def fromMatrix(stateObjs,*transProbsPerState):
''' returns a new Chain instance from given arguments
stateObjs is a sequence of objects representing states (strings, usually);
transProbsPerState arguments contain the transiiton probability weights;
there is one such argument per state, it is a tuple (stateObj,transProbs)
where transProbs is the sequence of probability weights of transition from
stateObj to each declared state, in the order of their declarations
'''
nextStateLeas = (Alea.fromValFreqs(*zip(stateObjs,transProbs)) for (stateObj,transProbs) in transProbsPerState)
nextStateLeaPerState = tuple(zip(stateObjs,nextStateLeas))
return Chain(nextStateLeaPerState)
def __init__(self,*args):
object.__init__(self)
aClass = self.__class__
for (attrName,arg) in zip(aClass.__slots__,args):
setattr(self,attrName,arg)
length = len(str(arg))
maxLengthAttrName = '__maxLength' + attrName
if length > getattr(aClass,maxLengthAttrName):
setattr(aClass,maxLengthAttrName,length)
aClass.__templateStr = "<%s>" % ', '.join("%s=%%%ds"%(attrName,getattr(aClass,'__maxLength'+attrName)) for attrName in aClass.__slots__)
def fromSeq(stateObjSeq):
''' returns a new Chain instance from given sequence of state objects
the probabilities of state transitions are set according to transition
frequencies in the given sequence
'''
(fromStateObjIter,toStateObjIter) = tee(stateObjSeq);
for _ in toStateObjIter:
break
nextStateObjsDict = dict()
for (fromStateObj,toStateObj) in zip(fromStateObjIter,toStateObjIter):
nextStateObjs = nextStateObjsDict.get(fromStateObj)
if nextStateObjs is None:
nextStateObjs = []
nextStateObjsDict[fromStateObj] = nextStateObjs
nextStateObjs.append(toStateObj)
nextStateNameAndObjs = list(nextStateObjsDict.items())
nextStateNameAndObjs.sort()
nextStateLeaPerState = tuple((stateObj,Alea.fromVals(*nextStateObjs)) for (stateObj,nextStateObjs) in nextStateNameAndObjs)
return Chain(nextStateLeaPerState)
def __str(self,formatFunc):
''' returns a string representation of the Markov chain
with each state S followed by the indented representation of probability distribution
of transition from S to next state
formatFunc is a function to represent probability distribution, taking a Lea instance
as argument and returning a string
'''
nextStateLeas = (nextStateILea._lea1 for nextStateILea in self._nextStateBlea._iLeas)
formattedNextStateLeas = (' -> ' + nextStateLea.map(str) for nextStateLea in nextStateLeas)
return '\n'.join('%s\n%s'%(stateObj,formatFunc(formattedNextStateLea)) for (stateObj,formattedNextStateLea) in zip(self._stateObjs,formattedNextStateLeas))
def _genVPs(self):
for (leaArg, factor) in zip(self._leaArgs, self._factors):
for (v, p) in leaArg.genVPs():
yield (v, p * factor)
