def test_getstrictlyminimummis(self):
inputData = {'T': [[[1]], [[2]], [[2]], [[1], [1, 2]]], 'MS': {1: 0.09600845652974467, 2: 0.2357830588199925}, 'SDC': 0.056047812216985904}
pymsgsp = pyMSGSP(inputData["T"], inputData["MS"], inputData["SDC"])
self.assertEqual(pymsgsp.getStrictlyMinimumMIS([]), sys.maxint)
self.assertTrue(abs(pymsgsp.getStrictlyMinimumMIS([[1]]) - 0.09600845652974467) <= 0.0001)
self.assertTrue(abs(pymsgsp.getStrictlyMinimumMIS([[1, 2]]) - 0.09600845652974467) <= 0.0001)
self.assertTrue(abs(pymsgsp.getStrictlyMinimumMIS([[1], [2]]) - 0.09600845652974467) <= 0.0001)
self.assertEqual(pymsgsp.getStrictlyMinimumMIS([[1], [1]]), sys.maxint)
def test_extendsequence(self):
inputData = {'T': [[[1]], [[2]], [[2]], [[1], [1, 2]]], 'MS': {1: 0.09600845652974467, 2: 0.2357830588199925}, 'SDC': 0.056047812216985904}
pymsgsp = pyMSGSP(inputData["T"], inputData["MS"], inputData["SDC"])
s1 = [[1]]
s2 = [[2]]
self.assertEqual(pymsgsp.extendSequence(s1, s2, MSCandidateJoinCriteria.FORWARD), [[[1], [2]]])
s1 = [[1], [2]]
s2 = [[2], [2]]
self.assertEqual(pymsgsp.extendSequence(s1, s2, MSCandidateJoinCriteria.FORWARD), [[[1], [2], [2]]])
def test_level2candidategenspm(self):
inputData = {'T': [[[1]], [[2], [2]], [[1], [1, 2]]], 'MS': {1: 0.09600845652974467, 2: 0.2357830588199925}, 'SDC': 0.056047812216985904}
pymsgsp = pyMSGSP(inputData["T"], inputData["MS"], inputData["SDC"])
M = Utils.getUniqueItems(inputData["T"])
M.sort(key=lambda item: inputData["MS"][item])
SUP = Utils.genItemSupportCount(M, inputData["T"])
L = [(M[m], SUP[m]) for m in range(len(M))]
out1 = pymsgsp.level2CandidateGenSPM(L)
out2 = [[[1, 1]], [[1], [1]], [[1, 2]], [[1], [2]], [[2], [1]], [[2, 2]], [[2], [2]]]
out1 = sorted(out1, cmp=self.comparator)
out2 = sorted(out2, cmp=self.comparator)
self.assertEqual(out1, out2)
def test_sample(self):
algo1 = pyMSGSP(self.inputData["T"], self.inputData["MS"], self.inputData["SDC"], logging.ERROR)
algo2 = BruteForceSPM(self.inputData["T"], self.inputData["MS"], self.inputData["SDC"])
output1 = algo1.run()
output2 = algo2.run()
for s in output1:
for it in s:
it.sort()
for s in output2:
for it in s:
it.sort()
for t in output1:
self.assertTrue(t in output2)
def test_mscandidategenspm(self):
inputData = {'T': [[[1]], [[2], [2]], [[1], [1, 2]]], 'MS': {1: 0.09600845652974467, 2: 0.2357830588199925}, 'SDC': 0.056047812216985904}
pymsgsp = pyMSGSP(inputData["T"], inputData["MS"], inputData["SDC"])
T = inputData["T"]
M = Utils.getUniqueItems(T)
M.sort(key=lambda item: inputData["MS"][item])
SUP = Utils.genItemSupportCount(M, T)
L = [(M[m], SUP[m]) for m in range(len(M))]
C2 = pymsgsp.level2CandidateGenSPM(L)
cSUP = Utils.genSupportCount(C2, T)
F2 = [C2[c] for c in range(len(C2)) if float(cSUP[c])/len(T) >= pymsgsp.getMinMIS(C2[c])]
out1 = pymsgsp.MSCandidateGenSPM(F2)
out2 = [[[1], [1], [1]], [[1], [1, 2]], [[1], [1], [2]], [[1, 2], [2]], [[1], [2], [2]], [[2], [2], [2]]]
out1 = sorted(out1, cmp=self.comparator)
out2 = sorted(out2, cmp=self.comparator)
self.assertEqual(out1, out2)
def test_random(self):
RDG = RandomDataGenerator()
for i in range(10):
randomData = RDG.genData()
print 'randomData', i+1, ':', randomData
algo1 = BruteForceSPM(randomData["T"], randomData["MS"], randomData["SDC"])
algo2 = pyMSGSP(randomData["T"], randomData["MS"], randomData["SDC"], logging.ERROR)
output1 = algo1.run()
output2 = algo2.run()
for s in output1:
for it in s:
it.sort()
for s in output2:
for it in s:
it.sort()
for t in output1:
self.assertTrue(t in output2)
def test_canprune(self):
inputData = {'T': [[[1]], [[2]], [[2]], [[1], [1, 2]]], 'MS': {1: 0.09600845652974467, 2: 0.2357830588199925}, 'SDC': 0.056047812216985904}
pymsgsp = pyMSGSP(inputData["T"], inputData["MS"], inputData["SDC"])
s = [[1], [1]]
self.assertFalse(pymsgsp.canPrune(s))
