.Net: Add prompt execution settings to AutoFunctionInvocationContext (#…

…10551)

### Motivation, Context and Description
Today, the prompt execution settings supplied to the chat completion
service are not available in `AutoFunctionInvocationContext`. This PR
adds the new `AutoFunctionInvocationContext.ExecutionSettings` and
assigns it to the settings with which the chat completion service was
invoked.

Closes: #10475

dotnet/src/Connectors/Connectors.OpenAI.UnitTests/Core/AutoFunctionInvocationFilterTests.cs

@@ -661,6 +661,65 @@ public async Task FilterContextHasValidStreamingFlagAsync(bool isStreaming)
         Assert.Equal(isStreaming, actualStreamingFlag);
     }
 
+    [Fact]
+    public async Task PromptExecutionSettingsArePropagatedFromInvokePromptToFilterContextAsync()
+    {
+        // Arrange
+        this._messageHandlerStub.ResponsesToReturn = GetFunctionCallingResponses();
+
+        var plugin = KernelPluginFactory.CreateFromFunctions("MyPlugin", [KernelFunctionFactory.CreateFromMethod(() => { }, "Function1")]);
+
+        AutoFunctionInvocationContext? actualContext = null;
+
+        var kernel = this.GetKernelWithFilter(plugin, (context, next) =>
+        {
+            actualContext = context;
+            return Task.CompletedTask;
+        });
+
+        var expectedExecutionSettings = new OpenAIPromptExecutionSettings
+        {
+            ToolCallBehavior = ToolCallBehavior.AutoInvokeKernelFunctions
+        };
+
+        // Act
+        var result = await kernel.InvokePromptAsync("Test prompt", new(expectedExecutionSettings));
+
+        // Assert
+        Assert.NotNull(actualContext);
+        Assert.Same(expectedExecutionSettings, actualContext!.ExecutionSettings);
+    }
+
+    [Fact]
+    public async Task PromptExecutionSettingsArePropagatedFromInvokePromptStreamingToFilterContextAsync()
+    {
+        // Arrange
+        this._messageHandlerStub.ResponsesToReturn = GetFunctionCallingStreamingResponses();
+
+        var plugin = KernelPluginFactory.CreateFromFunctions("MyPlugin", [KernelFunctionFactory.CreateFromMethod(() => { }, "Function1")]);
+
+        AutoFunctionInvocationContext? actualContext = null;
+
+        var kernel = this.GetKernelWithFilter(plugin, (context, next) =>
+        {
+            actualContext = context;
+            return Task.CompletedTask;
+        });
+
+        var expectedExecutionSettings = new OpenAIPromptExecutionSettings
+        {
+            ToolCallBehavior = ToolCallBehavior.AutoInvokeKernelFunctions
+        };
+
+        // Act
+        await foreach (var item in kernel.InvokePromptStreamingAsync("Test prompt", new(expectedExecutionSettings)))
+        { }
+
+        // Assert
+        Assert.NotNull(actualContext);
+        Assert.Same(expectedExecutionSettings, actualContext!.ExecutionSettings);
+    }
+
     public void Dispose()
     {
         this._httpClient.Dispose();

dotnet/src/Connectors/Connectors.OpenAI/Core/ClientCore.ChatCompletion.cs

@@ -200,6 +200,7 @@ internal async Task<IReadOnlyList<ChatMessageContent>> GetChatMessageContentsAsy
             // In such cases, we'll return the last message in the chat history.
             var lastMessage = await this.FunctionCallsProcessor.ProcessFunctionCallsAsync(
                 chatMessageContent,
+                chatExecutionSettings,
                 chatHistory,
                 requestIndex,
                 (FunctionCallContent content) => IsRequestableTool(chatOptions.Tools, content),
@@ -384,6 +385,7 @@ internal async IAsyncEnumerable<OpenAIStreamingChatMessageContent> GetStreamingC
             // In such cases, we'll return the last message in the chat history.
             var lastMessage = await this.FunctionCallsProcessor.ProcessFunctionCallsAsync(
                 chatMessageContent,
+                chatExecutionSettings,
                 chatHistory,
                 requestIndex,
                 (FunctionCallContent content) => IsRequestableTool(chatOptions.Tools, content),

dotnet/src/InternalUtilities/connectors/AI/FunctionCalling/FunctionCallsProcessor.cs

@@ -119,6 +119,7 @@ public FunctionCallsProcessor(ILogger? logger = null)
     /// Processes AI function calls by iterating over the function calls, invoking them and adding the results to the chat history.
     /// </summary>
     /// <param name="chatMessageContent">The chat message content representing AI model response and containing function calls.</param>
+    /// <param name="executionSettings">The prompt execution settings.</param>
     /// <param name="chatHistory">The chat history to add function invocation results to.</param>
     /// <param name="requestIndex">AI model function(s) call request sequence index.</param>
     /// <param name="checkIfFunctionAdvertised">Callback to check if a function was advertised to AI model or not.</param>
@@ -129,6 +130,7 @@ public FunctionCallsProcessor(ILogger? logger = null)
     /// <returns>Last chat history message if function invocation filter requested processing termination, otherwise null.</returns>
     public async Task<ChatMessageContent?> ProcessFunctionCallsAsync(
         ChatMessageContent chatMessageContent,
+        PromptExecutionSettings? executionSettings,
         ChatHistory chatHistory,
         int requestIndex,
         Func<FunctionCallContent, bool> checkIfFunctionAdvertised,
@@ -177,7 +179,8 @@ public FunctionCallsProcessor(ILogger? logger = null)
                     FunctionCount = functionCalls.Length,
                     CancellationToken = cancellationToken,
                     IsStreaming = isStreaming,
-                    ToolCallId = functionCall.Id
+                    ToolCallId = functionCall.Id,
+                    ExecutionSettings = executionSettings
                 };
 
             s_inflightAutoInvokes.Value++;

dotnet/src/SemanticKernel.Abstractions/Filters/AutoFunctionInvocation/AutoFunctionInvocationContext.cs

@@ -1,5 +1,6 @@
 // Copyright (c) Microsoft. All rights reserved.
 
+using System.Diagnostics.CodeAnalysis;
 using System.Threading;
 using Microsoft.SemanticKernel.ChatCompletion;
 
@@ -79,6 +80,12 @@ public AutoFunctionInvocationContext(
     /// </summary>
     public ChatMessageContent ChatMessageContent { get; }
 
+    /// <summary>
+    /// The execution settings associated with the operation.
+    /// </summary>
+    [Experimental("SKEXP0001")]
+    public PromptExecutionSettings? ExecutionSettings { get; init; }
+
     /// <summary>
     /// Gets the <see cref="Microsoft.SemanticKernel.ChatCompletion.ChatHistory"/> associated with automatic function invocation.
     /// </summary>

dotnet/src/SemanticKernel.Abstractions/Filters/Prompt/PromptRenderContext.cs

@@ -1,5 +1,6 @@
 // Copyright (c) Microsoft. All rights reserved.
 
+using System.Diagnostics.CodeAnalysis;
 using System.Threading;
 
 namespace Microsoft.SemanticKernel;
@@ -54,6 +55,12 @@ internal PromptRenderContext(Kernel kernel, KernelFunction function, KernelArgum
     /// </summary>
     public KernelArguments Arguments { get; }
 
+    /// <summary>
+    /// The execution settings associated with the operation.
+    /// </summary>
+    [Experimental("SKEXP0001")]
+    public PromptExecutionSettings? ExecutionSettings { get; init; }
+
     /// <summary>
     /// Gets or sets the rendered prompt.
     /// </summary>

dotnet/src/SemanticKernel.Abstractions/Kernel.cs

@@ -349,13 +349,15 @@ internal async Task<PromptRenderContext> OnPromptRenderAsync(
         KernelFunction function,
         KernelArguments arguments,
         bool isStreaming,
+        PromptExecutionSettings? executionSettings,
         Func<PromptRenderContext, Task> renderCallback,
         CancellationToken cancellationToken)
     {
         PromptRenderContext context = new(this, function, arguments)
         {
             CancellationToken = cancellationToken,
-            IsStreaming = isStreaming
+            IsStreaming = isStreaming,
+            ExecutionSettings = executionSettings
         };
 
         await InvokeFilterOrPromptRenderAsync(this._promptRenderFilters, renderCallback, context).ConfigureAwait(false);

dotnet/src/SemanticKernel.Core/Functions/KernelFunctionFromPrompt.cs

@@ -500,7 +500,7 @@ private async Task<PromptRenderingResult> RenderPromptAsync(
 
         Verify.NotNull(aiService);
 
-        var renderingContext = await kernel.OnPromptRenderAsync(this, arguments, isStreaming, async (context) =>
+        var renderingContext = await kernel.OnPromptRenderAsync(this, arguments, isStreaming, executionSettings, async (context) =>
         {
             renderedPrompt = await this._promptTemplate.RenderAsync(kernel, context.Arguments, cancellationToken).ConfigureAwait(false);
 

dotnet/src/SemanticKernel.UnitTests/Filters/PromptRenderFilterTests.cs

@@ -321,4 +321,42 @@ public async Task FilterContextHasValidStreamingFlagAsync(bool isStreaming)
         // Assert
         Assert.Equal(isStreaming, actualStreamingFlag);
     }
+
+    [Theory]
+    [InlineData(true)]
+    [InlineData(false)]
+    public async Task PromptExecutionSettingsArePropagatedToFilterContextAsync(bool isStreaming)
+    {
+        // Arrange
+        PromptExecutionSettings? actualExecutionSettings = null;
+
+        var mockTextGeneration = this.GetMockTextGeneration();
+
+        var function = KernelFunctionFactory.CreateFromPrompt("Prompt");
+
+        var kernel = this.GetKernelWithFilters(textGenerationService: mockTextGeneration.Object,
+            onPromptRender: (context, next) =>
+            {
+                actualExecutionSettings = context.ExecutionSettings;
+                return next(context);
+            });
+
+        var expectedExecutionSettings = new PromptExecutionSettings();
+
+        var arguments = new KernelArguments(expectedExecutionSettings);
+
+        // Act
+        if (isStreaming)
+        {
+            await foreach (var item in kernel.InvokeStreamingAsync(function, arguments))
+            { }
+        }
+        else
+        {
+            await kernel.InvokeAsync(function, arguments);
+        }
+
+        // Assert
+        Assert.Same(expectedExecutionSettings, actualExecutionSettings);
+    }
 }