kedro.framework.hooks.specs.NodeSpecs

class kedro.framework.hooks.specs.NodeSpecs[source]

Namespace that defines all specifications for a node’s lifecycle hooks.

Methods

after_node_run(node, catalog, inputs, ...)

Hook to be invoked after a node runs.

before_node_run(node, catalog, inputs, ...)

Hook to be invoked before a node runs.

on_node_error(error, node, catalog, inputs, ...)

Hook to be invoked if a node run throws an uncaught error.
after_node_run(node, catalog, inputs, outputs, is_async, session_id)[source]

Hook to be invoked after a node runs. The arguments received are the same as those used by kedro.runner.run_node as well as the outputs of the node run.

Parameters:

  • node (Node) – The Node that ran.

  • catalog (CatalogProtocol) – An implemented instance of CatalogProtocol containing the node’s inputs and outputs.

  • inputs (dict[str, Any]) – The dictionary of inputs dataset. The keys are dataset names and the values are the actual loaded input data, not the dataset instance.

  • outputs (dict[str, Any]) – The dictionary of outputs dataset. The keys are dataset names and the values are the actual computed output data, not the dataset instance.

  • is_async (bool) – Whether the node was run in async mode.

  • session_id (str) – The id of the session.

Return type:

None

before_node_run(node, catalog, inputs, is_async, session_id)[source]

Hook to be invoked before a node runs. The arguments received are the same as those used by kedro.runner.run_node

Parameters:

  • node (Node) – The Node to run.

  • catalog (CatalogProtocol) – An implemented instance of CatalogProtocol containing the node’s inputs and outputs.

  • inputs (dict[str, Any]) – The dictionary of inputs dataset. The keys are dataset names and the values are the actual loaded input data, not the dataset instance.

  • is_async (bool) – Whether the node was run in async mode.

  • session_id (str) – The id of the session.

Return type:

dict[str, Any] | None

Returns:

Either None or a dictionary mapping dataset name(s) to new value(s).

If returned, this dictionary will be used to update the node inputs, which allows to overwrite the node inputs.

on_node_error(error, node, catalog, inputs, is_async, session_id)[source]

Hook to be invoked if a node run throws an uncaught error. The signature of this error hook should match the signature of before_node_run along with the error that was raised.

Parameters:

  • error (Exception) – The uncaught exception thrown during the node run.

  • node (Node) – The Node to run.

  • catalog (CatalogProtocol) – An implemented instance of CatalogProtocol containing the node’s inputs and outputs.

  • inputs (dict[str, Any]) – The dictionary of inputs dataset. The keys are dataset names and the values are the actual loaded input data, not the dataset instance.

  • is_async (bool) – Whether the node was run in async mode.

  • session_id (str) – The id of the session.

Return type:

None