goulib.graph

efficient Euclidian Graphs for networkx and related algorithms

requires:

• networkx

• matplotlib

optional:

• scipy for delauney triangulation

• rtree for faster GeoGraph algorithms

Functions

delauney_triangulation(nodes[, ...])	https://en.wikipedia.org/wiki/Delaunay_triangulation :param nodes: _Geo graph or list of (x,y) or (x,y,z) node positions :param qhull_options: string passed to scipy.spatial.Delaunay(), which passes it to Qhull ( http://www.qhull.org/ ) *'Qt' ensures all points are connected *'Qz' required when nodes lie on a sphere *'QJ' solves some singularity situations
draw_networkx(g[, pos])	improves nx.draw_networkx :param g: NetworkX Graph :param pos: can be either :
euclidean_minimum_spanning_tree(nodes, **kwargs)
figure(g[, box])
points_on_sphere(N)
read_json(filename[, directed, multigraph, ...])
to_drawing(g[, d, edges])	draws Graph to a Drawing :param g: Graph :param d: existing Drawing to draw onto, or None to create a new Drawing :param edges: iterable of edges (with data) that will be added, in the same order.
to_json(g, **kwargs)
to_networkx_graph(data[, create_using, ...])	Make a NetworkX graph from a known data structure.
write_dot(g, filename)
write_dxf(g, filename)	writes networkx.Graph in .dxf format
write_json(g, filename, **kwargs)	write a JSON file, suitable for D*.js representation

Classes

AGraph()
DiGraph([data, nodes])	directed graph with nodes positions can be set to non multiedges anytime with attribute multi=False
GeoGraph([data, nodes])	Undirected graph with nodes positions can be set to non multiedges anytime with attribute multi=False
_Geo(parent[, data, nodes])	base class for graph with nodes at specified positions.
index()