Is it possible to create ChatCompletionAgent with data source?

[PanClifton]

hello

i am trying to create a ChatCompletionAgent with data source as follows

    var dataSource =  new AzureSearchChatDataSource
            {
                Endpoint = new Uri(_configuration.Endpoint),
                Authentication = DataSourceAuthentication.FromApiKey(_configuration.ApiKey),
                IndexName = indexName,
             Filter = SearchFilter.Create($"source_id eq 'generalknowledge'")
            };

    _logger.LogDebug("Building agent.");
    return new ChatCompletionAgent()
    {
        Description = "You are a frendly assisant for user",
        Instructions = promptConfigurationQueryResult.Value.SystemMessage,
        Name = "open_ai",
        Kernel = _provideOpenAiKernel.Kernel.,
        Arguments = new KernelArguments(new AzureOpenAIPromptExecutionSettings()
        {
            TopP = chatAgentConfiurationQueryResult.Value.TopP,
            Temperature = chatAgentConfiurationQueryResult.Value.Temperature,
            MaxTokens = chatAgentConfiurationQueryResult.Value.MaxTokens,
            AzureChatDataSource = dataSource,
        }),
    };

when i use method InvokeStreamingAsync(...) on agent i am getting exception

[2024-11-20T10:57:36.626Z] Unhandled excpetion thrown.
[2024-11-20T10:57:36.629Z] Azure.AI.OpenAI: Service request failed.
[2024-11-20T10:57:36.629Z] Status: 400 (Bad Request)

could you kindly advise?

[PanClifton]

i got it working

#pragma warning disable SKEXP0001
#pragma warning disable SKEXP0010
#pragma warning disable SKEXP0110
#pragma warning disable SKEXP0020
#pragma warning disable AOAI001

using Azure.AI.OpenAI.Chat;
using Azure.Search.Documents;
using HeiDi.Sandbox.Utils.Providers;
using Microsoft.SemanticKernel;
using Microsoft.SemanticKernel.Agents;
using Microsoft.SemanticKernel.ChatCompletion;
using Microsoft.SemanticKernel.Connectors.AzureOpenAI;
using System.Text.Json;

var configurationProvider = ConfigurationsProvider.CreateNew();

var aiSearchConfigration = configurationProvider.AiSearchConfiguration;
var openAiConfigration = configurationProvider.OpenAiConfigration;

var datasource = new AzureSearchChatDataSource
{
    Endpoint = new Uri(aiSearchConfigration.Endpoint),
    Authentication = DataSourceAuthentication.FromApiKey(aiSearchConfigration.ApiKey),
    IndexName = "contracts-index",
    Filter = SearchFilter.Create($"source_id eq 'generalknowledge'"),
};

var kernel = Kernel.CreateBuilder()
      .AddAzureOpenAIChatCompletion(
          openAiConfigration.ModelDelpoymentName,
         openAiConfigration.Endpoint,
         openAiConfigration.ApiKey)
      .Build();

var chatHistory = new ChatHistory();
var question = "Could you describe the change request process including all steps?";
//var question = "who are you?";
//var question = "Could you name a document where I can find a test strategy?";
//var question = "Propose the list of quality gates and their objective as defined in the Delivery Rule Book";

Console.WriteLine($"User > {question}");

chatHistory.AddUserMessage(question);

var agent = new ChatCompletionAgent()
{
    Description = "You are a frendly assisant for user",
    Instructions = "",
    Name = "open_ai",
    Kernel = kernel,
    Arguments = new KernelArguments(new AzureOpenAIPromptExecutionSettings()
    {
        TopP = 0.5,
        Temperature = 0.5,
        MaxTokens = 1000,
        AzureChatDataSource = datasource,
    }),
};

var citations = new List<ChatCitation>();

await foreach (var content in agent.InvokeStreamingAsync(chatHistory))
{
    var contentReult = content.Content ?? string.Empty;
    if (!string.IsNullOrEmpty(contentReult))
        Console.WriteLine($"Asistatnt > {contentReult}");

    var update = content.InnerContent as OpenAI.Chat.StreamingChatCompletionUpdate;
    var messageContext = update?.GetMessageContext();

    if (messageContext?.Citations is not null)
    {        
        citations.AddRange(messageContext?.Citations);
    }
}

var options = new JsonSerializerOptions { WriteIndented = true };
var json = JsonSerializer.Serialize(citations.Select(x => new Citation(x.Title, x.FilePath, x.Content)).Distinct(), options);
Console.WriteLine($"Tool > Citations: {Environment.NewLine}{json}");

Console.WriteLine("end of the line");

Console.ReadKey();

public record Citation(string Titile, string FileSource, string Content);

[PanClifton]

update it is failing when i add instruction

var agent = new ChatCompletionAgent()
{
    Description = "You are a frendly assisant for user",
   // failing because of this
    Instructions = "Use in your answers the same language than the user has used in the question, also use the language that the user request",
    Name = "open_ai",
    Kernel = kernel,
    Arguments = new KernelArguments(new AzureOpenAIPromptExecutionSettings()
    {
        TopP = 0.5,
        Temperature = 0.5,
        MaxTokens = 1000,
        AzureChatDataSource = datasource,
    }),
};

[sophialagerkranspandey]

Adding @RogerBarret0 @RogerBarreto and @markwallace-microsoft

[moonbox3]

Tagging @crickman as it involves agents.

[crickman]

@PanClifton - Glad to hear you got this working. My first thought on reading the initial question was that the Kernel required a chat-completion service to be configured. Glad to see you were able to address this.

Regarding the failure with the agent instructions, can you share the exception you are getting. In the meantime, I'll see if I can repro the same failure.

[crickman]

@PanClifton - I am unable to reproduce a failure with the code you shared. Looking forward to more details.