fix: update inference/agent example scripts to be more robust

examples/agents.test.ts

@@ -13,7 +13,7 @@ describe('RAG Integration Tests', () => {
   it('should create an agent and handle conversation successfully', async () => {
     // Get available models
     const models = await client.models.list();
-    const llmModel = models.find(model => model.model_type === 'llm');
+    const llmModel = models.find(model => model.model_type === 'llm' && !model.identifier.includes('guard') && !model.identifier.includes('405'));
     expect(llmModel).toBeTruthy();
 
     // Create agent with configuration

examples/agents.ts

@@ -7,7 +7,11 @@ const client = new LlamaStackClient({ baseURL: 'http://localhost:8321' });
 
 async function main() {
   const availableModels = (await client.models.list())
-    .filter((model: any) => model.model_type === 'llm')
+    .filter((model: any) => 
+      model.model_type === 'llm' && 
+      !model.identifier.includes('guard') &&
+      !model.identifier.includes('405')
+    )
     .map((model: any) => model.identifier);
 
   if (availableModels.length === 0) {
@@ -31,12 +35,13 @@ async function main() {
     },
     toolgroups: process.env["TAVILY_SEARCH_API_KEY"] ? ['builtin::websearch'] : [],
     tool_choice: 'auto',
-    tool_prompt_format: 'json',
+    tool_prompt_format: 'python_list',
     input_shields: [],
     output_shields: [],
     enable_session_persistence: false,
     max_infer_iters: 10,
   };
+  console.log('Agent Configuration:', JSON.stringify(agentConfig, null, 2));
 
   const agentic_system_create_response = await client.agents.create({agent_config: agentConfig});
   const agent_id = agentic_system_create_response.agent_id;

examples/inference.test.ts

@@ -2,9 +2,20 @@ import LlamaStackClient from 'llama-stack-client';
 
 describe('LlamaStack Client Integration Tests', () => {
   let client: LlamaStackClient;
+  let availableModels: string[];
 
-  beforeAll(() => {
+  beforeAll(async () => {
     client = new LlamaStackClient({ baseURL: 'http://localhost:8321' });
+
+    // Fetch available models once
+    const models = await client.models.list();
+    availableModels = models
+      .filter((model: any) => 
+        model.model_type === 'llm' && 
+        !model.identifier.includes('guard') &&
+        !model.identifier.includes('405')
+      )
+      .map((model: any) => model.identifier);
   });
 
   test('should list available models', async () => {
@@ -14,18 +25,28 @@ describe('LlamaStack Client Integration Tests', () => {
   });
 
   test('should perform non-streaming chat completion', async () => {
+    // Skip test if no models available
+    if (availableModels.length === 0) {
+      console.warn('Skipping test: No available models');
+      return;
+    }
     const chatCompletion = await client.inference.chatCompletion({
       messages: [{ content: 'Hello, how are you?', role: 'user' }],
-      model_id: 'meta-llama/Llama-3.2-3B-Instruct',
+      model_id: availableModels[0] as string,
     });
     expect(chatCompletion).toBeDefined();
     expect(chatCompletion.completion_message).toBeDefined();
   }, 30000);
 
   test('should perform streaming chat completion', async () => {
+    // Skip test if no models available
+    if (availableModels.length === 0) {
+      console.warn('Skipping test: No available models');
+      return;
+    }
     const stream = await client.inference.chatCompletion({
       messages: [{ content: 'Hello, how are you?', role: 'user' }],
-      model_id: 'meta-llama/Llama-3.2-3B-Instruct',
+      model_id: availableModels[0] as string,
       stream: true,
     });
 

examples/inference.ts

@@ -5,20 +5,34 @@ const client = new LlamaStackClient({ baseURL: 'http://localhost:8321' });
 
 async function main() {
   // list models
-  const models = await client.models.list();
-  console.log(models);
+  const availableModels = (await client.models.list())
+    .filter((model: any) => 
+      model.model_type === 'llm' && 
+      !model.identifier.includes('guard') &&
+      !model.identifier.includes('405')
+    )
+    .map((model: any) => model.identifier);
+
+  console.log(availableModels);
+
+  if (availableModels.length === 0) {
+    console.log('No available models. Exiting.');
+    return;
+  }
+  const selectedModel = availableModels[0];
+  console.log(`Using model: ${selectedModel}`);
 
   // non-streaming chat-completion
   const chatCompletion = await client.inference.chatCompletion({
     messages: [{ content: 'Hello, how are you?', role: 'user' }],
-    model_id: 'meta-llama/Llama-3.2-3B-Instruct',
+    model_id: selectedModel,
   });
   console.log(chatCompletion);
 
   // streaming chat-completion
   const stream = await client.inference.chatCompletion({
     messages: [{ content: 'Hello, how are you?', role: 'user' }],
-    model_id: 'meta-llama/Llama-3.2-3B-Instruct',
+    model_id: selectedModel,
     stream: true,
   });
   for await (const chunk of stream) {