graph

The DGM Graph module contains tools for creating graphs that describe the excitation wave propagation in the heart. The DGM Graph is a core part of the DGM toolchain.

This min graph class features:

• A default graph builder which creates connections between all nodes
• Edge weights which are the difference between the source and target scalars (usually delta LAT)
• Filtering of connections based on properties
• Merging graphs
• Calculating various graph measures

Other features:

• A graph class which creates connections based on a voronoi neighborhood
• A graph class which uses correlations and delays as weights
• A class for converting between DGM graphs and networkit graphs

CorrelationGraph

Bases: Graph

Create a graph with delays instead of DelataLAT for weights

correlations property

delay_cv_max = None instance-attribute

delay_cv_min = None instance-attribute

delay_selection = CvDelaySelection instance-attribute

delay_selections = {'cvdelayselection': CvDelaySelection, 'mindelayselection': MinDelaySelection} instance-attribute

delays property

filters = copy.deepcopy(Graph.filters) class-attribute instance-attribute

max_correlation = None instance-attribute

min_correlation = None instance-attribute

min_peak_distance = None instance-attribute

sampling_frequency = None instance-attribute

__init__(coords, scalars, indices=None)

reset_propeties()

set_correlation_range(min_correlation, max_correlation)

set_delay_cv_max(delay_cv_max)

set_delay_cv_min(delay_cv_min)

set_delay_cv_range(delay_cv_min, delay_cv_max)

set_delay_selection(selection)

Sets the delay selection parameter. A delay selector selects a single delay and correlation value for each edge based on the value of said delay.

Parameters:

Name	Type	Description	Default
selection	Type[DelaySelector] | DelaySelector | str	a selector class	required

Returns:

Type	Description
Self	self

set_max_correlation(max_correlation)

set_min_correlation(min_correlation)

set_min_peak_distance(min_peak_distance)

set_sampling_frequency(sampling_frequency)

Graph

Bases: PropertyMixin

A class representing a DGM graph.

A graph consists of two main attributes:

• nodes: a node in the graph consists of coordinates and a scalar value
• edges: the edges of a node define the connections between nodes.

adjacency_matrix = None instance-attribute

coords property

cv_max = None instance-attribute

cv_min = None instance-attribute

degree = None instance-attribute

degree_diff = None instance-attribute

edges = None instance-attribute

filters = {} class-attribute instance-attribute

in_degree = None instance-attribute

max_delta_lat = None instance-attribute

max_edge_length = None instance-attribute

maximal_clique_sizes = None instance-attribute

min_delta_lat = None instance-attribute

min_edge_length = None instance-attribute

nodes = Nodes(coords, scalars, indices) instance-attribute

nodes_number property

The number of nodes inside the graph

out_degree = None instance-attribute

period = None instance-attribute

scalars property

sources property

Sources are the coordinates of the starting points of the edges.

Returns:

Type	Description
ndarray	ndarray[N, 3]: Arrays of coordinates for each graph edge start.

strong_components = None instance-attribute

targets property

Targets are the coordinates of the end points of the edges.

Returns:

Type	Description
ndarray	ndarray[N, 3]: Arrays of coordinates for each graph edge end.

vectors property

Calculate vectors (arrows) based on the directed graph's edges.

Returns:

Type	Description
ndarray	ndarray[N, 3]: Arrays of vectors for each graph's edge.

weak_components = None instance-attribute

__init__(coords, scalars, indices=None)

Parameters:

Name	Type	Description	Default
coords	ndarray	3D coordinates of the nodes.	required
scalars	ndarray	3D (LAT) scalars of the nodes.	required
indices	ndarray	Indices for each node. Defaults to None.	None

add_filter(filter) classmethod

Manually register a filter.

Parameters:

Name	Type	Description	Default
filter	GraphFilter	Filter name.	required

add_mask(name, mask)

Add new boolean mask of specific property.

Parameters:

Name	Type	Description	Default
name	str	Name of the mask to add.	required
mask	ndarray[bool]	Boolean mask to add.	required

apply_filter(*filter_)

Apply filter(s) to the graph. Available filters can be found in Graph.filters. You can pass either a single filter, a list of filters or multiple filters passed as separate arguments. A filter van be a string registered in Graph.filters`, or a filter class inheriting from GraphFilter. Combining both is also possible.

Parameters:

Name	Type	Description	Default
filter_	(str | list[str] | GraphFilter, list[GraphFilter])	The filter(s) to apply.	()

Returns:

Type	Description
Self	self

apply_filters(*filter_)

apply_period()

Apply the period of the graph to connect nodes with LAT x and lat x + period with each other.

Returns:

Type	Description
Self	self

get_adjacency_matrix()

Generate and return the adjacency matrix of the graph.

The adjacency matrix is a square 2D NumPy array of shape (nodes_number, nodes_number) that represents the connections between nodes in a graph. Each entry (i, j) in the matrix is a boolean value:

• True indicates there is an edge from node i to node j.
• False indicates no edge exists from node i to node j.

For example, if there are 3 nodes and edges connecting:

• Node 0 to Node 1
• Node 1 to Node 2
• Node 2 to Node 0

The adjacency matrix would look like this:

[[False  True  False]
 [False False  True]
 [ True False False]]

Returns:

Type	Description
ndarray	np.ndarray: A 2D boolean NumPy array of shape (nodes_number, nodes_number), representing the adjacency matrix of the graph.

get_connections_by_nodes(nodes_ids)

Returns the edge indices that connect to the nodes defined by node_ids. You can pass either a single ID or a list of IDs.

Parameters:

Name	Type	Description	Default
nodes_ids	(ndarray[int], int)	Node index/indices.	required

Returns:

Type	Description
ndarray	np.ndarray[int, int]: edge indices

get_degree()

Calculates the degree of the graph The total degree is, for each node in the graph, the number of directed edges going in and out of that node.

Returns:

Type	Description
ndarray	np.ndarray: the degree of the graph

get_degree_diff()

Calculates the degree difference of the graph. i.e. the difference between the in degree and out degree, or the difference between the number of directed edges going in and out of each node

Returns:

Type	Description
ndarray	np.ndarray: the degree difference

get_edges_mask(edges_ids)

get all connections defined by edge_ids You can pass either a single edge or a list of edges.

Parameters:

Name	Type	Description	Default
edges_ids	ndarray[int, int]	Edge index/indices.	required

Returns:

Type	Description
ndarray[bool]	np.ndarray[bool]: edge mask

get_in_degree()

Calculates the in degree of the graph. The in degree is, for each node in the graph, the number of directed edges going into that node.

Returns:

Type	Description
ndarray	np.ndarray: the in degree of the graph

get_maximal_clique_sizes()

A graph measure calculating clique sizes of all maximal cliques in a graph. A maximal clique is a subset of vertices in a graph that forms a clique (where every pair of vertices is directly connected) and cannot be extended by including any additional vertices.

Returns:

Type	Description
ndarray	np.ndarray: the maximal clique sizes

get_nodes_mask(nodes_ids)

Returns a boolean mask that when applied to the graph's edges would result in only the edges that are connected to the node(s) defined by the given node_ids. You can pass either a single ID or a list of IDs.

Parameters:

Name	Type	Description	Default
nodes_ids	(ndarray[int], int)	Node index/indices.	required

Returns:

Type	Description
ndarray[bool]	np.ndarray[bool]: Boolean mask for the edges

get_out_degree()

Calculates the out degree of the graph. The out degree is, for each node in the graph, the number of directed edges going into that node.

Returns:

Type	Description
ndarray	np.ndarray: the out degree of the graph

get_strong_components()

Calculates the sizes of the strongly connected graph components. A connected component is a subset of a graph's vertices where each pair of vertices is connected by a path; it is called strongly connected if the paths respect edge direction

Returns:

Type	Description
ndarray	np.ndarray: the strongly connected components sizes

get_weak_components()

Calculates the sizes of the weakly connected graph components. A connected component is a subset of a graph's vertices where each pair of vertices is connected by a path; it is called weakly connected if edge direction is ignored.

Returns:

Type	Description
ndarray	np.ndarray: the weakly connected components sizes

mask_connections_by_edges(edges_ids, mask_name='edge_mask')

Remove all connections defined by edge_ids You can pass either a single edge or a list of edges.

Parameters:

Name	Type	Description	Default
edges_ids	(ndarray[int], int)	Edge index/indices.	required
mask_name	str	Name of the mask.	'edge_mask'

mask_connections_by_nodes(nodes_ids, mask_name='node_mask')

Remove all connections that connect to the nodes defined by node_ids. You can pass either a single ID or a list of IDs.

Parameters:

Name	Type	Description	Default
nodes_ids	(ndarray[int], int)	Node index/indices.	required
mask_name	str	Name of the mask.	'node_mask'

merge(graphs, mode='or', compare_scalars=True)

Merge graphs into one graph. If no merging order is set, the merging will be performed in the order of the graphs appearance.

Parameters:

Name	Type	Description	Default
graphs	list	List of graph objects.	required
mode	str	'or' or 'and'. In mode 'or', all edges of both graphs are combined, in mode 'and', only edges that appear in both graphs are kept.	'or'
compare_scalars	bool	if True, only edges connecting to nodes with the same scalar value in both graphs are merged	True

Returns:

Type	Description
Self	self

remove_mask(name)

Remove a boolean mask by name. A new mask is constructed by taking the bitwise and between all remaining masks

Returns:

Type	Description
Self	self

reset_measures()

Resets the measures of the graph. This can be useful when these have been previously calculated, but after calculation, the graph properties were changed.

Returns:

Type	Description
Self	self

reset_properties()

Reset all graph properties.

Returns:

Type	Description
Self	self

set_cv_max(cv_max)

set_cv_min(cv_min)

set_cv_range(cv_min, cv_max)

set_delta_lat_range(min_delta_lat, max_delta_lat)

set_edge_length_range(min_edge_length, max_edge_length)

set_max_delta_lat(max_delta_lat)

set_max_edge_length(max_edge_length)

set_min_delta_lat(min_delta_lat)

set_min_edge_length(min_edge_length)

set_period(period)

to_df()

NKGraph

Base class for networkit-based graph algorithms.

const_weights = None instance-attribute

dgm_graph = dgm_graph instance-attribute

directed = False instance-attribute

nk_graph = None instance-attribute

weighted = False instance-attribute

__init__(dgm_graph)

build()

Build networkit graph based on dgm graph.

set_const_weights(const_weights)

Sets a constant weight for each edge in the graph.

Parameters:

Name	Type	Description	Default
const_weights	number	Constant weights.	required

set_directed()

Sets if the networkkit graph is directed.

set_number_of_threads(number_of_threads)

Set number of threads used by Networkit.

Parameters:

Name	Type	Description	Default
number_of_threads	int	Number of threads.	required

set_weighted()

Sets if the networkkit graph can have edge weights not equal to 1.

VoronoiGraph

Bases: Graph

triangles = triangles instance-attribute

vertices = vertices instance-attribute

__init__(coords, scalars, vertices, triangles, indices=None)

from_surface_poly(coords, scalars, surface_poly, indices=None) classmethod

Create a VoronoiGraph from a surface polygon.