Options
All
  • Public
  • Public/Protected
  • All
Menu

Interface OpNode

Hierarchy

Implemented by

Index

Properties

attr

attr: object[]

Type declaration

  • key: string
  • value: any

cardinality

cardinality: number

The number of nodes this node represents. For OpNodes, this will be 1, and for GroupNodes it will be a count of the total number of descendents it contains.

inEmbeddings

inEmbeddings: OpNode[]

include

include: InclusionType

If the node is to be included or excluded from the main graph when rendered. Defaults to UNSPECIFIED, which means that the rendering algorithm determines if it will be included or not. Then can be set to INCLUDE or EXCLUDE manually by the user.

inputs

inputs: NormalizedInput[]

isGroupNode

isGroupNode: boolean

Whether this node is a type that may contain other nodes. Those types should extend from GroupNode.

For an OpNode, isGroupNode will be false, even though it may have embeddings. These embedding Nodes will have their parentNode set to the OpNode. However, embeddings are later rendered as annotations, not as children to be made visible on expansion (like a Metanode or SeriesNode).

name

name: string

The name of the node, used frequently to look up nodes by name.

nodeAttributes

nodeAttributes: object

Node attributes specify customizable visual aspects of a node and application-specific metadata associated with a node. The name 'nodeAttributes' is meant to avoid naming-conflicts with the 'attr' in subclasses of Node.

Type declaration

  • [key: string]: any

op

op: string

outEmbeddings

outEmbeddings: OpNode[]

outputShapes

outputShapes: TensorShape[]

Array of tensor shapes. Null if the number of output tensors is unknown, otherwise the length will equal the number of output tensors.

Each tensor shape is an array of numbers, or null. Details:

  • null means unknown rank, and therefore entire shape is unknown.
  • [4, 2, 1] means rank-3 tensor of size 4x2x1.
  • [] means a scalar (rank-0 tensor).
  • [1] means rank-1 tensor of size 1 (not the same as scalar).
  • [5, -1, 3] means rank-3 tensor of shape is 5x?x3. The size
    of the middle dimension is unknown (encoded as -1).
    

owningSeries

owningSeries: string

parentNode

parentNode: Node

The Node which is this Node's parent. This is of type Node and not GroupNode because of embeddings, which will have a parent OpNode.

type

type: NodeType

Which type of node this is.

Generated using TypeDoc