Basic graph generation using hypothesis, compare outputs with NX

[MridulS]

To extend tests https://github.com/python-graphblas/graphblas-algorithms/issues/10 Using hypothesis to generate a combinations of graphs to compare the outputs of nx implementation and graphblas ones. I'm not a 100% sure if this is the best way of doing this (this is the first time I'm writing hypothesis tests) and the coverage would be pretty poor as there are a lot of assumptions. Currently the graph generator gives a sample of graphs with the toggles:

• directed: bool generate a directed graph
• self_loops: bool to control if self edges are allowed
• sym_digraph: bool to control the symmetric digraphs.
• edges: generates a 3-tuple of edge data, all our integers
• edge_data: bool to control if the edge should be weighted or not.