updated iOS Swift SDK for LS0.1.3 (#20)

* updated iOS Swift SDK for LS0.1.3

* README update

* LocalAgents.swift update

Author: jeffxtang

README.md

@@ -4,7 +4,7 @@
 
 llama-stack-client-swift brings the inference and agents APIs of [Llama Stack](https://github.com/meta-llama/llama-stack) to iOS.
 
-**Update: January 27, 2025** The llama-stack-client-swift SDK version has been updated to 0.1.0, working with Llama Stack 0.1.0 ([release note](https://github.com/meta-llama/llama-stack/releases/tag/v0.1.0)).
+**Update: February 18, 2025** The llama-stack-client-swift SDK version has been updated to 0.1.3, working with Llama Stack 0.1.3 ([release note](https://github.com/meta-llama/llama-stack/releases/tag/v0.1.3)).
 
 ## Features
 
@@ -21,76 +21,81 @@ For a more advanced demo using the Llama Stack Agent API and custom tool calling
 
 1. Click "Xcode > File > Add Package Dependencies...".
 
-2. Add this repo URL at the top right: `https://github.com/meta-llama/llama-stack-client-swift` and 0.1.0 in the Dependency Rule, then click Add Package.
+2. Add this repo URL at the top right: `https://github.com/meta-llama/llama-stack-client-swift` and 0.1.3 in the Dependency Rule, then click Add Package.
 
 3. Select and add `llama-stack-client-swift` to your app target.
 
 4. On the first build: Enable & Trust the OpenAPIGenerator extension when prompted.
 
-5. Set up a remote Llama Stack distributions, assuming you have a [Fireworks](https://fireworks.ai/account/api-keys) or [Together](https://api.together.ai/) API key, which you can get easily by clicking the link:
+5. The quickest way to try out the demo for remote inference is using Together.ai's Llama Stack distro at https://llama-stack.together.ai - you can skip Step 6 unless you want to build your own distro. 
+
+6. (Optional) Set up a remote Llama Stack distributions, assuming you have a [Fireworks](https://fireworks.ai/account/api-keys) or [Together](https://api.together.ai/) API key, which you can get easily by clicking the link:
 
 ```
 conda create -n llama-stack python=3.10
 conda activate llama-stack
-pip install --no-cache llama-stack==0.1.0 llama-models==0.1.0 llama-stack-client==0.1.0
+pip install --no-cache llama-stack==0.1.3 llama-models==0.1.3 llama-stack-client==0.1.3
 ```
 
 Then, either:
 ```
-PYPI_VERSION=0.1.0 llama stack build --template fireworks --image-type conda
+PYPI_VERSION=0.1.3 llama stack build --template fireworks --image-type conda
 export FIREWORKS_API_KEY="<your_fireworks_api_key>"
 llama stack run fireworks
 ```
 or
 ```
-PYPI_VERSION=0.1.0 llama stack build --template together --image-type conda
+PYPI_VERSION=0.1.3 llama stack build --template together --image-type conda
 export TOGETHER_API_KEY="<your_together_api_key>"
 llama stack run together
 ```
 
 The default port is 5000 for `llama stack run` and you can specify a different port by adding `--port <your_port>` to the end of `llama stack run fireworks|together`.
 
-6. Replace the `RemoteInference` url string below with the host IP and port of the remote Llama Stack distro in Step 5:
+Replace the `RemoteInference` url string below with the host IP and port of the remote Llama Stack distro in Step 6:
 
 ```swift
 import LlamaStackClient
 
-let inference = RemoteInference(url: URL(string: "http://127.0.0.1:5000")!)
+let inference = RemoteInference(url: URL(string: "https://llama-stack.together.ai")!)
 ```
+
+7. Build and Run the iOS demo.
+
 Below is an example code snippet to use the Llama Stack inference API. See the iOS Demos above for complete code.
 
 ```swift
 for await chunk in try await inference.chatCompletion(
     request:
         Components.Schemas.ChatCompletionRequest(
+        model_id: "meta-llama/Llama-3.1-8B-Instruct",
         messages: [
             .user(
             Components.Schemas.UserMessage(
+                role: .user,
                 content:
                     .InterleavedContentItem(
                         .text(Components.Schemas.TextContentItem(
-                            text: userInput,
-                            _type: .text
+                            _type: .text,
+                            text: userInput
                         )
                     )
-                ),
-                role: .user
+                )
             )
         )
         ],
-        model_id: "meta-llama/Llama-3.1-8B-Instruct",
         stream: true)
     ) {
         switch (chunk.event.delta) {
-            case .text(let s):
-                message += s.text
-                break
-            case .image(let s):
-                print("> \(s)")
-                break
-            case .tool_call(let s):
-                print("> \(s)")
-                break
+        case .text(let s):
+            message += s.text
+            break
+        case .image(let s):
+            print("> \(s)")
+            break
+        case .tool_call(let s):
+            print("> \(s)")
+            break
         }
     }
 ```

Sources/LlamaStackClient/Agents/ChatAgent.swift

@@ -22,8 +22,8 @@ class ChatAgent {
     let session = Components.Schemas.Session(
       session_id: sessionId,
       session_name: name,
-      started_at: Date(),
-      turns: []
+      turns: [],
+      started_at: Date()
     )
 
     sessions[sessionId] = session
@@ -82,13 +82,12 @@ class ChatAgent {
           continuation.yield(chunk)
         }
         let turn = Components.Schemas.Turn(
-          input_messages: request.messages.map { $0.toAgenticSystemTurnCreateRequest() },
-          output_attachments: [],
-          output_message: outputMessage!,
+          turn_id: turnId,
           session_id: session_id,
-          started_at: Date(),
+          input_messages: request.messages.map { $0.toAgenticSystemTurnCreateRequest() },
           steps: steps,
-          turn_id: turnId
+          output_message: outputMessage!,
+          started_at: Date()
         )
         await MainActor.run {
           var s = self.sessions[session_id]
@@ -122,10 +121,10 @@ class ChatAgent {
         do {
           for try await chunk: Components.Schemas.ChatCompletionResponseStreamChunk in try await inferenceApi.chatCompletion(
             request: Components.Schemas.ChatCompletionRequest(
-              messages: inputMessages,
               model_id: agentConfig.model,
-              stream: true,
-              tools: [] //agentConfig.client_tools
+              messages: inputMessages,
+              tools: [], //agentConfig.client_tools
+              stream: true
             )
           ) {
               continuation.yield(
@@ -134,10 +133,10 @@ class ChatAgent {
                     payload:
                         .step_progress(
                           Components.Schemas.AgentTurnResponseStepProgressPayload(
-                            delta: chunk.event.delta,
                             event_type: .step_progress,
+                            step_type: .inference,
                             step_id: UUID().uuidString,
-                            step_type: .inference
+                            delta: chunk.event.delta
                           )
                         )
                   )

Sources/LlamaStackClient/Agents/CustomTools.swift

@@ -6,54 +6,53 @@ public class CustomTools {
   // for chat completion (inference) tool calling
   public class func getCreateEventTool() -> Components.Schemas.ToolDefinition {
     return Components.Schemas.ToolDefinition(
+      tool_name: Components.Schemas.ToolDefinition.tool_namePayload.case2( "create_event"),
       description: "Create a calendar event",
       parameters: Components.Schemas.ToolDefinition.parametersPayload(
         additionalProperties: [
           "event_name": Components.Schemas.ToolParamDefinition(
-            description: "The name of the meeting",
             param_type: "string",
+            description: "The name of the meeting",
             required: true
           ),
           "start": Components.Schemas.ToolParamDefinition(
-            description: "Start date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 13:00'",
             param_type: "string",
+            description: "Start date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 13:00'",
             required: true
           ),
           "end": Components.Schemas.ToolParamDefinition(
-            description: "End date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 14:00'",
             param_type: "string",
+            description: "End date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 14:00'",
             required: true
-          ),
+          )
         ]
-      ),
-      tool_name: Components.Schemas.ToolDefinition.tool_namePayload.case2( "create_event")
-
+      )
     )
   }
 
   // for agent tool calling
   public class func getCreateEventToolForAgent() -> Components.Schemas.ToolDef {
     return Components.Schemas.ToolDef(
-      description: "Create a calendar event",
-      metadata: nil,
       name: "create_event",
+      description: "Create a calendar event",
       parameters: [
         Components.Schemas.ToolParameter(
-            description: "The name of the meeting",
             name: "event_name",
             parameter_type: "string",
+            description: "The name of the meeting",
             required: true),
         Components.Schemas.ToolParameter(
-            description: "Start date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 13:00'",
             name: "start",
             parameter_type: "string",
+            description: "Start date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 13:00'",
             required: true),
         Components.Schemas.ToolParameter(
-            description: "End date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 14:00'",
             name: "end",
             parameter_type: "string",
+            description: "End date in yyyy-MM-dd HH:mm format, eg. '2024-01-01 14:00'",
             required: true)
-      ]
+      ],
+      metadata: nil
     )
   }
 }

Sources/LlamaStackClient/Agents/LocalAgents.swift

@@ -17,17 +17,14 @@ public class LocalAgents: Agents {
     let createSystemResponse = try await create(
       request: Components.Schemas.CreateAgentRequest(
         agent_config: Components.Schemas.AgentConfig(
-          enable_session_persistence: false,
+          sampling_params: nil,
           input_shields: [],
-          instructions: "You are a helpful assistant",
-          max_infer_iters: 1,
+          output_shields: [],
+          client_tools: [ CustomTools.getCreateEventToolForAgent() ],
+          max_infer_iters: 1, 
           model: "Meta-Llama3.1-8B-Instruct",
-          output_shields: []
-//          tools: [
-//            Components.Schemas.AgentConfig.toolsPayloadPayload.FunctionCallToolDefinition(
-//              CustomTools.getCreateEventTool()
-//              )
-//          ]
+          instructions: "You are a helpful assistant",
+          enable_session_persistence: false
         )
       )
     )

Sources/LlamaStackClient/Agents/RemoteAgents.swift

@@ -24,13 +24,13 @@ public class RemoteAgents: Agents {
     let createSystemResponse = try await create(
       request: Components.Schemas.CreateAgentRequest(
         agent_config: Components.Schemas.AgentConfig(
-          client_tools: [ CustomTools.getCreateEventToolForAgent() ],
-          enable_session_persistence: false,
           input_shields: ["llama_guard"],
-          instructions: "You are a helpful assistant",
+          output_shields: ["llama_guard"],
+          client_tools: [ CustomTools.getCreateEventToolForAgent() ],
           max_infer_iters: 1,
           model: "Meta-Llama3.1-8B-Instruct",
-          output_shields: ["llama_guard"]
+          instructions: "You are a helpful assistant",
+          enable_session_persistence: false
         )
       )
     )