yandex_gpt_rest_api 0.3.1

Communicating with the Yandex Foundation Models API.

yandex_gpt_rest_api

Library for communicating with the Yandex Foundation Models API.

UNDER WORK

Receiving data from the API is now working.

Asynchronous generation does not work.

The signature of the methods may change.

Getting started

Create YandexGptApi instance.

final api = YandexGptApiClient(
  token: "your_token",
  catalog: "your_catalog_id", // Not necessary
);

Now you can use the Foundation Models API.

API capabilities

Text Generation

Generate text

void main() async {
  final response = await api.generateText(
    TextGenerationRequest(
      modelUri: GModel.yandexGpt('folder_id'),
      messages: const [
        Message.system("Some joke"),
        Message.user("Generate joke"),
      ],
    ),
  );
  print(response.alternatives.first.message);
  print(response.usage.totalTokens);
}

Generate async text

UNDER_WORK

void main() async {
  final response = await api.generateAsyncText(
    TextGenerationRequest(
      model: GModel.yandexGpt('folder_id'),
      messages: const [
        Message.system("Some joke"),
        Message.user("Generate joke"),
      ],
    ),
  );
  print(response.done);
}

Tokenize

Tokenize completion

void main() async {
  final response = await api.tokenizeCompletion(
    TextGenerationRequest(
      model: GModel.yandexGpt('folder_id'),
      messages: const [
        Message.system("Some joke"),
        Message.user("Generate joke"),
      ],
    ),
  );
  print(response.tokens.length);
}

Tokenize text

void main() async {
  final response = await api.tokenizeText(
    TokenizeTextRequest(
      model: GModel.yandexGpt('folder_id'),
      text: 'some_response_text',
    ),
  );
  print(response.tokens.length);
}

Embeddings

void main() async {
  final response = await api.getTextEmbedding(
    EmbeddingRequest(
      model: VModel.documentation('folder_id'),
      text: 'Some text',
    ),
  );
  print(response.embedding);
}