llmp.components.generator.concurrent.AsyncGenerator

class llmp.components.generator.concurrent.AsyncGenerator(job: JobRecord, job_settings: dict = None, num_runs: int = 5, **kwargs)

Generates a job multiple times asynchronous.

generate()

generate an output based on the job + job_setting and input data.

run_engines(async)

run the engines in parallel with identical job setup. This is an async method so it can be used with asyncio.gather()

Initialize the generator with a job and job settings.

Parameters

• job – JobRecord

• job_settings – dict

• num_runs – int

Attributes

verification_type

Return the validation type for the generator.

Methods

__init__(job[, job_settings, num_runs])	Initialize the generator with a job and job settings.
generate(input_data, **kwargs)	Generate an output based on the job + job_setting and input data.
log_generation(input_object, ...)	Log the generation result.
run_engines(input_data, **kwargs)	Run the engines in parallel with identical job setup.

__init__(job: JobRecord, job_settings: dict = None, num_runs: int = 5, **kwargs)

Initialize the generator with a job and job settings.

Parameters

• job – JobRecord

• job_settings – dict

• num_runs – int

generate(input_data: Union[dict, list[dict]], **kwargs) → list[Tuple[dict, dict]]

Generate an output based on the job + job_setting and input data.

log_generation(input_object: dict, generated_object: dict, run_metrics: dict, **kwargs) → Optional[Event]

Log the generation result.

Parameters

• input_object (dict) – the input object

• generated_object (dict) – the generated object

• run_metrics (dict) – the run metrics

• **kwargs – any - passed to JobRecord.log_generation

async run_engines(input_data: Union[dict, list[dict]], **kwargs) → list[Tuple[dict, dict]]

Run the engines in parallel with identical job setup.

Return a list of Tuple[output, run_metrics] for each run.