def subsets_embed_number(cseg1, cseg2):
"""Returns the number of time the normal form of a csubseg appears
in cseg subsets. Marvin and Laprade (1987).
>>> c1, c2 = Contour([0, 1, 2, 3]), Contour([1, 0, 2])
>>> subsets_embed_number(c1, c2)
0
"""
cseg, csubseg = utils.greatest_first(cseg1, cseg2)
dic = contour.Contour(cseg).subsets_normal(len(csubseg))
if tuple(csubseg) in dic:
return len(dic[tuple(csubseg)])
else:
return 0
def subsets_embed_total_number(cseg1, cseg2):
"""Returns the number of subsets with csubseg_size in a set with
cseg_size. Marvin and Laprade (1987, p. 237).
>>> c1, c2 = Contour([0, 1, 2, 3]), Contour([1, 0, 2])
>>> subsets_embed_total_number(c1, c2)
4
"""
cseg, csubseg = utils.greatest_first(cseg1, cseg2)
cseg_size = len(cseg)
csubseg_size = len(csubseg)
a = math.factorial(cseg_size)
b = math.factorial(csubseg_size)
c = math.factorial(cseg_size - csubseg_size)
return a / (b * c)
def contour_embed(cseg1, cseg2):
"""Returns similarity between contours with different
cardinalities. 1 for greater similarity. Marvin and Laprade
(1987).
>>> contour_embed(Contour([0, 1, 2, 3]), Contour([0, 1, 2]))
1.0
"""
cseg, csubseg = utils.greatest_first(cseg1, cseg2)
n_csubseg = contour.Contour(csubseg).translation()
cseg_size = len(cseg)
csubseg_size = len(csubseg)
embed_times = subsets_embed_number(cseg, n_csubseg)
total_subsets = subsets_embed_total_number(cseg, csubseg)
return 1.0 * embed_times / total_subsets
def test_greatest_first(self):
result = utils.greatest_first([0, 1], [3, 2, 1])
self.assertEqual(result, [[3, 2, 1], [0, 1]])
