''' TM '''
tm = TemporalMemory(
  columnDimensions=(2048, ),
  cellsPerColumn=32,
  activationThreshold=16,
  initialPermanence=0.21,
  connectedPermanence=0.5,
  minThreshold=12,
  maxNewSynapseCount=20,
  permanenceIncrement=0.1,
  permanenceDecrement=0.1,
  predictedSegmentDecrement=0.0,
  maxSegmentsPerCell=128,
  maxSynapsesPerSegment=32,
  seed=1960
)
''' SP '''
sp = SpatialPooler(
  inputDimensions=(encodingWidth),
  columnDimensions=(2048),
  potentialPct=0.85,
  globalInhibition=True,
  localAreaDensity=-1.0,
  numActiveColumnsPerInhArea=40.0,
  synPermInactiveDec=0.005,
  synPermActiveInc=0.04,
  synPermConnected=0.1,
  boostStrength=3.0,
  seed=1956,
  wrapAround=False
)
''' Category ENC '''
cat_enc = CategoryEncoder(w,list(set(TEST_SEQ_1)))


''' Main Function '''
predictiveCells = list()
TM_predCols = list()
AnomScores = list()
precisionRatios = list()

for inputVal in TEST_SEQ_1:
    ### Get Category Encoding
    encoding =  cat_enc.encode(inputVal)

   ### Convert encoding bits to column indices
    encoding_indices = []
    for i in range(len(encoding)):
        if list(encoding)[i] == 1:
            encoding_indices.append(i)

    ### TM compute
    tm.compute(encoding_indices,learn=True)
    activeCells = tm.getActiveCells()

    ### Get active & correctly predicted columns
    activeCols = list(set([tm.columnForCell(cell) for cell in activeCells]))
    correctly_predictedCols = [col for col in activeCols if col in TM_predCols]

    ### Get Anomaly Score & Precisoin
    ANOM = 1.0 - (len(correctly_predictedCols)/float(len(activeCols)))
    AnomScores.append(ANOM)
    precisionRatio = len(TM_predCols) / float(w)  #len(activeCells) / float(len(activeCols))
    precisionRatios.append(precisionRatio)

    ### Get Predictive Cells & Columns
    predictiveCells = tm.getPredictiveCells()
    TM_predCols = list(set([tm.columnForCell(cell) for cell in predictiveCells]))

    ### Print to screen: Input, Anomaly Score, Precision
    if print_screen:
        print('INPUT: ',inputVal,' -- ','ANOM: ',ANOM,' -- ','Precision: ',precisionRatio)