操作說明：聊天完成代理程式（實驗性）

發行項
11/03/2024

警告

語意核心代理程式架構是實驗性的，仍在開發中，而且可能會變更。

概觀

在此範例中，我們將探索設定外掛程式以存取 GitHub API，並提供聊天完成代理程式的範本化指示，以回答 GitHub 存放庫的相關問題。此方法會逐步細分，以高光撰寫程序的關鍵部分。在工作中，代理程式會在回應中提供檔引文。

串流將用來傳遞代理程序的回應。這會在工作進行時提供即時更新。

快速入門

繼續進行功能程式代碼撰寫之前，請確定您的開發環境已完全設定和設定。

從建立 主控台 項目開始。然後，請包含下列套件參考，以確保所有必要的相依性都可供使用。

若要從命令行新增套件相依性， dotnet 請使用下列命令：

dotnet add package Azure.Identity
dotnet add package Microsoft.Extensions.Configuration
dotnet add package Microsoft.Extensions.Configuration.Binder
dotnet add package Microsoft.Extensions.Configuration.UserSecrets
dotnet add package Microsoft.Extensions.Configuration.EnvironmentVariables
dotnet add package Microsoft.SemanticKernel.Connectors.AzureOpenAI
dotnet add package Microsoft.SemanticKernel.Agents.Core --prerelease

項目檔（.csproj）應包含下列 PackageReference 定義：

  <ItemGroup>
    <PackageReference Include="Azure.Identity" Version="<stable>" />
    <PackageReference Include="Microsoft.Extensions.Configuration" Version="<stable>" />
    <PackageReference Include="Microsoft.Extensions.Configuration.Binder" Version="<stable>" />
    <PackageReference Include="Microsoft.Extensions.Configuration.UserSecrets" Version="<stable>" />
    <PackageReference Include="Microsoft.Extensions.Configuration.EnvironmentVariables" Version="<stable>" />
    <PackageReference Include="Microsoft.SemanticKernel.Agents.Core" Version="<latest>" />
    <PackageReference Include="Microsoft.SemanticKernel.Connectors.AzureOpenAI" Version="<latest>" />
  </ItemGroup>

Agent Framework 是實驗性的，需要警告隱藏。這可能會在中當做項目檔中的屬性來尋址（.csproj）：

  <PropertyGroup>
    <NoWarn>$(NoWarn);CA2007;IDE1006;SKEXP0001;SKEXP0110;OPENAI001</NoWarn>
  </PropertyGroup>

此外，請從語意核心LearnResources項目複製 GitHub 外掛程式和模型（GitHubPlugin.cs和 GitHubModels.cs）。在項目資料夾中新增這些檔案。

首先，建立一個資料夾來保存您的腳本（.py 檔案）和範例資源。在檔案頂端 .py 包含下列匯入：

import asyncio
import os
import sys
from datetime import datetime

from semantic_kernel.agents import ChatCompletionAgent
from semantic_kernel.connectors.ai.function_choice_behavior import FunctionChoiceBehavior
from semantic_kernel.connectors.ai.open_ai import AzureChatCompletion
from semantic_kernel.contents.chat_history import ChatHistory
from semantic_kernel.contents.chat_message_content import ChatMessageContent
from semantic_kernel.contents.utils.author_role import AuthorRole
from semantic_kernel.kernel import Kernel

# Adjust the sys.path so we can use the GitHubPlugin and GitHubSettings classes
# This is so we can run the code from the samples/learn_resources/agent_docs directory
# If you are running code from your own project, you may not need need to do this.
sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))

from plugins.GithubPlugin.github import GitHubPlugin, GitHubSettings  # noqa: E402

此外，從語意核心LearnResources項目複製 GitHub 外掛程式和模型（github.py）。 在項目資料夾中新增這些檔案。

組態

此範例需要組態設定，才能連線到遠端服務。您必須定義 Open AI 或 Azure Open AI 以及 GitHub 的設定。

# Open AI
dotnet user-secrets set "OpenAISettings:ApiKey" "<api-key>"
dotnet user-secrets set "OpenAISettings:ChatModel" "gpt-4o"

# Azure Open AI
dotnet user-secrets set "AzureOpenAISettings:ApiKey" "<api-key>" # Not required if using token-credential
dotnet user-secrets set "AzureOpenAISettings:Endpoint" "<model-endpoint>"
dotnet user-secrets set "AzureOpenAISettings:ChatModelDeployment" "gpt-4o"

# GitHub
dotnet user-secrets set "GitHubSettings:BaseUrl" "https://api.github.com"
dotnet user-secrets set "GitHubSettings:Token" "<personal access token>"

下列類別用於所有 Agent 範例中。請務必將它包含在專案中，以確保適當的功能。這個類別可作為後續範例的基礎元件。

using System.Reflection;
using Microsoft.Extensions.Configuration;

namespace AgentsSample;

public class Settings
{
    private readonly IConfigurationRoot configRoot;

    private AzureOpenAISettings azureOpenAI;
    private OpenAISettings openAI;

    public AzureOpenAISettings AzureOpenAI => this.azureOpenAI ??= this.GetSettings<Settings.AzureOpenAISettings>();
    public OpenAISettings OpenAI => this.openAI ??= this.GetSettings<Settings.OpenAISettings>();

    public class OpenAISettings
    {
        public string ChatModel { get; set; } = string.Empty;
        public string ApiKey { get; set; } = string.Empty;
    }

    public class AzureOpenAISettings
    {
        public string ChatModelDeployment { get; set; } = string.Empty;
        public string Endpoint { get; set; } = string.Empty;
        public string ApiKey { get; set; } = string.Empty;
    }

    public TSettings GetSettings<TSettings>() =>
        this.configRoot.GetRequiredSection(typeof(TSettings).Name).Get<TSettings>()!;

    public Settings()
    {
        this.configRoot =
            new ConfigurationBuilder()
                .AddEnvironmentVariables()
                .AddUserSecrets(Assembly.GetExecutingAssembly(), optional: true)
                .Build();
    }
}

若要開始使用適當的組態來執行範例程序代碼，最快的方式是在專案的根目錄建立 .env 檔案（執行腳本的位置）。

在 .env Azure OpenAI 或 OpenAI 的檔案中設定下列設定：

AZURE_OPENAI_API_KEY="..."
AZURE_OPENAI_ENDPOINT="https://..."
AZURE_OPENAI_CHAT_DEPLOYMENT_NAME="..."
AZURE_OPENAI_API_VERSION="..."

OPENAI_API_KEY="sk-..."
OPENAI_ORG_ID=""
OPENAI_CHAT_MODEL_ID=""

設定之後，個別的 AI 服務類別會挑選必要的變數，並在具現化期間使用這些變數。

撰寫程式碼

這個範例的編碼程式牽涉到：

安裝程式 - 初始化設定和外掛程式。
代理程式 定義 - 使用範本化指示和外掛程式建立 聊天完成代理程式 。
聊天迴圈 - 撰寫可驅動使用者/代理程式互動的迴圈。

完整範例程式代碼會在 Final 區段中提供。如需完整的實作，請參閱該區段。

設定

建立 聊天完成代理程式之前，必須初始化組態設定、外掛程式和核心。

Settings初始化上一節中所參考的類別。

Settings settings = new();

使用外掛程式的設定來初始化外掛程式。

在這裡，會顯示訊息以指出進度。

Console.WriteLine("Initialize plugins...");
GitHubSettings githubSettings = settings.GetSettings<GitHubSettings>();
GitHubPlugin githubPlugin = new(githubSettings);

gh_settings = GitHubSettings(
    token="<PAT value>"
)
kernel.add_plugin(GitHubPlugin(settings=gh_settings), plugin_name="github")

現在，使用 Kernel IChatCompletionService 和先前建立的 GitHubPlugin ，初始化實例。

Console.WriteLine("Creating kernel...");
IKernelBuilder builder = Kernel.CreateBuilder();

builder.AddAzureOpenAIChatCompletion(
    settings.AzureOpenAI.ChatModelDeployment,
    settings.AzureOpenAI.Endpoint,
    new AzureCliCredential());

builder.Plugins.AddFromObject(githubPlugin);

Kernel kernel = builder.Build();

kernel = Kernel()

# Add the AzureChatCompletion AI Service to the Kernel
service_id = "agent"
kernel.add_service(AzureChatCompletion(service_id=service_id))

settings = kernel.get_prompt_execution_settings_from_service_id(service_id=service_id)
# Configure the function choice behavior to auto invoke kernel functions
settings.function_choice_behavior = FunctionChoiceBehavior.Auto()

代理程式定義

最後，我們已準備好使用聊天完成代理程式及其指示、相關聯的核心，以及預設的自變數和執行設定來具現化聊天完成代理程式。在此情況下，我們想要自動執行任何外掛程式函式。

Console.WriteLine("Defining agent...");
ChatCompletionAgent agent =
    new()
    {
        Name = "SampleAssistantAgent",
        Instructions =
            """
            You are an agent designed to query and retrieve information from a single GitHub repository in a read-only manner.
            You are also able to access the profile of the active user.

            Use the current date and time to provide up-to-date details or time-sensitive responses.

            The repository you are querying is a public repository with the following name: {{$repository}}

            The current date and time is: {{$now}}. 
            """,
        Kernel = kernel,
        Arguments =
            new KernelArguments(new AzureOpenAIPromptExecutionSettings() { FunctionChoiceBehavior = FunctionChoiceBehavior.Auto() })
            {
                { "repository", "microsoft/semantic-kernel" }
            }
    };

Console.WriteLine("Ready!");

agent = ChatCompletionAgent(
    service_id="agent",
    kernel=kernel,
    name="SampleAssistantAgent",
    instructions=f"""
        You are an agent designed to query and retrieve information from a single GitHub repository in a read-only 
        manner.
        You are also able to access the profile of the active user.

        Use the current date and time to provide up-to-date details or time-sensitive responses.

        The repository you are querying is a public repository with the following name: microsoft/semantic-kernel

        The current date and time is: {current_time}. 
        """,
    execution_settings=settings,
)

聊天迴圈

最後，我們能夠協調使用者與 代理程式之間的互動。首先，建立 Chat History 對象來維護交談狀態，並建立空迴圈。

ChatHistory history = [];
bool isComplete = false;
do
{
    // processing logic here
} while (!isComplete);

history = ChatHistory()
is_complete: bool = False
while not is_complete:
    # processing logic here

現在讓我們在上一個迴圈中擷取用戶輸入。在此情況下，將會忽略空的輸入，而字詞 EXIT 會發出交談已完成的訊號。有效的輸入將會新增至 聊天記錄 作為 使用者 訊息。

Console.WriteLine();
Console.Write("> ");
string input = Console.ReadLine();
if (string.IsNullOrWhiteSpace(input))
{
    continue;
}
if (input.Trim().Equals("EXIT", StringComparison.OrdinalIgnoreCase))
{
    isComplete = true;
    break;
}

history.Add(new ChatMessageContent(AuthorRole.User, input));

Console.WriteLine();

user_input = input("User:> ")
if not user_input:
    continue

if user_input.lower() == "exit":
    is_complete = True
    break

history.add_message(ChatMessageContent(role=AuthorRole.USER, content=user_input))

若要產生對使用者輸入的 Agent 回應，請使用 Arguments 叫用代理程式，以提供指定目前日期和時間的最終範本參數。

然後，代理程式回應會顯示給使用者。

DateTime now = DateTime.Now;
KernelArguments arguments =
    new()
    {
        { "now", $"{now.ToShortDateString()} {now.ToShortTimeString()}" }
    };
await foreach (ChatMessageContent response in agent.InvokeAsync(history, arguments))
{
    Console.WriteLine($"{response.Content}");
}

即將推出

完成

將所有步驟結合在一起，我們有此範例的最終程序代碼。以下提供完整的實作。

using System;
using System.Threading.Tasks;
using Azure.Identity;
using Microsoft.SemanticKernel;
using Microsoft.SemanticKernel.Agents;
using Microsoft.SemanticKernel.ChatCompletion;
using Microsoft.SemanticKernel.Connectors.AzureOpenAI;
using Plugins;

namespace AgentsSample;

public static class Program
{
    public static async Task Main()
    {
        // Load configuration from environment variables or user secrets.
        Settings settings = new();

        Console.WriteLine("Initialize plugins...");
        GitHubSettings githubSettings = settings.GetSettings<GitHubSettings>();
        GitHubPlugin githubPlugin = new(githubSettings);

        Console.WriteLine("Creating kernel...");
        IKernelBuilder builder = Kernel.CreateBuilder();

        builder.AddAzureOpenAIChatCompletion(
            settings.AzureOpenAI.ChatModelDeployment,
            settings.AzureOpenAI.Endpoint,
            new AzureCliCredential());

        builder.Plugins.AddFromObject(githubPlugin);

        Kernel kernel = builder.Build();

        Console.WriteLine("Defining agent...");
        ChatCompletionAgent agent =
            new()
            {
                Name = "SampleAssistantAgent",
                Instructions =
                        """
                        You are an agent designed to query and retrieve information from a single GitHub repository in a read-only manner.
                        You are also able to access the profile of the active user.

                        Use the current date and time to provide up-to-date details or time-sensitive responses.

                        The repository you are querying is a public repository with the following name: {{$repository}}

                        The current date and time is: {{$now}}. 
                        """,
                Kernel = kernel,
                Arguments =
                    new KernelArguments(new AzureOpenAIPromptExecutionSettings() { FunctionChoiceBehavior = FunctionChoiceBehavior.Auto() })
                    {
                        { "repository", "microsoft/semantic-kernel" }
                    }
            };

        Console.WriteLine("Ready!");

        ChatHistory history = [];
        bool isComplete = false;
        do
        {
            Console.WriteLine();
            Console.Write("> ");
            string input = Console.ReadLine();
            if (string.IsNullOrWhiteSpace(input))
            {
                continue;
            }
            if (input.Trim().Equals("EXIT", StringComparison.OrdinalIgnoreCase))
            {
                isComplete = true;
                break;
            }

            history.Add(new ChatMessageContent(AuthorRole.User, input));

            Console.WriteLine();

            DateTime now = DateTime.Now;
            KernelArguments arguments =
                new()
                {
                    { "now", $"{now.ToShortDateString()} {now.ToShortTimeString()}" }
                };
            await foreach (ChatMessageContent response in agent.InvokeAsync(history, arguments))
            {
                // Display response.
                Console.WriteLine($"{response.Content}");
            }

        } while (!isComplete);
    }
}

import asyncio
import os
import sys
from datetime import datetime

from semantic_kernel.agents import ChatCompletionAgent
from semantic_kernel.connectors.ai.function_choice_behavior import FunctionChoiceBehavior
from semantic_kernel.connectors.ai.open_ai import AzureChatCompletion
from semantic_kernel.contents.chat_history import ChatHistory
from semantic_kernel.contents.chat_message_content import ChatMessageContent
from semantic_kernel.contents.utils.author_role import AuthorRole
from semantic_kernel.kernel import Kernel

# Adjust the sys.path so we can use the GitHubPlugin and GitHubSettings classes
# This is so we can run the code from the samples/learn_resources/agent_docs directory
# If you are running code from your own project, you may not need need to do this.
sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))

from plugins.GithubPlugin.github import GitHubPlugin, GitHubSettings  # noqa: E402

###################################################################
# The following sample demonstrates how to create a simple,       #
# ChatCompletionAgent to use a GitHub plugin to interact          #
# with the GitHub API.                                            #
###################################################################


async def main():
    kernel = Kernel()

    # Add the AzureChatCompletion AI Service to the Kernel
    service_id = "agent"
    kernel.add_service(AzureChatCompletion(service_id=service_id))

    settings = kernel.get_prompt_execution_settings_from_service_id(service_id=service_id)
    # Configure the function choice behavior to auto invoke kernel functions
    settings.function_choice_behavior = FunctionChoiceBehavior.Auto()

    # Set your GitHub Personal Access Token (PAT) value here
    gh_settings = GitHubSettings(token="<PAT value>")
    kernel.add_plugin(plugin=GitHubPlugin(gh_settings), plugin_name="GithubPlugin")

    current_time = datetime.now().isoformat()

    # Create the agent
    agent = ChatCompletionAgent(
        service_id="agent",
        kernel=kernel,
        name="SampleAssistantAgent",
        instructions=f"""
            You are an agent designed to query and retrieve information from a single GitHub repository in a read-only 
            manner.
            You are also able to access the profile of the active user.

            Use the current date and time to provide up-to-date details or time-sensitive responses.

            The repository you are querying is a public repository with the following name: microsoft/semantic-kernel

            The current date and time is: {current_time}. 
            """,
        execution_settings=settings,
    )

    history = ChatHistory()
    is_complete: bool = False
    while not is_complete:
        user_input = input("User:> ")
        if not user_input:
            continue

        if user_input.lower() == "exit":
            is_complete = True
            break

        history.add_message(ChatMessageContent(role=AuthorRole.USER, content=user_input))

        async for response in agent.invoke(history=history):
            print(f"{response.content}")


if __name__ == "__main__":
    asyncio.run(main())

操作說明： 開啟 AI 助理代理程式 代碼解釋器

共用方式為

操作說明：聊天完成代理程式（實驗性）

概觀

快速入門

組態

撰寫程式碼

設定

代理程式定義

聊天迴圈

完成

其他資源

共用方式為

操作說明： 聊天完成代理程式 （實驗性）

概觀

快速入門

組態

撰寫程式碼

設定

代理程式定義

聊天迴圈

完成

其他資源

操作說明：聊天完成代理程式（實驗性）