client.task

Task object defines a single executable object that will be added to a Workflow.

TaskInstances will be created from it for every execution.

Attributes

logger

Classes

Task	Task object defines a single executable object that will be added to a Workflow.

Functions

validate_task_resource_scales(→ None)

Validate resource scales are expected types.

Module Contents

client.task.logger

client.task.validate_task_resource_scales(resource_scales: Dict[str, Any]) → None

Validate resource scales are expected types.

class client.task.Task(node: jobmon.client.node.Node, task_args: Dict[str, Any], op_args: Dict[str, Any], array: jobmon.client.array.Array | None = None, cluster_name: str = '', compute_resources: Dict[str, Any] | None = None, compute_resources_callable: Callable | None = None, resource_scales: Dict[str, float] | None = None, fallback_queues: List[str] | None = None, name: str | None = None, max_attempts: int | None = None, upstream_tasks: List[Task] | None = None, task_attributes: List | dict | None = None, requester: jobmon.core.requester.Requester | None = None)

Task object defines a single executable object that will be added to a Workflow.

Task Instances will be created from it for every execution.

static is_valid_job_name(name: str) → bool

If the name is invalid it will raises an exception.

Primarily based on the restrictions SGE places on job names. The list of illegal characters might not be complete, I could not find an official list.

Must:

• Not be null or the empty string

• being with a digit

• contain am illegal character

Parameters:

name

Returns:

True (or raises)

Raises:

ValueError – if the name is not valid.

requester

node

task_args

mapped_task_args

task_args_hash

op_args

command

name

upstream_tasks: Set[Task]

downstream_tasks: Set[Task]

task_attributes: dict

_instance_max_attempts

_instance_cluster_name

_instance_compute_resources

_instance_compute_resources_callable

_instance_resource_scales

fallback_queues: List[str]

_errors: None | Dict[str, int | List[Dict[str, str | int]]] = None

property compute_resources: Dict[str, Any]

property requested_resources: Dict[str, Any]

A dictionary that includes the users requested resources for the current run.

E.g. {cores: 1, mem: 1, runtime: 60, queue: all.q}.

property resource_scales: Dict[str, float]

A dictionary that includes the users requested resource scales for the current run.

E.g. {memory: 0.1, runtime: 0.7}.

property cluster_name: str

The name of the cluster the user wants to run their task on.

property max_attempts: int

Get the max_attempts.

property compute_resources_callable: Callable | None

A callable that returns a compute resources dict.

property queue_name: str

property original_task_resources: jobmon.client.task_resources.TaskResources

Get the id of the task if it has been bound to the db otherwise raise an error.

property is_bound: bool

If the task template version has been bound to the database.

property task_id: int

Get the id of the task if it has been bound to the db otherwise raise an error.

property initial_status: str

Get initial status of the task if it has been bound to the db; else raise error.

property final_status: str

Get initial status of the task if it has been bound, otherwise raise error.

property array: jobmon.client.array.Array

Get the array the task has been added to or else raise an AttributeError.

property workflow: jobmon.client.workflow.Workflow

Get the workflow the task has been added to or else raise an AttributeError.

add_upstream(ancestor: Task) → None

Add an upstream (ancestor) Task.

This has Set semantics, an upstream task will only be added once. Symmetrically, this method also adds this Task as a downstream on the ancestor.

add_upstreams(tasks: List[Task]) → None

Add all Tasks in user provided list as upstreams.

add_downstream(descendent: Task) → None

Add a downstream (ancestor) Task.

This has Set semantics, a downstream task will only be added once. Symmetrically, this method also adds this Task as an upstream on the ancestor.

add_downstreams(tasks: List[Task]) → None

Add all Tasks in user provided list as downstreams.

add_attribute(attribute: str, value: str) → None

Function that users can call to add a single attribute for a task.

get_errors() → None | Dict[str, int | List[Dict[str, str | int]]]

Return all errors for each task, with the recent task_instance_id actually used.

set_compute_resources_from_yaml(cluster_name: str, yaml_file: str) → None

Set default compute resources from a user provided yaml file for task level.

TODO: Implement this method.

Parameters:

• cluster_name – name of cluster to set default values for.

• yaml_file – the yaml file that is providing the compute resource values.

update_compute_resources(**kwargs: Any) → None

Function that allows users to update their compute resources.

update_resource_scales(**kwargs: Any) → None

Function that allows users to update their resource scales.

_hash_task_args() → int

A hash of the encoded result of the args and values concatenated together.

__eq__(other: object) → bool

Check if the hashes of two tasks are equivalent.

__lt__(other: Task) → bool

Check if one hash is less than the has of another Task.

__hash__() → int

Create the hash for a task to determine if it is unique within a dag.

__repr__() → str

A representation string for a Task instance.

resource_usage() → dict

Get the resource usage for the successful TaskInstance of a Task.