wirewalk 0.0.2

Python3 Implementation of the WireWalk Algorithm

wirewalk

Install

pip install wirewalk

Prerequisites

You will need

• Python3
• Networkx
• Numpy
• Gensim
• editdistance

I highly recommend installing an Anaconda environment. Future versions of WireWalk will be available on PyPI and conda.

How to use

import networkx as nx
from wirewalk.core import WireWalk, jaccard_coefficient, max_flow

# Create a graph
graph = nx.fast_gnp_random_graph(n=10, p=0.5)

# Instantiate a WireWalk object
wireWalk = WireWalk(graph, dimensions = 128, window = 10, walk_length = 80, num_walks = 10, workers = 1)

# Compute transition probabilities using jaccard coefficient transformation, generate walks, and embed nodes
model = wireWalk.fit(jaccard_coefficient)

# **MAX_FLOW and MIN_COST_MAX_FLOW ONLY WORK WITH GIVEN capacity**
# If weight exists, then  
# nx.set_edge_attributes(graph, nx.get_edge_attributes(graph, "weight"), "capacity").
# Otherwise,
nx.set_edge_attributes(graph, 1, "capacity")
model = wireWalk.fit(max_flow)