The Python implementation of the algorithm is in the sklearn library. The whole process is much simpler. First, load the iris dataset:

from sklearn import datasets 
irisb = datasets.load_iris() 
iris = irisb['data'] 
iris.shape 

Call upon the built-in Gaussian naive Bayes estimator for a model and prediction in one step:

from sklearn.naive_bayes import GaussianNB 
gnb = GaussianNB() 
y_pred = gnb.fit(irisb.data, irisb.target).predict(irisb.data) 

Determine the accuracy of the model:

print("Number of errors out of a total %d points : %d"  
      % (irisb.data.shape[0],(irisb.target != y_pred).sum())) 
Number of errors out of a total 150 points : 6

We end up with very similar results for estimation accuracy.