def test_to_numpy_recarray(self):
G=nx.Graph()
G.add_edge(1,2,weight=7.0,cost=5)
A=nx.to_numpy_recarray(G,dtype=[('weight',float),('cost',int)])
assert_equal(sorted(A.dtype.names),['cost','weight'])
assert_equal(A.weight[0,1],7.0)
assert_equal(A.weight[0,0],0.0)
assert_equal(A.cost[0,1],5)
assert_equal(A.cost[0,0],0)
def test_to_numpy_recarray(self):
G = nx.Graph()
G.add_edge(1, 2, weight=7.0, cost=5)
A = nx.to_numpy_recarray(G, dtype=[('weight', float), ('cost', int)])
assert_equal(sorted(A.dtype.names), ['cost', 'weight'])
assert_equal(A.weight[0, 1], 7.0)
assert_equal(A.weight[0, 0], 0.0)
assert_equal(A.cost[0, 1], 5)
assert_equal(A.cost[0, 0], 0)
def test_to_numpy_recarray(self):
G = nx.Graph()
G.add_edge(1, 2, weight=7.0, cost=5)
A = nx.to_numpy_recarray(G, dtype=[("weight", float), ("cost", int)])
assert sorted(A.dtype.names) == ["cost", "weight"]
assert A.weight[0, 1] == 7.0
assert A.weight[0, 0] == 0.0
assert A.cost[0, 1] == 5
assert A.cost[0, 0] == 0
def test_to_numpy_recarray(recarray_test_graph):
A = nx.to_numpy_recarray(recarray_test_graph,
dtype=[("weight", float), ("cost", int)])
assert sorted(A.dtype.names) == ["cost", "weight"]
assert A.weight[0, 1] == 7.0
assert A.weight[0, 0] == 0.0
assert A.cost[0, 1] == 5
assert A.cost[0, 0] == 0
with pytest.raises(AttributeError, match="has no attribute"):
A.color[0, 1]
def _get_matdict(self):
if not self.loaded:
self.load()
import networkx as nx
matdict = {}
g = self._relabel_to_int(self.graph)
# grab keys from the first edge, discarding id
dl = self._get_edge_values()
# create numpy matrix for each key using recarray
matrec = nx.to_numpy_recarray(g, dtype=zip(dl, [float]*len(dl)) )
for k in dl:
matdict[k] = matrec[k]
return matdict
def nx_to_weighted_array(self, G):
labels=G.nodes()
matrix=nx.to_numpy_recarray(G)
distances=numpy.zeros((len(G),len(G)))
i=0
for x in matrix:
j=0
for y in x:
distances[i][j]=y[0]
distances[j][i]=y[0]
if i==j: distances[i][j]=0
j+=1
i+=1
return distances
def test_to_numpy_recarray_bad_nodelist(recarray_nodelist_test_graph, nodelist,
errmsg):
with pytest.raises(nx.NetworkXError, match=errmsg):
A = nx.to_numpy_recarray(recarray_nodelist_test_graph,
nodelist=nodelist)
def test_to_numpy_recarray_nodelist(recarray_nodelist_test_graph):
A = nx.to_numpy_recarray(recarray_nodelist_test_graph, nodelist=[0, 1])
np.testing.assert_array_equal(A.weight, np.array([[0, 1], [1, 0]]))
def test_to_numpy_recarray_directed(recarray_test_graph):
G = recarray_test_graph.to_directed()
G.remove_edge(2, 1)
A = nx.to_numpy_recarray(G, dtype=[("weight", float), ("cost", int)])
np.testing.assert_array_equal(A.weight, np.array([[0, 7.0], [0, 0]]))
np.testing.assert_array_equal(A.cost, np.array([[0, 5], [0, 0]]))
def __init__(
self,
num_nodes=None, # Number of nodes
num_items=None, # Number of items
item_distribution='uniform', # Different item distributions
# 1. uniform
# 2. direct
# 3. inverse
# 4. ego
# 5. custom
initial_item_distribution=None, # Initial item distribution for custom
ego_graph_radius=1, # Radius of the ego graph
current_time=1, # Total time
G=None): # Networkx DiGraph such that:
# 1. Each node has an attribute 'num_items'
# 2. Each edge has an attribute 'weight'
self.num_nodes = num_nodes
self.num_items = num_items
self.current_time = current_time
self.item_distribution = item_distribution
self.ego_graph_radius = ego_graph_radius
# self.initial_transition_matrix = np.array([[ 0., 1., 0. ],
# [ 0.21128204, 0., 0.78871796],
# [ 0., 1., 0. ]]
# self.initial_item_distribution = {0: 1, 1:1, 2:1}
# self.G = nx.from_numpy_matrix(self.initial_transition_matrix, create_using=nx.DiGraph())
# nx.set_node_attributes(self.G, 'num_items', self.initial_item_distribution)
if G is None: # No networkx graph provided
while True:
try:
self.initial_transition_matrix = self.initialize_transition_matrix(
)
except:
continue
break
self.G = nx.from_numpy_matrix(self.initial_transition_matrix,
create_using=nx.DiGraph())
self.initial_item_distribution = self.initialize_item_distribution(
initial_item_distribution)
nx.set_node_attributes(self.G, 'num_items',
self.initial_item_distribution)
else:
self.G = G
self.initial_transition_matrix = nx.to_numpy_recarray(
self.G, dtype=[('weight', float)]).weight
# Test that resultant transition matrix is row stochastic
absdiff = np.abs(
self.initial_transition_matrix.sum(axis=1)) - np.ones(
(self.num_nodes))
assert absdiff.max() <= 10 * np.spacing(
np.float64(1)), "Not row stochastic"
# self.initial_item_distribution = nx.get_node_attributes(self.G, 'num_items')
self.initial_item_distribution = self.initialize_item_distribution(
initial_item_distribution)
nx.set_node_attributes(self.G, 'num_items',
self.initial_item_distribution)
