@forbocai/node 0.4.10 → 0.5.1

package/dist/index.d.mts

@@ -96,12 +96,27 @@ declare const generateEmbedding: (text: string) => Promise<number[]>;
  */
 declare const getVectorTable: (dbPath: string, tableName?: string) => Promise<any>;
 /**
- * Creates a new table in LanceDB.
+ * Creates a new LanceDB table with the memory schema.
  * @param dbPath - Path to the LanceDB directory.
  * @param tableName - Name of the table.
- * @param data - Initial data to populate the table.
- * @returns The created table object.
+ * @param initialData - Initial data rows.
+ * @returns The created table.
  */
-declare const createTable: (dbPath: string, tableName: string, data: any[]) => Promise<any>;
+declare const createTable: (dbPath: string, tableName?: string, initialData?: any[]) => Promise<any>;
+/**
+ * Creates a memory item with an embedding.
+ * @param text - The text to embed and store.
+ * @param type - Memory type.
+ * @param importance - Importance score (0-1).
+ * @returns A memory item object with embedding.
+ */
+declare const createMemoryItemWithVector: (text: string, type?: string, importance?: number) => Promise<{
+    id: string;
+    text: string;
+    timestamp: number;
+    type: string;
+    importance: number;
+    vector: number[];
+}>;
 
-export { type IMemory, type MemoryModuleConfig, type VectorStatus, createCortex, createMemory, createMemoryItem, createTable, generateEmbedding, getVectorTable, initVectorEngine };
+export { type IMemory, type MemoryModuleConfig, type VectorStatus, createCortex, createMemory, createMemoryItem, createMemoryItemWithVector, createTable, generateEmbedding, getVectorTable, initVectorEngine };

package/dist/index.d.ts

@@ -96,12 +96,27 @@ declare const generateEmbedding: (text: string) => Promise<number[]>;
  */
 declare const getVectorTable: (dbPath: string, tableName?: string) => Promise<any>;
 /**
- * Creates a new table in LanceDB.
+ * Creates a new LanceDB table with the memory schema.
  * @param dbPath - Path to the LanceDB directory.
  * @param tableName - Name of the table.
- * @param data - Initial data to populate the table.
- * @returns The created table object.
+ * @param initialData - Initial data rows.
+ * @returns The created table.
  */
-declare const createTable: (dbPath: string, tableName: string, data: any[]) => Promise<any>;
+declare const createTable: (dbPath: string, tableName?: string, initialData?: any[]) => Promise<any>;
+/**
+ * Creates a memory item with an embedding.
+ * @param text - The text to embed and store.
+ * @param type - Memory type.
+ * @param importance - Importance score (0-1).
+ * @returns A memory item object with embedding.
+ */
+declare const createMemoryItemWithVector: (text: string, type?: string, importance?: number) => Promise<{
+    id: string;
+    text: string;
+    timestamp: number;
+    type: string;
+    importance: number;
+    vector: number[];
+}>;
 
-export { type IMemory, type MemoryModuleConfig, type VectorStatus, createCortex, createMemory, createMemoryItem, createTable, generateEmbedding, getVectorTable, initVectorEngine };
+export { type IMemory, type MemoryModuleConfig, type VectorStatus, createCortex, createMemory, createMemoryItem, createMemoryItemWithVector, createTable, generateEmbedding, getVectorTable, initVectorEngine };

package/dist/index.js

@@ -33,6 +33,7 @@ __export(index_exports, {
   createCortex: () => createCortex,
   createMemory: () => createMemory,
   createMemoryItem: () => createMemoryItem,
+  createMemoryItemWithVector: () => createMemoryItemWithVector,
   createTable: () => createTable,
   generateEmbedding: () => generateEmbedding,
   getVectorTable: () => getVectorTable,
@@ -41,9 +42,68 @@ __export(index_exports, {
 module.exports = __toCommonJS(index_exports);
 
 // src/cortex.ts
-var fs = __toESM(require("fs"));
-var path = __toESM(require("path"));
+var fs2 = __toESM(require("fs"));
+var path2 = __toESM(require("path"));
 var https = __toESM(require("https"));
+
+// src/vector.ts
+var path = __toESM(require("path"));
+var fs = __toESM(require("fs"));
+var embedder = null;
+var lancedb;
+var initVectorEngine = async () => {
+  if (embedder) return;
+  try {
+    console.log("> Initializing Vector Engine (Transformers)...");
+    const transformers = await import("@huggingface/transformers");
+    transformers.env.cacheDir = path.join(process.env.HOME || ".", ".forbocai", "cache");
+    embedder = await transformers.pipeline("feature-extraction", "Xenova/all-MiniLM-L6-v2");
+    console.log("> Initializing LanceDB...");
+    lancedb = await import("@lancedb/lancedb");
+  } catch (e) {
+    console.error("Failed to init Vector Engine:", e);
+  }
+};
+var generateEmbedding = async (text) => {
+  if (!embedder) await initVectorEngine();
+  try {
+    const output = await embedder(text, { pooling: "mean", normalize: true });
+    return Array.from(output.data);
+  } catch (e) {
+    console.error("Embedding failed:", e);
+    return new Array(384).fill(0);
+  }
+};
+var getVectorTable = async (dbPath, tableName = "memories") => {
+  if (!lancedb) await initVectorEngine();
+  if (!fs.existsSync(dbPath)) fs.mkdirSync(dbPath, { recursive: true });
+  const db = await lancedb.connect(dbPath);
+  const tables = await db.tableNames();
+  if (tables.includes(tableName)) {
+    return await db.openTable(tableName);
+  } else {
+    return null;
+  }
+};
+var createTable = async (dbPath, tableName = "memories", initialData = []) => {
+  if (!lancedb) await initVectorEngine();
+  if (!fs.existsSync(dbPath)) fs.mkdirSync(dbPath, { recursive: true });
+  const db = await lancedb.connect(dbPath);
+  return await db.createTable(tableName, initialData);
+};
+var createMemoryItemWithVector = async (text, type = "observation", importance = 0.5) => {
+  const embedding = await generateEmbedding(text);
+  return {
+    id: `mem_${Date.now()}_${Math.random().toString(36).substring(7)}`,
+    text,
+    timestamp: Date.now(),
+    type,
+    importance,
+    vector: embedding
+  };
+};
+
+// src/cortex.ts
 var isNode = typeof process !== "undefined" && process.versions != null && process.versions.node != null;
 if (!isNode) {
   throw new Error("ForbocAI SDK requires Node.js environment - browser support has been removed");
@@ -55,7 +115,7 @@ var MODEL_URLS = {
 var DEFAULT_MODEL = "smollm2-135m";
 var downloadFile = (url, destPath) => {
   return new Promise((resolve, reject) => {
-    const file = fs.createWriteStream(destPath);
+    const file = fs2.createWriteStream(destPath);
     https.get(url, (response) => {
       if (response.statusCode === 302 || response.statusCode === 301) {
         downloadFile(response.headers.location, destPath).then(resolve).catch(reject);
@@ -71,7 +131,7 @@ var downloadFile = (url, destPath) => {
         resolve();
       });
     }).on("error", (err) => {
-      fs.unlink(destPath, () => {
+      fs2.unlink(destPath, () => {
       });
       reject(err);
     });
@@ -88,24 +148,22 @@ var createNativeCortex = (config) => {
   let model;
   let context;
   let session;
-  let featureExtraction = null;
-  const MODELS_DIR = path.join(process.cwd(), "local_infrastructure", "models");
+  const MODELS_DIR = path2.join(process.cwd(), "local_infrastructure", "models");
   const init = async () => {
     if (status.ready) return status;
     try {
       console.log("> Initializing Native Cortex...");
       console.log("> Loading Embedding Model (all-MiniLM-L6-v2)...");
-      const { pipeline: pipeline2 } = await import("@xenova/transformers");
-      featureExtraction = await pipeline2("feature-extraction", "Xenova/all-MiniLM-L6-v2");
+      await initVectorEngine();
       const ensureDirectoryExists = (dirPath) => {
-        if (!fs.existsSync(dirPath)) fs.mkdirSync(dirPath, { recursive: true });
+        if (!fs2.existsSync(dirPath)) fs2.mkdirSync(dirPath, { recursive: true });
       };
       ensureDirectoryExists(MODELS_DIR);
       const modelKey = config.model || DEFAULT_MODEL;
       const modelUrl = MODEL_URLS[modelKey] || MODEL_URLS[DEFAULT_MODEL];
-      const modelFileName = path.basename(modelUrl);
-      const modelPath = path.join(MODELS_DIR, modelFileName);
-      if (!fs.existsSync(modelPath)) {
+      const modelFileName = path2.basename(modelUrl);
+      const modelPath = path2.join(MODELS_DIR, modelFileName);
+      if (!fs2.existsSync(modelPath)) {
         console.log(`> Downloading SLM: ${modelKey}...`);
         await downloadFile(modelUrl, modelPath);
         console.log(`> Download complete.`);
@@ -183,13 +241,7 @@ var createNativeCortex = (config) => {
   };
   const embed = async (text) => {
     if (!status.ready) await init();
-    try {
-      const output = await featureExtraction(text, { pooling: "mean", normalize: true });
-      return Array.from(output.data);
-    } catch (e) {
-      console.error("Embedding failed:", e);
-      return new Array(384).fill(0);
-    }
+    return generateEmbedding(text);
   };
   return {
     init,
@@ -202,66 +254,6 @@ var createCortex = (config) => createNativeCortex(config);
 
 // src/memory.ts
 var path3 = __toESM(require("path"));
-
-// src/vector.ts
-var path2 = __toESM(require("path"));
-var fs2 = __toESM(require("fs"));
-var pipeline;
-var lancedb;
-var initVectorEngine = async () => {
-  if (pipeline) return;
-  try {
-    console.log("> Initializing Vector Engine (Transformers)...");
-    const transformers = await import("@xenova/transformers");
-    transformers.env.cacheDir = path2.join(process.env.HOME || ".", ".forbocai", "cache");
-    pipeline = await transformers.pipeline("feature-extraction", "Xenova/all-MiniLM-L6-v2");
-    console.log("> Initializing LanceDB...");
-    lancedb = await import("@lancedb/lancedb");
-  } catch (e) {
-    console.error("Failed to init Vector Engine:", e);
-  }
-};
-var generateEmbedding = async (text) => {
-  try {
-    if (!pipeline) await initVectorEngine();
-    const output = await pipeline(text, { pooling: "mean", normalize: true });
-    return Array.from(output.data);
-  } catch (e) {
-    console.warn("Transformers embedding failed, falling back to Cortex:", e);
-    const cortexConfig = {
-      model: "smollm2-135m",
-      temperature: 0.7,
-      maxTokens: 128,
-      gpu: true
-    };
-    const cortex = createCortex(cortexConfig);
-    try {
-      await cortex.init();
-      return await cortex.embed(text);
-    } catch (cortexError) {
-      console.error("All embedding methods failed:", cortexError);
-      return new Array(384).fill(0);
-    }
-  }
-};
-var getVectorTable = async (dbPath, tableName = "memories") => {
-  if (!lancedb) await initVectorEngine();
-  if (!fs2.existsSync(dbPath)) fs2.mkdirSync(dbPath, { recursive: true });
-  const db = await lancedb.connect(dbPath);
-  const tables = await db.tableNames();
-  if (tables.includes(tableName)) {
-    return await db.openTable(tableName);
-  } else {
-    return null;
-  }
-};
-var createTable = async (dbPath, tableName, data) => {
-  if (!lancedb) await initVectorEngine();
-  const db = await lancedb.connect(dbPath);
-  return await db.createTable(tableName, data);
-};
-
-// src/memory.ts
 var createMemoryItem = (text, type = "observation", importance = 0.5) => ({
   id: `mem_${Date.now()}_${Math.random().toString(36).substring(7)}`,
   text,
@@ -373,6 +365,7 @@ var createMemory = (config = {}) => createLanceDBMemory(config);
   createCortex,
   createMemory,
   createMemoryItem,
+  createMemoryItemWithVector,
   createTable,
   generateEmbedding,
   getVectorTable,

package/dist/index.mjs

@@ -1,7 +1,66 @@
 // src/cortex.ts
-import * as fs from "fs";
-import * as path from "path";
+import * as fs2 from "fs";
+import * as path2 from "path";
 import * as https from "https";
+
+// src/vector.ts
+import * as path from "path";
+import * as fs from "fs";
+var embedder = null;
+var lancedb;
+var initVectorEngine = async () => {
+  if (embedder) return;
+  try {
+    console.log("> Initializing Vector Engine (Transformers)...");
+    const transformers = await import("@huggingface/transformers");
+    transformers.env.cacheDir = path.join(process.env.HOME || ".", ".forbocai", "cache");
+    embedder = await transformers.pipeline("feature-extraction", "Xenova/all-MiniLM-L6-v2");
+    console.log("> Initializing LanceDB...");
+    lancedb = await import("@lancedb/lancedb");
+  } catch (e) {
+    console.error("Failed to init Vector Engine:", e);
+  }
+};
+var generateEmbedding = async (text) => {
+  if (!embedder) await initVectorEngine();
+  try {
+    const output = await embedder(text, { pooling: "mean", normalize: true });
+    return Array.from(output.data);
+  } catch (e) {
+    console.error("Embedding failed:", e);
+    return new Array(384).fill(0);
+  }
+};
+var getVectorTable = async (dbPath, tableName = "memories") => {
+  if (!lancedb) await initVectorEngine();
+  if (!fs.existsSync(dbPath)) fs.mkdirSync(dbPath, { recursive: true });
+  const db = await lancedb.connect(dbPath);
+  const tables = await db.tableNames();
+  if (tables.includes(tableName)) {
+    return await db.openTable(tableName);
+  } else {
+    return null;
+  }
+};
+var createTable = async (dbPath, tableName = "memories", initialData = []) => {
+  if (!lancedb) await initVectorEngine();
+  if (!fs.existsSync(dbPath)) fs.mkdirSync(dbPath, { recursive: true });
+  const db = await lancedb.connect(dbPath);
+  return await db.createTable(tableName, initialData);
+};
+var createMemoryItemWithVector = async (text, type = "observation", importance = 0.5) => {
+  const embedding = await generateEmbedding(text);
+  return {
+    id: `mem_${Date.now()}_${Math.random().toString(36).substring(7)}`,
+    text,
+    timestamp: Date.now(),
+    type,
+    importance,
+    vector: embedding
+  };
+};
+
+// src/cortex.ts
 var isNode = typeof process !== "undefined" && process.versions != null && process.versions.node != null;
 if (!isNode) {
   throw new Error("ForbocAI SDK requires Node.js environment - browser support has been removed");
@@ -13,7 +72,7 @@ var MODEL_URLS = {
 var DEFAULT_MODEL = "smollm2-135m";
 var downloadFile = (url, destPath) => {
   return new Promise((resolve, reject) => {
-    const file = fs.createWriteStream(destPath);
+    const file = fs2.createWriteStream(destPath);
     https.get(url, (response) => {
       if (response.statusCode === 302 || response.statusCode === 301) {
         downloadFile(response.headers.location, destPath).then(resolve).catch(reject);
@@ -29,7 +88,7 @@ var downloadFile = (url, destPath) => {
         resolve();
       });
     }).on("error", (err) => {
-      fs.unlink(destPath, () => {
+      fs2.unlink(destPath, () => {
       });
       reject(err);
     });
@@ -46,24 +105,22 @@ var createNativeCortex = (config) => {
   let model;
   let context;
   let session;
-  let featureExtraction = null;
-  const MODELS_DIR = path.join(process.cwd(), "local_infrastructure", "models");
+  const MODELS_DIR = path2.join(process.cwd(), "local_infrastructure", "models");
   const init = async () => {
     if (status.ready) return status;
     try {
       console.log("> Initializing Native Cortex...");
       console.log("> Loading Embedding Model (all-MiniLM-L6-v2)...");
-      const { pipeline: pipeline2 } = await import("@xenova/transformers");
-      featureExtraction = await pipeline2("feature-extraction", "Xenova/all-MiniLM-L6-v2");
+      await initVectorEngine();
       const ensureDirectoryExists = (dirPath) => {
-        if (!fs.existsSync(dirPath)) fs.mkdirSync(dirPath, { recursive: true });
+        if (!fs2.existsSync(dirPath)) fs2.mkdirSync(dirPath, { recursive: true });
       };
       ensureDirectoryExists(MODELS_DIR);
       const modelKey = config.model || DEFAULT_MODEL;
       const modelUrl = MODEL_URLS[modelKey] || MODEL_URLS[DEFAULT_MODEL];
-      const modelFileName = path.basename(modelUrl);
-      const modelPath = path.join(MODELS_DIR, modelFileName);
-      if (!fs.existsSync(modelPath)) {
+      const modelFileName = path2.basename(modelUrl);
+      const modelPath = path2.join(MODELS_DIR, modelFileName);
+      if (!fs2.existsSync(modelPath)) {
         console.log(`> Downloading SLM: ${modelKey}...`);
         await downloadFile(modelUrl, modelPath);
         console.log(`> Download complete.`);
@@ -141,13 +198,7 @@ var createNativeCortex = (config) => {
   };
   const embed = async (text) => {
     if (!status.ready) await init();
-    try {
-      const output = await featureExtraction(text, { pooling: "mean", normalize: true });
-      return Array.from(output.data);
-    } catch (e) {
-      console.error("Embedding failed:", e);
-      return new Array(384).fill(0);
-    }
+    return generateEmbedding(text);
   };
   return {
     init,
@@ -160,66 +211,6 @@ var createCortex = (config) => createNativeCortex(config);
 
 // src/memory.ts
 import * as path3 from "path";
-
-// src/vector.ts
-import * as path2 from "path";
-import * as fs2 from "fs";
-var pipeline;
-var lancedb;
-var initVectorEngine = async () => {
-  if (pipeline) return;
-  try {
-    console.log("> Initializing Vector Engine (Transformers)...");
-    const transformers = await import("@xenova/transformers");
-    transformers.env.cacheDir = path2.join(process.env.HOME || ".", ".forbocai", "cache");
-    pipeline = await transformers.pipeline("feature-extraction", "Xenova/all-MiniLM-L6-v2");
-    console.log("> Initializing LanceDB...");
-    lancedb = await import("@lancedb/lancedb");
-  } catch (e) {
-    console.error("Failed to init Vector Engine:", e);
-  }
-};
-var generateEmbedding = async (text) => {
-  try {
-    if (!pipeline) await initVectorEngine();
-    const output = await pipeline(text, { pooling: "mean", normalize: true });
-    return Array.from(output.data);
-  } catch (e) {
-    console.warn("Transformers embedding failed, falling back to Cortex:", e);
-    const cortexConfig = {
-      model: "smollm2-135m",
-      temperature: 0.7,
-      maxTokens: 128,
-      gpu: true
-    };
-    const cortex = createCortex(cortexConfig);
-    try {
-      await cortex.init();
-      return await cortex.embed(text);
-    } catch (cortexError) {
-      console.error("All embedding methods failed:", cortexError);
-      return new Array(384).fill(0);
-    }
-  }
-};
-var getVectorTable = async (dbPath, tableName = "memories") => {
-  if (!lancedb) await initVectorEngine();
-  if (!fs2.existsSync(dbPath)) fs2.mkdirSync(dbPath, { recursive: true });
-  const db = await lancedb.connect(dbPath);
-  const tables = await db.tableNames();
-  if (tables.includes(tableName)) {
-    return await db.openTable(tableName);
-  } else {
-    return null;
-  }
-};
-var createTable = async (dbPath, tableName, data) => {
-  if (!lancedb) await initVectorEngine();
-  const db = await lancedb.connect(dbPath);
-  return await db.createTable(tableName, data);
-};
-
-// src/memory.ts
 var createMemoryItem = (text, type = "observation", importance = 0.5) => ({
   id: `mem_${Date.now()}_${Math.random().toString(36).substring(7)}`,
   text,
@@ -330,6 +321,7 @@ export {
   createCortex,
   createMemory,
   createMemoryItem,
+  createMemoryItemWithVector,
   createTable,
   generateEmbedding,
   getVectorTable,

package/package.json

@@ -1,23 +1,25 @@
 {
   "name": "@forbocai/node",
-  "version": "0.4.10",
+  "version": "0.5.1",
   "license": "UNLICENSED",
   "description": "Node.js native implementation for ForbocAI SDK",
   "main": "dist/index.js",
   "module": "dist/index.mjs",
   "types": "dist/index.d.ts",
+  "bin": {
+    "forbocai": "dist/cli.js"
+  },
   "scripts": {
-    "build": "tsup src/index.ts --format cjs,esm --dts",
-    "dev": "tsup src/index.ts --watch",
+    "build": "tsup src/index.ts src/cli.ts --format cjs,esm --dts",
+    "dev": "tsup src/index.ts src/cli.ts --watch",
     "test": "vitest"
   },
   "dependencies": {
-    "@forbocai/core": "^0.4.10",
+    "@forbocai/core": "^0.5.1",
     "@lancedb/lancedb": "^0.23.0",
-    "@xenova/transformers": "^2.17.2",
+    "@huggingface/transformers": "^3.0.0",
     "apache-arrow": "^18.1.0",
-    "node-llama-cpp": "^3.15.1",
-    "onnxruntime-node": "^1.23.2"
+    "node-llama-cpp": "^3.15.1"
   },
   "devDependencies": {
     "tsup": "^8.5.1",