nodes

Classes

AbstractNode	Abstract node model use to represent different components of a Spark model.
SourceNode	Node representing the input to the system.
SinkNode	Node representing the output of the system.
SparkModuleNode	Abstract node representing a SparkModule.

Functions

module_to_nodegraph(entry)

Factory function that creates a new NodeGraphQt node class from an Spark module class.

Module Contents

class nodes.AbstractNode

Bases: NodeGraphQt.BaseNode, abc.ABC

Abstract node model use to represent different components of a Spark model.

__identifier__ = 'spark'

NODE_NAME = 'Abstract Node'

input_specs: dict[str, spark.core.specs.InputSpec]

output_specs: dict[str, spark.core.specs.OutputSpec]

graph: spark.graph_editor.models.graph.SparkNodeGraph

update_io_spec(spec, port_name, payload_type=None, shape=None, dtype=None, description=None)

Method to update node IO specs. Valid updates are broadcasted.

Parameters:

• spec (str) – str, target spec {input, output}

• port_name (str) – str, the name of the port to update

• payload_type (type[spark.core.payloads.SparkPayload] | None) – type[SparkPayload] | None, the new payload_type

• shape (tuple[int, Ellipsis] | None) – tuple[int, …] | None, the new shape

• dtype (jax.typing.DTypeLike | None) – DTypeLike | None, the new dtype

• description (str | None) – str | None, the new description

property node_config_metadata: dict

Abstractmethod:

Return type:

dict

class nodes.SourceNode

Bases: AbstractNode

Node representing the input to the system.

NODE_NAME = 'Source Node'

output_specs

property node_config_metadata: dict

Return type:

dict

class nodes.SinkNode

Bases: AbstractNode

Node representing the output of the system.

NODE_NAME = 'Sink Node'

input_specs

property node_config_metadata: dict

Return type:

dict

class nodes.SparkModuleNode

Bases: AbstractNode, abc.ABC

Abstract node representing a SparkModule.

NODE_NAME = 'SparkModule'

module_cls: type[spark.core.module.SparkModule]

node_config: spark.core.config.BaseSparkConfig

input_specs

output_specs

property node_config_metadata: dict

Return type:

dict

nodes.module_to_nodegraph(entry)

Factory function that creates a new NodeGraphQt node class from an Spark module class.

Parameters:

entry (spark.core.registry.RegistryEntry)

Return type:

type[SparkModuleNode]