def test_insert(self):
self.assertEqual(compare('abc', 'xabc'), 1)
self.assertEqual(compare('abc', 'axbc'), 1)
self.assertEqual(compare('abc', 'abxc'), 1)
self.assertEqual(compare('abc', 'abcx'), 1)
self.assertEqual(compare('abc', 'xxabc'), 2)
self.assertEqual(compare('abc', 'axxbc'), 2)
self.assertEqual(compare('abc', 'abxxc'), 2)
self.assertEqual(compare('abc', 'abcxx'), 2)
self.assertEqual(compare('abc', 'xabcx'), 2)
def test_delete(self):
self.assertEqual(compare('abc', 'ab'), 1)
self.assertEqual(compare('abc', 'ac'), 1)
self.assertEqual(compare('abc', 'bc'), 1)
self.assertEqual(compare('a', 'abc'), 2)
self.assertEqual(compare('b', 'abc'), 2)
self.assertEqual(compare('c', 'abc'), 2)
def test_replace(self):
self.assertEqual(compare('abc', 'xbc'), 1)
self.assertEqual(compare('abc', 'axc'), 1)
self.assertEqual(compare('abc', 'abx'), 1)
self.assertEqual(compare('abc', 'xxc'), 2)
self.assertEqual(compare('abc', 'axx'), 2)
self.assertEqual(compare('abc', 'xbx'), 2)
def compress(self):
# Merging to-be-merged list with template list
for seq in tqdm(self.tbmergedList):
mergeCandidatesD1 = [
] # List containing templates that each sequence might be merged with distance of 1,
mergeCandidatesD2 = [] # and distance of 2
seqReadCount = seq[1][0]
ampID1, ampID2 = extractAmpID(seq[1][2])
self.numMergeAttempts += seqReadCount
for template in self.templateNestedList[
ampID1]: # Get edit distance between sequence and every applicable template
dist = compare(seq[0], template[0])
if dist != -1: # If distance is not more than 2, put template in consideration for merge
if dist == 1:
mergeCandidatesD1.append(template)
else:
mergeCandidatesD2.append(template)
numCandidates = len(mergeCandidatesD1) + len(mergeCandidatesD2)
if numCandidates > 0:
self.mergedCount += seqReadCount
if numCandidates > 1:
self.mergedUnsureCount += seqReadCount
if mergeCandidatesD1: # Prioritize templates that are a distance of 1, rather than 2, from the current sequence
splitValue = seqReadCount / len(
mergeCandidatesD1
) # Allocate read count equally among templates equally similar to sequence
self.mergedD1Count += seqReadCount
for template in mergeCandidatesD1:
template[1][
0] += splitValue # Increase total read count
template[1][
1] += splitValue # Increase read count of merges
else:
splitValue = seqReadCount / len(mergeCandidatesD2)
self.mergedD2Count += seqReadCount
for template in mergeCandidatesD2:
template[1][0] += splitValue
template[1][1] += splitValue
else:
if seqReadCount >= self.j3x_readDeletorThreshold: # If we can't merge the sequence but it has a high read depth
self.leftoverList.append(seq)
else:
self.discardedList.append(seq)
self.discardCountList[ampID1] += 1 # Discard
self.failedMergeAndDiscarded += 1
# Combine the newly reinforced templates with the leftovers for inclusion in j3x
j3xSeqs = self.templateFlatList + self.leftoverList
return j3xSeqs
def test_damerau_levenshtein(self):
for i, s1 in enumerate(PATTERNS):
for j, s2 in enumerate(PATTERNS):
self.assertEqual(compare(s1, s2, True), MATRIX_DL[i][j])
def test_beyond(self):
self.assertEqual(compare('abc', 'def'), -1)
def test_emptystr(self):
self.assertEqual(compare('', ''), 0)
self.assertEqual(compare('', 'a'), 1)
self.assertEqual(compare('', 'ab'), 2)
self.assertEqual(compare('', 'abc'), -1)
self.assertEqual(compare('abc', ''), -1)
def test_transpose(self):
self.assertEqual(compare('abc', 'bac', True), 1)
self.assertEqual(compare('abc', 'acb', True), 1)
self.assertEqual(compare('abc', 'cba', True), 2)
self.assertEqual(compare('abc', 'ba', True), 2)
self.assertEqual(compare('abc', 'ca', True), 2)
def test_equal(self):
self.assertEqual(compare('abc', 'abc'), 0)
def test_insert_delete(self):
self.assertEqual(compare('abcde', 'eabcd'), 2)
self.assertEqual(compare('abcde', 'acdeb'), 2)
self.assertEqual(compare('abcde', 'abdec'), 2)
self.assertEqual(compare('ababa', 'babab'), 2)
