flytekit.extend.PythonCustomizedContainerTask

class flytekit.extend.PythonCustomizedContainerTask(*args, **kwargs)[source]

Please take a look at the comments for :py:class`flytekit.extend.ExecutableTemplateShimTask` as well. This class should be subclassed and a custom Executor provided as a default to this parent class constructor when building a new external-container flytekit-only plugin.

This class provides authors of new task types the basic scaffolding to create task-template based tasks. In order to write such a task, authors need to

  • subclass the ShimTaskExecutor class and override the execute_from_model function. This function is where all the business logic should go. Keep in mind though that you, the plugin author, will not have access to anything that’s not serialized within the TaskTemplate which is why you’ll also need to

  • subclass this class, and override the get_custom function to include all the information the executor will need to run.

  • Also pass the executor you created as the executor_type argument of this class’s constructor.

Keep in mind that the total size of the TaskTemplate still needs to be small, since these will be accessed frequently by the Flyte engine.

Parameters
  • name – unique name for the task, usually the function’s module and name.

  • task_config – Configuration object for Task. Should be a unique type for that specific Task

  • container_image – This is the external container image the task should run at platform-run-time.

  • executor – This is an executor which will actually provide the business logic.

  • task_resolver – Custom resolver - if you don’t make one, use the default task template resolver.

  • task_type – String task type to be associated with this Task

  • requests – custom resource request settings.

  • limits – custom resource limit settings.

  • environment – Environment variables you want the task to have when run.

  • secret_requests (List[Secret]) –

    Secrets that are requested by this container execution. These secrets will be mounted based on the configuration in the Secret and available through the SecretManager using the name of the secret as the group Ideally the secret keys should also be semi-descriptive. The key values will be available from runtime, if the backend is configured to provide secrets and if secrets are available in the configured secrets store. Possible options for secret stores are

__init__(name, task_config, container_image, executor_type, task_resolver=None, task_type='python-task', requests=None, limits=None, environment=None, secret_requests=None, **kwargs)[source]
Parameters
  • name (str) – unique name for the task, usually the function’s module and name.

  • task_config (flytekit.core.python_customized_container_task.TC) – Configuration object for Task. Should be a unique type for that specific Task

  • container_image (str) – This is the external container image the task should run at platform-run-time.

  • executor – This is an executor which will actually provide the business logic.

  • task_resolver (Optional[flytekit.core.python_customized_container_task.TaskTemplateResolver]) – Custom resolver - if you don’t make one, use the default task template resolver.

  • task_type – String task type to be associated with this Task

  • requests (Optional[flytekit.core.resources.Resources]) – custom resource request settings.

  • limits (Optional[flytekit.core.resources.Resources]) – custom resource limit settings.

  • environment (Optional[Dict[str, str]]) – Environment variables you want the task to have when run.

  • secret_requests (List[Secret]) –

    Secrets that are requested by this container execution. These secrets will be mounted based on the configuration in the Secret and available through the SecretManager using the name of the secret as the group Ideally the secret keys should also be semi-descriptive. The key values will be available from runtime, if the backend is configured to provide secrets and if secrets are available in the configured secrets store. Possible options for secret stores are

  • executor_type (Type[flytekit.core.shim_task.ShimTaskExecutor]) –

Methods

__init__(name, task_config, container_image, ...)

param name

unique name for the task, usually the function's module and name.

compile(ctx, *args, **kwargs)

Generates a node that encapsulates this task in a workflow definition.

construct_node_metadata()

Used when constructing the node that encapsulates this task as part of a broader workflow definition.

dispatch_execute(ctx, input_literal_map)

This function is mostly copied from the base PythonTask, but differs in that we have to infer the Python interface before executing.

execute(**kwargs)

Send things off to the executor instead of running here.

find_lhs()

get_command(settings)

get_config(settings)

Returns the task config as a serializable dictionary.

get_container(settings)

Returns the container definition (if any) that is used to run the task on hosted Flyte.

get_custom(settings)

Return additional plugin-specific custom data (if any) as a serializable dictionary.

get_input_types()

Returns the names and python types as a dictionary for the inputs of this task.

get_k8s_pod(settings)

Returns the kubernetes pod definition (if any) that is used to run the task on hosted Flyte.

get_sql(settings)

Returns the Sql definition (if any) that is used to run the task on hosted Flyte.

get_type_for_input_var(k, v)

Returns the python type for an input variable by name.

get_type_for_output_var(k, v)

Returns the python type for the specified output variable by name.

local_execute(ctx, **kwargs)

This function is used only in the local execution path and is responsible for calling dispatch execute.

post_execute(user_params, rval)

This function is a stub, just here to keep dispatch_execute compatibility between this class and PythonTask.

pre_execute(user_params)

This function is a stub, just here to keep dispatch_execute compatibility between this class and PythonTask.

serialize_to_model(settings)

Attributes

SERIALIZE_SETTINGS

container_image

environment

Any environment variables that supplied during the execution of the task.

executor

executor_type

instantiated_in

interface

lhs

location

metadata

name

python_interface

Returns this task's python interface.

resources

security_context

task_config

Returns the user-specified task config which is used for plugin-specific handling of the task.

task_resolver

task_template

Override the base class implementation to serialize on first call.

task_type

task_type_version