TaskEnvironment

Define an execution environment for a set of tasks.

Attributes

Attribute	Type	Description
cache	`CacheRequest` = "disable"	Cache policy — "auto", "override", "disable", or a `Cache` object. Also settable in `@env.task(cache=...)` and `task.override(cache=...)`.
reusable	`ReusePolicy	None` = null
plugin_config	`Optional[Any]` = null	Plugin configuration for custom task types (e.g., Ray, Spark). Cannot be combined with `reusable`. TaskEnvironment level only.
queue	`Optional[str]` = null	Queue name for scheduling. Queues identify specific partitions of your compute infrastructure (e.g., a particular cluster in a multi-cluster deployment) and are configured as part of your Flyte/Union deployment. Also settable in `@env.task` and `task.override`.

Constructor

Signature

def TaskEnvironment(
    name: str,
    image: Optional[Union[str, [Image](../../image/image.md?sid=flyte__image_image), Literal["auto"]]] = None,
    depends_on: Optional[List[[Environment](../../environment/environment.md?sid=flyte__environment_environment)]] = None,
    description: Optional[str] = None,
    plugin_config: Optional[Any] = None,
    resources: Optional[[Resources](../../resources/resources.md?sid=flyte__resources_resources)] = None,
    env_vars: Optional[Dict[str, str]] = None,
    secrets: Optional[SecretRequest] = None,
    cache: CacheRequest = "disable",
    reusable: Optional[[ReusePolicy](../../reusable/environment/reusepolicy.md?sid=flyte__reusable_environment_reusepolicy)] = None,
    queue: Optional[str] = None,
    pod_template: Optional[Union[str, [PodTemplate](../../pod/podtemplate.md?sid=flyte__pod_podtemplate)]] = None,
    interruptible: Optional[bool] = None
)

Parameters

Name	Type	Description
name	`str`	Name of the environment (required). Must be snake_case or kebab-case. TaskEnvironment level only. The fully-qualified name of each task is `< env_name >.< function_name >` (e.g., environment `"my_env"` containing function `my_task` produces FQN `"my_env.my_task"`). Neither component is overridable.
image	`Optional[Union[str, [Image](../../image/image.md?sid=flyte__image_image), Literal["auto"]]]` = None	Docker image for the environment. Can be a string (image URI), an `Image` object, or `"auto"` to use the default image. TaskEnvironment level only.
depends_on	`Optional[List[[Environment](../../environment/environment.md?sid=flyte__environment_environment)]]` = None	List of other environments this one depends on. Used at deploy time to ensure dependencies are also deployed. TaskEnvironment level only.
description	`Optional[str]` = None	Human-readable description (max 255 characters). TaskEnvironment level only.
plugin_config	`Optional[Any]` = None	Plugin configuration for custom task types (e.g., Ray, Spark). Cannot be combined with `reusable`. TaskEnvironment level only.
resources	`Optional[[Resources](../../resources/resources.md?sid=flyte__resources_resources)]` = None	Compute resources (CPU, memory, GPU, disk). Overridable via `task.override(resources=...)` when not using reusable containers.
env_vars	`Optional[Dict[str, str]]` = None	Environment variables as `dict[str, str]`. Overridable via `task.override(env_vars=...)` when not using reusable containers.
secrets	`Optional[SecretRequest]` = None	Secrets to inject. Overridable via `task.override(secrets=...)` when not using reusable containers.
cache	`CacheRequest` = "disable"	Cache policy — `"auto"`, `"override"`, `"disable"`, or a `Cache` object. Also settable in `@env.task(cache=...)` and `task.override(cache=...)`.
reusable	`Optional[[ReusePolicy](../../reusable/environment/reusepolicy.md?sid=flyte__reusable_environment_reusepolicy)]` = None	`ReusePolicy` for container reuse. Also overridable via `task.override(reusable=...)`. Note: when `reusable` is set on the environment, overriding `resources`, `env_vars`, or `secrets` in `task.override()` requires passing `reusable="off"` in the same call. Additionally, `secrets` cannot be overridden at the `@env.task` decorator level when the environment has `reusable` set.
queue	`Optional[str]` = None	Queue name for scheduling. Queues identify specific partitions of your compute infrastructure (e.g., a particular cluster in a multi-cluster deployment) and are configured as part of your Flyte/Union deployment. Also settable in `@env.task` and `task.override`.
pod_template	`Optional[Union[str, [PodTemplate](../../pod/podtemplate.md?sid=flyte__pod_podtemplate)]]` = None	Kubernetes pod template for advanced configuration (sidecars, volumes, etc.). Also settable in `@env.task` and `task.override`.
interruptible	`Optional[bool]` = None	Whether tasks can run on spot/preemptible instances. Also settable in `@env.task` and `task.override`.

Methods

`clone_with()`

@classmethod
def clone_with(
    name: str,
    image: Optional[Union[str, [Image](../../image/image.md?sid=flyte__image_image), Literal["auto"]]] = None,
    resources: Optional[[Resources](../../resources/resources.md?sid=flyte__resources_resources)] = None,
    env_vars: Optional[Dict[str, str]] = None,
    secrets: Optional[SecretRequest] = None,
    depends_on: Optional[List[[Environment](../../environment/environment.md?sid=flyte__environment_environment)]] = None,
    description: Optional[str] = None,
    interruptible: Optional[bool] = None,
    include: Optional[Tuple[str, ...]] = None,
    kwargs: Any
) - > [TaskEnvironment](taskenvironment.md?sid=flyte__task_environment_taskenvironment)

Create a new TaskEnvironment that shares most settings with this one but differs in name and selected overrides. Use clone_with when you need several environments that share a common base configuration (image, resources, secrets, etc.) but vary in one or two settings, avoiding repetition.

Parameters

Name	Type	Description
name	`str`	Name for the new environment (required).
image	`Optional[Union[str, [Image](../../image/image.md?sid=flyte__image_image), Literal["auto"]]]` = None	Override the container image.
resources	`Optional[[Resources](../../resources/resources.md?sid=flyte__resources_resources)]` = None	Override compute resources.
env_vars	`Optional[Dict[str, str]]` = None	Override environment variables.
secrets	`Optional[SecretRequest]` = None	Override secrets.
depends_on	`Optional[List[[Environment](../../environment/environment.md?sid=flyte__environment_environment)]]` = None	Override deployment dependencies.
description	`Optional[str]` = None	Override the description.
interruptible	`Optional[bool]` = None	Override the interruptible setting.
include	`Optional[Tuple[str, ...]]` = None	This parameter is not described in the docstring, but based on its type, it likely specifies a tuple of strings to include.
kwargs	`Any`	Additional `TaskEnvironment`-specific overrides (e.g., `cache`, `reusable`, `plugin_config`).

Returns

Type	Description
`[TaskEnvironment](taskenvironment.md?sid=flyte__task_environment_taskenvironment)`	A new `TaskEnvironment` instance with the specified overrides.

`task()`

@classmethod
def task(
    _func: F | None = None,
    short_name: Optional[str] = None,
    cache: CacheRequest | None = None,
    retries: Union[int, [RetryStrategy](../../retry/retrystrategy.md?sid=flyte__retry_retrystrategy)] = 0,
    timeout: Union[timedelta, int] = 0,
    docs: Optional[Documentation] = None,
    pod_template: Optional[Union[str, [PodTemplate](../../pod/podtemplate.md?sid=flyte__pod_podtemplate)]] = None,
    report: bool = False,
    interruptible: bool | None = None,
    max_inline_io_bytes: int = MAX_INLINE_IO_BYTES,
    queue: Optional[str] = None,
    triggers: Tuple[[Trigger](../../remote/trigger/trigger.md?sid=flyte_remote__trigger_trigger), ...]| Trigger = (),
    links: Tuple[[Link](../../app/types/link.md?sid=flyte_app__types_link), ...]| Link = (),
    task_resolver: Any | None = None,
    entrypoint: bool = False
) - > Callable[[F], [AsyncFunctionTaskTemplate](../asyncfunctiontasktemplate.md?sid=flyte__task_asyncfunctiontasktemplate)[P, R, F]]| AsyncFunctionTaskTemplate[P, R, F]

Decorate a function to be a task.

Parameters

Name	Type	Description
_func	`F	None` = None
short_name	`Optional[str]` = None	Optional friendly name for the task or action, used in parts of the UI (defaults to the function name). Overriding `short_name` does not change the task's fully-qualified name.
cache	`CacheRequest	None` = None
retries	`Union[int, [RetryStrategy](../../retry/retrystrategy.md?sid=flyte__retry_retrystrategy)]` = 0	Number of retries (`int`) or a `RetryStrategy` object that defines retry behavior. Defaults to `0` (no retries).
timeout	`Union[timedelta, int]` = 0	Task timeout, as a `timedelta` object or an integer number of seconds. `0` means no timeout.
docs	`Optional[Documentation]` = None	Optional The documentation for the task, if not provided the function docstring will be used.
pod_template	`Optional[Union[str, [PodTemplate](../../pod/podtemplate.md?sid=flyte__pod_podtemplate)]]` = None	Optional The pod template for the task, if not provided the default pod template will be used.
report	`bool` = False	Optional Whether to generate the html report for the task, defaults to False.
interruptible	`bool	None` = None
max_inline_io_bytes	`int` = MAX_INLINE_IO_BYTES	Maximum allowed size (in bytes) for all inputs and outputs passed directly to the task (e.g., primitives, strings, dicts). Does not apply to files, directories, or dataframes. Default is 10 MiB.
queue	`Optional[str]` = None	Optional queue name to use for this task. If not set, the environment's queue will be used.
triggers	`Tuple[Trigger, ...]	Trigger` = ()
links	`Tuple[Link, ...]	Link` = ()
task_resolver	`Any	None` = None
entrypoint	`bool` = False	Optionally mark a task as an entrypoint task, defaults to False. This serves as a hint to the UI.

Returns

Type	Description
`Callable[[F], AsyncFunctionTaskTemplate[P, R, F]]	AsyncFunctionTaskTemplate[P, R, F]`

`sandbox()`

@classmethod
def sandbox() - > _SandboxNamespace

Access the sandbox namespace for creating sandboxed tasks.

Returns

Type	Description
`_SandboxNamespace`	The sandbox namespace object.

`tasks()`

@classmethod
def tasks() - > Dict[str, [TaskTemplate](../tasktemplate.md?sid=flyte__task_tasktemplate)]

Get all tasks defined in the environment.

Returns

Type	Description
`Dict[str, [TaskTemplate](../tasktemplate.md?sid=flyte__task_tasktemplate)]`	A dictionary mapping task names to their `TaskTemplate` objects.

`from_task()`

@classmethod
def from_task(
    name: str,
    tasks: [TaskTemplate](../tasktemplate.md?sid=flyte__task_tasktemplate),
    depends_on: Optional[List["Environment"]] = None
) - > [TaskEnvironment](taskenvironment.md?sid=flyte__task_environment_taskenvironment)

Create a TaskEnvironment from a list of tasks. All tasks should have the same image or no Image defined. Similarity of Image is determined by the python reference, not by value.

Parameters

Name	Type	Description
name	`str`	The name of the environment.
tasks	`[TaskTemplate](../tasktemplate.md?sid=flyte__task_tasktemplate)`	The list of tasks to create the environment from.
depends_on	`Optional[List["Environment"]]` = None	Optional list of environments that this environment depends on.

Returns

Type	Description
`[TaskEnvironment](taskenvironment.md?sid=flyte__task_environment_taskenvironment)`	The created TaskEnvironment.

Attributes​

Constructor​

Signature​

Parameters​

Methods​

clone_with()​

Parameters​

Returns​

task()​

Parameters​

Returns​

sandbox()​

Returns​

tasks()​

Returns​

from_task()​

Parameters​

Returns​

Attributes

Constructor

Signature

Parameters

Methods

`clone_with()`

Parameters

Returns

`task()`

Parameters

Returns

`sandbox()`

Returns

`tasks()`

Returns

`from_task()`

Parameters

Returns