Extracting the network using porespy

[sara-madi]

I generated the artificial image coded in the "snow_advanced" and extracted the network using PoreSpy. I then used pickle to save the extracted network python dictionary as follows:

import pickle
dict_file = open('temp.dict', 'wb')
pickle.dump(net, dict_file)
dict_file.close()

After, I used OpenPNM to load the network:

prj = op.io.PoreSpy.load('temp')
net = prj.network
op.io.XDMF.save(network=net, filename="out")
print(prj.geometries())
geo = prj.geometries()["geo_01"]
net_health = net.check_network_health()
op.topotools.trim(network=net, pores=net_health["trim_pores"])
air = op.phases.Air(network=net)
phys_air = op.physics.Standard(network=net, phase=air, geometry=geo)
perm = op.algorithms.StokesFlow(network=net)
perm.setup(phase=air)

However, when I am assigning the boundary conditions, I am getting the following error:

Traceback (most recent call last):
File "", line 1, in
File "c:\users\legion\desktop\sara\openpnm-dev\openpnm\core\Base.py", line 856, in pores
ind = self._get_indices(element='pore', labels=labels, mode=mode)
File "c:\users\legion\desktop\sara\openpnm-dev\openpnm\core\Base.py", line 747, in _get_indices
union = union + self[element+'.'+item.split('.')[-1]]
File "c:\users\legion\desktop\sara\openpnm-dev\openpnm\network\GenericNetwork.py", line 155, in getitem
vals = super().getitem(key)
File "c:\users\legion\desktop\sara\openpnm-dev\openpnm\core\Base.py", line 272, in getitem
raise KeyError(key)

I think the problem is that openpnm can't find the pores(left, right,..). How can I solve this issue?

[jgostick]

You try to do print(net) and see what labels are available on your extracted network. You can add your own boundary pores using this example

[sara-madi]

My labels are as follows:

Labels Assigned Locations
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
1 pore.all 668
2 throat.all 3268
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
So, I should divide the total pores into other labels defined as: pore.left, pore.right, pore.bottom, pore.top, pore.front, pore.back.
I don't think the example you provided(using clone) will allow me to create these new labels.

[jgostick]

You can have labels applied during the extraction process...I think they should be applied by default actually, so I'm surprized they're not here. You can also add them to the network using spatial info, so:

Ps = net['pore.coords'][:, 0] < lower_lim
net.set_label(pores=Ps, label='pore.left')
Ps = net['pore.coords'][:, 0] > upper_lim
net.set_label(pores=Ps, label='pore.right')

Where upper_lim and lower_lim are physical distances that you consider to be 'on the right or left face' of the domain. For instance:

lower_lim = pn['pore.coords'][:, 0].min() * 1.1
upper_lim = pn['pore.coords'][:, 0].min() * 0.9`

[sara-madi]

The "Snow_Advanced" extracted network does not give the labels of the pores; however, the older version of snow algorithm given the following example

https://github.com/PMEAL/OpenPNM/blob/master/examples/extractions/Working%20with%20Extracted%20Networks.ipynb

extracts the network specifying the labels of the pores as left, right,...