is it possible to split n2v to generate walks only?

[vhranger]

is it possible to split n2v to generate walks only?

[VHRanger]

Yeah, specify n2v = Node2Vec(keep_walks=True) in the node2vec object, then you have access to them in the object after fitting with n2v.walks

[amjass12]

Thank you, and is there a way if generating the walk without fitting the model? This would speed up the process even more, is there an easy way to do this?

[VHRanger]

You can do it from the CSRGraph object directly, but that returns the node IDs rather than the node names (even if node names are integers, the IDs aren't necessarily the same). So you'd have to map it back, here's a sketch of how to do it, taken from the Node2Vec fit method:

import csrgraph as cg
G =  cg.csrgraph(INPUT_DATA)
walks = G.random_walks()
# Make into dataframe to map ID to names
walks = pd.DataFrame(walks)
# mapping dict node ID -> node name
node_dict = dict(zip(np.arange(len(G.names)), G.names))
for col in walks.columns:
      walks[col] = walks[col].map(node_dict).astype(str)

[amjass12]

Thank you! This is a good workaround. Appreciate your time