@nebutra/knowledge-rag 0.2.0 → 0.2.1

package/dist/embedder.d.ts

@@ -4,7 +4,6 @@ export declare class LocalHashEmbedder implements Embedder {
     private readonly dim;
     constructor(dim?: number);
     embed(texts: string[]): Promise<number[][]>;
-    private embedOne;
 }
 /**
  * Wraps the Sailor provider abstraction (`@nebutra/agents` → Vercel AI SDK

package/dist/embedder.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"embedder.d.ts","sourceRoot":"","sources":["../src/embedder.ts"],"names":[],"mappings":"AAYA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAmBxC,qBAAa,iBAAkB,YAAW,QAAQ;IAChD,QAAQ,CAAC,IAAI,gBAAgB;IAC7B,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAS;gBAEjB,GAAG,SAAM;IAWf,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;IAIjD,OAAO,CAAC,QAAQ;CA0BjB;AAED;;;;GAIG;AACH,qBAAa,gBAAiB,YAAW,QAAQ;IAC/C,QAAQ,CAAC,IAAI,oBAAoB;IACjC,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAqB;gBAE/B,KAAK,CAAC,EAAE,MAAM;IAIpB,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;CA0BlD"}
+{"version":3,"file":"embedder.d.ts","sourceRoot":"","sources":["../src/embedder.ts"],"names":[],"mappings":"AAaA,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAExC,qBAAa,iBAAkB,YAAW,QAAQ;IAChD,QAAQ,CAAC,IAAI,gBAAgB;IAC7B,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAS;gBAEjB,GAAG,SAAM;IAWf,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;CAGlD;AAED;;;;GAIG;AACH,qBAAa,gBAAiB,YAAW,QAAQ;IAC/C,QAAQ,CAAC,IAAI,oBAAoB;IACjC,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAqB;gBAE/B,KAAK,CAAC,EAAE,MAAM;IAIpB,KAAK,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;CA0BlD"}

package/dist/embedder.js

@@ -8,22 +8,8 @@
 // - ProviderEmbedder: wraps @nebutra/agents `embedMany()` (Vercel AI SDK
 //   fallback chain) when real embeddings are configured.
 // =============================================================================
+import { embedTextLocal } from "@nebutra/local-embedding";
 import { KnowledgeRagError } from "./errors";
-/** FNV-1a 32-bit hash — fast, deterministic, well-distributed. */
-function fnv1a(input) {
-    let h = 0x811c9dc5;
-    for (let i = 0; i < input.length; i++) {
-        h ^= input.charCodeAt(i);
-        h = Math.imul(h, 0x01000193);
-    }
-    return h >>> 0;
-}
-function tokenize(text) {
-    return text
-        .toLowerCase()
-        .split(/[^a-z0-9]+/)
-        .filter((t) => t.length > 0);
-}
 export class LocalHashEmbedder {
     name = "local-hash";
     dim;
@@ -38,35 +24,7 @@ export class LocalHashEmbedder {
     }
     // eslint-disable-next-line @typescript-eslint/require-await
     async embed(texts) {
-        return texts.map((t) => this.embedOne(t));
-    }
-    embedOne(text) {
-        const vec = new Array(this.dim).fill(0);
-        const tokens = tokenize(text);
-        for (const tok of tokens) {
-            // Signed feature hashing reduces collision bias.
-            const h = fnv1a(tok);
-            const idx = h % this.dim;
-            const sign = (h & 1) === 0 ? 1 : -1;
-            vec[idx] += sign;
-            // Character bigrams add sub-word locality (typo / morphology robustness).
-            for (let i = 0; i < tok.length - 1; i++) {
-                const bg = fnv1a(`#${tok.slice(i, i + 2)}`);
-                vec[bg % this.dim] += (bg & 1) === 0 ? 0.5 : -0.5;
-            }
-        }
-        let norm = 0;
-        for (const v of vec)
-            norm += v * v;
-        norm = Math.sqrt(norm);
-        if (norm === 0) {
-            // Empty / symbol-only text — deterministic unit vector on axis 0.
-            vec[0] = 1;
-            return vec;
-        }
-        for (let i = 0; i < vec.length; i++)
-            vec[i] /= norm;
-        return vec;
+        return texts.map((text) => embedTextLocal(text, { dimensions: this.dim }));
     }
 }
 /**

package/dist/scoring.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"scoring.d.ts","sourceRoot":"","sources":["../src/scoring.ts"],"names":[],"mappings":"AASA;;;GAGG;AACH,wBAAgB,gBAAgB,CAAC,CAAC,EAAE,SAAS,MAAM,EAAE,EAAE,CAAC,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,CAoBnF;AAED;;;GAGG;AACH,wBAAgB,eAAe,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,EAAE,CAWnE;AAOD;;;GAGG;AACH,wBAAgB,WAAW,CACzB,WAAW,EAAE,MAAM,EACnB,YAAY,EAAE,MAAM,EACpB,YAAY,EAAE,MAAM,GACnB,MAAM,CAGR"}
+{"version":3,"file":"scoring.d.ts","sourceRoot":"","sources":["../src/scoring.ts"],"names":[],"mappings":"AAUA;;;GAGG;AACH,wBAAgB,gBAAgB,CAAC,CAAC,EAAE,SAAS,MAAM,EAAE,EAAE,CAAC,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,CASnF;AAED;;;GAGG;AACH,wBAAgB,eAAe,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,EAAE,CAWnE;AAED;;;GAGG;AACH,wBAAgB,WAAW,CACzB,WAAW,EAAE,MAAM,EACnB,YAAY,EAAE,MAAM,EACpB,YAAY,EAAE,MAAM,GACnB,MAAM,CAGR"}

package/dist/scoring.js

@@ -4,6 +4,7 @@
 // Pure functions: cosine similarity, min-max normalisation, weighted hybrid
 // blend of the vector and keyword retrieval legs.
 // =============================================================================
+import { clamp, cosineSimilarity as primitiveCosineSimilarity } from "@nebutra/ai-primitives";
 import { KnowledgeRagError } from "./errors";
 /**
  * Cosine similarity in [-1, 1]. Returns 0 for a zero-magnitude vector
@@ -16,19 +17,7 @@ export function cosineSimilarity(a, b) {
             suggestion: "Ensure all chunks were embedded with the same embedder/model as the query. Re-ingest after switching embedder.",
         });
     }
-    let dot = 0;
-    let na = 0;
-    let nb = 0;
-    for (let i = 0; i < a.length; i++) {
-        const x = a[i];
-        const y = b[i];
-        dot += x * y;
-        na += x * x;
-        nb += y * y;
-    }
-    if (na === 0 || nb === 0)
-        return 0;
-    return dot / (Math.sqrt(na) * Math.sqrt(nb));
+    return primitiveCosineSimilarity(a, b, { onMismatch: "throw" });
 }
 /**
  * Min-max normalise to [0, 1]. All-equal input → all 1 (avoids divide by
@@ -50,10 +39,6 @@ export function normalizeScores(scores) {
         return scores.map(() => 1);
     return scores.map((s) => (s - min) / range);
 }
-/** Clamp a number into [lo, hi]. */
-function clamp(x, lo, hi) {
-    return Math.min(hi, Math.max(lo, x));
-}
 /**
  * Weighted hybrid blend. `vectorWeight` is clamped to [0, 1]; the keyword
  * leg receives `1 - vectorWeight`.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@nebutra/knowledge-rag",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "private": false,
   "license": "MIT",
   "type": "module",
@@ -16,7 +16,11 @@
     "gaps": [
       "pgvector store is interface-only (Prisma snippet in README, not migrated)",
       "reranker interface ships with identity impl only"
-    ]
+    ],
+    "featureId": "knowledge-rag",
+    "category": "ai",
+    "surface": "semantic-index",
+    "summary": "Knowledge retrieval grammar with zero-config local embedder and pluggable vector/keyword/rerank ports"
   },
   "main": "./dist/index.js",
   "types": "./dist/index.d.ts",
@@ -32,11 +36,13 @@
   ],
   "dependencies": {
     "zod": "^4.3.6",
-    "@nebutra/agents": "1.1.0",
-    "@nebutra/errors": "0.1.0",
-    "@nebutra/logger": "0.1.0",
-    "@nebutra/search": "0.1.0",
-    "@nebutra/tenant-store": "0.2.0"
+    "@nebutra/ai-primitives": "0.1.1",
+    "@nebutra/agents": "1.1.1",
+    "@nebutra/errors": "0.1.1",
+    "@nebutra/local-embedding": "0.1.1",
+    "@nebutra/logger": "0.1.1",
+    "@nebutra/search": "0.1.1",
+    "@nebutra/tenant-store": "0.2.1"
   },
   "devDependencies": {
     "typescript": "^5.9.3",