,IAAI,EAAE,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC;YAAE,OAAO,KAAK,CAAC;QAC/E,OAAO,IAAI,CAAC;IACd,CAAC,CAAC,CAAC;IAEH,MAAM,MAAM,GAAwE,EAAE,CAAC;IACvF,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAC1B,MAAM,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC;QACxC,MAAM,IAAI,GAAG,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC;QAEhD,IAAI,IAAI,KAAK,EAAE,IAAI,SAAS,KAAK,EAAE,IAAI,IAAI,KAAK,EAAE,EAAE,CAAC;YACnD,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC;YAChD,SAAS;QACX,CAAC;QACD,8DAA8D;QAC9D,MAAM,MAAM,GAAG,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC,IAAI,SAAS,CAAC,QAAQ,CAAC,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,IAAI,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;QACtG,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,GAAG,QAAQ,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC;QACvF,IAAI,OAAO,GAAG,CAAC,CAAC;QAChB,KAAK,MAAM,CAAC,IAAI,OAAO;YAAE,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC;gBAAE,OAAO,EAAE,CAAC;QACvE,IAAI,MAAM,IAAI,OAAO,GAAG,CAAC,EAAE,CAAC;YAC1B,qCAAqC;YACrC,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,GAAG,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC;YACtE,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,SAAS,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;QAC/C,CAAC;IACH,CAAC;IAED,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,EAAE,SAAS,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC;AAChF,CAAC"}

package/dist/scaffold-kb/entity-util.d.ts

@@ -0,0 +1,11 @@
+/**
+ * F192 — 实体共享工具（id 归一 + confidence clamp），供 heuristic/LLM 抽取 + matcher 复用
+ */
+/**
+ * 实体稳定唯一 id：qualified_name + kind (+ overload_key) 归一。
+ * 同名重载靠 overload_key 区分（W-2）。
+ */
+export declare function makeEntityId(qualifiedName: string, kind: string, overloadKey?: string | null): string;
+/** confidence clamp 到 [0,1]；非数值 → fallback（I-1） */
+export declare function clampConfidence(v: unknown, fallback?: number): number;
+//# sourceMappingURL=entity-util.d.ts.map

package/dist/scaffold-kb/entity-util.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"entity-util.d.ts","sourceRoot":"","sources":["../../src/scaffold-kb/entity-util.ts"],"names":[],"mappings":"AAAA;;GAEG;AAOH;;;GAGG;AACH,wBAAgB,YAAY,CAC1B,aAAa,EAAE,MAAM,EACrB,IAAI,EAAE,MAAM,EACZ,WAAW,CAAC,EAAE,MAAM,GAAG,IAAI,GAC1B,MAAM,CAGR;AAED,mDAAmD;AACnD,wBAAgB,eAAe,CAAC,CAAC,EAAE,OAAO,EAAE,QAAQ,SAAI,GAAG,MAAM,CAIhE"}

package/dist/scaffold-kb/entity-util.js

@@ -0,0 +1,23 @@
+/**
+ * F192 — 实体共享工具（id 归一 + confidence clamp），供 heuristic/LLM 抽取 + matcher 复用
+ */
+/** 归一字符串：trim + 小写 + 折叠空白 */
+function norm(s) {
+    return s.trim().toLowerCase().replace(/\s+/g, ' ');
+}
+/**
+ * 实体稳定唯一 id：qualified_name + kind (+ overload_key) 归一。
+ * 同名重载靠 overload_key 区分（W-2）。
+ */
+export function makeEntityId(qualifiedName, kind, overloadKey) {
+    const base = `${norm(qualifiedName)}#${kind}`;
+    return overloadKey && overloadKey.trim() ? `${base}#${norm(overloadKey)}` : base;
+}
+/** confidence clamp 到 [0,1]；非数值 → fallback（I-1） */
+export function clampConfidence(v, fallback = 0) {
+    const n = typeof v === 'number' ? v : Number(v);
+    if (!Number.isFinite(n))
+        return fallback;
+    return Math.max(0, Math.min(1, n));
+}
+//# sourceMappingURL=entity-util.js.map

package/dist/scaffold-kb/entity-util.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"entity-util.js","sourceRoot":"","sources":["../../src/scaffold-kb/entity-util.ts"],"names":[],"mappings":"AAAA;;GAEG;AAEH,6BAA6B;AAC7B,SAAS,IAAI,CAAC,CAAS;IACrB,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;AACrD,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,YAAY,CAC1B,aAAqB,EACrB,IAAY,EACZ,WAA2B;IAE3B,MAAM,IAAI,GAAG,GAAG,IAAI,CAAC,aAAa,CAAC,IAAI,IAAI,EAAE,CAAC;IAC9C,OAAO,WAAW,IAAI,WAAW,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;AACnF,CAAC;AAED,mDAAmD;AACnD,MAAM,UAAU,eAAe,CAAC,CAAU,EAAE,QAAQ,GAAG,CAAC;IACtD,MAAM,CAAC,GAAG,OAAO,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;IAChD,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;QAAE,OAAO,QAAQ,CAAC;IACzC,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;AACrC,CAAC"}

package/dist/scaffold-kb/evidence-envelope.d.ts

@@ -0,0 +1,17 @@
+/**
+ * F190/F191 — untrusted-evidence envelope 共享工具（从 kb-search 抽出，kb-search + injection-format 同源）
+ *
+ * 信任边界（F190 FR-011 / SC-010）：把 KB 检索内容包成定界 evidence 区，
+ * 中和内嵌 sentinel（防提前闭合逃逸）、编码属性值、surrogate-safe 截断。
+ * 依赖方向：本模块在 scaffold-kb，不 import kb-mcp（SourceKind 取自 ./types）。
+ */
+import type { SourceKind } from './types.js';
+/** envelope 属性值编码：去除可破坏头部的 `]` `"` 换行（防注入逃逸） */
+export declare function safeAttr(v: string): string;
+/** 中和正文内的 envelope sentinel，防止提前闭合把注入文本放到 envelope 外（trust boundary） */
+export declare function defangSentinel(content: string): string;
+/** 按 UTF-16 code unit 截断但不切开代理对（避免孤立 surrogate） */
+export declare function safeTruncate(s: string, max: number): string;
+/** 把单条内容包成 `[KB-EVIDENCE …]…[/KB-EVIDENCE]` evidence 块 */
+export declare function buildEvidenceEnvelope(content: string, docId: string, src: SourceKind, builtAt: string): string;
+//# sourceMappingURL=evidence-envelope.d.ts.map

package/dist/scaffold-kb/evidence-envelope.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"evidence-envelope.d.ts","sourceRoot":"","sources":["../../src/scaffold-kb/evidence-envelope.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AAE7C,gDAAgD;AAChD,wBAAgB,QAAQ,CAAC,CAAC,EAAE,MAAM,GAAG,MAAM,CAE1C;AAED,wEAAwE;AACxE,wBAAgB,cAAc,CAAC,OAAO,EAAE,MAAM,GAAG,MAAM,CAItD;AAED,mDAAmD;AACnD,wBAAgB,YAAY,CAAC,CAAC,EAAE,MAAM,EAAE,GAAG,EAAE,MAAM,GAAG,MAAM,CAM3D;AAED,0DAA0D;AAC1D,wBAAgB,qBAAqB,CACnC,OAAO,EAAE,MAAM,EACf,KAAK,EAAE,MAAM,EACb,GAAG,EAAE,UAAU,EACf,OAAO,EAAE,MAAM,GACd,MAAM,CAER"}

package/dist/scaffold-kb/evidence-envelope.js

@@ -0,0 +1,32 @@
+/**
+ * F190/F191 — untrusted-evidence envelope 共享工具（从 kb-search 抽出，kb-search + injection-format 同源）
+ *
+ * 信任边界（F190 FR-011 / SC-010）：把 KB 检索内容包成定界 evidence 区，
+ * 中和内嵌 sentinel（防提前闭合逃逸）、编码属性值、surrogate-safe 截断。
+ * 依赖方向：本模块在 scaffold-kb，不 import kb-mcp（SourceKind 取自 ./types）。
+ */
+/** envelope 属性值编码：去除可破坏头部的 `]` `"` 换行（防注入逃逸） */
+export function safeAttr(v) {
+    return v.replace(/[\]"\r\n]/g, ' ');
+}
+/** 中和正文内的 envelope sentinel，防止提前闭合把注入文本放到 envelope 外（trust boundary） */
+export function defangSentinel(content) {
+    return content
+        .replace(/\[\s*\/\s*KB-EVIDENCE\s*\]/gi, '[ /KB-EVIDENCE ]')
+        .replace(/\[\s*KB-EVIDENCE/gi, '[ KB-EVIDENCE');
+}
+/** 按 UTF-16 code unit 截断但不切开代理对（避免孤立 surrogate） */
+export function safeTruncate(s, max) {
+    if (s.length <= max)
+        return s;
+    let end = max;
+    const code = s.charCodeAt(end - 1);
+    if (code >= 0xd800 && code <= 0xdbff)
+        end -= 1; // 末位是高代理 → 回退一位
+    return s.slice(0, end);
+}
+/** 把单条内容包成 `[KB-EVIDENCE …]…[/KB-EVIDENCE]` evidence 块 */
+export function buildEvidenceEnvelope(content, docId, src, builtAt) {
+    return `[KB-EVIDENCE doc_id="${safeAttr(docId)}" src="${src}" built_at="${safeAttr(builtAt)}"]\n${defangSentinel(content)}\n[/KB-EVIDENCE]`;
+}
+//# sourceMappingURL=evidence-envelope.js.map

package/dist/scaffold-kb/evidence-envelope.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"evidence-envelope.js","sourceRoot":"","sources":["../../src/scaffold-kb/evidence-envelope.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,gDAAgD;AAChD,MAAM,UAAU,QAAQ,CAAC,CAAS;IAChC,OAAO,CAAC,CAAC,OAAO,CAAC,YAAY,EAAE,GAAG,CAAC,CAAC;AACtC,CAAC;AAED,wEAAwE;AACxE,MAAM,UAAU,cAAc,CAAC,OAAe;IAC5C,OAAO,OAAO;SACX,OAAO,CAAC,8BAA8B,EAAE,kBAAkB,CAAC;SAC3D,OAAO,CAAC,oBAAoB,EAAE,eAAe,CAAC,CAAC;AACpD,CAAC;AAED,mDAAmD;AACnD,MAAM,UAAU,YAAY,CAAC,CAAS,EAAE,GAAW;IACjD,IAAI,CAAC,CAAC,MAAM,IAAI,GAAG;QAAE,OAAO,CAAC,CAAC;IAC9B,IAAI,GAAG,GAAG,GAAG,CAAC;IACd,MAAM,IAAI,GAAG,CAAC,CAAC,UAAU,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC;IACnC,IAAI,IAAI,IAAI,MAAM,IAAI,IAAI,IAAI,MAAM;QAAE,GAAG,IAAI,CAAC,CAAC,CAAC,gBAAgB;IAChE,OAAO,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;AACzB,CAAC;AAED,0DAA0D;AAC1D,MAAM,UAAU,qBAAqB,CACnC,OAAe,EACf,KAAa,EACb,GAAe,EACf,OAAe;IAEf,OAAO,wBAAwB,QAAQ,CAAC,KAAK,CAAC,UAAU,GAAG,eAAe,QAAQ,CAAC,OAAO,CAAC,OAAO,cAAc,CAAC,OAAO,CAAC,kBAAkB,CAAC;AAC9I,CAAC"}

package/dist/scaffold-kb/index.d.ts

@@ -0,0 +1,20 @@
+/**
+ * F190 scaffold-kb — buildKb 主流程编排
+ *
+ * ingest → split → doc-graph + chunks → 落盘 kb/{doc-graph.json, chunks.sqlite}
+ * 原子性：ingest 失败前未写任何文件；所有产物在内存算好后再一次性落盘
+ * （EC-008：不留中间态残片）。
+ *
+ * 同一 CLI 指向项目路径即写项目库（FR-016：`--output <project>/.{tool}/kb`）。
+ */
+import type { BuildKbOptions } from './types.js';
+export interface BuildKbResult {
+    docCount: number;
+    chunkCount: number;
+    entityCount: number;
+    extractionMethod: 'llm' | 'heuristic' | 'mixed';
+    outputPath: string;
+    builtAt: string;
+}
+export declare function buildKb(opts: BuildKbOptions): Promise<BuildKbResult>;
+//# sourceMappingURL=index.d.ts.map

package/dist/scaffold-kb/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/scaffold-kb/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,KAAK,EAAiB,cAAc,EAA8B,MAAM,YAAY,CAAC;AAS5F,MAAM,WAAW,aAAa;IAC5B,QAAQ,EAAE,MAAM,CAAC;IACjB,UAAU,EAAE,MAAM,CAAC;IACnB,WAAW,EAAE,MAAM,CAAC;IACpB,gBAAgB,EAAE,KAAK,GAAG,WAAW,GAAG,OAAO,CAAC;IAChD,UAAU,EAAE,MAAM,CAAC;IACnB,OAAO,EAAE,MAAM,CAAC;CACjB;AAqBD,wBAAsB,OAAO,CAAC,IAAI,EAAE,cAAc,GAAG,OAAO,CAAC,aAAa,CAAC,CAqE1E"}

package/dist/scaffold-kb/index.js

@@ -0,0 +1,102 @@
+/**
+ * F190 scaffold-kb — buildKb 主流程编排
+ *
+ * ingest → split → doc-graph + chunks → 落盘 kb/{doc-graph.json, chunks.sqlite}
+ * 原子性：ingest 失败前未写任何文件；所有产物在内存算好后再一次性落盘
+ * （EC-008：不留中间态残片）。
+ *
+ * 同一 CLI 指向项目路径即写项目库（FR-016：`--output <project>/.{tool}/kb`）。
+ */
+import { ingestDocuments } from './ingester.js';
+import { splitDocument } from './chunk-splitter.js';
+import { buildDocGraph } from './doc-graph-builder.js';
+import { buildChunksDbBytes } from './sqlite-writer.js';
+import { aggregateSections, extractEntities } from './entity-extractor.js';
+import { serializeApiEntities } from './api-entities-serializer.js';
+import { writeKbArtifactsAtomic } from './kb-writer.js';
+/** 把内存 DocGraph 序列化为 spec §3.2 的 on-disk JSON（snake_case，含 schema_version 供 EC-009） */
+function serializeDocGraph(g) {
+    return {
+        schema_version: g.schemaVersion,
+        source: g.source,
+        built_at: g.builtAt,
+        sdk_version: g.sdkVersion,
+        nodes: g.nodes.map((n) => ({
+            id: n.id,
+            title: n.title,
+            summary: n.summary ?? null,
+            tags: n.tags ?? [],
+            lang: n.lang,
+            source_url: n.sourceUrl,
+        })),
+        edges: g.edges.map((e) => ({ source: e.source, target: e.target, relation: e.relation })),
+    };
+}
+export async function buildKb(opts) {
+    const outputPath = opts.outputPath ?? './kb';
+    const builtAt = opts.builtAt ?? new Date().toISOString();
+    const sdkVersion = opts.sdkVersion ?? null;
+    const source = opts.llmsTxtUrl ? 'llms.txt' : 'directory';
+    // 1. ingest（参数校验 + 抓取/扫描；失败抛错，此前未落盘 → 原子性）
+    const ingestOpts = {};
+    if (opts.dirPath !== undefined)
+        ingestOpts.dirPath = opts.dirPath;
+    if (opts.llmsTxtUrl !== undefined)
+        ingestOpts.llmsTxtUrl = opts.llmsTxtUrl;
+    if (opts.lang !== undefined)
+        ingestOpts.lang = opts.lang;
+    const docs = await ingestDocuments(ingestOpts);
+    // 2. 切分 + 组装 chunk_meta（doc_title/source_url 冗余自 doc，R-003）
+    const chunks = [];
+    const meta = [];
+    for (const doc of docs) {
+        for (const c of splitDocument(doc)) {
+            chunks.push(c);
+            meta.push({
+                chunkId: c.chunkId,
+                docId: c.docId,
+                docTitle: doc.title,
+                sourceUrl: doc.sourceUrl,
+                anchor: c.anchor,
+                sdkVersion,
+                builtAt,
+            });
+        }
+    }
+    // 3. doc-graph
+    const graph = buildDocGraph(docs, { source, sdkVersion, builtAt });
+    // 4. API 实体抽取（FR-001：LLM 提质 + heuristic 兜底；--no-llm 跳过 LLM）
+    const docLang = new Map(docs.map((d) => [d.id, d.lang]));
+    const sections = aggregateSections(chunks, docLang);
+    const extractOpts = {};
+    if (opts.noLlm === true)
+        extractOpts.noLlm = true;
+    const extraction = await extractEntities(sections, extractOpts);
+    const entityFile = {
+        schemaVersion: '1.0',
+        builtAt,
+        sdkVersion,
+        sourceKind: opts.sourceKind ?? 'vendor',
+        entities: extraction.entities,
+        coverage: extraction.coverage,
+    };
+    // 5. 内存算好全部产物，再一次性原子落盘（三文件：doc-graph + chunks + api-entities，EC-008）
+    const sqliteBytes = await buildChunksDbBytes(chunks, meta);
+    const graphJson = JSON.stringify(serializeDocGraph(graph), null, 2);
+    const entitiesJson = JSON.stringify(serializeApiEntities(entityFile), null, 2);
+    // 三文件近似原子落盘（C-5），build 与 ingest 共用 kb-writer
+    writeKbArtifactsAtomic(outputPath, [
+        { name: 'doc-graph.json', data: graphJson },
+        { name: 'chunks.sqlite', data: sqliteBytes },
+        { name: 'api-entities.json', data: entitiesJson },
+    ]);
+    return {
+        docCount: docs.length,
+        chunkCount: chunks.length,
+        entityCount: extraction.entities.length,
+        extractionMethod: extraction.method,
+        outputPath,
+        builtAt,
+    };
+}
+//# sourceMappingURL=index.js.map

package/dist/scaffold-kb/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/scaffold-kb/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAGH,OAAO,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AAChD,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,aAAa,EAAE,MAAM,wBAAwB,CAAC;AACvD,OAAO,EAAE,kBAAkB,EAAE,MAAM,oBAAoB,CAAC;AACxD,OAAO,EAAE,iBAAiB,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;AAC3E,OAAO,EAAE,oBAAoB,EAAE,MAAM,8BAA8B,CAAC;AACpE,OAAO,EAAE,sBAAsB,EAAE,MAAM,gBAAgB,CAAC;AAWxD,uFAAuF;AACvF,SAAS,iBAAiB,CAAC,CAAW;IACpC,OAAO;QACL,cAAc,EAAE,CAAC,CAAC,aAAa;QAC/B,MAAM,EAAE,CAAC,CAAC,MAAM;QAChB,QAAQ,EAAE,CAAC,CAAC,OAAO;QACnB,WAAW,EAAE,CAAC,CAAC,UAAU;QACzB,KAAK,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;YACzB,EAAE,EAAE,CAAC,CAAC,EAAE;YACR,KAAK,EAAE,CAAC,CAAC,KAAK;YACd,OAAO,EAAE,CAAC,CAAC,OAAO,IAAI,IAAI;YAC1B,IAAI,EAAE,CAAC,CAAC,IAAI,IAAI,EAAE;YAClB,IAAI,EAAE,CAAC,CAAC,IAAI;YACZ,UAAU,EAAE,CAAC,CAAC,SAAS;SACxB,CAAC,CAAC;QACH,KAAK,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAC;KAC1F,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAoB;IAChD,MAAM,UAAU,GAAG,IAAI,CAAC,UAAU,IAAI,MAAM,CAAC;IAC7C,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,IAAI,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACzD,MAAM,UAAU,GAAG,IAAI,CAAC,UAAU,IAAI,IAAI,CAAC;IAC3C,MAAM,MAAM,GAA6B,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,WAAW,CAAC;IAEpF,2CAA2C;IAC3C,MAAM,UAAU,GAA0C,EAAE,CAAC;IAC7D,IAAI,IAAI,CAAC,OAAO,KAAK,SAAS;QAAE,UAAU,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC;IAClE,IAAI,IAAI,CAAC,UAAU,KAAK,SAAS;QAAE,UAAU,CAAC,UAAU,GAAG,IAAI,CAAC,UAAU,CAAC;IAC3E,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS;QAAE,UAAU,CAAC,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC;IACzD,MAAM,IAAI,GAAG,MAAM,eAAe,CAAC,UAAU,CAAC,CAAC;IAE/C,4DAA4D;IAC5D,MAAM,MAAM,GAAY,EAAE,CAAC;IAC3B,MAAM,IAAI,GAAgB,EAAE,CAAC;IAC7B,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,KAAK,MAAM,CAAC,IAAI,aAAa,CAAC,GAAG,CAAC,EAAE,CAAC;YACnC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACf,IAAI,CAAC,IAAI,CAAC;gBACR,OAAO,EAAE,CAAC,CAAC,OAAO;gBAClB,KAAK,EAAE,CAAC,CAAC,KAAK;gBACd,QAAQ,EAAE,GAAG,CAAC,KAAK;gBACnB,SAAS,EAAE,GAAG,CAAC,SAAS;gBACxB,MAAM,EAAE,CAAC,CAAC,MAAM;gBAChB,UAAU;gBACV,OAAO;aACR,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,eAAe;IACf,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,EAAE,EAAE,MAAM,EAAE,UAAU,EAAE,OAAO,EAAE,CAAC,CAAC;IAEnE,4DAA4D;IAC5D,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACzD,MAAM,QAAQ,GAAG,iBAAiB,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACpD,MAAM,WAAW,GAA0C,EAAE,CAAC;IAC9D,IAAI,IAAI,CAAC,KAAK,KAAK,IAAI;QAAE,WAAW,CAAC,KAAK,GAAG,IAAI,CAAC;IAClD,MAAM,UAAU,GAAG,MAAM,eAAe,CAAC,QAAQ,EAAE,WAAW,CAAC,CAAC;IAChE,MAAM,UAAU,GAAkB;QAChC,aAAa,EAAE,KAAK;QACpB,OAAO;QACP,UAAU;QACV,UAAU,EAAE,IAAI,CAAC,UAAU,IAAI,QAAQ;QACvC,QAAQ,EAAE,UAAU,CAAC,QAAQ;QAC7B,QAAQ,EAAE,UAAU,CAAC,QAAQ;KAC9B,CAAC;IAEF,qEAAqE;IACrE,MAAM,WAAW,GAAG,MAAM,kBAAkB,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;IAC3D,MAAM,SAAS,GAAG,IAAI,CAAC,SAAS,CAAC,iBAAiB,CAAC,KAAK,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACpE,MAAM,YAAY,GAAG,IAAI,CAAC,SAAS,CAAC,oBAAoB,CAAC,UAAU,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IAE/E,6CAA6C;IAC7C,sBAAsB,CAAC,UAAU,EAAE;QACjC,EAAE,IAAI,EAAE,gBAAgB,EAAE,IAAI,EAAE,SAAS,EAAE;QAC3C,EAAE,IAAI,EAAE,eAAe,EAAE,IAAI,EAAE,WAAW,EAAE;QAC5C,EAAE,IAAI,EAAE,mBAAmB,EAAE,IAAI,EAAE,YAAY,EAAE;KAClD,CAAC,CAAC;IAEH,OAAO;QACL,QAAQ,EAAE,IAAI,CAAC,MAAM;QACrB,UAAU,EAAE,MAAM,CAAC,MAAM;QACzB,WAAW,EAAE,UAAU,CAAC,QAAQ,CAAC,MAAM;QACvC,gBAAgB,EAAE,UAAU,CAAC,MAAM;QACnC,UAAU;QACV,OAAO;KACR,CAAC;AACJ,CAAC"}

package/dist/scaffold-kb/ingest/ingest-core.d.ts

@@ -0,0 +1,47 @@
+/**
+ * F192 T016 — 三方导入核心（FR-009/014/015）
+ *
+ * 源（url/office/minutes）→ ParsedDoc(+provenance) → splitDocument → 抽实体 →
+ * 与现有项目库**合并去重**（内容级 dedup chunk + merge key 覆盖 entity）→ 三件套原子写项目库。
+ * 默认网络实现 = SSRF 安全 fetcher（禁 F190 默认 fetch）。单源失败不阻断其他源。
+ * 导入内容全 untrusted：入库即落 provenance，检索经 envelope（复用既有路径）。
+ */
+import type { IngestSourceType } from '../types.js';
+import { safeFetchUrl } from './url-fetcher.js';
+export declare class IngestError extends Error {
+    constructor(message: string);
+}
+export interface IngestSource {
+    kind: 'url' | 'file' | 'minutes';
+    value: string;
+}
+export interface IngestSourceResult {
+    origin: string;
+    type: IngestSourceType | null;
+    ok: boolean;
+    reason?: string;
+}
+export interface IngestPlan {
+    newDocs: number;
+    newChunks: number;
+    newEntities: number;
+    totalChunks: number;
+    totalEntities: number;
+    sources: IngestSourceResult[];
+    artifacts: {
+        graphJson: string;
+        sqliteBytes: Uint8Array;
+        entitiesJson: string;
+    };
+}
+export interface IngestOptions {
+    noLlm?: boolean;
+    builtAt?: string;
+    /** 注入安全 fetcher（测试用），缺省 safeFetchUrl */
+    fetchUrl?: typeof safeFetchUrl;
+}
+/** 准备导入（处理所有源 + 合并去重，不落盘）；commitIngest 才写 */
+export declare function prepareIngest(sources: IngestSource[], projectKbPath: string, opts?: IngestOptions): Promise<IngestPlan>;
+/** 落盘项目库三件套（原子） */
+export declare function commitIngest(projectKbPath: string, plan: IngestPlan): void;
+//# sourceMappingURL=ingest-core.d.ts.map

package/dist/scaffold-kb/ingest/ingest-core.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ingest-core.d.ts","sourceRoot":"","sources":["../../../src/scaffold-kb/ingest/ingest-core.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAIH,OAAO,KAAK,EACoD,gBAAgB,EAC/E,MAAM,aAAa,CAAC;AAQrB,OAAO,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAGhD,qBAAa,WAAY,SAAQ,KAAK;gBACxB,OAAO,EAAE,MAAM;CAI5B;AAED,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,KAAK,GAAG,MAAM,GAAG,SAAS,CAAC;IACjC,KAAK,EAAE,MAAM,CAAC;CACf;AAED,MAAM,WAAW,kBAAkB;IACjC,MAAM,EAAE,MAAM,CAAC;IACf,IAAI,EAAE,gBAAgB,GAAG,IAAI,CAAC;IAC9B,EAAE,EAAE,OAAO,CAAC;IACZ,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB;AAED,MAAM,WAAW,UAAU;IACzB,OAAO,EAAE,MAAM,CAAC;IAChB,SAAS,EAAE,MAAM,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,WAAW,EAAE,MAAM,CAAC;IACpB,aAAa,EAAE,MAAM,CAAC;IACtB,OAAO,EAAE,kBAAkB,EAAE,CAAC;IAC9B,SAAS,EAAE;QAAE,SAAS,EAAE,MAAM,CAAC;QAAC,WAAW,EAAE,UAAU,CAAC;QAAC,YAAY,EAAE,MAAM,CAAA;KAAE,CAAC;CACjF;AAED,MAAM,WAAW,aAAa;IAC5B,KAAK,CAAC,EAAE,OAAO,CAAC;IAChB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,wCAAwC;IACxC,QAAQ,CAAC,EAAE,OAAO,YAAY,CAAC;CAChC;AAwID,6CAA6C;AAC7C,wBAAsB,aAAa,CACjC,OAAO,EAAE,YAAY,EAAE,EACvB,aAAa,EAAE,MAAM,EACrB,IAAI,GAAE,aAAkB,GACvB,OAAO,CAAC,UAAU,CAAC,CA+ErB;AAED,mBAAmB;AACnB,wBAAgB,YAAY,CAAC,aAAa,EAAE,MAAM,EAAE,IAAI,EAAE,UAAU,GAAG,IAAI,CAM1E"}

package/dist/scaffold-kb/ingest/ingest-core.js

@@ -0,0 +1,231 @@
+/**
+ * F192 T016 — 三方导入核心（FR-009/014/015）
+ *
+ * 源（url/office/minutes）→ ParsedDoc(+provenance) → splitDocument → 抽实体 →
+ * 与现有项目库**合并去重**（内容级 dedup chunk + merge key 覆盖 entity）→ 三件套原子写项目库。
+ * 默认网络实现 = SSRF 安全 fetcher（禁 F190 默认 fetch）。单源失败不阻断其他源。
+ * 导入内容全 untrusted：入库即落 provenance，检索经 envelope（复用既有路径）。
+ */
+import { readFileSync, existsSync } from 'node:fs';
+import { basename, join } from 'node:path';
+import { splitDocument } from '../chunk-splitter.js';
+import { aggregateSections, extractEntities } from '../entity-extractor.js';
+import { buildChunksDbBytes } from '../sqlite-writer.js';
+import { serializeApiEntities, deserializeApiEntities } from '../api-entities-serializer.js';
+import { provenanceSelectFragment } from '../schema-compat.js';
+import { writeKbArtifactsAtomic } from '../kb-writer.js';
+import { loadDbFromBytes, queryRows } from '../sqlite-engine.js';
+import { safeFetchUrl } from './url-fetcher.js';
+import { parseOfficeFile, detectOfficeFormat } from './office-parser.js';
+export class IngestError extends Error {
+    constructor(message) {
+        super(message);
+        this.name = 'IngestError';
+    }
+}
+const HAN = /\p{Script=Han}/u;
+function detectLang(text) {
+    return HAN.test(text) ? 'zh' : 'en';
+}
+function firstH1(md) {
+    for (const line of md.split('\n')) {
+        const m = /^#\s+(.+)$/.exec(line.trim());
+        if (m)
+            return m[1].trim();
+    }
+    return undefined;
+}
+function urlToId(url) {
+    try {
+        const u = new URL(url);
+        return (u.host + u.pathname).replace(/\/$/, '');
+    }
+    catch {
+        return url;
+    }
+}
+/** 单源 → ParsedDoc + provenance；失败抛错（调用方记 reason，不阻断其他源） */
+async function sourceToDoc(source, fetchUrl) {
+    if (source.kind === 'url') {
+        const { finalUrl, markdown } = await fetchUrl(source.value);
+        const id = urlToId(finalUrl);
+        return {
+            doc: { id, title: firstH1(markdown) ?? id, content: markdown, sourceUrl: finalUrl, lang: detectLang(markdown) },
+            type: 'url',
+            origin: finalUrl,
+        };
+    }
+    if (source.kind === 'file') {
+        const fmt = detectOfficeFormat(source.value);
+        if (!fmt)
+            throw new Error(`不支持的文件类型: ${source.value}`);
+        const { text } = await parseOfficeFile(new Uint8Array(readFileSync(source.value)), fmt);
+        const id = basename(source.value);
+        const type = fmt === 'md' ? 'markdown-dir' : `office-${fmt}`;
+        return {
+            doc: { id, title: firstH1(text) ?? id, content: text, sourceUrl: source.value, lang: detectLang(text) },
+            type,
+            origin: id,
+        };
+    }
+    // minutes：自由文本/markdown
+    const text = readFileSync(source.value, 'utf-8');
+    const id = basename(source.value);
+    return {
+        doc: { id, title: firstH1(text) ?? id, content: text, sourceUrl: source.value, lang: detectLang(text) },
+        type: 'minutes',
+        origin: id,
+    };
+}
+/**
+ * 加载现有项目库三件套（不存在 → 空）。
+ * C-2 fail-closed：sqlite 存在但读取失败 → 抛 IngestError（拒绝导入，绝不静默覆盖丢数据）。
+ * 坏的 api-entities.json / doc-graph.json 隔离处理（按空，保留 sqlite 数据）。
+ */
+async function loadExisting(projectKbPath) {
+    const empty = { chunks: [], meta: [], entities: [], nodes: [], edges: [] };
+    const sqlitePath = join(projectKbPath, 'chunks.sqlite');
+    if (!existsSync(sqlitePath))
+        return empty; // 首次导入
+    const chunks = [];
+    const meta = [];
+    try {
+        const { db } = await loadDbFromBytes(readFileSync(sqlitePath));
+        const prov = provenanceSelectFragment(db);
+        const rows = queryRows(db, `SELECT chunks.chunk_id, chunks.doc_id, chunks.content_raw, chunk_meta.doc_title,
+        chunk_meta.source_url, chunk_meta.anchor, chunk_meta.sdk_version, chunk_meta.built_at${prov}
+       FROM chunks JOIN chunk_meta ON chunk_meta.chunk_id = chunks.chunk_id`, []);
+        for (const r of rows) {
+            const s = (v) => (v == null ? '' : String(v));
+            const sn = (v) => (v == null ? null : String(v));
+            chunks.push({ chunkId: s(r[0]), docId: s(r[1]), contentRaw: s(r[2]), anchor: sn(r[5]) });
+            meta.push({
+                chunkId: s(r[0]), docId: s(r[1]), docTitle: s(r[3]), sourceUrl: sn(r[4]), anchor: sn(r[5]),
+                sdkVersion: sn(r[6]), builtAt: s(r[7]),
+                ingestSourceType: sn(r[8]), ingestOrigin: sn(r[9]), ingestedAt: sn(r[10]),
+            });
+        }
+    }
+    catch (e) {
+        // fail-closed：现有 sqlite 损坏不可读 → 拒绝导入，避免 commit 覆盖丢数据（C-2）
+        throw new IngestError(`既有项目库 chunks.sqlite 读取失败，已拒绝导入以防数据丢失（请人工修复后重试）: ${e.message}`);
+    }
+    // entities：坏文件隔离（按空，不阻断、不清空 sqlite）
+    let entities = [];
+    try {
+        const entPath = join(projectKbPath, 'api-entities.json');
+        if (existsSync(entPath)) {
+            const f = deserializeApiEntities(JSON.parse(readFileSync(entPath, 'utf-8')));
+            if (f)
+                entities = f.entities;
+        }
+    }
+    catch {
+        entities = [];
+    }
+    // doc-graph：读取已有 nodes/edges 并保留（W-2，防覆盖丢旧 graph）
+    let nodes = [];
+    let edges = [];
+    try {
+        const graphPath = join(projectKbPath, 'doc-graph.json');
+        if (existsSync(graphPath)) {
+            const g = JSON.parse(readFileSync(graphPath, 'utf-8'));
+            nodes = (g.nodes ?? []).map((n) => ({ id: n.id, title: n.title ?? n.id, lang: n.lang ?? 'en', sourceUrl: n.source_url ?? '' }));
+            edges = Array.isArray(g.edges) ? g.edges : [];
+        }
+    }
+    catch {
+        nodes = [];
+        edges = [];
+    }
+    return { chunks, meta, entities, nodes, edges };
+}
+/** 准备导入（处理所有源 + 合并去重，不落盘）；commitIngest 才写 */
+export async function prepareIngest(sources, projectKbPath, opts = {}) {
+    const builtAt = opts.builtAt ?? new Date().toISOString();
+    const fetchUrl = opts.fetchUrl ?? safeFetchUrl;
+    const results = [];
+    const newDocs = [];
+    for (const source of sources) {
+        try {
+            newDocs.push(await sourceToDoc(source, fetchUrl));
+            results.push({ origin: source.value, type: newDocs[newDocs.length - 1].type, ok: true });
+        }
+        catch (e) {
+            results.push({ origin: source.value, type: null, ok: false, reason: e.message });
+        }
+    }
+    const existing = await loadExisting(projectKbPath);
+    const newDocIds = new Set(newDocs.map((d) => d.doc.id));
+    // W-3 doc-level replace：移除既有同 doc id 的 chunk/meta/entity/node（支持更新 + 防 chunk_id 撞 PRIMARY KEY）
+    const keptChunks = existing.chunks.filter((c) => !newDocIds.has(c.docId));
+    const keptMeta = existing.meta.filter((m) => !newDocIds.has(m.docId));
+    const keptEntities = existing.entities.filter((e) => !newDocIds.has(e.sourceDocId));
+    const keptNodes = existing.nodes.filter((n) => !newDocIds.has(n.id));
+    // 去重计数基准 = 导入前全集（使同内容 re-ingest 计 0，幂等）
+    const originalContent = new Set(existing.chunks.map((c) => c.contentRaw));
+    const originalEntityKeys = new Set(existing.entities.map((e) => e.id));
+    const seenContent = new Set(keptChunks.map((c) => c.contentRaw));
+    const newChunks = [];
+    const newMeta = [];
+    let newChunkCount = 0;
+    const nodes = [...keptNodes];
+    for (const { doc, type, origin } of newDocs) {
+        nodes.push({ id: doc.id, title: doc.title, lang: doc.lang, sourceUrl: doc.sourceUrl });
+        for (const c of splitDocument(doc)) {
+            if (c.contentRaw.trim().length < 20)
+                continue; // 过滤空/超短
+            if (seenContent.has(c.contentRaw))
+                continue; // 内容级 dedup（FR-014，防同内容重复存）
+            seenContent.add(c.contentRaw);
+            newChunks.push(c);
+            newMeta.push({
+                chunkId: c.chunkId, docId: c.docId, docTitle: doc.title, sourceUrl: doc.sourceUrl, anchor: c.anchor,
+                sdkVersion: null, builtAt, ingestSourceType: type, ingestOrigin: origin, ingestedAt: builtAt,
+            });
+            if (!originalContent.has(c.contentRaw))
+                newChunkCount++; // 仅统计导入前不存在的内容
+        }
+    }
+    // 抽实体（仅新内容）→ merge key 覆盖
+    const extractOpts = {};
+    if (opts.noLlm === true)
+        extractOpts.noLlm = true;
+    const docLang = new Map(newDocs.map((d) => [d.doc.id, d.doc.lang]));
+    const extraction = await extractEntities(aggregateSections(newChunks, docLang), extractOpts);
+    const entityByKey = new Map();
+    for (const e of keptEntities)
+        entityByKey.set(e.id, e);
+    let newEntityCount = 0;
+    for (const e of extraction.entities) {
+        if (!originalEntityKeys.has(e.id))
+            newEntityCount++;
+        entityByKey.set(e.id, e);
+    }
+    const allChunks = [...keptChunks, ...newChunks];
+    const allMeta = [...keptMeta, ...newMeta];
+    const allEntities = [...entityByKey.values()];
+    const sqliteBytes = await buildChunksDbBytes(allChunks, allMeta);
+    const entityFile = {
+        schemaVersion: '1.0', builtAt, sdkVersion: null, sourceKind: 'project', entities: allEntities,
+    };
+    const graphJson = JSON.stringify({ schema_version: '1.0', source: 'directory', built_at: builtAt, sdk_version: null, nodes: nodes.map((n) => ({ id: n.id, title: n.title, summary: null, tags: [], lang: n.lang, source_url: n.sourceUrl })), edges: existing.edges }, null, 2);
+    return {
+        newDocs: newDocs.length,
+        newChunks: newChunkCount,
+        newEntities: newEntityCount,
+        totalChunks: allChunks.length,
+        totalEntities: allEntities.length,
+        sources: results,
+        artifacts: { graphJson, sqliteBytes, entitiesJson: JSON.stringify(serializeApiEntities(entityFile), null, 2) },
+    };
+}
+/** 落盘项目库三件套（原子） */
+export function commitIngest(projectKbPath, plan) {
+    writeKbArtifactsAtomic(projectKbPath, [
+        { name: 'doc-graph.json', data: plan.artifacts.graphJson },
+        { name: 'chunks.sqlite', data: plan.artifacts.sqliteBytes },
+        { name: 'api-entities.json', data: plan.artifacts.entitiesJson },
+    ]);
+}
+//# sourceMappingURL=ingest-core.js.map