Merge branch 'main' into issues/9711-ollama-jsonelement-fc

microsoft · Nov 21, 2024 · 69d19d9 · 69d19d9
2 parents 8d5db3f + d8acb75
commit 69d19d9
Show file tree

Hide file tree

Showing 67 changed files with 1,819 additions and 606 deletions.
diff --git a/dotnet/Directory.Packages.props b/dotnet/Directory.Packages.props
@@ -16,7 +16,7 @@
     <PackageVersion Include="Microsoft.Extensions.Configuration.Abstractions" Version="8.0.0" />
     <PackageVersion Include="Microsoft.ML.Tokenizers.Data.Cl100kBase" Version="1.0.0" />
     <PackageVersion Include="Microsoft.IdentityModel.JsonWebTokens" Version="6.34.0" />
-    <PackageVersion Include="Microsoft.VisualStudio.Threading" Version="17.11.20" />
+    <PackageVersion Include="Microsoft.VisualStudio.Threading" Version="17.12.19" />
     <PackageVersion Include="MSTest.TestFramework" Version="3.6.1" />
     <PackageVersion Include="Newtonsoft.Json" Version="13.0.3" />
     <PackageVersion Include="Npgsql" Version="8.0.5" />
@@ -62,9 +62,9 @@
     <PackageVersion Include="Microsoft.DeepDev.TokenizerLib" Version="1.3.3" />
     <PackageVersion Include="SharpToken" Version="2.0.3" />
     <!-- Microsoft.Extensions.* -->
-    <PackageVersion Include="Microsoft.Extensions.AI" Version="9.0.0-preview.9.24525.1" />
-    <PackageVersion Include="Microsoft.Extensions.AI.Abstractions" Version="9.0.0-preview.9.24525.1" />
-    <PackageVersion Include="Microsoft.Extensions.AI.AzureAIInference" Version="9.0.0-preview.9.24525.1" />
+    <PackageVersion Include="Microsoft.Extensions.AI" Version="9.0.0-preview.9.24556.5" />
+    <PackageVersion Include="Microsoft.Extensions.AI.Abstractions" Version="9.0.0-preview.9.24556.5" />
+    <PackageVersion Include="Microsoft.Extensions.AI.AzureAIInference" Version="9.0.0-preview.9.24556.5" />
     <PackageVersion Include="Microsoft.Extensions.Configuration" Version="8.0.0" />
     <PackageVersion Include="Microsoft.Extensions.Configuration.Binder" Version="8.0.2" />
     <PackageVersion Include="Microsoft.Extensions.Configuration.EnvironmentVariables" Version="8.0.0" />
@@ -106,8 +106,8 @@
     <PackageVersion Include="Microsoft.Identity.Client.Extensions.Msal" Version="[2.28.0, )" />
     <PackageVersion Include="Microsoft.OpenApi" Version="1.6.22" />
     <PackageVersion Include="Microsoft.OpenApi.Readers" Version="1.6.22" />
-    <PackageVersion Include="Microsoft.OpenApi.ApiManifest" Version="0.5.5-preview" />
-    <PackageVersion Include="Microsoft.Plugins.Manifest" Version="1.0.0-preview3" />
+    <PackageVersion Include="Microsoft.OpenApi.ApiManifest" Version="0.5.6-preview" />
+    <PackageVersion Include="Microsoft.Plugins.Manifest" Version="1.0.0-rc2" />
     <PackageVersion Include="Google.Apis.CustomSearchAPI.v1" Version="[1.60.0.3001, )" />
     <PackageVersion Include="Grpc.Net.Client" Version="2.66.0" />
     <PackageVersion Include="protobuf-net" Version="3.2.45" />
@@ -130,7 +130,7 @@
       <PrivateAssets>all</PrivateAssets>
       <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
     </PackageReference>
-    <PackageVersion Include="Microsoft.VisualStudio.Threading.Analyzers" Version="17.11.20" />
+    <PackageVersion Include="Microsoft.VisualStudio.Threading.Analyzers" Version="17.12.19" />
     <PackageReference Include="Microsoft.VisualStudio.Threading.Analyzers">
       <PrivateAssets>all</PrivateAssets>
       <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>

diff --git a/dotnet/samples/Concepts/Filtering/TelemetryWithFilters.cs b/dotnet/samples/Concepts/Filtering/TelemetryWithFilters.cs
@@ -147,7 +147,17 @@ public async Task OnFunctionInvocationAsync(FunctionInvocationContext context, F
 
                 logger.LogInformation("Function {FunctionName} succeeded.", context.Function.Name);
 
-                await this.LogFunctionResultAsync(context);
+                if (context.IsStreaming)
+                {
+                    // Overriding the result in a streaming scenario enables the filter to stream chunks 
+                    // back to the operation's origin without interrupting the data flow.
+                    var enumerable = context.Result.GetValue<IAsyncEnumerable<StreamingChatMessageContent>>();
+                    context.Result = new FunctionResult(context.Result, ProcessFunctionResultStreamingAsync(enumerable!));
+                }
+                else
+                {
+                    ProcessFunctionResult(context.Result);
+                }
             }
             catch (Exception exception)
             {
@@ -167,34 +177,43 @@ public async Task OnFunctionInvocationAsync(FunctionInvocationContext context, F
             }
         }
 
-        private async Task LogFunctionResultAsync(FunctionInvocationContext context)
+        private void ProcessFunctionResult(FunctionResult functionResult)
+        {
+            string? result = functionResult.GetValue<string>();
+            object? usage = functionResult.Metadata?["Usage"];
+
+            if (!string.IsNullOrWhiteSpace(result))
+            {
+                logger.LogTrace("Function result: {Result}", result);
+            }
+
+            if (logger.IsEnabled(LogLevel.Information) && usage is not null)
+            {
+                logger.LogInformation("Usage: {Usage}", JsonSerializer.Serialize(usage));
+            }
+        }
+
+        private async IAsyncEnumerable<StreamingChatMessageContent> ProcessFunctionResultStreamingAsync(IAsyncEnumerable<StreamingChatMessageContent> data)
         {
-            string? result = null;
             object? usage = null;
 
-            if (context.IsStreaming)
+            var stringBuilder = new StringBuilder();
+
+            await foreach (var item in data)
             {
-                var stringBuilder = new StringBuilder();
+                yield return item;
 
-                await foreach (var item in context.Result.GetValue<IAsyncEnumerable<StreamingChatMessageContent>>()!)
+                if (item.Content is not null)
                 {
-                    if (item.Content is not null)
-                    {
-                        stringBuilder.Append(item.Content);
-                    }
-
-                    usage = item.Metadata?["Usage"];
+                    stringBuilder.Append(item.Content);
                 }
 
-                result = stringBuilder.ToString();
-            }
-            else
-            {
-                result = context.Result.GetValue<string>();
-                usage = context.Result.Metadata?["Usage"];
+                usage = item.Metadata?["Usage"];
             }
 
-            if (result is not null)
+            var result = stringBuilder.ToString();
+
+            if (!string.IsNullOrWhiteSpace(result))
             {
                 logger.LogTrace("Function result: {Result}", result);
             }

diff --git a/dotnet/src/Connectors/Connectors.OpenAI.UnitTests/Core/AutoFunctionInvocationFilterTests.cs b/dotnet/src/Connectors/Connectors.OpenAI.UnitTests/Core/AutoFunctionInvocationFilterTests.cs
@@ -78,6 +78,41 @@ public async Task FiltersAreExecutedCorrectlyAsync()
         Assert.Equal("Test chat response", result.ToString());
     }
 
+    [Fact]
+    public async Task FunctionSequenceIndexIsCorrectForConcurrentCallsAsync()
+    {
+        // Arrange
+        List<int> functionSequenceNumbers = [];
+        List<int> expectedFunctionSequenceNumbers = [0, 1, 0, 1];
+
+        var function1 = KernelFunctionFactory.CreateFromMethod((string parameter) => { return parameter; }, "Function1");
+        var function2 = KernelFunctionFactory.CreateFromMethod((string parameter) => { return parameter; }, "Function2");
+
+        var plugin = KernelPluginFactory.CreateFromFunctions("MyPlugin", [function1, function2]);
+
+        var kernel = this.GetKernelWithFilter(plugin, async (context, next) =>
+        {
+            functionSequenceNumbers.Add(context.FunctionSequenceIndex);
+
+            await next(context);
+        });
+
+        this._messageHandlerStub.ResponsesToReturn = GetFunctionCallingResponses();
+
+        // Act
+        var result = await kernel.InvokePromptAsync("Test prompt", new(new OpenAIPromptExecutionSettings
+        {
+            FunctionChoiceBehavior = FunctionChoiceBehavior.Auto(options: new()
+            {
+                AllowParallelCalls = true,
+                AllowConcurrentInvocation = true
+            })
+        }));
+
+        // Assert
+        Assert.Equal(expectedFunctionSequenceNumbers, functionSequenceNumbers);
+    }
+
     [Fact]
     public async Task FiltersAreExecutedCorrectlyOnStreamingAsync()
     {

diff --git a/dotnet/src/SemanticKernel.Abstractions/AI/ChatCompletion/ChatClientChatCompletionService.cs b/dotnet/src/SemanticKernel.Abstractions/AI/ChatCompletion/ChatClientChatCompletionService.cs
@@ -130,6 +130,11 @@ public async IAsyncEnumerable<StreamingChatMessageContent> GetStreamingChatMessa
                 {
                     options.TopK = topK;
                 }
+                else if (entry.Key.Equals("seed", StringComparison.OrdinalIgnoreCase) &&
+                    TryConvert(entry.Value, out long seed))
+                {
+                    options.Seed = seed;
+                }
                 else if (entry.Key.Equals("max_tokens", StringComparison.OrdinalIgnoreCase) &&
                     TryConvert(entry.Value, out int maxTokens))
                 {

diff --git a/dotnet/src/SemanticKernel.Abstractions/AI/ChatCompletion/ChatCompletionServiceChatClient.cs b/dotnet/src/SemanticKernel.Abstractions/AI/ChatCompletion/ChatCompletionServiceChatClient.cs
@@ -138,6 +138,11 @@ public void Dispose()
             settings.ExtensionData["top_k"] = options.TopK.Value;
         }
 
+        if (options.Seed is not null)
+        {
+            settings.ExtensionData["seed"] = options.Seed.Value;
+        }
+
         if (options.ResponseFormat is not null)
         {
             if (options.ResponseFormat is ChatResponseFormatText)

diff --git a/...manticKernel.Abstractions/Filters/AutoFunctionInvocation/AutoFunctionInvocationContext.cs b/...manticKernel.Abstractions/Filters/AutoFunctionInvocation/AutoFunctionInvocationContext.cs
@@ -1,6 +1,5 @@
 // Copyright (c) Microsoft. All rights reserved.
 
-using System.Diagnostics.CodeAnalysis;
 using System.Threading;
 using Microsoft.SemanticKernel.ChatCompletion;
 
@@ -9,7 +8,6 @@ namespace Microsoft.SemanticKernel;
 /// <summary>
 /// Class with data related to automatic function invocation.
 /// </summary>
-[Experimental("SKEXP0001")]
 public class AutoFunctionInvocationContext
 {
     /// <summary>

diff --git a/...manticKernel.Abstractions/Filters/AutoFunctionInvocation/IAutoFunctionInvocationFilter.cs b/...manticKernel.Abstractions/Filters/AutoFunctionInvocation/IAutoFunctionInvocationFilter.cs
@@ -1,7 +1,6 @@
 // Copyright (c) Microsoft. All rights reserved.
 
 using System;
-using System.Diagnostics.CodeAnalysis;
 using System.Threading.Tasks;
 
 namespace Microsoft.SemanticKernel;
@@ -11,7 +10,6 @@ namespace Microsoft.SemanticKernel;
 /// <summary>
 /// Interface for filtering actions during automatic function invocation.
 /// </summary>
-[Experimental("SKEXP0001")]
 public interface IAutoFunctionInvocationFilter
 {
     /// <summary>

diff --git a/dotnet/src/SemanticKernel.Abstractions/Filters/Function/FunctionInvocationContext.cs b/dotnet/src/SemanticKernel.Abstractions/Filters/Function/FunctionInvocationContext.cs
@@ -1,6 +1,5 @@
 // Copyright (c) Microsoft. All rights reserved.
 
-using System.Diagnostics.CodeAnalysis;
 using System.Threading;
 
 namespace Microsoft.SemanticKernel;
@@ -38,7 +37,6 @@ internal FunctionInvocationContext(Kernel kernel, KernelFunction function, Kerne
     /// <summary>
     /// Boolean flag which indicates whether a filter is invoked within streaming or non-streaming mode.
     /// </summary>
-    [Experimental("SKEXP0001")]
     public bool IsStreaming { get; init; }
 
     /// <summary>

diff --git a/dotnet/src/SemanticKernel.Abstractions/Filters/Prompt/PromptRenderContext.cs b/dotnet/src/SemanticKernel.Abstractions/Filters/Prompt/PromptRenderContext.cs
@@ -1,7 +1,7 @@
 // Copyright (c) Microsoft. All rights reserved.
 
-using System.Diagnostics.CodeAnalysis;
 using System.Threading;
+
 namespace Microsoft.SemanticKernel;
 
 /// <summary>
@@ -37,7 +37,6 @@ internal PromptRenderContext(Kernel kernel, KernelFunction function, KernelArgum
     /// <summary>
     /// Boolean flag which indicates whether a filter is invoked within streaming or non-streaming mode.
     /// </summary>
-    [Experimental("SKEXP0001")]
     public bool IsStreaming { get; init; }
 
     /// <summary>

diff --git a/dotnet/src/SemanticKernel.Abstractions/Kernel.cs b/dotnet/src/SemanticKernel.Abstractions/Kernel.cs
@@ -144,7 +144,6 @@ public Kernel Clone() =>
     /// <summary>
     /// Gets the collection of auto function invocation filters available through the kernel.
     /// </summary>
-    [Experimental("SKEXP0001")]
     public IList<IAutoFunctionInvocationFilter> AutoFunctionInvocationFilters =>
         this._autoFunctionInvocationFilters ??
         Interlocked.CompareExchange(ref this._autoFunctionInvocationFilters, [], null) ??

diff --git a/dotnet/src/SemanticKernel.UnitTests/AI/ServiceConversionExtensionsTests.cs b/dotnet/src/SemanticKernel.UnitTests/AI/ServiceConversionExtensionsTests.cs
@@ -526,7 +526,7 @@ public async Task AsChatCompletionServiceNonStreamingContentConvertedAsExpected(
         Assert.Equal(0.5f, actualOptions.FrequencyPenalty);
         Assert.Equal(0.75f, actualOptions.TopP);
         Assert.Equal(["hello"], actualOptions.StopSequences);
-        Assert.Equal(42, actualOptions.AdditionalProperties?.TryGetValue("Seed", out int seed) is true ? seed : 0);
+        Assert.Equal(42, actualOptions.Seed);
         Assert.Equal("user123", actualOptions.AdditionalProperties?["User"]);
     }
 
@@ -621,7 +621,7 @@ public async Task AsChatCompletionServiceStreamingContentConvertedAsExpected()
         Assert.Equal(0.5f, actualOptions.FrequencyPenalty);
         Assert.Equal(0.75f, actualOptions.TopP);
         Assert.Equal(["hello"], actualOptions.StopSequences);
-        Assert.Equal(42, actualOptions.AdditionalProperties?.TryGetValue("Seed", out int seed) is true ? seed : 0);
+        Assert.Equal(42, actualOptions.Seed);
         Assert.Equal("user123", actualOptions.AdditionalProperties?["User"]);
     }
 

diff --git a/python/samples/concepts/chat_completion/chat_bedrock_api.py b/python/samples/concepts/chat_completion/chat_bedrock_api.py
@@ -3,8 +3,7 @@
 import asyncio
 
 from semantic_kernel import Kernel
-from semantic_kernel.connectors.ai.bedrock.bedrock_prompt_execution_settings import BedrockChatPromptExecutionSettings
-from semantic_kernel.connectors.ai.bedrock.services.bedrock_chat_completion import BedrockChatCompletion
+from semantic_kernel.connectors.ai.bedrock import BedrockChatCompletion, BedrockChatPromptExecutionSettings
 from semantic_kernel.contents import ChatHistory
 
 system_message = """

diff --git a/python/samples/concepts/local_models/lm_studio_chat_completion.py b/python/samples/concepts/local_models/lm_studio_chat_completion.py
@@ -5,7 +5,7 @@
 
 from openai import AsyncOpenAI
 
-from semantic_kernel.connectors.ai.open_ai.services.open_ai_chat_completion import OpenAIChatCompletion
+from semantic_kernel.connectors.ai.open_ai import OpenAIChatCompletion
 from semantic_kernel.contents.chat_history import ChatHistory
 from semantic_kernel.functions.kernel_arguments import KernelArguments
 from semantic_kernel.kernel import Kernel

diff --git a/python/samples/concepts/local_models/lm_studio_text_embedding.py b/python/samples/concepts/local_models/lm_studio_text_embedding.py
@@ -4,7 +4,7 @@
 
 from openai import AsyncOpenAI
 
-from semantic_kernel.connectors.ai.open_ai.services.open_ai_text_embedding import OpenAITextEmbedding
+from semantic_kernel.connectors.ai.open_ai import OpenAITextEmbedding
 from semantic_kernel.core_plugins.text_memory_plugin import TextMemoryPlugin
 from semantic_kernel.kernel import Kernel
 from semantic_kernel.memory.semantic_text_memory import SemanticTextMemory

diff --git a/python/samples/concepts/local_models/ollama_chat_completion.py b/python/samples/concepts/local_models/ollama_chat_completion.py
@@ -5,7 +5,7 @@
 
 from openai import AsyncOpenAI
 
-from semantic_kernel.connectors.ai.open_ai.services.open_ai_chat_completion import OpenAIChatCompletion
+from semantic_kernel.connectors.ai.open_ai import OpenAIChatCompletion
 from semantic_kernel.contents.chat_history import ChatHistory
 from semantic_kernel.functions.kernel_arguments import KernelArguments
 from semantic_kernel.kernel import Kernel

diff --git a/python/samples/concepts/model_as_a_service/mmlu_model_eval.py b/python/samples/concepts/model_as_a_service/mmlu_model_eval.py
@@ -12,9 +12,7 @@
     formatted_question,
     formatted_system_message,
 )
-from semantic_kernel.connectors.ai.azure_ai_inference.services.azure_ai_inference_chat_completion import (
-    AzureAIInferenceChatCompletion,
-)
+from semantic_kernel.connectors.ai.azure_ai_inference import AzureAIInferenceChatCompletion
 from semantic_kernel.connectors.ai.chat_completion_client_base import ChatCompletionClientBase
 from semantic_kernel.contents.chat_history import ChatHistory
 from semantic_kernel.functions.kernel_arguments import KernelArguments