Reading edge list from csv line

[vadhoob]

Hi,

I am new to graph ML and have been trying to play with node2vec for the past couple of days. I have tried several implementations of which and they all seem quite slow when it comes to generating the baised walk. I am keen to use the pecanpy implementation however my edge list is in a csv file (attached) and I am not clear how to convert it to .edg format pecanpy is expecting it to be.

I have tried looking around but haven't found anything that works. Any help would be greatly appreciated!

Cheers!

musae_git_edges.csv

[RemyLau]

Hi @vadhoob, thanks for your interest in using PecanPy! A recent patch #66 made by @justinlboyer allows one to specify the delimiter type. So essentially, you can pass your csv file directly to the PecanPy program by specifying --delimiter ,.

One note I want to make is that the current implementation can not exclude the harder line automatically. So you will need to manually remove the header before running PecanPy, either using a text editor or using the following command

$ tail -n +2 csv_with_header.csv > csv_without_header.csv

Finally, here's an example script you can use the run PecanPy with your csv file

$ pecanpy --input csv_without_header.csv --delimiter , --mode SparseOTF --output out.emd

[vadhoob]

thanks @RemyLau that worked! I am impressed at how quick it runs!

For anyone using the API, after removing the header, this is what I did:

g.read_edg('_edges_without_headers.csv', delimiter=',', weighted=False, directed=False)
My file was separated by ',' .

Cheers!

[RemyLau]

Sounds great! I'm glad that it worked out for you!