Closed coleslaw481 closed 2 years ago
After doing some benchmarking the numpy.load/save
was pretty slow and generated giant files (multiple gigabyte) so just decided to go with documentation for the load/save, but did update the function to take a graph and internally call the get noramlized adjacency matrix
For easier use the following changes should be made:
Update docs to say
netprop.get_normalized_adjacency_matrix()
is no longer needed cause the networkx graph can be passed directly tonetprop.get_individual_heats_matrix()
Updates to
netprop.get_individual_heats_matrix()
w_prime
is a networkx graph. If it is a networkx graph then pass that graph tonetprop.get_normalized_adjacency_matrix()
to generate thew_prime
matrixconserve_heat
named parameter to this function in casew_prime
needs to be generatedcache_dir
named parameter with default value ofNone
cache_dir
is set to a directory then the code should look for cached versions of the matrix files and use those and if they do not exist the code should generate those filesnetcoloc_w_double_prime.npz
is thew_double_prime
filenetcoloc_w_prime.npz
is thew_prime
file