chemicalchecker.tool.node2vec.node2vec.Node2Vec

class Node2Vec(executable='node2vec', **kwargs)

Bases: object

Wrapper to run SNAP’s node2vec.

Check if executable is found.

Methods

emb_to_h5	Convert from native node2vec format to familiar h5.
heuristic_max_degree	Return the maximum degree.
merge_edgelists	Convert Nearest-neighbor to an edgelist.
run	Call external exe with given parameters.
split_edgelist	Split a graph in train and test.
to_edgelist	Convert Nearest-neighbor to an edgelist.

emb_to_h5(keys, in_file, out_file)

Convert from native node2vec format to familiar h5.

We need to map back from sign1 ids to inchikeys and sort.

static heuristic_max_degree(nr_nodes, max_mbytes=6000.0)

Return the maximum degree.

The heuristic is based on the limiting factor that is node2vec’s memory footprint. To precompute transition probabilities, node2vec needs 12 bytes for each triplet node-neighbor-neighbor.

Parameters:

• nr_nodes (int) – Number of nodes the network will contain.

• max_mbytes (float) – Maximum RAM consumption to allow (in MBs).

merge_edgelists(signs, edgefiles, out_file, **params)

Convert Nearest-neighbor to an edgelist.

Parameters:

• signs (list) – List of signature objects.

• edgefiles (list) – List of edge files.

• out_file (str) – Destination file.

• params (dict) – Parameters defining similarity network.

run(i, o, **kwargs)

Call external exe with given parameters.

Parameters:

• i – Input graph path (default:’graph/karate.edgelist’)

• o – Output graph path (default:’emb/karate.emb’)

• d – Number of dimensions. Default is 128 (default:128)

• l – Length of walk per source. Default is 80 (default:80)

• r – Number of walks per source. Default is 10 (default:10)

• k – Context size for optimization. Default is 10 (default:10)

• e – Number of epochs in SGD. Default is 1 (default:1)

• p – Return hyperparameter. Default is 1 (default:1)

• q – Inout hyperparameter. Default is 1 (default:1)

• output. (v Verbose) –

• directed. (dr Graph is) –

• weighted. (w Graph is) –

• embeddings. (ow Output random walks instead of) –

split_edgelist(full_graph, train, test, train_fraction=0.8)

Split a graph in train and test.

Give a Network object split it in two sets, train and test, so that train has train_fraction of edges for each node.

to_edgelist(sign1, neig1, out_file, **params)

Convert Nearest-neighbor to an edgelist.

Parameters:

• sign1 (str) – Signature 1, h5 file path.

• neig1 (str) – Nearest neighbors 1, h5 file path.

• out_file (str) – Destination file.

• params (dict) – Parameters defining similarity network.