Node

• Module for representing generic nodes in a graph
• Multiple node types can be inherited with this new implementation
• Base node attribute will contain the id for mapping to adjacency list index

Node Type

Attribute	Description
node	General node value structure for testing graph algorithms
variable	Variable structure for creating computation graphs

• Structure used for storing all generic types that a node can have
• Any potential structure that will be part of a graph can go here

Node

Attribute	Description
id	Numerical Identifier for node object
node_type	Structure that refers to generic node types
next	Pointer to next node for adjacency list representation
left	Left pointer for set implementation
right	Right pointer for set implementation

• General node structure that will be used for all graph representations
• Temporarily used for representing data structure values

Node Value

Attribute	Description
weight	Weight value to represent how close nodes are to eachother
counter	Used in certain algorithms for counting amount of times a node was visited
label	Used for identifying a node by character label instead of id

• This node value is used for general graph algorithms and demonstration purpose
• The node value structure is the default for testing out graph algorithms
• It inherits the base node type

Variable

Attribute	Description
label	Name of operation being performed on inputs
output	Generic value stored with variable structure
upstream_value	Gradients passed backward from other nodes
forward	Operation to perform on inputs during forward pass
backward	Operation to perform on inputs during backward pass
inputs	Amount of inputs if the variable is also an operation

• This structure is meant for creating nodes in a computation graph
• Each variable node uses it's neighbors as inputs

Edge

Attribute	Description
src	Source node value
dst	Destination node value
weight	Weight that represents how closely related the nodes are

• This is used for some searching/path finding