pylibmgm.GmModel

class pylibmgm.GmModel

Pairwise graph matching (GM) model.

Represents a matching problem between two graphs with unary assignment costs and pairwise edge costs. This is the core data structure for single pairwise graph matching problems.

Methods

__init__(*args, **kwargs)

Overloaded function.

add_assignment(node1, node2, cost)

Add an assignment with its unary cost.

add_edge(*args, **kwargs)

Overloaded function.

costs()

Access the cost hashmap.

no_assignments()

Get number of assignments.

no_edges()

Get number of edges.

Attributes

Name

Type

Description

assignment_list

list[tuple[int, int]]

List of all assignments. Index is assignment ID.

graph1

Graph

First graph in the matching.

graph2

Graph

Second graph in the matching.
__init__(graph1: Graph, graph2: Graph) → None

Create a GM model between two graphs.

Parameters:

  • graph1 (Graph) – First graph to match.

  • graph2 (Graph) – Second graph to match.

__init__(graph1: Graph, graph2: Graph, no_assignments: int, no_edges: int) → None

Create a GM model with pre-allocation.

Parameters:

  • graph1 (Graph) – First graph to match.

  • graph2 (Graph) – Second graph to match.

  • no_assignments (int) – Expected number of assignments (for pre-allocation).

  • no_edges (int) – Expected number of edges (for pre-allocation).

add_edge(assignment1: int, assignment2: int, cost: float) → None

Add an edge via two assignment IDs.

Parameters:

  • assignment1 (int) – First assignment ID.

  • assignment2 (int) – Second assignment ID.

  • cost (float) – Pairwise cost of this edge.

add_edge(assignment1_left: int, assignment1_right: int, assignment2_left: int, assignment2_right: int, cost: float) → None

Add an edge via four node IDs.

Parameters:

  • assignment1_left (int) – First node of first assignment.

  • assignment1_right (int) – Second node of first assignment.

  • assignment2_left (int) – First node of second assignment.

  • assignment2_right (int) – Second node of second assignment.

  • cost (float) – Pairwise cost of this edge.

add_assignment(node1: int, node2: int, cost: float) → None

Add an assignment with its unary cost.

Parameters:

  • node1 (int) – Node from graph1.

  • node2 (int) – Node from graph2.

  • cost (float) – Unary cost of this assignment.

costs() → pylibmgm.CostMap

Access the cost hashmap.

Returns:

The cost hashmap containing all unary and pairwise costs.

Return type:

CostMap

no_assignments() → int

Get number of assignments.

Returns:

Number of assignments in the model.

Return type:

int

no_edges() → int

Get number of edges.

Returns:

Number of edges in the model.

Return type:

int