def test_paths_empty():
"""
Test shortest paths on an empty graph.
"""
g = tg.TinyGraph(0, adj_type=np.bool)
with pytest.raises(TypeError, match='Graph weights are not numbers.'):
algs.get_shortest_paths(g, True)
np.testing.assert_equal(algs.get_shortest_paths(g, False),
np.zeros((0, 0), dtype=np.float64))
def test_negative_cycles():
"""
Test shortest path on a graph with a negative cycle (expect raise error).
"""
g = tg.TinyGraph(3)
g[0, 1] = 1
g[1, 2] = -5
g[2, 0] = 1
with pytest.raises(Exception, match='Graph has a negative cycle.'):
algs.get_shortest_paths(g, True)
assert np.array_equal(
algs.get_shortest_paths(g, False),
np.array([[0, 1, 1], [1, 0, 1], [1, 1, 0]], dtype=np.float64))
def test_paths_fully_connected_with_path():
"""
Test shortest paths on a fully connected graph.
"""
g = tg.TinyGraph(5)
g[0, 1] = 1
g[1, 2] = 1
g[2, 3] = 1
g[3, 4] = 1
g[4, 0] = 1
dists, next = algs.get_shortest_paths(g, True, True)
paths = algs.construct_all_shortest_paths(next)
for i in paths:
for j in i:
print(list(j))
np.testing.assert_equal(dists,\
np.array([[0,1,2,2,1],\
[1,0,1,2,2],\
[2,1,0,1,2],\
[2,2,1,0,1],\
[1,2,2,1,0]],dtype=np.float64))
np.testing.assert_equal(next,\
np.array([[0,1,1,4,4],\
[0,1,2,2,0],\
[1,1,2,3,3],\
[4,2,2,3,4],\
[0,0,3,3,4]],dtype=np.float64))
np.testing.assert_equal(paths,\
np.array([[[0,np.inf,np.inf,np.inf,np.inf],\
[0,1,np.inf,np.inf,np.inf],\
[0,1,2,np.inf,np.inf],\
[0,4,3,np.inf,np.inf],\
[0,4,np.inf,np.inf,np.inf]],\
[[1,0,np.inf,np.inf,np.inf],\
[1,np.inf,np.inf,np.inf,np.inf],\
[1,2,np.inf,np.inf,np.inf],\
[1,2,3,np.inf,np.inf],\
[1,0,4,np.inf,np.inf]],\
[[2,1,0,np.inf,np.inf],\
[2,1,np.inf,np.inf,np.inf],\
[2,np.inf,np.inf,np.inf,np.inf],\
[2,3,np.inf,np.inf,np.inf],\
[2,3,4,np.inf,np.inf]],\
[[3,4,0,np.inf,np.inf],\
[3,2,1,np.inf,np.inf],\
[3,2,np.inf,np.inf,np.inf],\
[3,np.inf,np.inf,np.inf,np.inf],\
[3,4,np.inf,np.inf,np.inf]],\
[[4,0,np.inf,np.inf,np.inf],\
[4,0,1,np.inf,np.inf],\
[4,3,2,np.inf,np.inf],\
[4,3,np.inf,np.inf,np.inf],\
[4,np.inf,np.inf,np.inf,np.inf]]],dtype=np.float64))
def test_paths_fully_connected():
"""
Test shortest paths on a fully connected graph.
"""
print("Fully connected")
g = tg.TinyGraph(5)
g[0, 1] = 1
g[1, 2] = 1
g[2, 3] = 1
g[3, 4] = 1
g[4, 0] = 1
np.testing.assert_equal(algs.get_shortest_paths(g,True),\
np.array([[0,1,2,2,1],\
[1,0,1,2,2],\
[2,1,0,1,2],\
[2,2,1,0,1],\
[1,2,2,1,0]],dtype=np.float64))
def test_nx_paths():
"""
Check that our shortest paths match those from networkx.
"""
for nx_g in [
nx.generators.classic.balanced_tree(i, j)
for (i, j) in zip(range(2, 6), range(1, 10))
]:
nx_sp = dict(nx.all_pairs_shortest_path(nx_g))
g = tg.io.from_nx(nx_g)
tg_sp, next = algs.get_shortest_paths(g, False, True)
tg_sp_p = algs.construct_all_shortest_paths(next)
for i in range(g.vert_N):
for j in range(g.vert_N):
if not j in nx_sp[i]:
assert np.all(tg_sp_p[i][j] == np.inf)
else:
for k, node in enumerate(nx_sp[i][j]):
assert node == tg_sp_p[i][j][k]
def test_random(test_name):
"""
Test randomly generated graphs against networkx algorithms.
"""
for g in suite[test_name]:
tg_cc = algs.get_connected_components(g)
tg_sp = algs.get_shortest_paths(g, False, False)
nx_g = to_nx(g, weight_prop="weight")
nx_cc = nx.connected_components(nx_g)
nx_sp = dict(nx.all_pairs_shortest_path_length(nx_g))
for cc in nx_cc:
assert cc in tg_cc
for i in range(g.vert_N):
for j in range(g.vert_N):
l = tg_sp[i][j]
if not j in nx_sp[i]:
assert l == np.inf
else:
assert l == nx_sp[i][j]
def test_paths_disjointed():
"""
Test shortest paths on a graph with at least two components.
"""
g = tg.TinyGraph(8)
g[0, 1] = 1
g[0, 2] = 1
g[1, 2] = 3
g[2, 3] = 1
g[3, 4] = 1
g[5, 6] = 4
g[6, 7] = 2
g[7, 5] = 1
dists, paths = algs.get_shortest_paths(g, True, True)
np.testing.assert_equal(dists,\
np.array([[0,1,1,2,3,np.inf,np.inf,np.inf],\
[1,0,2,3,4,np.inf,np.inf,np.inf],\
[1,2,0,1,2,np.inf,np.inf,np.inf],\
[2,3,1,0,1,np.inf,np.inf,np.inf],\
[3,4,2,1,0,np.inf,np.inf,np.inf],\
[np.inf,np.inf,np.inf,np.inf,np.inf,0,3,1],\
[np.inf,np.inf,np.inf,np.inf,np.inf,3,0,2],\
[np.inf,np.inf,np.inf,np.inf,np.inf,1,2,0]]\
,dtype=np.float64))
np.testing.assert_equal(paths,\
np.array([[0,1,2,2,2,np.inf,np.inf,np.inf],\
[0,1,0,0,0,np.inf,np.inf,np.inf],\
[0,0,2,3,3,np.inf,np.inf,np.inf],\
[2,2,2,3,4,np.inf,np.inf,np.inf],\
[3,3,3,3,4,np.inf,np.inf,np.inf],\
[np.inf,np.inf,np.inf,np.inf,np.inf,5,7,7],\
[np.inf,np.inf,np.inf,np.inf,np.inf,7,6,7],\
[np.inf,np.inf,np.inf,np.inf,np.inf,5,6,7]]\
,dtype=np.float64))
import networkx as nx
import tinygraph as tg
import tinygraph.algorithms as algs
from tinygraph.io import to_nx
import numpy as np
import pandas as pd
np.random.seed(0)
VERT_N = 128
res = []
for i in range(100):
g = graph_test_suite.gen_random(VERT_N, np.int32, [1], 0.02)
t1 = time.time()
tg_sp = algs.get_shortest_paths(g,False)
t2 = time.time()
t_uw = t2 - t1
t1 = time.time()
tg_sp_w = algs.get_shortest_paths(g,True)
t2 = time.time()
t_w = t2 - t1
nx_g = to_nx(g, weight_prop = "weight")
nx_sp = dict(nx.all_pairs_shortest_path_length(nx_g))
for i in range(VERT_N):
for j in range(VERT_N):
if not j in nx_sp[i]:
assert tg_sp[i][j] == np.inf
else: