rnaglib.algorithms.Hasher

class rnaglib.algorithms.Hasher(method='WL', string_hash_size=8, graphlet_hash_size=16, symmetric_edges=True, wl_hops=2, label='LW', directed=True)

The hasher object. Once created, it will hash new graphs and optionnaly, one can run it onto a whole graph dir to create the hashtable once and for all.

Parameters:

• method – for now only WL hashing is supported

• string_hash_size – The length of the hash, longer ones will yields less collision but more processing time

• graphlet_hash_size – Same thing for hashes of graphlets

• symmetric_edges – Whether to use symetric weights for the edges.

• wl_hops – The depth to consider for hashing

• label – How the data for the surrounding edges is encoded in the nodes.

• directed – To use directed graphs instead of undirected ones

__init__(method='WL', string_hash_size=8, graphlet_hash_size=16, symmetric_edges=True, wl_hops=2, label='LW', directed=True)

The hasher object. Once created, it will hash new graphs and optionnaly, one can run it onto a whole graph dir to create the hashtable once and for all.

Parameters:

• method – for now only WL hashing is supported

• string_hash_size – The length of the hash, longer ones will yields less collision but more processing time

• graphlet_hash_size – Same thing for hashes of graphlets

• symmetric_edges – Whether to use symetric weights for the edges.

• wl_hops – The depth to consider for hashing

• label – How the data for the surrounding edges is encoded in the nodes.

• directed – To use directed graphs instead of undirected ones

Methods

__init__([method, string_hash_size, ...])	The hasher object.
get_hash_table(graph_dir[, max_graphs])
get_node_hash(graph, n)	Get the correct node hashing from a node and a graph
hash(graph)	WL hash of a graph.