agent-swarm-kit 1.0.228 → 1.0.230

package/README.md

@@ -346,6 +346,31 @@ See the **[Test Cases](https://github.com/tripolskypetr/agent-swarm-kit/blob/mas
 
 ---
 
+## 🌍 Ecosystem
+
+The `agent-swarm-kit` ecosystem extends beyond the core library, offering complementary tools to enhance your multi-agent AI development experience. One such tool is:
+
+### Agent Tune Dataset Constructor
+
+> **[Explore Agent Tune](https://agent-tune.github.io)**
+
+The **Agent Tune Dataset Constructor** is a React-based tool designed for crafting fine-tuning datasets tailored for AI models, published on GitHub Pages at [agent-tune.github.io](https://agent-tune.github.io). It provides a dynamic, user-friendly interface with list and grid layouts, enabling you to define user inputs, preferred and non-preferred outputs, and multi-turn chat histories—complete with tool definitions and calls. Built with the `react-declarative` library and styled using Material-UI, it’s optimized for creating JSONL files compatible with OpenAI’s fine-tuning API.
+
+#### Key Features
+- **Dynamic Forms**: Build dataset entries with configurable user inputs, outputs, and tools (up to five per entry), featuring autocomplete for tool names and enum values.
+- **Chat History**: Include conversational context with tool calls, supporting up to five messages per entry.
+- **Data Management**: Import/export datasets as JSONL files (`SFT` and `DPO` both), with automatic `tool_call_id` generation, and persist changes to local storage.
+- **Validation**: Ensures tool consistency, message order, and data integrity.
+- **Ease of Use**: Navigate with breadcrumbs, save drafts, and export directly for fine-tuning with commands like `openai api fine_tunes.create`.
+
+#### Use Case
+Perfect for preparing training data to fine-tune agents within `agent-swarm-kit`, Agent Tune lets you define precise behaviors—like how a sales agent responds or a triage agent routes requests—before integrating them into your swarm. Export your dataset and fine-tune your models to enhance performance across your agent network.
+
+#### Get Started
+Visit [agent-tune.github.io](https://agent-tune.github.io) to try it out, or clone the repository to customize it further. Combine it with `agent-swarm-kit` for a seamless workflow from dataset creation to agent deployment.
+
+---
+
 ## 🤝 Contribute
 
 We’d love your input! Fork the repo, submit a PR, or open an issue on **[GitHub](https://github.com/tripolskypetr/agent-swarm-kit)**.

package/build/index.cjs

@@ -8757,16 +8757,34 @@ const CREATE_EMBEDDING_FN = async (item, self) => {
             id: item.id,
         });
     const index = await self.params.createIndex(item);
-    const hash = createSHA256Hash(index);
-    let embeddings = await self.params.readEmbeddingCache(self.params.embedding, hash);
-    if (!embeddings) {
-        embeddings = await self.params.createEmbedding(index, self.params.embedding);
-        await self.params.writeEmbeddingCache(embeddings, self.params.embedding, hash);
+    const result = [];
+    if (typeof index === "string") {
+        const hash = createSHA256Hash(index);
+        let embeddings = await self.params.readEmbeddingCache(self.params.embedding, hash);
+        if (!embeddings) {
+            embeddings = await self.params.createEmbedding(index, self.params.embedding);
+            await self.params.writeEmbeddingCache(embeddings, self.params.embedding, hash);
+        }
+        if (self.params.onCreate) {
+            self.params.onCreate(index, embeddings, self.params.clientId, self.params.embedding);
+        }
+        result.push([embeddings, index]);
     }
-    if (self.params.onCreate) {
-        self.params.onCreate(index, embeddings, self.params.clientId, self.params.embedding);
+    else {
+        for (const value of Object.values(index)) {
+            const hash = createSHA256Hash(value);
+            let embeddings = await self.params.readEmbeddingCache(self.params.embedding, hash);
+            if (!embeddings) {
+                embeddings = await self.params.createEmbedding(value, self.params.embedding);
+                await self.params.writeEmbeddingCache(embeddings, self.params.embedding, hash);
+            }
+            if (self.params.onCreate) {
+                self.params.onCreate(value, embeddings, self.params.clientId, self.params.embedding);
+            }
+            result.push([embeddings, value]);
+        }
     }
-    return [embeddings, index];
+    return result;
 };
 /**
  * Waits for the ClientStorage instance to initialize by loading initial data and creating embeddings.
@@ -8992,12 +9010,15 @@ class ClientStorage {
             this.params.onCreate(search, searchEmbeddings, this.params.clientId, this.params.embedding);
         }
         await Promise.all(Array.from(this._itemMap.values()).map(functoolsKit.execpool(async (item) => {
-            const [targetEmbeddings, index] = await this._createEmbedding(item);
-            const score = await this.params.calculateSimilarity(searchEmbeddings, targetEmbeddings);
-            if (this.params.onCompare) {
-                this.params.onCompare(search, index, score, this.params.clientId, this.params.embedding);
+            let maxScore = Number.NEGATIVE_INFINITY;
+            for (const [targetEmbeddings, index] of await this._createEmbedding(item)) {
+                const score = await this.params.calculateSimilarity(searchEmbeddings, targetEmbeddings);
+                if (this.params.onCompare) {
+                    this.params.onCompare(search, index, score, this.params.clientId, this.params.embedding);
+                }
+                maxScore = Math.max(score, maxScore);
             }
-            indexed.push(item, score);
+            indexed.push(item, maxScore);
         }, {
             delay: 10,
             maxExec: GLOBAL_CONFIG.CC_STORAGE_SEARCH_POOL,

package/build/index.mjs

@@ -8755,16 +8755,34 @@ const CREATE_EMBEDDING_FN = async (item, self) => {
             id: item.id,
         });
     const index = await self.params.createIndex(item);
-    const hash = createSHA256Hash(index);
-    let embeddings = await self.params.readEmbeddingCache(self.params.embedding, hash);
-    if (!embeddings) {
-        embeddings = await self.params.createEmbedding(index, self.params.embedding);
-        await self.params.writeEmbeddingCache(embeddings, self.params.embedding, hash);
+    const result = [];
+    if (typeof index === "string") {
+        const hash = createSHA256Hash(index);
+        let embeddings = await self.params.readEmbeddingCache(self.params.embedding, hash);
+        if (!embeddings) {
+            embeddings = await self.params.createEmbedding(index, self.params.embedding);
+            await self.params.writeEmbeddingCache(embeddings, self.params.embedding, hash);
+        }
+        if (self.params.onCreate) {
+            self.params.onCreate(index, embeddings, self.params.clientId, self.params.embedding);
+        }
+        result.push([embeddings, index]);
     }
-    if (self.params.onCreate) {
-        self.params.onCreate(index, embeddings, self.params.clientId, self.params.embedding);
+    else {
+        for (const value of Object.values(index)) {
+            const hash = createSHA256Hash(value);
+            let embeddings = await self.params.readEmbeddingCache(self.params.embedding, hash);
+            if (!embeddings) {
+                embeddings = await self.params.createEmbedding(value, self.params.embedding);
+                await self.params.writeEmbeddingCache(embeddings, self.params.embedding, hash);
+            }
+            if (self.params.onCreate) {
+                self.params.onCreate(value, embeddings, self.params.clientId, self.params.embedding);
+            }
+            result.push([embeddings, value]);
+        }
     }
-    return [embeddings, index];
+    return result;
 };
 /**
  * Waits for the ClientStorage instance to initialize by loading initial data and creating embeddings.
@@ -8990,12 +9008,15 @@ class ClientStorage {
             this.params.onCreate(search, searchEmbeddings, this.params.clientId, this.params.embedding);
         }
         await Promise.all(Array.from(this._itemMap.values()).map(execpool(async (item) => {
-            const [targetEmbeddings, index] = await this._createEmbedding(item);
-            const score = await this.params.calculateSimilarity(searchEmbeddings, targetEmbeddings);
-            if (this.params.onCompare) {
-                this.params.onCompare(search, index, score, this.params.clientId, this.params.embedding);
+            let maxScore = Number.NEGATIVE_INFINITY;
+            for (const [targetEmbeddings, index] of await this._createEmbedding(item)) {
+                const score = await this.params.calculateSimilarity(searchEmbeddings, targetEmbeddings);
+                if (this.params.onCompare) {
+                    this.params.onCompare(search, index, score, this.params.clientId, this.params.embedding);
+                }
+                maxScore = Math.max(score, maxScore);
             }
-            indexed.push(item, score);
+            indexed.push(item, maxScore);
         }, {
             delay: 10,
             maxExec: GLOBAL_CONFIG.CC_STORAGE_SEARCH_POOL,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agent-swarm-kit",
-  "version": "1.0.228",
+  "version": "1.0.230",
   "description": "A TypeScript library for building orchestrated framework-agnostic multi-agent AI systems",
   "author": {
     "name": "Petr Tripolsky",
@@ -72,6 +72,7 @@
     "rollup-plugin-peer-deps-external": "2.2.4",
     "tslib": "2.7.0",
     "typedoc": "0.27.9",
+    "undici": "6.21.2",
     "worker-testbed": "1.0.10"
   },
   "peerDependencies": {

package/types.d.ts

@@ -220,7 +220,7 @@ interface IStorageSchema<T extends IStorageData = IStorageData> {
      * @param {T} item - The storage item to index.
      * @returns {Promise<string> | string} The index string for the item, synchronously or asynchronously.
      */
-    createIndex(item: T): Promise<string> | string;
+    createIndex(item: T): Promise<string> | string | Record<string, string> | Promise<Record<string, string>>;
     /** The name of the embedding mechanism used for indexing and searching storage data. */
     embedding: EmbeddingName;
     /** The unique name of the storage within the swarm. */
@@ -6006,7 +6006,7 @@ declare class ClientStorage<T extends IStorageData = IStorageData> implements IS
      * @returns {Promise<readonly [Embeddings, string]>} A tuple of embeddings (from Embedding.interface) and index string.
      * @private
      */
-    _createEmbedding: ((item: T) => Promise<readonly [Embeddings, string]>) & functools_kit.IClearableMemoize<string | number> & functools_kit.IControlMemoize<string | number, Promise<readonly [Embeddings, string]>>;
+    _createEmbedding: ((item: T) => Promise<readonly [Embeddings, string][]>) & functools_kit.IClearableMemoize<string | number> & functools_kit.IControlMemoize<string | number, Promise<readonly [Embeddings, string][]>>;
     /**
      * Waits for the initialization of the storage, loading initial data and creating embeddings via WAIT_FOR_INIT_FN.
      * Ensures initialization happens only once using singleshot, supporting StorageConnectionService’s lifecycle.