Task Meta

Task identification `TaskPassport`

Task identifier has 3 components:

Task Namespace
Task Family
Task ID

While Task Namespace and Task Family are the code's identifiers, the ID is the instance identifier.
To override the default Namespace, you need to set Task.task_namespace property. Look at the PrepareData task in the example below:

prepared_data = PrepareData(data='just some data')
print(prepared_data)                 # --> "prepare_data__19d1172be0"

print(prepared_data.task_namespace)  # --> ""
print(prepared_data.task_family)     # --> "prepare_data"
print(prepared_data.task_id)         # --> "prepare_data__19d1172be0"

print(PrepareData.task_namespace)   # --> ""
print(PrepareData.task_family)      # --> "<property object at 0x1111e9548>"
print(PrepareData.task_id)          # --> Error!

Q: How to get access the current task's object?

A: You can access your current task (during pipeline build or task execution) via current_task, and the current run - via current_task_run.

from dbnd import current_task, task

@task
def calculate_alpha(alpha: int = 0.5):
    return current_task().task_version

Q: Is there a way to access the current env inside a task or pipeline?

A: Yes, you can access it via your current task.

from dbnd import current_task, task

@task
def calculate_alpha(alpha: int = 0.5):
    return current_task().task_env.name  # The environment of the task
                                  # See EnvConfig object for all properties

Q: How can I find the user running the pipeline in the `task_run`?

A: Using the task_run_env property of task_run that holds the user's context of the run, including with the user name.

task_run.task_run_env.user

See more about task_run_env here.

How to access all user facing task parameters programmatically?

from dbnd._core.parameter.parameter_value import ParameterFilters
from dbnd import task

all_params = {param.name: param.value for param in task.task_params.get_param_values(ParameterFilters.USER)}

How to access all input task parameters?

from dbnd._core.parameter.parameter_value import ParameterFilters

all_params = {param.name: param.value for param in task.task_params.get_param_values(ParameterFilters.INPUTS)}

See Run Info for accessing run environment info.