📊 Adding Your Data

Your functions probably create and return data. Runnable handles this automatically.

Start with your return values

Here's a function that creates some data:

def write_parameter():
    """A function that returns multiple parameters"""
    df = pd.DataFrame({"x": [1, 2, 3]})
    return df, 10, 3.14, "hello", SamplePydanticModel(x=10, foo="bar"), 0.95

Job mode - Returns work automatically

In job mode, you can name your return values for better tracking and storage:

from runnable import PythonJob, pickled, metric

def main():
    job = PythonJob(
        function=write_parameter,
        returns=[
            pickled("df"),        # For pandas DataFrames
            "integer",            # Simple types work as-is
            "floater",
            "stringer",
            "pydantic_param",     # Pydantic models handled automatically
            metric("score")       # Mark metrics for monitoring
        ]
    )
    job.execute()
    return job

if __name__ == "__main__":
    main()

See complete runnable code

examples/11-jobs/passing_parameters_python.py

"""
The below example shows how to set/get parameters in python
tasks of the pipeline.

The function, set_parameter, returns
    - JSON serializable types
    - pydantic models
    - pandas dataframe, any "object" type

pydantic models are implicitly handled by runnable
but "object" types should be marked as "pickled".

Use pickled even for python data types is advised for
reasonably large collections.

Run the below example as:
    python examples/03-parameters/passing_parameters_python.py

"""

from examples.common.functions import write_parameter
from runnable import PythonJob, metric, pickled


def main():
    job = PythonJob(
        function=write_parameter,
        returns=[
            pickled("df"),
            "integer",
            "floater",
            "stringer",
            "pydantic_param",
            metric("score"),
        ],
    )

    job.execute()

    return job


if __name__ == "__main__":
    main()

Try it now:

uv run examples/11-jobs/passing_parameters_python.py

Pipeline mode - Name your outputs

When your function is part of a workflow, name the outputs so other tasks can use them:

from runnable import PythonTask, pickled, metric

task = PythonTask(
    function=write_parameter,
    returns=[
        pickled("df"),        # For pandas DataFrames
        "integer",            # Simple types work as-is
        "floater",
        "stringer",
        "pydantic_param",     # Pydantic models handled automatically
        metric("score")       # Mark metrics for monitoring
    ]
)

Now downstream tasks can access df, integer, floater, etc.

See complete runnable code

examples/03-parameters/passing_parameters_python.py

"""
The below example shows how to set/get parameters in python
tasks of the pipeline.

The function, set_parameter, returns
    - JSON serializable types
    - pydantic models
    - pandas dataframe, any "object" type

pydantic models are implicitly handled by runnable
but "object" types should be marked as "pickled".

Use pickled even for python data types is advised for
reasonably large collections.

Run the below example as:
    python examples/03-parameters/passing_parameters_python.py

"""

from examples.common.functions import read_parameter, write_parameter
from runnable import Pipeline, PythonTask, metric, pickled


def main():
    write_parameters = PythonTask(
        function=write_parameter,
        returns=[
            pickled("df"),
            "integer",
            "floater",
            "stringer",
            "pydantic_param",
            metric("score"),
        ],
        name="set_parameter",
    )

    read_parameters = PythonTask(
        function=read_parameter,
        terminate_with_success=True,
        name="get_parameters",
    )

    pipeline = Pipeline(
        steps=[write_parameters, read_parameters],
    )

    _ = pipeline.execute()

    return pipeline


if __name__ == "__main__":
    main()

Try it now:

uv run examples/03-parameters/passing_parameters_python.py

Handle different data types

📦 Large or complex objects

Use pickled() for pandas DataFrames, models, or large collections:

returns=[pickled("df"), "score"]

📈 Track metrics

Use metric() for numbers you want to monitor:

returns=[metric("accuracy"), metric("loss")]

📋 Everything else

Simple types (strings, numbers, lists) work as-is:

returns=["count", "status", "results"]

Pro tip

Name your returns clearly. ["df", "score"] is better than ["output1", "output2"].

Parameter types across task types

The returns=[] syntax works across all task types, but parameter type support varies:

• Python functions (PythonTask, PythonJob): Full support for all parameter types (primitive, pickled, pydantic, metric)
• Jupyter notebooks (NotebookTask, NotebookJob): Can receive primitives only, can return all types
• Shell scripts (ShellTask, ShellJob): Limited to primitive types for both input and output

Choose your parameter types based on the task types in your workflow!

Next: Learn how to connect functions in workflows.