def test_gd_resetAllUtility(self):
gd = Linking.Grid(3, 3)
gd.grid_dict[0].utility = 3
gd.grid_dict[2].utility = 4
gd.resetAllUtility()
self.assertEqual(gd.grid_dict[0].utility, 0)
self.assertEqual(gd.grid_dict[2].utility, 0)
def test_valueIteration(self):
g = Linking.Grid(2, 2)
g.grid_dict[0].reward = -4
g.grid_dict[1].utility = 100
g.grid_dict[1].reward = 100
g.grid_dict[2].utility = -100
g.grid_dict[2].reward = -100
g.setObstacle(3)
al = Algorithms.algorithms(g.grid_dict)
action = {'forward': 0.8, 'left': 0.1, 'right': 0.1}
discount = 0.9
expexted = {0: 64.83478199000001, 1: 100, 2: -100, 3: 0}
# print(al.valueIteration(discount,action,g))
self.assertDictEqual(al.valueIteration(discount, action, g), expexted)
#another test
g = Linking.Grid(4, 3)
for i in g.grid_dict:
if i != 5:
g.grid_dict[i].reward = -4
g.grid_dict[3].utility = 100
g.grid_dict[3].reward = 100
g.grid_dict[7].utility = -100
g.grid_dict[7].reward = -100
g.setObstacle(5)
al = Algorithms.algorithms(g.grid_dict)
action = {'forward': 0.8, 'left': 0.1, 'right': 0.1}
discount = 0.9
expexted = {
0: 50.94115412930875,
1: 64.95855729696827,
2: 79.5362031683467,
3: 100,
4: 39.85041277059019,
5: 0,
6: 48.6439148544741,
7: -100,
8: 29.64382885103273,
9: 25.391792612055916,
10: 34.476553020311734,
11: 12.99228937944113
}
# print(al.valueIteration(discount,action,g))
self.assertDictEqual(al.valueIteration(discount, action, g), expexted)
def test_gd_setWall(self):
gd = Linking.Grid(3, 3)
self.assertTrue(4 in gd.grid_dict[3].adjacent.keys()
) #before setWall, they are connected to each other
self.assertTrue(3 in gd.grid_dict[4].adjacent.keys())
gd.setWall(3, 4)
self.assertFalse(4 in gd.grid_dict[3].adjacent.keys(
)) #after setWall, they are not connected to each other
self.assertFalse(3 in gd.grid_dict[4].adjacent.keys())
def test_gd_setManhattanDist(self):
gd = Linking.Grid(3, 3)
dict = {0: 1, 1: 0, 2: 1, 3: 2, 4: 1, 5: 2, 6: 3, 7: 2, 8: 3}
gd.setManhattanDist(1)
for i in range(9):
self.assertEqual(gd.grid_dict[i].heuristic, dict[i])
dict = {0: 2, 1: 1, 2: 2, 3: 1, 4: 0, 5: 1, 6: 2, 7: 1, 8: 2}
gd.setManhattanDist(4)
for i in range(9):
self.assertEqual(gd.grid_dict[i].heuristic, dict[i])
def test_gd_setObstacle(self):
gd = Linking.Grid(3, 3)
for i in [1, 7, 3, 5]: #before setObstacle, 4 is connecting to 1,7,3,5
self.assertTrue(4 in gd.grid_dict[i].adjacent.keys())
self.assertTrue(i in gd.grid_dict[4].adjacent.keys())
gd.setObstacle(4)
for i in [1, 7, 3,
5]: #after setObstacle, they are not connected to each other
self.assertFalse(4 in gd.grid_dict[i].adjacent.keys())
self.assertFalse(i in gd.grid_dict[4].adjacent.keys())
def test_gd_physicalNeighbor(self):
gd = Linking.Grid(3, 3)
self.assertDictEqual(gd.physicalNeighbor(3), {
'upperCell': 0,
'underCell': 6,
'rightCell': 4
})
gd.setWall(3, 4) # the result won't be changed by adding a wall
self.assertDictEqual(gd.physicalNeighbor(3), {
'upperCell': 0,
'underCell': 6,
'rightCell': 4
})
def test_gd_setGridWeight(self):
gd = Linking.Grid(3, 3)
self.assertTrue(
gd.grid_weight[6] == 1) #cost of unit grid6 is 1 (default)
self.assertTrue(
gd.grid_dict[3].adjacent[6] ==
1) #the weight of physical neighbors connecting to it is also 1
self.assertTrue(gd.grid_dict[7].adjacent[6] == 1)
gd.setGridWeight(6, 2) #after setting cost to 2
self.assertTrue(gd.grid_weight[6] == 2) #cost of unit grid6 is 2
self.assertTrue(
gd.grid_dict[3].adjacent[6] ==
2) #the weight of physical neighbors connecting to it is also 2
self.assertTrue(gd.grid_dict[7].adjacent[6] == 2)
def test_gd_init(self):
gd = Linking.Grid(3, 3)
self.assertTrue(8 in gd.grid_dict.keys())
self.assertEqual([i for i in gd.grid_dict[4].adjacent.keys()],
[3, 5, 1, 7]) #the order doesn't matter
self.assertEqual([i for i in gd.grid_dict[8].adjacent.keys()], [7, 5])
class Test_algorithm(unittest.TestCase):
g = Linking.Graph()
gd = Linking.Grid(3, 3) #3*3 grid
# 3 layer absolutely complete perfect tree
for i in range(7):
g.add_vertex(i)
g.add_edge(1, 4, 1)
g.add_edge(1, 3, 1)
g.add_edge(2, 6, 1)
g.add_edge(2, 5, 1)
g.add_edge(0, 2, 1)
g.add_edge(0, 1, 1)
AL = Algorithms.algorithms(g.vert_dict)
AL_gd = Algorithms.algorithms(gd.grid_dict)
#For Graph data structure
def test_bdfs_bfs(self):
#BFS
self.assertEqual([0, 1, 4], self.AL.bdfs(0, 4,
'BFS')) #Test BFS final path
qsExpect = [[0, [1, 2]], [1, [2, 3, 4]], [2, [3, 4, 5, 6]],
[3, [4, 5, 6]], [4, [5, 6]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test BFS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 2], [0, 1, 2, 3], [0, 1, 2, 3, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test BFS visitedLog
def test_bdfs_dfs(self):
#DFS
self.assertEqual([0, 1, 4], self.AL.bdfs(0, 4,
'DFS')) #Test DFS final path
qsExpect = [[0, [1, 2]], [2, [1, 5, 6]], [6, [1, 5]], [5, [1]],
[1, [3, 4]], [4, [3]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test DFS qsLog
visitedExpect = [[0], [0, 2], [0, 2, 6], [0, 2, 6, 5], [0, 2, 6, 5, 1],
[0, 2, 6, 5, 1, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test DFS visitedLog
def test_ucsAStar_ucs_dif_Cost(self):
g = Linking.Graph()
for i in range(7):
g.add_vertex(i)
g.add_edge(0, 2, 2)
g.add_edge(0, 1, 1)
g.add_edge(1, 4, 4)
g.add_edge(1, 3, 1)
g.add_edge(2, 6, 1)
g.add_edge(2, 5, 1)
g.add_edge(2, 4, 2)
al = Algorithms.algorithms(g.vert_dict)
self.assertListEqual([0, 2, 4], al.ucsAStar(0, 4, 'UCS'))
expect = [[0], [0, 1], [0, 1, 2], [0, 1, 2, 3], [0, 1, 2, 3, 5],
[0, 1, 2, 3, 5, 6], [0, 1, 2, 3, 5, 6, 4]]
self.assertListEqual(expect, al.getVisitedLog())
def test_ucsAStar_ucs(self):
#UCS
self.assertEqual([0, 1, 4],
self.AL.ucsAStar(0, 4, 'UCS')) #Test UCS final path
qsExpect = [[0, [1, 2], [1, 1]], [1, [2, 3, 4], [1, 2, 2]],
[2, [3, 4, 5, 6], [2, 2, 2, 2]], [3, [4, 5, 6], [2, 2, 2]],
[4, [5, 6], [2, 2]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test UCS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 2], [0, 1, 2, 3], [0, 1, 2, 3, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test UCS visitedLog
def test_ucsAStar_aStar(self):
#A* (manual heuristic)
self.g.get_vertex(1).heuristic = 1
self.g.get_vertex(3).heuristic = 1
self.assertEqual([0, 1, 4],
self.AL.ucsAStar(0, 4, 'aStar')) #Test A* final path
qsExpect = [[0, [2, 1], [1, 2]], [2, [1, 5, 6], [2, 2, 2]],
[1, [4, 5, 6, 3], [2, 2, 2, 3]], [4, [5, 3, 6], [2, 3, 2]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test A* qsLog
visitedExpect = [[0], [0, 2], [0, 2, 1], [0, 2, 1, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test A* visitedLog
self.g.resetAllHeuristic()
#A* (default heuristic=0)
self.assertEqual([0, 1, 4],
self.AL.ucsAStar(0, 4, 'aStar')) #Test A* final path
qsExpect = [[0, [1, 2], [1, 1]], [1, [2, 3, 4], [1, 2, 2]],
[2, [3, 4, 5, 6], [2, 2, 2, 2]], [3, [4, 5, 6], [2, 2, 2]],
[4, [5, 6], [2, 2]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test A* qsLog
visitedExpect = [[0], [0, 1], [0, 1, 2], [0, 1, 2, 3], [0, 1, 2, 3, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test A* visitedLog
def test_bdfs_bfs_iter(self):
#BFS
self.assertEqual([0, 1, 4], self.AL.iterative(0, 4, 'BFS',
3)) #Test BFS final path
qsExpect = [[0, [1, 2]], [1, [2]], [2, []], [0, [1, 2]],
[1, [2, 3, 4]], [2, [3, 4, 5, 6]], [3, [4, 5, 6]],
[4, [5, 6]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test BFS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 2], [0], [0, 1], [0, 1, 2],
[0, 1, 2, 3], [0, 1, 2, 3, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test BFS visitedLog
def test_bdfs_dfs_iter(self):
#DFS
self.assertEqual([0, 2, 5], self.AL.iterative(0, 5, 'DFS',
3)) #Test DFS final path
qsExpect = [[0, [1, 2]], [2, [1]], [1, []], [0, [1, 2]],
[2, [1, 5, 6]], [6, [1, 5]], [5, [1]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test DFS qsLog
visitedExpect = [[0], [0, 2], [0, 2, 1], [0], [0, 2], [0, 2, 6],
[0, 2, 6, 5]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test DFS visitedLog
def test_ucsAStar_ucs_iter(self):
#UCS
self.assertEqual([0, 1, 4], self.AL.iterative(0, 4, 'UCS',
3)) #Test UCS final path
qsExpect = [[0, [1, 2], [1, 1]], [1, [2], [1]], [2, [], []],
[0, [1, 2], [1, 1]], [1, [2, 3, 4], [1, 2, 2]],
[2, [3, 4, 5, 6], [2, 2, 2, 2]], [3, [4, 5, 6], [2, 2, 2]],
[4, [5, 6], [2, 2]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test UCS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 2], [0], [0, 1], [0, 1, 2],
[0, 1, 2, 3], [0, 1, 2, 3, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test UCS visitedLog
def test_ucsAStar_aStar_iter(self):
#A* (manual heuristic)
self.g.get_vertex(1).heuristic = 1
self.g.get_vertex(3).heuristic = 1
self.assertEqual([0, 1, 4], self.AL.iterative(0, 4, 'aStar',
3)) #Test A* final path
qsExpect = [[0, [2, 1], [1, 2]], [2, [1], [2]], [1, [], []],
[0, [2, 1], [1, 2]], [2, [1, 5, 6], [2, 2, 2]],
[1, [4, 5, 6, 3], [2, 2, 2, 3]], [4, [5, 3, 6], [2, 3, 2]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test A* qsLog
visitedExpect = [[0], [0, 2], [0, 2, 1], [0], [0, 2], [0, 2, 1],
[0, 2, 1, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test A* visitedLog
self.g.resetAllHeuristic()
#A* (default heuristic=0)
self.assertEqual([0, 1, 4], self.AL.iterative(0, 4, 'aStar',
3)) #Test A* final path
qsExpect = [[0, [1, 2], [1, 1]], [1, [2], [1]], [2, [], []],
[0, [1, 2], [1, 1]], [1, [2, 3, 4], [1, 2, 2]],
[2, [3, 4, 5, 6], [2, 2, 2, 2]], [3, [4, 5, 6], [2, 2, 2]],
[4, [5, 6], [2, 2]]]
self.assertEqual(qsExpect, self.AL.getQsLog()) #Test A* qsLog
visitedExpect = [[0], [0, 1], [0, 1, 2], [0], [0, 1], [0, 1, 2],
[0, 1, 2, 3], [0, 1, 2, 3, 4]]
self.assertEqual(visitedExpect,
self.AL.getVisitedLog()) #Test A* visitedLog
#For Grid data structure
def test_bdfs_bfs_grid(self):
#BFS
self.assertEqual([0, 1, 4],
self.AL_gd.bdfs(0, 4, 'BFS')) #Test BFS final path
qsExpect = [[0, [1, 3]], [1, [3, 2, 4]], [3, [2, 4, 6]],
[2, [4, 6, 5]], [4, [6, 5]]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test BFS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 3], [0, 1, 3, 2], [0, 1, 3, 2, 4]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test BFS visitedLog
def test_bdfs_dfs_grid(self):
#DFS
self.assertEqual([0, 3, 4],
self.AL_gd.bdfs(0, 4, 'DFS')) #Test DFS final path
qsExpect = [[0, [1, 3]], [3, [1, 4, 6]], [6, [1, 4,
7]], [7, [1, 4, 8]],
[8, [1, 4, 5]], [5, [1, 4, 2]], [2, [1, 4]], [4, [1]]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test DFS qsLog
visitedExpect = [[0], [0, 3], [0, 3, 6], [0, 3, 6, 7], [0, 3, 6, 7, 8],
[0, 3, 6, 7, 8, 5], [0, 3, 6, 7, 8, 5, 2],
[0, 3, 6, 7, 8, 5, 2, 4]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test DFS visitedLog
def test_ucsAStar_ucs_grid(self):
#UCS
self.assertEqual([0, 1, 4],
self.AL_gd.ucsAStar(0, 4,
'UCS')) #Test UCS final path
qsExpect = [[0, [1, 3], [1, 1]], [1, [3, 2, 4], [1, 2, 2]],
[3, [2, 4, 6], [2, 2, 2]], [2, [4, 6, 5], [2, 2, 3]],
[4, [6, 5], [2, 3]]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test UCS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 3], [0, 1, 3, 2], [0, 1, 3, 2, 4]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test UCS visitedLog
def test_ucsAStar_aStar_grid(self):
#A* (getManhattanDist)
self.gd.setManhattanDist(4)
self.assertEqual([0, 1, 4],
self.AL_gd.ucsAStar(0, 4,
'aStar')) #Test A* final path
qsExpect = [[0, [1, 3], [2, 2]], [1, [3, 4, 2], [2, 2, 4]],
[3, [4, 2, 6], [2, 4, 4]], [4, [2, 6], [4, 4]]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test A* qsLog
visitedExpect = [[0], [0, 1], [0, 1, 3], [0, 1, 3, 4]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test A* visitedLog
def test_bdfs_bfs_iter_grid(self):
#BFS
self.assertEqual([0, 3, 6],
self.AL_gd.iterative(0, 6, 'BFS',
it=3)) #Test BFS final path
qsExpect = [[0, [1, 3]], [1, [3]], [3, []], [0, [1, 3]],
[1, [3, 2, 4]], [3, [2, 4, 6]], [2, [4, 6]], [4, [6]],
[6, []]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test BFS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 3], [0], [0, 1], [0, 1, 3],
[0, 1, 3, 2], [0, 1, 3, 2, 4], [0, 1, 3, 2, 4, 6]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test BFS visitedLog
def test_bdfs_dfs_iter_grid(self):
#DFS
self.assertEqual([0, 3, 6],
self.AL_gd.iterative(0, 6, 'DFS',
it=3)) #Test DFS final path
qsExpect = [[0, [1, 3]], [3, [1]], [1, []], [0, [1, 3]],
[3, [1, 4, 6]], [6, [1, 4]]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test DFS qsLog
visitedExpect = [[0], [0, 3], [0, 3, 1], [0], [0, 3], [0, 3, 6]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test DFS visitedLog
def test_ucsAStar_ucs_iter_grid(self):
#UCS
self.assertEqual([0, 3, 6],
self.AL_gd.iterative(0, 6, 'UCS',
it=3)) #Test UCS final path
qsExpect = [[0, [1, 3], [1, 1]], [1, [3], [1]], [3, [], []],
[0, [1, 3], [1, 1]], [1, [3, 2, 4], [1, 2, 2]],
[3, [2, 4, 6], [2, 2, 2]], [2, [4, 6], [2, 2]],
[4, [6], [2]], [6, [], []]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test UCS qsLog
visitedExpect = [[0], [0, 1], [0, 1, 3], [0], [0, 1], [0, 1, 3],
[0, 1, 3, 2], [0, 1, 3, 2, 4], [0, 1, 3, 2, 4, 6]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test UCS visitedLog
def test_ucsAStar_aStar_iter_grid(self):
#A* (getManhattanDist)
self.gd.setManhattanDist(4)
self.assertEqual([0, 1, 4],
self.AL_gd.iterative(0, 4, 'aStar',
it=3)) #Test A* final path
qsExpect = [[0, [1, 3], [2, 2]], [1, [3], [2]], [3, [], []],
[0, [1, 3], [2, 2]], [1, [3, 4, 2], [2, 2, 4]],
[3, [4, 2, 6], [2, 4, 4]], [4, [2, 6], [4, 4]]]
self.assertEqual(qsExpect, self.AL_gd.getQsLog()) #Test A* qsLog
visitedExpect = [[0], [0, 1], [0, 1, 3], [0], [0, 1], [0, 1, 3],
[0, 1, 3, 4]]
self.assertEqual(visitedExpect,
self.AL_gd.getVisitedLog()) #Test A* visitedLog
def test_miniMaxAlphaBeta_miniMax(self):
self.g.resetAllUtility()
self.g.get_vertex(3).utility = 20
self.g.get_vertex(4).utility = 10
self.g.get_vertex(5).utility = 5
self.g.get_vertex(6).utility = 15
expect = [[0, -math.inf, math.inf], [1, -math.inf, math.inf],
[3, 20, 20], [1, -math.inf, 20], [4, 10, 10], [1, 10, 10],
[0, 10, math.inf], [2, -math.inf, math.inf], [5, 5, 5],
[2, -math.inf, 5], [6, 15, 15], [2, 5, 5], [0, 10, 10]]
self.assertListEqual(self.AL.miniMaxAlphaBeta(0, 3, 'miniMax'), expect)
self.assertListEqual(self.AL.finalPath, [0, 1, 4])
def test_miniMaxAlphaBeta_alphaBeta(self):
self.g.resetAllUtility()
self.g.get_vertex(3).utility = 20
self.g.get_vertex(4).utility = 1
self.g.get_vertex(5).utility = 5
self.g.get_vertex(6).utility = 30
#without prune
expect = [[0, -math.inf, math.inf], [1, -math.inf, math.inf],
[3, 20, 20], [1, -math.inf, 20], [4, 1, 1], [1, 1, 1],
[0, 1, math.inf], [2, -math.inf, math.inf], [5, 5, 5],
[2, -math.inf, 5], [6, 30, 30], [2, 5, 5], [0, 5, 5]]
self.assertListEqual(self.AL.miniMaxAlphaBeta(0, 3, 'alphaBeta'),
expect)
self.assertListEqual(self.AL.finalPath, [0, 2, 5])
# #with prune
self.g.resetAllUtility()
self.g.get_vertex(3).utility = 20
self.g.get_vertex(4).utility = 10
self.g.get_vertex(5).utility = 5
self.g.get_vertex(6).utility = 15
expect = [[0, -math.inf, math.inf], [1, -math.inf, math.inf],
[3, 20, 20], [1, -math.inf, 20], [4, 10, 10], [1, 10, 10],
[0, 10, math.inf], [2, -math.inf, math.inf], [5, 5, 5],
[2, -math.inf, 5], [0, 10, 10]]
self.assertListEqual(self.AL.miniMaxAlphaBeta(0, 3, 'alphaBeta'),
expect) #node 6 is pruned
self.assertListEqual(self.AL.finalPath, [0, 1, 4])
def test_miniMaxAlphaBeta_exMiniMax(self):
self.g.resetAllUtility()
self.g.get_vertex(3).utility = 20
self.g.get_vertex(4).utility = 1
self.g.get_vertex(5).utility = 10
self.g.get_vertex(6).utility = 4
self.g.add_objEdge('a0')
self.g.add_edge(2, 'a0')
self.g.delete_edge(2, 5)
self.g.delete_edge(2, 6)
self.g.add_objEdge_vert_connection(
'a0', 5, 0.3) #utility of node 5 =10, 10*0.3=3
self.g.add_objEdge_vert_connection('a0', 6,
0.5) #utility of node 6 =4, 4*0.5=2
#so now the total utility of action0 = 5
expect = [[0, -math.inf, math.inf], [1, -math.inf, math.inf],
[3, 20, 20], [1, -math.inf, 20], [4, 1, 1], [1, 1, 1],
[0, 1, math.inf], [2, -math.inf, math.inf], [5, 10, 10],
[6, 4, 4], [2, 5.0, 5.0], [0, 5.0, 5.0]]
self.assertListEqual(self.AL.miniMaxAlphaBeta(0, 3, 'exMiniMax'),
expect)
self.assertListEqual(self.AL.finalPath, [0, 2])
self.g.add_edge(2, 5)
self.g.add_edge(2, 6)
self.g.delete_objEdge('a0')
def test_generateProbabilityTable(self):
self.g.add_edge(1, 2)
self.g.add_edge(1, 5)
self.g.add_edge(4, 5)
self.AL.generateProbabilityTable()
expect = {
2: [0, 1],
'TT': [0, 0, 0, 0],
'TF': [0, 0, 0, 0],
'FT': [0, 0, 0, 0],
'FF': [0, 0, 0, 0]
}
self.assertDictEqual(self.g.get_vertex(2).probabilityTable, expect)
expect = {
5: [1, 2, 4],
'TTT': [0, 0, 0, 0],
'TTF': [0, 0, 0, 0],
'TFT': [0, 0, 0, 0],
'TFF': [0, 0, 0, 0],
'FTT': [0, 0, 0, 0],
'FTF': [0, 0, 0, 0],
'FFT': [0, 0, 0, 0],
'FFF': [0, 0, 0, 0]
}
self.assertDictEqual(self.g.get_vertex(5).probabilityTable, expect)
self.g.delete_edge(1, 2)
self.g.delete_edge(1, 5)
self.g.delete_edge(4, 5)
def test_simulateData(self):
g = Linking.Graph()
for i in range(5):
g.add_vertex(i)
# 0 1
# \ / \
# 2 3
# /
# 4
g.add_edge(0, 2)
g.add_edge(1, 2)
g.add_edge(1, 3)
g.add_edge(2, 4)
al = Algorithms.algorithms(g.vert_dict)
al.generateProbabilityTable()
al.setProbabilityTable(0, 'T', 0.1)
al.setProbabilityTable(1, 'T', 0.5)
al.setProbabilityTable(2, 'TT', 1)
al.setProbabilityTable(2, 'TF', 1)
al.setProbabilityTable(2, 'FT', 0.5)
al.setProbabilityTable(2, 'FF', 0.01)
al.setProbabilityTable(3, 'T', 0.8)
al.setProbabilityTable(3, 'F', 0.01)
al.setProbabilityTable(4, 'T', 0.1)
al.setProbabilityTable(4, 'F', 0.01)
al.simulateData(10000)
v = g.get_vertex(0).probabilityTable['T'][0] #node 0
self.assertAlmostEqual(v, 0.1, 1) #expect v =0.1
v = g.get_vertex(1).probabilityTable['T'][0] #node 1
self.assertAlmostEqual(v, 0.5, 1) #expect v = 0.5
v = g.get_vertex(2).probabilityTable['TT'][0] #node 2
self.assertAlmostEqual(v, 1, 1) #expect v =1
v = g.get_vertex(3).probabilityTable['T'][0] #node 3
self.assertAlmostEqual(v, 0.8, 1) #expect v = 0.8
v = g.get_vertex(4).probabilityTable['T'][0] #node 4
self.assertAlmostEqual(0.1, v, 1) #expect v = 0.1
def test_setProbabilityTable(self):
g = Linking.Graph()
for i in range(5):
g.add_vertex(i)
# 0 1
# \ / \
# 2 3
# /
# 4
g.add_edge(0, 2)
g.add_edge(1, 2)
g.add_edge(1, 3)
g.add_edge(2, 4)
al = Algorithms.algorithms(g.vert_dict)
al.generateProbabilityTable()
al.setProbabilityTable(0, 'T', 0.1)
al.setProbabilityTable(1, 'T', 0.5)
al.setProbabilityTable(2, 'TT', 1)
al.setProbabilityTable(2, 'TF', 1)
al.setProbabilityTable(2, 'FT', 0.5)
al.setProbabilityTable(2, 'FF', 0.01)
al.setProbabilityTable(3, 'T', 0.8)
al.setProbabilityTable(3, 'F', 0.01)
al.setProbabilityTable(4, 'T', 0.1)
al.setProbabilityTable(4, 'F', 0.01)
v = g.get_vertex(2).probabilityTable['TT'][3] #node 2
self.assertTrue(v == 1) #expect v =1
v = g.get_vertex(3).probabilityTable['T'][3] #node 3
self.assertTrue(v == 0.8) #expect v = 0.8
v = g.get_vertex(4).probabilityTable['T'][3] #node 4
self.assertTrue(v == 0.1) #expect v = 0.1
v = g.get_vertex(0).probabilityTable['T'][3] #node 0
self.assertTrue(v == 0.1) #expect v = 0.1
v = g.get_vertex(1).probabilityTable['T'][3] #node 1
self.assertTrue(v == 0.5) #expect v = 0.5
def test_setKnownPT(self):
g = Linking.Graph()
for i in range(5):
g.add_vertex(i)
# 0 1
# \ / \
# 2 3
# /
# 4
g.add_edge(0, 2)
g.add_edge(1, 2)
g.add_edge(1, 3)
g.add_edge(2, 4)
al = Algorithms.algorithms(g.vert_dict)
al.generateProbabilityTable()
al.setKnownPT(0, 'T', 0.1)
al.setKnownPT(1, 'T', 0.5)
al.setKnownPT(2, 'TT', 1)
al.setKnownPT(2, 'TF', 1)
al.setKnownPT(2, 'FT', 0.5)
al.setKnownPT(2, 'FF', 0.01)
al.setKnownPT(3, 'T', 0.8)
al.setKnownPT(3, 'F', 0.01)
al.setKnownPT(4, 'T', 0.1)
al.setKnownPT(4, 'F', 0.01)
v = g.get_vertex(2).probabilityTable['TT'][0] #node 2
self.assertTrue(v == 1) #expect v =1
v = g.get_vertex(3).probabilityTable['T'][0] #node 3
self.assertTrue(v == 0.8) #expect v = 0.8
v = g.get_vertex(4).probabilityTable['T'][0] #node 4
self.assertTrue(v == 0.1) #expect v = 0.1
v = g.get_vertex(0).probabilityTable['T'][0] #node 0
self.assertTrue(v == 0.1) #expect v = 0.1
v = g.get_vertex(1).probabilityTable['T'][0] #node 1
self.assertTrue(v == 0.5) #expect v = 0.5
def test_generatePriorProbability(self):
g = Linking.Graph()
for i in range(5):
g.add_vertex(i)
# 0 1
# \ / \
# 2 3
# /
# 4
g.add_edge(0, 2)
g.add_edge(1, 2)
g.add_edge(1, 3)
g.add_edge(2, 4)
al = Algorithms.algorithms(g.vert_dict)
al.generateProbabilityTable()
al.setKnownPT(0, 'T', 0.1)
al.setKnownPT(1, 'T', 0.5)
al.setKnownPT(2, 'TT', 1)
al.setKnownPT(2, 'TF', 1)
al.setKnownPT(2, 'FT', 0.5)
al.setKnownPT(2, 'FF', 0.01)
al.setKnownPT(3, 'T', 0.8)
al.setKnownPT(3, 'F', 0.01)
al.setKnownPT(4, 'T', 0.1)
al.setKnownPT(4, 'F', 0.01)
al.generatePriorProbability()
self.assertTrue(g.get_vertex(0).probability == 0.1)
self.assertTrue(g.get_vertex(1).probability == 0.5)
self.assertTrue(g.get_vertex(2).probability == 0.3295)
self.assertTrue(g.get_vertex(3).probability == 0.405)
self.assertTrue(g.get_vertex(4).probability == 0.039655)
def test_refreshP_simulating(self):
g = Linking.Graph()
for i in range(5):
g.add_vertex(i)
# 0 1
# \ / \
# 2 3
# /
# 4
g.add_edge(0, 2)
g.add_edge(1, 2)
g.add_edge(1, 3)
g.add_edge(2, 4)
al = Algorithms.algorithms(g.vert_dict)
al.generateProbabilityTable()
al.setProbabilityTable(0, 'T', 0.1)
al.setProbabilityTable(1, 'T', 0.5)
al.setProbabilityTable(2, 'TT', 1)
al.setProbabilityTable(2, 'TF', 1)
al.setProbabilityTable(2, 'FT', 0.5)
al.setProbabilityTable(2, 'FF', 0.01)
al.setProbabilityTable(3, 'T', 0.8)
al.setProbabilityTable(3, 'F', 0.01)
al.setProbabilityTable(4, 'T', 0.1)
al.setProbabilityTable(4, 'F', 0.01)
al.simulateData(10000)
obs = {3: 'T', 4: 'T'}
self.assertTrue(math.fabs(al.query(obs, 0) - 0.17) <= 0.05)
self.assertTrue(math.fabs(al.query(obs, 1) - 1) <= 0.05)
self.assertTrue(math.fabs(al.query(obs, 2) - 0.92) <= 0.05)
self.assertTrue(math.fabs(al.query(obs, 3) - 1) <= 0.05)
self.assertTrue(math.fabs(al.query(obs, 4) - 1) <= 0.05)
# print(al.refreshP(obs))
def test_refreshP_known(self):
g = Linking.Graph()
for i in range(5):
g.add_vertex(i)
# 0 1
# \ / \
# 2 3
# /
# 4
g.add_edge(0, 2)
g.add_edge(1, 2)
g.add_edge(1, 3)
g.add_edge(2, 4)
al = Algorithms.algorithms(g.vert_dict)
al.generateProbabilityTable()
al.setKnownPT(0, 'T', 0.1)
al.setKnownPT(1, 'T', 0.5)
al.setKnownPT(2, 'TT', 1)
al.setKnownPT(2, 'TF', 1)
al.setKnownPT(2, 'FT', 0.5)
al.setKnownPT(2, 'FF', 0.01)
al.setKnownPT(3, 'T', 0.8)
al.setKnownPT(3, 'F', 0.01)
al.setKnownPT(4, 'T', 0.1)
al.setKnownPT(4, 'F', 0.01)
al.generatePriorProbability()
obs = {3: 'T', 4: 'T'}
expected = {
0: 0.1694628029147602,
1: 0.9958554837953809,
2: 0.9228191078724887,
3: 1,
4: 1
}
self.assertDictEqual(al.refreshP(obs), expected)
obs = {3: 'T'}
expected = {
0: 0.1,
1: 0.9876543209876543,
2: 0.5445555555555555,
3: 1,
4: 0.05900999999999999
}
self.assertDictEqual(al.refreshP(obs), expected)
obs = {3: 'F', 4: 'F'}
expected = {
0: 0.09244807952710092,
1: 0.16236679513583266,
2: 0.1692887479810971,
3: 0,
4: 0
}
self.assertDictEqual(al.refreshP(obs), expected)
obs = {3: 'T', 4: 'F'}
expected = {
0: 0.09564394945748625,
1: 0.9871400215612162,
2: 0.520834440323489,
3: 1,
4: 0
}
self.assertDictEqual(al.refreshP(obs), expected)
obs = {1: 'T', 3: 'T', 4: 'F'} #3(true) is conditional independant
expected = {
0: 0.09665052244976859,
1: 1,
2: 0.5263157894736843,
3: 1,
4: 0
}
self.assertDictEqual(al.refreshP(obs), expected)
obs = {1: 'T', 4: 'F'} #3(not given) is conditional independant
expected = {
0: 0.1597316508266113,
1: 1,
2: 0.5263157894736843,
3: 0.8,
4: 0
}
self.assertDictEqual(al.refreshP(obs), expected)
obs = {1: 'T', 3: 'T', 4: 'F'} #3(False) is conditional independant
expected = {
0: 0.09665052244976859,
1: 1,
2: 0.5263157894736843,
3: 1,
4: 0
}
self.assertDictEqual(al.refreshP(obs), expected)
obs = {} #empty
expected = {0: 0.1, 1: 0.5, 2: 0.3295, 3: 0.405, 4: 0.039655}
self.assertDictEqual(al.refreshP(obs), expected)
def test_markov(self):
g = Linking.Graph()
g.add_vertex('rainy')
g.add_vertex('sunny')
g.add_edge('rainy', 'sunny', 0.5)
g.add_edge('rainy', 'rainy', 0.5)
g.add_edge('sunny', 'rainy', 0.1)
g.add_edge('sunny', 'sunny', 0.9)
g.get_vertex('rainy').probability = 1 #initial probability
g.get_vertex('sunny').probability = 0 #initial probability
model = g.vert_dict
gg = Linking.Graph()
gg.add_vertex(0)
gg.add_vertex(1)
gg.add_vertex(2)
gg.add_vertex(3)
gg.add_edge(0, 1)
gg.add_edge(1, 2)
gg.add_edge(2, 3)
al = Algorithms.algorithms(gg.vert_dict)
expected = {
0: {
'rainy': 1,
'sunny': 0
},
1: {
'rainy': 0.5,
'sunny': 0.5
},
2: {
'rainy': 0.3,
'sunny': 0.7
},
3: {
'rainy': 0.21999999999999997,
'sunny': 0.78
}
}
self.assertDictEqual(al.markov(model, 0), expected)
expected = {
1: {
'rainy': 1,
'sunny': 0
},
2: {
'rainy': 0.5,
'sunny': 0.5
},
3: {
'rainy': 0.3,
'sunny': 0.7
}
}
self.assertDictEqual(al.markov(model, 1), expected)
def test_hiddenMarkov(self):
g = Linking.Graph()
g.add_vertex('rainy')
g.add_vertex('sunny')
g.add_vertex('hear')
g.add_vertex('not')
#transition probability
g.add_edge('rainy', 'sunny', 0.5)
g.add_edge('rainy', 'rainy', 0.5)
g.add_edge('sunny', 'rainy', 0.1)
g.add_edge('sunny', 'sunny', 0.9)
#emission probability
g.add_edge('rainy', 'hear', 0.7)
g.add_edge('rainy', 'not', 0.3)
g.add_edge('sunny', 'hear', 0.2)
g.add_edge('sunny', 'not', 0.8)
#initial probability
g.get_vertex('rainy').probability = 0.5
g.get_vertex('sunny').probability = 0.5
model = g.vert_dict
gg = Linking.Graph()
#node
gg.add_vertex(0)
gg.add_vertex(1)
gg.add_vertex(2)
gg.add_vertex(3)
#obs
gg.add_vertex(4)
gg.add_vertex(5)
gg.add_vertex(6)
gg.add_edge(0, 1)
gg.add_edge(1, 2)
gg.add_edge(2, 3)
gg.add_edge(0, 4) #obs
gg.add_edge(1, 5)
gg.add_edge(2, 6)
al = Algorithms.algorithms(gg.vert_dict)
obs = {4: 'hear', 5: 'hear', 6: 'not'}
states = ['rainy', 'sunny']
expected = {
0: {
'rainy': 0.7777777777777778,
'sunny': 0.22222222222222227
},
1: {
'rainy': 0.7095890410958903,
'sunny': 0.2904109589041096
},
2: {
'rainy': 0.18936697454381615,
'sunny': 0.8106330254561839
},
3: {
'rainy': 0.17574678981752645,
'sunny': 0.8242532101824737
}
}
# print(al.hiddenMarkov(model,0,obs,states))
self.assertDictEqual(al.hiddenMarkov(model, 0, obs, states), expected)
def test_valueIteration(self):
g = Linking.Grid(2, 2)
g.grid_dict[0].reward = -4
g.grid_dict[1].utility = 100
g.grid_dict[1].reward = 100
g.grid_dict[2].utility = -100
g.grid_dict[2].reward = -100
g.setObstacle(3)
al = Algorithms.algorithms(g.grid_dict)
action = {'forward': 0.8, 'left': 0.1, 'right': 0.1}
discount = 0.9
expexted = {0: 64.83478199000001, 1: 100, 2: -100, 3: 0}
# print(al.valueIteration(discount,action,g))
self.assertDictEqual(al.valueIteration(discount, action, g), expexted)
#another test
g = Linking.Grid(4, 3)
for i in g.grid_dict:
if i != 5:
g.grid_dict[i].reward = -4
g.grid_dict[3].utility = 100
g.grid_dict[3].reward = 100
g.grid_dict[7].utility = -100
g.grid_dict[7].reward = -100
g.setObstacle(5)
al = Algorithms.algorithms(g.grid_dict)
action = {'forward': 0.8, 'left': 0.1, 'right': 0.1}
discount = 0.9
expexted = {
0: 50.94115412930875,
1: 64.95855729696827,
2: 79.5362031683467,
3: 100,
4: 39.85041277059019,
5: 0,
6: 48.6439148544741,
7: -100,
8: 29.64382885103273,
9: 25.391792612055916,
10: 34.476553020311734,
11: 12.99228937944113
}
# print(al.valueIteration(discount,action,g))
self.assertDictEqual(al.valueIteration(discount, action, g), expexted)