Read MATLAB graph objects into Python

MATLAB has a representation of a directed/undirected graph. I would like to deserialize a graph serialized via MATLAB's save function into Python. I know about scipy.io.loadmat and h5py's File, but neither seems to produce a representation in Python that actually holds intelligible vertex/edge data.

How do I do this? I'm concerned with this and the inverse operation, i.e writing an object from Python to a format MATLAB load can read. Is there a bytewise data description of what a serialized MATLAB objects and a Graph looks like somewhere?

For example, in MATLAB I could:

    s = [1 1 2 2 3];
    t = [2 4 3 4 4];
    G = digraph(s,t);
    G.Edges.Rand = rand(size(G.Edges)); % Add an edge attribute
    G.Nodes.Val = rand(size(G.Nodes)); % Add an a node attribute

    save('loadmat.mat', 'G'); % Readable by scipy.io.loadmat
    save('h5py.mat', 'G', '-v7.3'); % Readable by h5py.File

then, in Python I could read these

    from scipy.io import loadmat
    G0 = loadmat('loadmat.mat')

    from h5py import File
    G1 = File('h5py.mat')

Thanks

Answer 1

[Christine Tobler]

I don't know much about python loadmat and h5py files, but I don't think loading the MATLAB object graph would realistically work. The classes graph/digraph are described in MATLAB code which will not be available in python, so even if you could load the internal data, this would not give you access to any of the graph object's methods and properties in python.

I would suggest that instead, you could save the edge list or the adjacency matrix of the graph in a .mat file:

[s, t] = findedge(g); M = adjacency(g);

Whichever python package you are using to represent graphs is likely to be able to construct a graph from one or both of these.

[disa]

I have attributed graphs so this solution won't work well for me. MATLAB is fundamentally bindings from a lower level language so the data should be binary readable. Also, loadmat works for matrix structs. Scipy figured out how to parse a serialized matlab struct. I'm fairly certain matlab save writes a header then data, but I need a bytewise description to write something to read the binary. Any body got that??

[Christine Tobler]

Unfortunately, I don't have a bytewise description, as I don't know what mat-serialization does internally. To just get all the graph attributes into python, you might try doing

edgeStruct = table2struct(g.Edges);
nodeStruct = table2struct(g.Nodes);

These two structs should contain all information in the graph object, and should be parseable as you mentioned.