@buggyorg/dynatype-network-graph

Dynatype Network Graph

Usage

function	description
addTypeConversion (processGraph, convertGraph)	Adds translator nodes for type conversion to the processGraph using only translations which are defined in the convertGraph

Input Process Graph

The process graph is a directed multigraph.

Each node which doesn't represent a port should have the label: processNode. Nodes which represent ports should have the datatype of the output/input signal as label. The id of port-nodes should be the id of the corresponding processNode + ('_OUT_' | '_IN_') + <a number> to differentiate between different input/output ports.

Edges from port to port should have the label edge and edges between process nodes and ports should have the label processNode_in resp. processEdge_out.

Convert Graph

The convert graph is a directed graph.

The nodes should have it's datatype as label and id.

An edge from a first node to a second node should have the label: id_of_the_first_node + ':' + id_of_the_second_node.

Returned Process Graph

The returned process graph is an expansion of the input process graph.

The edges between two ports with different datatypes are removed. Instead translator nodes with input and output port-nodes are inserted. The translator nodes have the id : id_of_port-node_with_first_datatype + ':' + id_of_port-node_with_second_datatype + <a number> (to identify the different type conversions). The label is first_type + ':' + second_type. The input and output port-nodes of the translator have the id of the corresponding translator_node + ('_OUT_' | '_IN_') + <a number>, and the label translator_in resp. translator_out.