#format python

"""andgate using NeuralNetwork

for practical example, see [AndGateShell.py]
"""

def f(x1,x2,w1,w2):
    return x1*w1 + x2*w2

thres = 0.5
isOver=lambda x1,x2,w1,w2:f(x1,x2,w1,w2)>thres
result=isOver
isSame=lambda x1,x2,w1,w2,expected:result(x1,x2,w1,w2)==expected

def printGenerations(w1,w2):
    testCase=((0,0),(0,1),(1,0),(1,1))
    for eachX1,eachX2 in testCase:
        print "(%d,%d) : %f"%(eachX1,eachX2,f(eachX1,eachX2,w1,w2))

def initialize():
    pass


def everythingDone():
    pass

def advanceGeneration():
    pass

def printParameters():
    pass

def calculateThisGeneration(w1,w2):
    testCases=((0,0,0),
               (0,1,0),
               (1,0,0),
               (1,1,1))
    for eachX1,eachX2,expected in testCases:
        #t=result(eachX1,eachX2,w1,w2)
        if isSame(eachX1,eachX2,w1,w2,expected):
            continue
        w1,w2=adjust(eachX1,eachX2,w1,w2,expected)
    return w1,w2

def adjust(x1,x2,w1,w2,expected):
    diffForW1=0.1*(expected-f(x1,x2,w1,w2))*x1
    diffForW2=0.1*(expected-f(x1,x2,w1,w2))*x2
    newW1=w1+diffForW1
    newW2=w2+diffForW2
    return newW1,newW2


def untilDone():
    initialize()
    iteration=0
    while 1:
        iteration+=1
        calculateThisGeneration()
        if everythingDone() or tooOld(iteration):
            break
        advanceGeneration()
    printParameters()

def tooOld(iteration):
    if iteration>=MAX_ITERATION:
        raise "Too many iterations!"