AIModelGenerator

class src.ai_graph.ai.ai_model_generator.AIModelGenerator(assistant_name: str, client: OpenAiClient, open_ai_model_version: OpenAIModelVersion, temperature: float, max_tokens: int, instructions: str, input_describer: ModelDescriber, retry_wait_min: int = 4, retry_wait_max: int = 128, retry_attempts: int = 20)

Bases: BaseAIModel, Generic

__init__(assistant_name: str, client: OpenAiClient, open_ai_model_version: OpenAIModelVersion, temperature: float, max_tokens: int, instructions: str, input_describer: ModelDescriber, retry_wait_min: int = 4, retry_wait_max: int = 128, retry_attempts: int = 20)

AIModelGenerator is responsible for generating AI models based on a given configuration.

Parameters:

• assistant_name – The name of the assistant using this model.

• client – The OpenAI client used for generating completions.

• open_ai_model_version – The version of the OpenAI model being used.

• temperature – The temperature setting for the model (affects randomness).

• max_tokens – The maximum number of tokens to generate.

• instructions – Instructions for the AI model.

• input_describer – A describer for generating the model’s input prompts.

• retry_wait_min – Minimum wait time between retries in case of failure.

• retry_wait_max – Maximum wait time between retries in case of failure.

• retry_attempts – Number of retry attempts in case of failure.

async get_parsed_completion(input_instance: BaseModel, output_type: type[BaseModel], *args, **kwargs) → OM

Returns the parsed completion result based on the input instance.

Parameters:

• input_instance – The input data to generate the completion.

• output_type – The expected output type for the completion.

Returns:

The generated output object.