import re
import numpy as np
from sklearn.linear_model import LinearRegression
with open('testi/collodi_pinocchio.txt', mode='r', encoding='iso-8859-1') as f: 
	testo = f.read()
testo=testo[1704:]
testo=re.sub("'","' ", testo)
testo=re.sub('[\.,;:\-?!]{1}','',testo)
testo=testo.lower()
testo=testo.split()
fd=[testo.count(i) for i in set(testo)]
fd.sort(reverse=True)
V=len(set(testo))
Y=np.array([np.log(i) for i in fd])
X=np.array(list(np.log(i) for i in range(1,V+1))).reshape((-1, 1))
model = LinearRegression().fit(X, Y)
C=np.exp(model.intercept_)
a=model.coef_[0]*-1
print('C:', C)
print('a:', a)