How to identify and remove dead end pores and isolated clusters after using reduce_coordination()

[harshpallam]

I wanted to generate a random network with a mean coordination number of 3.9.
The initial size of the cubic network is 51 * 20 * 20 nodes in x,y, and z directions with a coordination number of 26 for all pore bodies.
I used reduce_coordination() to reduce the mean connectivity to 3.9. However, after coordination number reduction, I see that 20400 pores are still present and none of the pores are eliminated. When I tried to trim the pores which are not connected to the network and used check_network_health(), it showed an empty set. Please see the attached image. Does it mean that there are no isolated nodes or clusters present in the network?
I request you to kindly help me with this.

import os
import warnings
import numpy as np
import openpnm as op
import scipy as sp
import random
import matplotlib.pyplot as plt
import openpnm.models as mods
import sys
# import openpnm.physics.models as pm
np.random.seed(11)
ws = op.Workspace()
ws.settings['loglevel'] = 40
warnings.filterwarnings('ignore')
#######################################################################
#1 setting up network
pn = op.network.Cubic(shape=[50, 20, 20], spacing=[0.001, 0.001, 0.001], connectivity=26) #add_boundary_pores=['left', 'right']
op.io.VTK.save(pn, filename="hp")
op.io.CSV.save(pn, filename="hp")
print(pn.am)
print(pn.conns)
print(pn.coords)
print(pn.im)
print("No of throats before elem =",pn.num_throats())
print("No of pores before elem =",pn.num_pores())
print(f"Does 'hp.vtk' exist? {os.path.isfile('hp.vtp')}")
print("A Cubic Lattice is created!")
#Elimination of network
op.topotools.reduce_coordination(network=pn, z=3.9)
print(pn['throat.conns'])
#checking the pore connectivity
h = pn.check_network_health()
op.topotools.trim(network=pn, pores=h['trim_pores'])
h = pn.check_network_health()
print(h)
op.io.VTK.save(pn, filename="hp_r")
op.io.CSV.save(pn, filename="hp_r")
print("No of throats after elem =",pn.num_throats())
print("No of pores after elem =",pn.num_pores())

[jgostick]

Isolate clusters is easy...these should show up on the check_network_health function. Deadend pores are not technically a problem for the simulations, so we consider those as healthy. However, if you really don't want deadend pores you can do Ps = pn.num_neighbors(pores=np.Ps) which will return a list of the number of neighbors to each pore. You can then trim any pores with only 1 neighbor (op.topotools.trim(pn, pores=Ps==1). You'll need to do this several times since you may create new deadend pores upon deletion of the previous ones (as would happen in the case of a chain of pores leading to a dead end).

[jgostick]

Oh, and I should also mention that trim coordination does not remove any pores, only throats, AND it does it in a way that maintains connectivity. It finds the "minimum spanning tree" of the network (assuming random throat weights), then only deletes throats that are NOT part of the spanning tree.

[harshpallam]

Thanks a lot, sir. It worked. Ps = pn.num_neighbors(pores=pn.Ps)