def __init__(self,
vertex1Indices,
vertex2Indices,
converters,
undirected=True):
"""
vertex1Indices is a list of fields for the first vertex, with the 1st index
being the ID.
"""
if len(vertex1Indices) < 1 or len(vertex1Indices) < 1:
raise ValueError("vertexIndices must have at least 1 index")
if len(vertex1Indices) != len(vertex2Indices):
raise ValueError("len(vertex1Indices)=" +
str(len(vertex1Indices)) +
"and len(vertex2Indices)=" + len(vertex2Indices))
Parameter.checkList(vertex1Indices, Parameter.checkInt,
[0, float('inf')])
Parameter.checkList(vertex2Indices, Parameter.checkInt,
[0, float('inf')])
self.vertex1IdIndex = vertex1Indices[0]
self.vertex2IdIndex = vertex2Indices[0]
self.vertex1Indices = copy.copy(vertex1Indices)
self.vertex2Indices = copy.copy(vertex2Indices)
self.vertex1Indices.remove(self.vertex1IdIndex)
self.vertex2Indices.remove(self.vertex2IdIndex)
self.converters = converters
self.undirected = undirected
self.edgeWeight = 1
def __init__(self, vertex1Indices, vertex2Indices, converters, undirected=True):
"""
vertex1Indices is a list of fields for the first vertex, with the 1st index
being the ID.
"""
if len(vertex1Indices) < 1 or len(vertex1Indices) < 1:
raise ValueError("vertexIndices must have at least 1 index")
if len(vertex1Indices) != len(vertex2Indices):
raise ValueError("len(vertex1Indices)=" + str(len(vertex1Indices)) + "and len(vertex2Indices)=" + len(vertex2Indices))
Parameter.checkList(vertex1Indices, Parameter.checkInt, [0, float('inf')])
Parameter.checkList(vertex2Indices, Parameter.checkInt, [0, float('inf')])
self.vertex1IdIndex = vertex1Indices[0]
self.vertex2IdIndex = vertex2Indices[0]
self.vertex1Indices = copy.copy(vertex1Indices)
self.vertex2Indices = copy.copy(vertex2Indices)
self.vertex1Indices.remove(self.vertex1IdIndex)
self.vertex2Indices.remove(self.vertex2IdIndex)
self.converters = converters
self.undirected = undirected
self.edgeWeight = 1
def setVertices(self, vertices, indices=None):
"""
Set the vertices to the given list of vertices. If indices = None then
all vertices are replaced, and if not the given indices are used.
:param vertices: a list of vertices..
:type vertices: :class:`list`
:param indices: a list of indices of the same length as vertices or None for all indices in this object.
:type indices: :class:`list`
"""
if indices != None:
Parameter.checkList(indices, Parameter.checkIndex,
[0, len(self.V)])
if len(vertices) != len(indices):
raise ValueError(
"Length of indices list must be same as that of vertices list"
)
if indices == None and len(vertices) != len(self.V):
raise ValueError("Incorrect number of vertices " +
str(len(vertices)) + ", expecting " +
str(len(self.V)))
if indices == None:
for i in range(len(vertices)):
self.V[i] = vertices[i]
else:
for i in range(len(indices)):
self.V[indices[i]] = vertices[i]
def sequenceVectorStats(self,
graph,
subgraphIndices,
treeStats=False,
eigenStats=True):
"""
Pass in a list of graphs are returns a series of statistics. Each list
element is a dict of vector statistics.
"""
Parameter.checkClass(graph, AbstractMatrixGraph)
for inds in subgraphIndices:
Parameter.checkList(inds, Parameter.checkInt,
[0, graph.getNumVertices()])
Parameter.checkBoolean(treeStats)
numGraphs = len(subgraphIndices)
statsDictList = []
for i in range(numGraphs):
Util.printIteration(i, self.vectorPrintStep, numGraphs)
subgraph = graph.subgraph(subgraphIndices[i])
statsDictList.append(
self.vectorStatistics(subgraph, treeStats, eigenStats))
return statsDictList
def sequenceScalarStats(self,
graph,
subgraphIndices,
slowStats=True,
treeStats=False):
"""
Pass in a graph and list of subgraph indices and returns a series of statistics. Each row
corresponds to the statistics on the subgraph.
"""
Parameter.checkClass(graph, AbstractMatrixGraph)
for inds in subgraphIndices:
Parameter.checkList(inds, Parameter.checkInt,
[0, graph.getNumVertices()])
Parameter.checkBoolean(slowStats)
Parameter.checkBoolean(treeStats)
numGraphs = len(subgraphIndices)
statsMatrix = numpy.zeros((numGraphs, self.numStats))
for i in range(numGraphs):
Util.printIteration(i, self.printStep, numGraphs)
#logging.debug("Subgraph size: " + str(len(subgraphIndices[i])))
subgraph = graph.subgraph(subgraphIndices[i])
statsMatrix[i, :] = self.scalarStatistics(subgraph, slowStats,
treeStats)
return statsMatrix
def subList(self, indices):
"""
Returns a subset of this object, indicated by the given indices.
"""
Parameter.checkList(indices, Parameter.checkIndex, (0, self.getNumVertices()))
vList = VertexList(len(indices), self.getNumFeatures())
vList.setVertices(self.getVertices(indices))
return vList
def subList(self, indices):
"""
Returns a subset of this object, indicated by the given indices.
"""
Parameter.checkList(indices, Parameter.checkIndex,
(0, self.getNumVertices()))
vList = GeneralVertexList(len(indices))
vList.setVertices(self.getVertices(indices))
return vList
def setOutDegSequence(self, outDegSequence):
"""
Set the (out)degree sequence of this object.
:param outDegSequence: a vector of degrees for each vertex in the graph.
:type outDegSequence: :class:`numpy.ndarray`
"""
Parameter.checkClass(outDegSequence, numpy.ndarray)
if outDegSequence.ndim != 1:
raise ValueError("Degree sequence must be one dimensional")
Parameter.checkList(outDegSequence, Parameter.checkInt, [0, outDegSequence.shape[0]])
self.outDegSequence = outDegSequence
def setOutDegSequence(self, outDegSequence):
'''
Set the (out)degree sequence of this object.
:param outDegSequence: a vector of degrees for each vertex in the graph.
:type outDegSequence: :class:`numpy.ndarray`
'''
Parameter.checkClass(outDegSequence, numpy.ndarray)
if outDegSequence.ndim != 1:
raise ValueError("Degree sequence must be one dimensional")
Parameter.checkList(outDegSequence, Parameter.checkInt,
[0, outDegSequence.shape[0]])
self.outDegSequence = outDegSequence
def setInDegSequence(self, inDegSequence):
"""
Set the (in)degree sequence of this object.
:param inDegSequence: a vector of degrees for each vertex in the graph.
:type inDegSequence: :class:`numpy.ndarray`
"""
Parameter.checkClass(inDegSequence, numpy.ndarray)
if inDegSequence.ndim != 1:
raise ValueError("Degree sequence must be one dimensional")
if inDegSequence.shape[0] != self.outDegSequence.shape[0]:
raise ValueError("In-degree sequence must be same length as out-degree sequence")
Parameter.checkList(inDegSequence, Parameter.checkInt, [0, inDegSequence.shape[0]])
self.inDegSequence = inDegSequence
def getVertices(self, vertexIndices=None):
"""
Returns a set of vertices specified by vertexIndices. If vertexIndices
is None then all vertices are returned.
:param vertexIndices: a list of vertex indices.
:type vertexIndices: :class:`list`
:returns: A set of vertices corresponding to the input indices.
"""
if vertexIndices == None:
return self.V
else:
Parameter.checkList(vertexIndices, Parameter.checkIndex, (0, self.V.shape[0]))
return self.V[vertexIndices]
def getVertices(self, vertexIndices=None):
"""
Returns a set of vertices specified by vertexIndices. If vertexIndices
is None then all vertices are returned.
:param vertexIndices: a list of vertex indices.
:type vertexIndices: :class:`list`
:returns: A set of vertices corresponding to the input indices.
"""
if vertexIndices == None:
return self.V
else:
Parameter.checkList(vertexIndices, Parameter.checkIndex,
(0, self.V.shape[0]))
return self.V[vertexIndices]
def subgraph(self, vertexIndices):
"""
Pass in a list or set of vertexIndices and returns the subgraph containing
those vertices only, and edges between them.
:param vertexIndices: the indices of the subgraph vertices.
:type vertexIndices: :class:`list`
"""
Parameter.checkList(vertexIndices, Parameter.checkIndex, (0, self.getNumVertices()))
vertexIndices = numpy.unique(numpy.array(vertexIndices)).tolist()
vList = self.vList.subList(vertexIndices)
subGraph = CsArrayGraph(vList, self.undirected, self.W.dtype)
subGraph.W = self.W[vertexIndices, :][:, vertexIndices]
return subGraph
def subgraph(self, vertexIndices):
"""
Pass in a list or set of vertexIndices and returns the subgraph containing
those vertices only, and edges between them.
:param vertexIndices: the indices of the subgraph vertices.
:type vertexIndices: :class:`list`
"""
Parameter.checkList(vertexIndices, Parameter.checkIndex,
(0, self.getNumVertices()))
vertexIndices = numpy.unique(numpy.array(vertexIndices)).tolist()
vList = self.vList.subList(vertexIndices)
subGraph = DenseGraph(vList, self.undirected, self.W.dtype)
subGraph.W = self.W[vertexIndices, :][:, vertexIndices]
return subGraph
def expandIntArray(v):
"""
Take a vector of integers and expand it into a vector with counts of the
corresponding integers. For example, with v = [1, 3, 2, 4], the expanded
vector is [0, 1, 1, 1, 2, 2, 3, 3, 3, 3].
"""
Parameter.checkClass(v, numpy.ndarray)
Parameter.checkList(v, Parameter.checkInt, [0, float('inf')])
w = numpy.zeros(numpy.sum(v), numpy.int)
currentInd = 0
for i in range(v.shape[0]):
w[currentInd:currentInd + v[i]] = i
currentInd += v[i]
return w
def expandIntArray(v):
"""
Take a vector of integers and expand it into a vector with counts of the
corresponding integers. For example, with v = [1, 3, 2, 4], the expanded
vector is [0, 1, 1, 1, 2, 2, 3, 3, 3, 3].
"""
Parameter.checkClass(v, numpy.ndarray)
Parameter.checkList(v, Parameter.checkInt, [0, float('inf')])
w = numpy.zeros(numpy.sum(v), numpy.int)
currentInd = 0
for i in range(v.shape[0]):
w[currentInd:currentInd+v[i]] = i
currentInd += v[i]
return w
def harmonicGeodesicDistance(self, P=None, vertexInds=None):
"""
Compute the "harmonic mean" geodesic distance for a graph. This is
denoted by the inverse of 1/(1/2 n(n+1)) \sum_{i<=j} d_ij^-1 where d_ij is the
shortest path length between i and j for an undirected graph. The distance from a
node to itself is infinite. For a directed graph, the inverse distance is
1/n^2 sum_{i,j} d_ij^-1.
:param P: An optional nxn matrix whose ijth entry is the shortest path from i to j.
:type P: :class:`ndarray`
:param vertexInds: An optional list of vertices used to compute the mean geodesic distance. If this list is none, then all vertices are used.
:type vertexInds: :class:`list`
:returns: The mean harmonic geodesic distance of this graph.
"""
if P!=None and (type(P) != numpy.ndarray or P.shape != (self.getNumVertices(), self.getNumVertices())):
raise ValueError("P must be array of same size as weight matrix of graph")
if vertexInds!=None:
Parameter.checkList(vertexInds, Parameter.checkInt, [0, self.getNumVertices()])
if self.getNumVertices() == 0 or (vertexInds != None and len(vertexInds)==0):
return 0
if P == None:
P = self.floydWarshall(True)
else:
P = P.copy()
if vertexInds != None:
P = P[vertexInds, :][:, vertexInds]
n = P.shape[0]
P = 1/(P + numpy.diag(numpy.ones(n)*numpy.inf))
if self.isUndirected():
distanceSum = numpy.sum(numpy.triu(P))
if distanceSum != 0:
return (n*(n+1))/(2*distanceSum)
else:
distanceSum = numpy.sum(P)
if distanceSum != 0:
return n**2/distanceSum
#Means that all vertices are disconnected
return float('inf')
def setInDegSequence(self, inDegSequence):
'''
Set the (in)degree sequence of this object.
:param inDegSequence: a vector of degrees for each vertex in the graph.
:type inDegSequence: :class:`numpy.ndarray`
'''
Parameter.checkClass(inDegSequence, numpy.ndarray)
if inDegSequence.ndim != 1:
raise ValueError("Degree sequence must be one dimensional")
if inDegSequence.shape[0] != self.outDegSequence.shape[0]:
raise ValueError(
"In-degree sequence must be same length as out-degree sequence"
)
Parameter.checkList(inDegSequence, Parameter.checkInt,
[0, inDegSequence.shape[0]])
self.inDegSequence = inDegSequence
def sequenceVectorStats(self, graph, subgraphIndices, treeStats=False, eigenStats=True):
"""
Pass in a list of graphs are returns a series of statistics. Each list
element is a dict of vector statistics.
"""
Parameter.checkClass(graph, AbstractMatrixGraph)
for inds in subgraphIndices:
Parameter.checkList(inds, Parameter.checkInt, [0, graph.getNumVertices()])
Parameter.checkBoolean(treeStats)
numGraphs = len(subgraphIndices)
statsDictList = []
for i in range(numGraphs):
Util.printIteration(i, self.vectorPrintStep, numGraphs)
subgraph = graph.subgraph(subgraphIndices[i])
statsDictList.append(self.vectorStatistics(subgraph, treeStats, eigenStats))
return statsDictList
def getVertices(self, vertexIndices=None):
"""
Returns a list of vertices specified by vertexIndices, or all vertices if
vertexIndices == None.
:param vertexIndices: a list of vertex indices.
:type vertexIndices: :class:`list`
:returns: A set of vertices corresponding to the input indices.
"""
if vertexIndices != None:
Parameter.checkList(vertexIndices, Parameter.checkIndex, (0, len(self.V)))
else:
vertexIndices = range(len(self.V))
vertices = []
for i in vertexIndices:
vertices.append(self.V[i])
return vertices
def subgraph(self, vertexIndices):
"""
Pass in a list or set of vertexIndices and returns the subgraph containing
those vertices only, and edges between them.
:param vertexIndices: the indices of the subgraph vertices.
:type vertexIndices: :class:`list`
:returns: A new PySparseGraph containing only vertices and edges from vertexIndices
"""
Parameter.checkList(vertexIndices, Parameter.checkIndex, (0, self.getNumVertices()))
vertexIndices = numpy.unique(numpy.array(vertexIndices))
vList = self.vList.subList(vertexIndices.tolist())
subGraph = PySparseGraph(vList, self.undirected)
if len(vertexIndices) != 0:
subGraph.W = self.W[vertexIndices, vertexIndices]
return subGraph
def __init__(self, initialGraph, k):
"""
Initialise with a starting graph, and number of iterations k. The weights
of the initial graph correspond to probabilities.
:param initialGraph: The intial graph to use.
:type initialGraph: :class:`apgl.graph.AbstractMatrixGraph`
:param k: The number of iterations.
:type k: :class:`int`
"""
Parameter.checkInt(k, 1, float('inf'))
edgeVals = initialGraph.getEdgeValues(initialGraph.getAllEdges())
Parameter.checkList(edgeVals, Parameter.checkFloat, [0.0, 1.0])
W = initialGraph.getWeightMatrix()
if (numpy.diag(W) == numpy.zeros(W.shape[0])).any():
raise ValueError("Initial graph must have all self-edges")
self.initialGraph = initialGraph
self.k = k
def getVertices(self, vertexIndices=None):
"""
Returns a list of vertices specified by vertexIndices, or all vertices if
vertexIndices == None.
:param vertexIndices: a list of vertex indices.
:type vertexIndices: :class:`list`
:returns: A set of vertices corresponding to the input indices.
"""
if vertexIndices != None:
Parameter.checkList(vertexIndices, Parameter.checkIndex,
(0, len(self.V)))
else:
vertexIndices = range(len(self.V))
vertices = []
for i in vertexIndices:
vertices.append(self.V[i])
return vertices
def __init__(self, initialGraph, k):
"""
Initialise with a starting graph, and number of iterations k. The weights
of the initial graph correspond to probabilities.
:param initialGraph: The intial graph to use.
:type initialGraph: :class:`apgl.graph.AbstractMatrixGraph`
:param k: The number of iterations.
:type k: :class:`int`
"""
Parameter.checkInt(k, 1, float('inf'))
edgeVals = initialGraph.getEdgeValues(initialGraph.getAllEdges())
Parameter.checkList(edgeVals, Parameter.checkFloat, [0.0, 1.0])
W = initialGraph.getWeightMatrix()
if (numpy.diag(W)==numpy.zeros(W.shape[0])).any():
raise ValueError("Initial graph must have all self-edges")
self.initialGraph = initialGraph
self.k = k
def subgraph(self, vertexIndices):
"""
Pass in a list or set of vertexIndices and returns the subgraph containing
those vertices only, and edges between them.
:param vertexIndices: the indices of the subgraph vertices.
:type vertexIndices: :class:`list`
:returns: A new PySparseGraph containing only vertices and edges from vertexIndices
"""
Parameter.checkList(vertexIndices, Parameter.checkIndex,
(0, self.getNumVertices()))
vertexIndices = numpy.unique(numpy.array(vertexIndices))
vList = self.vList.subList(vertexIndices.tolist())
subGraph = PySparseGraph(vList, self.undirected)
if len(vertexIndices) != 0:
subGraph.W = self.W[vertexIndices, vertexIndices]
return subGraph
def categoricalToIndicator(self, X, indices):
"""
Convert a set of categorical variables to indicator ones.
"""
Parameter.checkList(indices, Parameter.checkIndex, (0, X.shape[1]))
X2 = numpy.zeros((X.shape[0], 0))
for i in range(X.shape[1]):
if i in indices:
categories = numpy.unique(X[:, i])
Z = numpy.zeros((X.shape[0], categories.shape[0]))
for j in range(categories.shape[0]):
Z[:, j] = X[:, i] == categories[j]
X2 = numpy.c_[X2, Z]
else:
X2 = numpy.c_[X2, X[:, i]]
return X2
def sequenceScalarStats(self, graph, subgraphIndices, slowStats=True, treeStats=False):
"""
Pass in a graph and list of subgraph indices and returns a series of statistics. Each row
corresponds to the statistics on the subgraph.
"""
Parameter.checkClass(graph, AbstractMatrixGraph)
for inds in subgraphIndices:
Parameter.checkList(inds, Parameter.checkInt, [0, graph.getNumVertices()])
Parameter.checkBoolean(slowStats)
Parameter.checkBoolean(treeStats)
numGraphs = len(subgraphIndices)
statsMatrix = numpy.zeros((numGraphs, self.numStats))
for i in range(numGraphs):
Util.printIteration(i, self.printStep, numGraphs)
logging.debug("Subgraph size: " + str(len(subgraphIndices[i])))
subgraph = graph.subgraph(subgraphIndices[i])
statsMatrix[i, :] = self.scalarStatistics(subgraph, slowStats, treeStats)
return statsMatrix
def testCheckList(self):
lst = [1, 2, 3, 2, 2]
Parameter.checkList(lst, Parameter.checkInt, [1, 3])
lst = [1, 2, 3, 2, 4]
self.assertRaises(ValueError, Parameter.checkList, lst, Parameter.checkInt, [1, 3])
lst = [1, 2, 3, 2, 0]
self.assertRaises(ValueError, Parameter.checkList, lst, Parameter.checkInt, [1, 3])
lst = [1, 2, 3, 2, 1.2]
self.assertRaises(ValueError, Parameter.checkList, lst, Parameter.checkInt, [1, 3])
lst = "a"
self.assertRaises(ValueError, Parameter.checkList, lst, Parameter.checkInt, [1, 3])
lst = [0.1, 0.6, 1.4]
Parameter.checkList(lst, Parameter.checkFloat, [0.1, 3.0])
#Test use of array
lst = numpy.array([0.1, 0.6, 1.4])
Parameter.checkList(lst, Parameter.checkFloat, [0.1, 3.0])
lst = numpy.array([[0.1, 0.6, 1.4]])
self.assertRaises(ValueError, Parameter.checkList, lst, Parameter.checkFloat, [0.1, 3.0])
def geodesicDistance(self, P=None, vertexInds=None):
"""
Compute the mean geodesic distance for a graph. This is denoted for an
undirected graph by 1/(1/2 n(n+1)) \sum_{i<=j} d_ij where d_ij is the
shortest path length between i and j. Note that if i and j are not connected
we assume a path length of 0. If the graph is directed then the geodesic
distance is 1/(n^2) sum_{i, j} d_ij.
:param P: An optional nxn matrix whose ijth entry is the shortest path from i to j.
:type P: :class:`ndarray`
:param vertexInds: An optional list of vertices used to compute the mean geodesic distance. If this list is none, then all vertices are used.
:type vertexInds: :class:`list`
:returns: The mean geodesic distance of this graph.
"""
if P!=None and (type(P) != numpy.ndarray or P.shape != (self.getNumVertices(), self.getNumVertices())):
raise ValueError("P must be array of same size as weight matrix of graph")
if vertexInds!=None:
Parameter.checkList(vertexInds, Parameter.checkInt, [0, self.getNumVertices()])
if self.getNumVertices() == 0 or (vertexInds != None and len(vertexInds)==0):
return 0
if P == None:
P = self.floydWarshall(True)
else:
P = P.copy()
if vertexInds != None:
P = P[vertexInds, :][:, vertexInds]
n = P.shape[0]
P[P==numpy.inf] = 0
P[numpy.diag_indices(P.shape[0])] = 0.0
distanceSum = numpy.sum(P)
if self.isUndirected():
return distanceSum/(n*(n+1))
else:
return distanceSum/(n**2)
def testCheckList(self):
lst = [1, 2, 3, 2, 2]
Parameter.checkList(lst, Parameter.checkInt, [1, 3])
lst = [1, 2, 3, 2, 4]
self.assertRaises(ValueError, Parameter.checkList, lst,
Parameter.checkInt, [1, 3])
lst = [1, 2, 3, 2, 0]
self.assertRaises(ValueError, Parameter.checkList, lst,
Parameter.checkInt, [1, 3])
lst = [1, 2, 3, 2, 1.2]
self.assertRaises(ValueError, Parameter.checkList, lst,
Parameter.checkInt, [1, 3])
lst = "a"
self.assertRaises(ValueError, Parameter.checkList, lst,
Parameter.checkInt, [1, 3])
lst = [0.1, 0.6, 1.4]
Parameter.checkList(lst, Parameter.checkFloat, [0.1, 3.0])
#Test use of array
lst = numpy.array([0.1, 0.6, 1.4])
Parameter.checkList(lst, Parameter.checkFloat, [0.1, 3.0])
lst = numpy.array([[0.1, 0.6, 1.4]])
self.assertRaises(ValueError, Parameter.checkList, lst,
Parameter.checkFloat, [0.1, 3.0])
def setVertices(self, vertices, indices=None):
"""
Set the vertices to the given list of vertices. If indices = None then
all vertices are replaced, and if not the given indices are used.
:param vertices: a list of vertices..
:type vertices: :class:`list`
:param indices: a list of indices of the same length as vertices or None for all indices in this object.
:type indices: :class:`list`
"""
if indices!=None:
Parameter.checkList(indices, Parameter.checkIndex, [0, len(self.V)])
if len(vertices) != len(indices):
raise ValueError("Length of indices list must be same as that of vertices list")
if indices==None and len(vertices) != len(self.V):
raise ValueError("Incorrect number of vertices " + str(len(vertices)) + ", expecting " + str(len(self.V)))
if indices == None:
for i in range(len(vertices)):
self.V[i] = vertices[i]
else:
for i in range(len(indices)):
self.V[indices[i]] = vertices[i]
