[FIXED] NetworkX: list of directed edges in a triad census

Issue

In NetworkX, I can use the nx.triadic_census(g) to give me a list of triads in my network.

Out:
{'003': 1434217,
 '012': 282025,
 '102': 32640,
 '021D': 21,
 '021U': 267246,
 '021C': 445}

I wanted to know how I can list the directed edges belonging to a triad. For example, in Pyspark, to list the nodes/edges in triad "102" : g.find("(a)-[e]->(b); (b)-[e2]->(a)").show()

and it will give you

|                   a|            e|                   b|            e2|
+--------------------+-------------+--------------------+--------------+
|[US, United State...|  [US, AZ, 0]|    [AZ, Azerbaijan]| [AZ, US, 637]|
|    [LU, Luxembourg]|[LU, BE, 213]|       [BE, Belgium]|[BE, LU, 1470]|
|       [FI, Finland]| [FI, CZ, 24]|[CZ, Czech Republic]|  [CZ, FI, 51]|
|       [HU, Hungary]|  [HU, PL, 0]|        [PL, Poland]| [PL, HU, 231]|
|[RU, Russian Fede...|  [RU, UA, 0]|       [UA, Ukraine]|   [UA, RU, 0]|

Is there a way for me to do this in NetworkX?

Solution

There is no function in networkx that allow you to do it, so you should implement it manually. I modified the networkx.algorithms.triads code to return triad nodes, not their count. You can find it here. It can be modified with replacing tuple in census['...'].add(tuple(sorted([u, v, w]))) lines to add edges instead of nodes.

Answered By - vurmux