def findAllMotifPattern(self, maxMergeCnt, hasFilter=False):
"""
ABANDONED!!
calculate all the Patterns(mergeCnt <= maxMergeCnt) by merging iteratively
from initial pattern(mergeCnt = 0, i.e. path)
The patterns are stored in self.pattern which is a dict
key : (k, mergeCnt)
value : [(newPattern, pattern1, pattern2)]
:param hasFilter:
whether the filter works
:param maxMergeCnt:
how many times do we merge
"""
for k, pathLis in self.path.iteritems():
patternLis2 = set([(BioinfoComm.Kmer2Str(path.nodeLis), tuple(path.visitLis)) for path in pathLis])
for mergeCnt in range(maxMergeCnt):
print "searching motif pattern: length = %s, mergeCnt = %s " % (k + 1, mergeCnt + 1)
Comm.PrintTime()
# fetch two pattern list(contain all the patterns with a certain length and mergeCnt) to merge
# patternLis2 is a list of paths and never change
if (k, mergeCnt) in self.pattern: # (after the first loop) patternLis1 to be merged is pattern
mergedPatternLis = map(lambda x:x[0], self.pattern[(k, mergeCnt)])
# 'tuple' conversion is necessary otherwise there will be an error because list is unhashable
patternLis1 = set([(x.patternStr, tuple(x.visitLis)) for x in mergedPatternLis])
patternLis1 = [(x[0], x[1]) for x in patternLis1]
patternLis2 = patternLis2.union(patternLis1)
else: # (in the first loop) patternLis1 to be merged is paths
patternLis1 = patternLis2
# merge two pattern lists
for (patternStr1, visitLis1) in patternLis1:
pattern1 = Pattern(patternStr1, list(visitLis1))
if hasFilter and self.IsPatternFiltered(visitLis1): continue
for (patternStr2, visitLis2) in patternLis2:
# TODO: to be improved
if patternStr1 >= patternStr2: continue
if self.IsPatternFiltered(visitLis1): continue
visitLis2 = list(visitLis2)
pattern2 = Pattern(patternStr2, visitLis2)
newPattern = Pattern.mergePattern(pattern1, pattern2)
if not newPattern: continue
self.pattern.setdefault((k, mergeCnt + 1), [])
self.pattern[(k, mergeCnt + 1)].append((newPattern, pattern1, pattern2))
