def test_empty():
l1 = 'abcde'
l2 = 'fghij'
assert(allcommonsubstrings(l1, l2) == [])
l1 = []
l2 = []
assert(allcommonsubstrings(l1, l2) == [])
def extract_single(tokens1, tokens2, minlength, maxlength, same):
"""Extract gold alignments between two phone lists.
Parameters
----------
tokens1, tokens2 : list of FragmentTokens
minlength : int
Minimum number of symbols in a fragment
same : boolean
Whether `tokens1` and `tokens2` are identical.
Returns
-------
l : list of (FragmentToken, FragmentToken)
List of token pairs containing the cooccurring fragments
"""
ids1, intervals1, phones1 = zip(*tokens1)
ids2, intervals2, phones2 = zip(*tokens2)
id1 = ids1[0] # ids are all the same
id2 = ids2[0]
css = allcommonsubstrings(phones1, phones2,
minlength=minlength, maxlength=maxlength,
same=same)
if css is None:
return []
r = []
for slice1, slice2 in css:
r.append((FragmentToken(id1,
Interval(intervals1[slice1.start].start,
intervals1[slice1.stop - 1].end),
phones1[slice1]),
FragmentToken(id2,
Interval(intervals2[slice2.start].start,
intervals2[slice2.stop - 1].end),
phones2[slice2])))
return r
def test_uniqueness(l1, l2):
results = set()
ss = allcommonsubstrings(l1, l2, minlength=4)
for slice1, slice2 in ss:
results.add((slice1.start, slice1.stop, slice2.start, slice2.stop))
assert(len(results) == len(ss))
def test_equality(l1, l2):
for slice1, slice2 in allcommonsubstrings(l1, l2, minlength=4):
assert(l1[slice1] == l2[slice2])
def test_length(l1, l2):
n = 4
for slice1, slice2 in allcommonsubstrings(l1, l2, minlength=n):
assert(slice1.stop - slice1.start >= n)
assert(slice2.stop - slice2.start >= n)
def test_simple():
l1 = 'abcde'
l2 = 'fghijabc'
m1, m2 = allcommonsubstrings(l1, l2, minlength=3)[0]
assert(l1[m1] == l2[m2])
def test_single_nonempty():
l1 = 'abcdeabc'
expected = [(slice(0, 3, None), slice(5, 8, None)),
(slice(5, 8, None), slice(0, 3, None))]
assert(allcommonsubstrings(l1, minlength=3, same=True) == expected)
def test_single_empty():
l1 = 'abcde'
assert(allcommonsubstrings(l1, same=True) == [])
