plugins.h_narhwals.NarwhalsAdapter¶

Provides a convenience wrapper for the Narwhals library; use the Narwhals decorator underneath. Must have Narwhals installed to use it:

pip install “sf-hamilton[narwhals]”

class hamilton.plugins.h_narwhals.NarwhalsAdapter¶

Adapter to make it simpler to use narwhals with Hamilton.

from hamilton import base, driver
from hamilton.plugins import h_narwhals
import example

# pandas
dr = (
    driver.Builder()
    .with_config({"load": "pandas"})
    .with_modules(example)
    .with_adapters(
        h_narwhals.NarwhalsAdapter(),
        h_narwhals.NarwhalsDataFrameResultBuilder(
            base.PandasDataFrameResult()
        ),
    )
    .build()
)
result = dr.execute(
     [example.group_by_mean, example.example1],
     inputs={"col_name": "a"}
)

do_node_execute(*, run_id: str, node_: Node, kwargs: Dict[str, Any], task_id: str | None = None) → Any¶

Method that is called to implement node execution. This can replace the execution of a node with something all together, augment it, or delegate it.

Parameters:

run_id – ID of the run, unique in scope of the driver.
node – Node that is being executed
kwargs – Keyword arguments that are being passed into the node
task_id – ID of the task, defaults to None if not in a task setting

run_to_execute_node(*, node_name: str, node_tags: Dict[str, Any], node_callable: Any, node_kwargs: Dict[str, Any], task_id: str | None, **future_kwargs: Any) → Any¶

This method is responsible for executing the node and returning the result.

It uses nw_kwargs from the node tags to know if any special flags should be passed to the narwhals decorator function.

Parameters:

node_name – Name of the node.
node_tags – Tags of the node.
node_callable – Callable of the node.
node_kwargs – Keyword arguments to pass to the node.
task_id – The ID of the task, none if not in a task-based environment
future_kwargs – Additional keyword arguments – this is kept for backwards compatibility

Returns:

The result of the node execution – up to you to return this.

plugins.h_narhwals.NarwhalsDataFrameResultBuilder¶

Result builder to be used with the NarwhalsAdapter. Must have Narwhals installed to use it:

pip install “sf-hamilton[narwhals]”

class hamilton.plugins.h_narwhals.NarwhalsDataFrameResultBuilder(result_builder: ResultBuilder | LegacyResultMixin)¶

Builds the result. It unwraps the narwhals parts of it and delegates to the passed in result builder.

from hamilton import base, driver
from hamilton.plugins import h_narwhals, h_polars
import example

# polars
dr = (
    driver.Builder()
    .with_config({"load": "polars"})
    .with_modules(example)
    .with_adapters(
        h_narwhals.NarwhalsAdapter(),
        h_narwhals.NarwhalsDataFrameResultBuilder(
           h_polars.PolarsDataFrameResult()
        ),
    )
    .build()
)
result = dr.execute(
    ["group_by_mean", "example1"],
    inputs={"col_name": "a"}
)

__init__(result_builder: ResultBuilder | LegacyResultMixin)¶

build_result(**outputs: Any) → Any¶

Given a set of outputs, build the result.

Parameters:: outputs – the outputs from the execution of the graph.
Returns:: the result of the execution of the graph.

do_build_result(outputs: Dict[str, Any]) → Any¶: Implements the do_build_result method from the BaseDoBuildResult class. This is kept from the user as the public-facing API is build_result, allowing us to change the API/implementation of the internal set of hooks

input_types() → List[Type[Type]]¶

Gives the applicable types to this result builder. This is optional for backwards compatibility, but is recommended.

Returns:: A list of types that this can apply to.

output_type() → Type¶: Returns the output type of this result builder :return: the type that this creates