trinity.explorer
Submodules
trinity.explorer.explorer module
The explorer module
trinity.explorer.runner_pool module
Runner pool for running tasks in parallel. Modified from ray.util.actor_pool.ActorPool.
- class trinity.explorer.runner_pool.RunnerPool(config: Config, models: List[InferenceModel], auxiliary_models: List[List[InferenceModel]] | None = None)[source]
Bases:
object
A pool of WorkflowRunner.
The RunnerPool will automatically handle the exceptions during the workflow and retry when the workflow fails or timeout. The number of max retries is set in config.explorer.max_retry_times and the max timeout is set in config.explorer.max_timeout.
- __init__(config: Config, models: List[InferenceModel], auxiliary_models: List[List[InferenceModel]] | None = None)[source]
- run_tasks(tasks: List[Task] | Task) None [source]
Schedule a list of tasks to run in the pool.
- Parameters:
tasks – A list of tasks.
- has_next()[source]
Returns whether there are any pending results to return.
- Returns:
True if there are any pending results not yet returned.
- get_next_unorder() List[Status] [source]
Returns the next pending result unorder.
- Returns:
The return status of the next task.
- get_next() Status [source]
Returns the next pending result in order.
This returns the next task result, blocking for up to the specified timeout until it is available.
- Returns:
The return status of the next task.
trinity.explorer.workflow_runner module
The Workflow Runner Moudle.
- class trinity.explorer.workflow_runner.Status(ok: bool, metric: dict[str, float], message: str | None = None)[source]
Bases:
object
Status of the task running result.
- ok: bool
- metric: dict[str, float]
- message: str | None = None
- __init__(ok: bool, metric: dict[str, float], message: str | None = None) None
Module contents
- class trinity.explorer.RunnerPool(config: Config, models: List[InferenceModel], auxiliary_models: List[List[InferenceModel]] | None = None)[source]
Bases:
object
A pool of WorkflowRunner.
The RunnerPool will automatically handle the exceptions during the workflow and retry when the workflow fails or timeout. The number of max retries is set in config.explorer.max_retry_times and the max timeout is set in config.explorer.max_timeout.
- __init__(config: Config, models: List[InferenceModel], auxiliary_models: List[List[InferenceModel]] | None = None)[source]
- run_tasks(tasks: List[Task] | Task) None [source]
Schedule a list of tasks to run in the pool.
- Parameters:
tasks – A list of tasks.
- has_next()[source]
Returns whether there are any pending results to return.
- Returns:
True if there are any pending results not yet returned.
- get_next_unorder() List[Status] [source]
Returns the next pending result unorder.
- Returns:
The return status of the next task.
- get_next() Status [source]
Returns the next pending result in order.
This returns the next task result, blocking for up to the specified timeout until it is available.
- Returns:
The return status of the next task.