Edmond's optimal branching algorithm in C++ wrapped by Python. Original version is from https://github.com/xiaohan2012/pyedmond.
As it's in C++ internally, it's faster and more memory-efficient than that of NetworkX.
import numpy as np
from pyedmond import find_minimum_branching
n_nodes, edges = 10, []
weights = np.abs(np.random.rand(n_nodes, n_nodes))
for u in range(n_nodes):
for v in range(n_nodes):
if u != v:
edges.append((u, v, weights[u, v]))
edges = find_minimum_branching(n_nodes, edges, roots=[0, 1]) # returns a list of (int, int) edges- Boost 1.34 or later
- Python 3.7 or later
pip install git+https://github.com/kmkurn/pyedmond.git#egg=pyedmond
MIT
Copyright (c) 2017 Han Xiao, 2019 Kemal Kurniawan.