Support for Gemini Embedding model

[doggy8088]

Context / Scenario

Google's Gemini has a models/embedding-001 Embedding model for embedding. Hope Kernel Memory can support it. Is it on the roadmap?

The problem

N/A

Proposed solution

N/A

Importance

would be great to have

[dluc]

hi @doggy8088 you should be able to reuse SK Gemini connector, via WithSemanticKernelTextEmbeddingGenerationService

[SymoHTL]

like this?

setup.WithSemanticKernelTextEmbeddingGenerationService(
                    new GoogleAITextEmbeddingGenerationService(googleConfig.ModelId, googleConfig.ApiKey),
                    semanticKernelConfig);

if i want to use Gemini for Text Generation too, do i have to write an adapter that just calls the GoogleAIGeminiChatCompletionService methods?
like this?

 public async Task<IReadOnlyList<TextContent>> GetTextContentsAsync(string prompt,
        PromptExecutionSettings? executionSettings = null, Kernel? kernel = null,
        CancellationToken cancellationToken = new CancellationToken()) {
        var chatHistory = new ChatHistory();
        chatHistory.AddUserMessage(prompt);
        var chatMessageContents =
            await _chatCompletionService.GetChatMessageContentsAsync(chatHistory, executionSettings, kernel,
                cancellationToken);
        var textContents = new List<TextContent>();
        foreach (var chatMessageContent in chatMessageContents) {
            var textContent = new TextContent(chatMessageContent.Content);
            textContents.Add(textContent);
        }
        return textContents;
    }