astrovascpy.utils¶

Functions

`comm`()	returns: The MPI communicator (mpi4py.MPI.Comm)
`convert_from_networkx`(nx_graph)	Convert a networkx graph to a vasculature graph.
`convert_to_networkx`(graph)	Convert a vasculatureAPI graph to a networkx graph (mostly for plotting purposes).
`convert_to_temporal`(data, frequencies, times)	Convert spectral data to temporal data.
`create_box`(graph)	Create a box to compute metrics inside.
`create_entry_largest_nodes`(graph, params)	Get the largest nodes of degree 1 for input in the largest connected components.
`create_input_speed`(T, step[, A, f, C, ...])	Creates an input speed v(t) according to the model:
`ensure_ids`(a)	Convert a numpy array dtype into IDS_DTYPE.
`ensure_list`(v)	Convert iterable / wrap scalar into list (strings are considered scalar).
`find_degrees_of_neighbors`(graph, node_id)	Given an edge described by node_id, find the neighbors edges.
`find_neighbors`(graph, section_id, segment_id)	Given an edge described by section_id and segment_id, find the neighbors edges.
`fit_sine_model`(signal[, window])	Helper function for fitting a sine signal.
`get_large_nodes`(graph[, min_radius, ...])	Get degree 1 nodes that are larger than min_radius.
`get_largest_nodes`(graph[, n_nodes, ...])	Get largest nodes of degree 1 for input in the largest connected components.
`get_local_neighboors`(graph, iterations, ...)	Get local neighbours.
`get_subset`(graph[, iterations, ...])	Get subset of graph as connected component from a starting node, using random walk.
`is_iterable`(v)	Check if v is any iterable (strings are considered scalar and CircuitNode/EdgeId also).
`mpi`()	returns: The mpi4py.MPI module
`rank`()	returns: MPI rank of the current process (int)
`rank0`()	returns: True if the current process has MPI rank 0, False otherwise
`reduce_graph`(graph, to_keep_labels)	Return a subgraph with to_keep_labels.
`reduce_to_largest_cc`(graph)	Return a new graph with only the largest CC.
`size`()	returns: The size of the MPI communicator (int)

Classes

`Graph`(args, *kwargs)	Wrapper on top of a vascpy.PointVasculator providing additional graph-related capabilities
`mpi_mem`()	A simple mpi memory profiler class
`mpi_timer`()	A simple mpi timer class

class astrovascpy.utils.Graph(*args, **kwargs)¶

Bases: PointVasculature

Wrapper on top of a vascpy.PointVasculator providing additional graph-related capabilities

Concerns: This class is part of the public API. Any change of signature or functional behavior may be done thoroughly.

property cc_mask¶

ids (numpy.array) of the main connected component.

Computation is made once, next calls for this property will returned the cached result.

Type:: Returns

property degrees¶

degrees (numpy.array) of each node

Computation is made once, next calls for this property will returned the cached result.

Type:: Returns

classmethod from_point_vasculature(point_vasculature)¶

Instantiate a Graph from a PointVasculature instance

Parameters:: graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.

get_main_connected_component()¶

Build a graph with only the largest Connected Component (CC).

Returns: largest CC point graph (Graph)

astrovascpy.utils.comm()¶

Returns:: The MPI communicator (mpi4py.MPI.Comm)

astrovascpy.utils.convert_from_networkx(nx_graph)¶

Convert a networkx graph to a vasculature graph.

Parameters:: nx_graph (networkx graph) – input networkx graph
Returns:: graph containing point vasculature skeleton.
Return type:: Graph

astrovascpy.utils.convert_to_networkx(graph)¶

Convert a vasculatureAPI graph to a networkx graph (mostly for plotting purposes).

Parameters:: graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
Returns:: it returns a graph.
Return type:: networkx graph

astrovascpy.utils.convert_to_temporal(data, frequencies, times)¶: Convert spectral data to temporal data.

astrovascpy.utils.create_box(graph)¶

Create a box to compute metrics inside.

Parameters:: graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
Returns:: edge index to filter edges.
Return type:: pandas.Series

astrovascpy.utils.create_entry_largest_nodes(graph: Graph, params: VasculatureParams)¶

Get the largest nodes of degree 1 for input in the largest connected components.

Parameters:

graph (Graph) – graph containing point vasculature skeleton.
params (VasculatureParams) – general parameters for vasculature.

Returns:

(1,) Ids of the largest nodes.

Return type:

numpy.array

Raises:

BloodFlowError – if n_nodes <= 0 or if vasc_axis is not 0, 1 or 2.

Concerns: This function is part of the public API. Any change of signature or functional behavior may be done thoroughly.

astrovascpy.utils.create_input_speed(T, step, A=1, f=1, C=0, read_from_file=None)¶

Creates an input speed v(t) according to the model:

v(t) = A sin( 2 pi f t ) + C

Parameters:

T (float) – simulation time
step (float) – time step size
A (float) – sine amplitude.
f (float) – sine frequency.
C (float) – offset.
read_from_file (path) – path to the data file to read.

Returns:

vector of speed for any time point.

Return type:

numpy.array

astrovascpy.utils.ensure_ids(a)¶

Convert a numpy array dtype into IDS_DTYPE.

This function is here due to the https://github.com/numpy/numpy/issues/15084 numpy issue. It is quite unsafe to the use uint64 for the ids due to this problem where : numpy.uint64 + int –> float64 numpy.uint64 += int –> float64

This function needs to be used everywhere node_ids or edge_ids are returned.

astrovascpy.utils.ensure_list(v)¶: Convert iterable / wrap scalar into list (strings are considered scalar).

astrovascpy.utils.find_degrees_of_neighbors(graph, node_id)¶

Given an edge described by node_id, find the neighbors edges.

Parameters:

graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
node_id (int) – id of the selected node

Returns:

neighbors_mask to filter edges with 1 node in common with the original edge set: nodes ids of the connected nodes to the node id. numpy.array: list of nodes close to the original node id

Return type:

pandas.Series

astrovascpy.utils.find_neighbors(graph, section_id, segment_id)¶

Given an edge described by section_id and segment_id, find the neighbors edges.

Parameters:

graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
section_id (int) – id of the corresponding section.
segment_id (int) – id of the corresponding segment.

Returns:

neighbors_mask to filter edges with 1 node in common with the original edge.

Return type:

pandas.Series

astrovascpy.utils.fit_sine_model(signal, window=None)¶

Helper function for fitting a sine signal.

Computes the parameters [A,f,C] of the model:

f(t) = A sin( 2 pi f t ) + C

Parameters:

signal (numpy.array) – sinusoidal signal.
window (float) – window for smoothing the signal

Returns:

sine amplitude. f (float): sine frequency. C (float): offset.

Return type:

A (float)

astrovascpy.utils.get_large_nodes(graph, min_radius=6, depth_ratio=1.0, vasc_axis=1)¶

Get degree 1 nodes that are larger than min_radius.

Parameters:

graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
min_radius (float) – minimal radius in µm.
depth_ratio (float) – corresponding to the y-axis (depth of points).
vasculature. (Portion of the)
vasc_axis (int) – axis along which we consider the top/bottom of vasculature.

Returns:

(ids_largest_nodes,) Ids of the largest nodes. Node ids are sorted according to their diameters.

Return type:

numpy.array

Raises:

BloodFlowError – if min_radius <= 0 or if vasc_axis is not 0, 1 or 2.

astrovascpy.utils.get_largest_nodes(graph, n_nodes=1, depth_ratio=1.0, vasc_axis=1)¶

Get largest nodes of degree 1 for input in the largest connected components.

Parameters:

graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
n_nodes (int) – number of nodes to return
depth_ratio (float) – corresponding to the y-axis (depth of points).
vasculature. (Portion of the)
vasc_axis (int) – axis along which we consider the top/bottom of vasculature.

Returns:

(1,) Ids of the largest nodes.

Return type:

numpy.array

Raises:

BloodFlowError – if n_nodes <= 0 or if vasc_axis is not 0, 1 or 2.

astrovascpy.utils.get_local_neighboors(graph, iterations, starting_nodes_id)¶

Get local neighbours.

Parameters:

graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
iterations (int) – number of iterations.
starting_nodes_id (int) – start node id.

Returns:

get local neighbourg of selected nodes ids.

Return type:

numpy.array

astrovascpy.utils.get_subset(graph, iterations=100, starting_nodes_id=None)¶

Get subset of graph as connected component from a starting node, using random walk.

Parameters:

graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
iterations (int) – number of iterations.
starting_nodes_id (int) – start node id.

Returns:

reduced to the largest cc point graph.

Return type:

Graph

astrovascpy.utils.is_iterable(v)¶: Check if v is any iterable (strings are considered scalar and CircuitNode/EdgeId also).

astrovascpy.utils.mpi()¶

Returns:: The mpi4py.MPI module

class astrovascpy.utils.mpi_mem¶

Bases: object

A simple mpi memory profiler class

class astrovascpy.utils.mpi_timer¶

Bases: object

A simple mpi timer class

astrovascpy.utils.rank()¶

Returns:: MPI rank of the current process (int)

astrovascpy.utils.rank0()¶

Returns:: True if the current process has MPI rank 0, False otherwise

astrovascpy.utils.reduce_graph(graph, to_keep_labels)¶: Return a subgraph with to_keep_labels.

astrovascpy.utils.reduce_to_largest_cc(graph)¶

Return a new graph with only the largest CC.

Parameters:: graph (vasculatureAPI.PointVasculature) – graph containing point vasculature skeleton.
Returns:: reduced to the largest cc point graph.
Return type:: vasculatureAPI.PointVasculature

astrovascpy.utils.size()¶

Returns:: The size of the MPI communicator (int)