GenericNetwork¶
- class GenericNetwork(*args, **kwargs)[source]¶
Bases:
openpnm.core.Base.Base
,openpnm.core.ModelsMixin.ModelsMixin
This generic class contains the main functionality used by all networks
- Parameters
coords (array_like) – An Np-by-3 array of [x, y, z] coordinates for each pore.
conns (array_like) – An Nt-by-2 array of [head, tail] connections between pores.
Notes
The GenericNetwork class houses a number of methods used for querying and managing the network’s spatial and topological information. The following table gives a very short overview of the methods added those already found on the
openpnm.core.Base
class.Method or Attribute
Functionality
create_adjacency_matrix
Create an adjacency matrix using given weights in a specified format
create_incidence_matrix
Create an incidence matrix using given weights in a specified format
get_adjacency_matrix
Retrieve an existing adjacency matrix in the specified format (from
am
)get_incidence_matrix
Retrieve an existing incidence matrix in the specified format (from
im
)am
Returns the adjacency matrix in COO format
im
Returns the incidence matrix in COO format
find_neighbor_pores
For a given set of pores, find all neighboring pores
find_neighbor_throats
For a given set of pores, find all neighboring throats
find_connecting_throat
For each pair of throats find the pores they connect
find_connected_pores
For each throat, find the pores which it connects
num_neighbors
For a given set of pores find the number of neighbors for each
find_nearby_pores
For a given set of pores, find pores that are within a certain distance
check_network_health
Check the topology for any problems such as isolated pores
Examples
>>> import openpnm as op
Create some pore coordinates and connections manually and assign to a GenericNetwork instance. Consider a linear network of 4 pores and 3 throats:
0 ―― 1 ―― 3 ―― 2
>>> coords = [[0, 0, 0], [1, 0, 0], [2, 0, 0], [3, 0, 0]] >>> conns = [[0, 1], [1, 3], [2, 3]] >>> pn = op.network.GenericNetwork(conns=conns, coords=coords)
Networks have two required properties: ‘pore.coords’ and ‘throat.conns’. These arrays indicate the spatial location of each pore, and which pores are connected to which. Without these the Network object cannot function.
>>> print(pn.props()) ―――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――― 1 : pore.coords 2 : throat.conns ――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
The GenericNetwork class has several methods for querying the topology.
>>> Ps = pn.find_neighbor_pores(pores=1) >>> print(Ps) [0 3] >>> Ts = pn.find_neighbor_throats(pores=[0, 1]) >>> print(Ts) [0 1] >>> print(pn.num_neighbors(2)) [1]
All of the topological queries are accomplished by inspecting the adjacency and incidence matrices. They are created on demand, and are stored for future use to save construction time.
- property am¶
Returns an adjacency matrix in the specified sparse format, with throat IDs indicating the non-zero values.
- Parameters
fmt (string, optional) –
The sparse storage format to return. Options are:
’coo’ : (default) This is the native format of OpenPNM data
’lil’ : Enables row-wise slice of the matrix
’csr’ : Favored by most linear algebra routines
’dok’ : Enables subscript access of locations
Notes
This method will only create the requested matrix in the specified format if one is not already saved on the object. If not present, this method will create and return the matrix, as well as store it for future use.
To obtain a matrix with weights other than throat IDs at each non-zero location use
create_adjacency_matrix
.To obtain the non-directed graph, with only upper-triangular entries, use
sp.sparse.triu(am, k=1)
.
- check_network_health()[source]¶
This method check the network topological health by checking for:
Isolated pores
Islands or isolated clusters of pores
Duplicate throats
Bidirectional throats (ie. symmetrical adjacency matrix)
Headless throats
- Returns
A dictionary containing the offending pores or throat numbers under
each named key.
It also returns a list of which pores and throats should be trimmed
from the network to restore health. This list is a suggestion only,
and is based on keeping the largest cluster and trimming the others.
Notes
Does not yet check for duplicate pores
Does not yet suggest which throats to remove
This is just a ‘check’ and does not ‘fix’ the problems it finds
- create_adjacency_matrix(weights=None, fmt='coo', triu=False, drop_zeros=False)[source]¶
Generates a weighted adjacency matrix in the desired sparse format
- Parameters
weights (array_like, optional) –
An array containing the throat values to enter into the matrix (in graph theory these are known as the ‘weights’).
If the array is Nt-long, it implies that the matrix is symmetric, so the upper and lower triangular regions are mirror images. If it is 2*Nt-long then it is assumed that the first Nt elements are for the upper triangle, and the last Nt element are for the lower triangular.
If omitted, ones are used to create a standard adjacency matrix representing connectivity only.
fmt (string, optional) –
The sparse storage format to return. Options are:
’coo’ : (default) This is the native format of OpenPNM data
’lil’ : Enables row-wise slice of the matrix
’csr’ : Favored by most linear algebra routines
’dok’ : Enables subscript access of locations
triu (boolean (default is
False
)) – IfTrue
, the returned sparse matrix only contains the upper- triangular elements. This argument is ignored if theweights
array is 2*Nt-long.drop_zeros (boolean (default is
False
)) – IfTrue
, applies theeliminate_zeros
method of the sparse array to remove all zero locations.
- Returns
- Return type
An adjacency matrix in the specified Scipy sparse format.
Notes
The adjacency matrix is used by OpenPNM for finding the pores connected to a give pore or set of pores. Specifically, an adjacency matrix has Np rows and Np columns. Each row represents a pore, containing non-zero values at the locations corresponding to the indices of the pores connected to that pore. The
weights
argument indicates what value to place at each location, with the default being 1’s to simply indicate connections. Another useful option is throat indices, such that the data values on each row indicate which throats are connected to the pore.Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[5, 5, 5]) >>> weights = np.random.rand(pn.num_throats(), ) < 0.5 >>> am = pn.create_adjacency_matrix(weights=weights, fmt='csr')
- create_incidence_matrix(weights=None, fmt='coo', drop_zeros=False)[source]¶
Creates a weighted incidence matrix in the desired sparse format
- Parameters
weights (array_like, optional) – An array containing the throat values to enter into the matrix (In graph theory these are known as the ‘weights’). If omitted, ones are used to create a standard incidence matrix representing connectivity only.
fmt (string, optional) –
The sparse storage format to return. Options are:
’coo’ : (default) This is the native format of OpenPNMs data
’lil’ : Enables row-wise slice of the matrix
’csr’ : Favored by most linear algebra routines
’dok’ : Enables subscript access of locations
drop_zeros (boolean (default is
False
)) – IfTrue
, applies theeliminate_zeros
method of the sparse array to remove all zero locations.
- Returns
- Return type
An incidence matrix in the specified sparse format
Notes
The incidence matrix is a cousin to the adjacency matrix, and used by OpenPNM for finding the throats connected to a give pore or set of pores. Specifically, an incidence matrix has Np rows and Nt columns, and each row represents a pore, containing non-zero values at the locations corresponding to the indices of the throats connected to that pore. The
weights
argument indicates what value to place at each location, with the default being 1’s to simply indicate connections. Another useful option is throat indices, such that the data values on each row indicate which throats are connected to the pore, though this is redundant as it is identical to the locations of non-zeros.Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[5, 5, 5]) >>> weights = np.random.rand(pn.num_throats(), ) < 0.5 >>> im = pn.create_incidence_matrix(weights=weights, fmt='csr')
- find_connected_pores(throats=[], flatten=False, mode='union')[source]¶
Return a list of pores connected to the given list of throats
- Parameters
throats (array_like) – List of throats numbers
flatten (boolean, optional) – If
True
(default) a 1D array of unique pore numbers is returned. IfFalse
each location in the the returned array contains a sub-arras of neighboring pores for each input throat, in the order they were sent.mode (string) –
Specifies logic to filter the resulting list. Options are:
’or’ : (default) All neighbors of the input throats. This is also known as the ‘union’ in set theory or ‘any’ in boolean logic. Both keywords are accepted and treated as ‘or’.
’xor’ : Only neighbors of one and only one input throat. This is useful for finding the sites that are not shared by any of the input throats.
’xnor’ : Neighbors that are shared by two or more input throats. This is equivalent to finding all neighbors with ‘or’, minus those found with ‘xor’, and is useful for finding neighbors that the inputs have in common.
’and’ : Only neighbors shared by all input throats. This is also known as ‘intersection’ in set theory and (somtimes) as ‘all’ in boolean logic. Both keywords are accepted and treated as ‘and’.
- Returns
1D array (if
flatten
isTrue
) or ndarray of arrays (ifflatten
isFalse
)
Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[5, 5, 5]) >>> Ps = pn.find_connected_pores(throats=[0, 1]) >>> print(Ps) [[0 1] [1 2]] >>> Ps = pn.find_connected_pores(throats=[0, 1], flatten=True) >>> print(Ps) [0 1 2]
- find_connecting_throat(P1, P2)[source]¶
Return the throat index connecting pairs of pores
- Parameters
P1 (array_like) – The indices of the pores whose throats are sought. These can be vectors of indices, but must be the same length
P2 (array_like) – The indices of the pores whose throats are sought. These can be vectors of indices, but must be the same length
- Returns
Returns a list the same length as P1 (and P2) with the each element
containing the throat index that connects the corresponding pores,
or None` if pores are not connected.
Notes
The returned list can be converted to an ND-array, which will convert the
None
values tonan
. These can then be found usingnumpy.isnan
.Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[5, 5, 5]) >>> Ts = pn.find_connecting_throat([0, 1, 2], [2, 2, 2]) >>> print(Ts) [None, 1, None]
- find_nearby_pores(pores, r, flatten=False, include_input=False)[source]¶
Find all pores within a given radial distance of the input pore(s) regardless of whether or not they are toplogically connected.
- Parameters
pores (array_like) – The list of pores for whom nearby neighbors are to be found
r (scalar) – The maximum radius within which the search should be performed
include_input (bool) – Controls whether the input pores should be included in the list of pores nearby the other pores in the input list. So if
pores=[1, 2]
and 1 and 2 are withinr
of each other, then 1 will be included in the returned for pores near 2, and vice-versa if this argument isTrue
. The default isFalse
.flatten (bool) – If
True
returns a single list of all pores that match the criteria, otherwise returns an array containing a sub-array for each input pore, where each sub-array contains the pores that are nearby to each given input pore. The default is False.
- Returns
A list of pores which are within the given spatial distance. If a
list of N pores is supplied, then a an N-long list of such lists is
returned. The returned lists each contain the pore for which the
neighbors were sought.
Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[3, 3, 3]) >>> Ps = pn.find_nearby_pores(pores=[0, 1], r=1) >>> print(Ps[0]) [3 9] >>> print(Ps[1]) [ 2 4 10] >>> Ps = pn.find_nearby_pores(pores=[0, 1], r=0.5) >>> print(Ps) [array([], dtype=int64), array([], dtype=int64)] >>> Ps = pn.find_nearby_pores(pores=[0, 1], r=1, flatten=True) >>> print(Ps) [ 2 3 4 9 10]
- find_neighbor_pores(pores, mode='union', flatten=True, include_input=False)[source]¶
Returns a list of pores that are direct neighbors to the given pore(s)
- Parameters
pores (array_like) – Indices of the pores whose neighbors are sought
flatten (boolean) – If
True
(default) the returned result is a compressed array of all neighbors. IfFalse
, a list of lists with each sub-list containing the neighbors for each input site. Note that an unflattened list might be slow to generate since it is a Pythonlist
rather than a Numpyarray
.include_input (bool) – If
False
(default) then the input pores are not included in the returned list(s). Note that since pores are not neighbors of themselves, the neighbors of pore N will not include N, even if this flag isTrue
.mode (string) –
Specifies logic to filter the resulting list. Options are:
’or’ : (default) All neighbors of the input pores. This is also known as the ‘union’ in set theory or ‘any’ in boolean logic. Both keywords are accepted and treated as ‘or’.
’xor’ : Only neighbors of one and only one input pore. This is useful for finding the pores that are not shared by any of the input pores. This is known as ‘exclusive_or’ in set theory, and is an accepted input.
’xnor’ : Neighbors that are shared by two or more input pores. This is equivalent to finding all neighbors with ‘or’, minus those found with ‘xor’, and is useful for finding neighbors that the inputs have in common.
’and’ : Only neighbors shared by all input pores. This is also known as ‘intersection’ in set theory and (somtimes) as ‘all’ in boolean logic. Both keywords are accepted and treated as ‘and’.
- Returns
If
flatten
isTrue
, returns a 1D array of pore indices filteredaccording to the specified mode. If
flatten
isFalse
, returnsa list of lists, where each list contains the neighbors of the
corresponding input pores.
Notes
The
logic
options are applied to neighboring pores only, thus it is not possible to obtain pores that are part of the global set but not neighbors. This is because (a) the list of global pores might be very large, and (b) it is not possible to return a list of neighbors for each input pores if global pores are considered.Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[5, 5, 5]) >>> Ps = pn.find_neighbor_pores(pores=[0, 2]) >>> print(Ps) [ 1 3 5 7 25 27] >>> Ps = pn.find_neighbor_pores(pores=[0, 1]) >>> print(Ps) [ 2 5 6 25 26] >>> Ps = pn.find_neighbor_pores(pores=[0, 1], mode='union', ... include_input=True) >>> print(Ps) [ 0 1 2 5 6 25 26] >>> Ps = pn.find_neighbor_pores(pores=[0, 2], flatten=False) >>> print(Ps[0]) [ 1 5 25] >>> print(Ps[1]) [ 1 3 7 27] >>> Ps = pn.find_neighbor_pores(pores=[0, 2], mode='xnor') >>> print(Ps) [1] >>> Ps = pn.find_neighbor_pores(pores=[0, 2], mode='xor') >>> print(Ps) [ 3 5 7 25 27]
- find_neighbor_throats(pores, mode='union', flatten=True)[source]¶
Returns a list of throats neighboring the given pore(s)
- Parameters
pores (array_like) – Indices of pores whose neighbors are sought
flatten (boolean, optional) – If
True
(default) a 1D array of unique throat indices is returned. IfFalse
the returned array contains arrays of neighboring throat indices for each input pore, in the order they were sent.mode (string) –
Specifies logic to filter the resulting list. Options are:
’or’ : (default) All neighbors of the input pores. This is also known as the ‘union’ in set theory or ‘any’ in boolean logic. Both keywords are accepted and treated as ‘or’.
’xor’ : Only neighbors of one and only one input pore. This is useful for finding the thraots that are not shared by any of the input pores.
’xnor’ : Neighbors that are shared by two or more input pores. This is equivalent to finding all neighbors with ‘or’, minus those found with ‘xor’, and is useful for finding neighbors that the inputs have in common.
’and’ : Only neighbors shared by all input pores. This is also known as ‘intersection’ in set theory and (somtimes) as ‘all’ in boolean logic. Both keywords are accepted and treated as ‘and’.
- Returns
If
flatten
isTrue
, returns a 1D array of throat indicesfiltered according to the specified mode. If
flatten
isFalse
,returns a list of lists, where each list contains the neighbors of the
corresponding input pores.
Notes
The
logic
options are applied to neighboring bonds only, thus it is not possible to obtain bonds that are part of the global set but not neighbors. This is because (a) the list of global bonds might be very large, and (b) it is not possible to return a list of neighbors for each input site if global sites are considered.Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[5, 5, 5]) >>> Ts = pn.find_neighbor_throats(pores=[0, 1]) >>> print(Ts) [ 0 1 100 101 200 201] >>> Ts = pn.find_neighbor_throats(pores=[0, 1], flatten=False) >>> print(Ts[0]) [ 0 100 200] >>> print(Ts[1]) [ 0 1 101 201]
- get_adjacency_matrix(fmt='coo')[source]¶
Returns an adjacency matrix in the specified sparse format, with throat IDs indicating the non-zero values.
- Parameters
fmt (string, optional) –
The sparse storage format to return. Options are:
’coo’ : (default) This is the native format of OpenPNM data
’lil’ : Enables row-wise slice of the matrix
’csr’ : Favored by most linear algebra routines
’dok’ : Enables subscript access of locations
Notes
This method will only create the requested matrix in the specified format if one is not already saved on the object. If not present, this method will create and return the matrix, as well as store it for future use.
To obtain a matrix with weights other than throat IDs at each non-zero location use
create_adjacency_matrix
.To obtain the non-directed graph, with only upper-triangular entries, use
sp.sparse.triu(am, k=1)
.
- get_incidence_matrix(fmt='coo')[source]¶
Returns an incidence matrix in the specified sparse format, with pore IDs indicating the non-zero values.
- Parameters
fmt (string, optional) –
The sparse storage format to return. Options are:
’coo’ : (default) This is the native format of OpenPNM data
’lil’ : Enables row-wise slice of the matrix
’csr’ : Favored by most linear algebra routines
’dok’ : Enables subscript access of locations
Notes
This method will only create the requested matrix in the specified format if one is not already saved on the object. If not present, this method will create and return the matrix, as well as store it for future use.
To obtain a matrix with weights other than pore IDs at each non-zero location use
create_incidence_matrix
.
- property im¶
Returns an incidence matrix in the specified sparse format, with pore IDs indicating the non-zero values.
- Parameters
fmt (string, optional) –
The sparse storage format to return. Options are:
’coo’ : (default) This is the native format of OpenPNM data
’lil’ : Enables row-wise slice of the matrix
’csr’ : Favored by most linear algebra routines
’dok’ : Enables subscript access of locations
Notes
This method will only create the requested matrix in the specified format if one is not already saved on the object. If not present, this method will create and return the matrix, as well as store it for future use.
To obtain a matrix with weights other than pore IDs at each non-zero location use
create_incidence_matrix
.
- num_neighbors(pores, mode='or', flatten=False)[source]¶
Returns the number of neigbhoring pores for each given input pore
- Parameters
pores (array_like) – Pores whose neighbors are to be counted
flatten (boolean (optional)) – If
False
(default) the number of pores neighboring each input pore as an array the same length aspores
. IfTrue
the sum total number of is counted.mode (string) –
The logic to apply to the returned count of pores.
’or’ : (default) All neighbors of the input pores. This is also known as the ‘union’ in set theory or ‘any’ in boolean logic. Both keywords are accepted and treated as ‘or’.
’xor’ : Only neighbors of one and only one input pore. This is useful for counting the pores that are not shared by any of the input pores. This is known as ‘exclusive_or’ in set theory, and is an accepted input.
’xnor’ : Neighbors that are shared by two or more input pores. This is equivalent to counting all neighbors with ‘or’, minus those found with ‘xor’, and is useful for finding neighbors that the inputs have in common.
’and’ : Only neighbors shared by all input pores. This is also known as ‘intersection’ in set theory and (somtimes) as ‘all’ in boolean logic. Both keywords are accepted and treated as ‘and’.
- Returns
If
flatten
is False, a 1D array with number of neighbors in eachelement, otherwise a scalar value of the number of neighbors.
Notes
This method literally just counts the number of elements in the array returned by
find_neighbor_pores
using the same logic. Explore those methods if uncertain about the meaning of themode
argument here.See also
Examples
>>> import openpnm as op >>> pn = op.network.Cubic(shape=[5, 5, 5]) >>> Np = pn.num_neighbors(pores=[0, 1], flatten=False) >>> print(Np) [3 4] >>> Np = pn.num_neighbors(pores=[0, 2], flatten=True) >>> print(Np) 6 >>> Np = pn.num_neighbors(pores=[0, 2], mode='and', flatten=True) >>> print(Np) 1