Calculate TF-IDF from a count matrix

tfidf.py

import numpy as np
from sklearn.preprocessing import normalize

def tf_idf(X):  # corresponds to smooth=True and norm="l2" in sklearn.feature_extraction.text.TfidfVectorizer
    tf = normalize(X, norm="l1", axis=1)
    N = len(X)
    df = np.count_nonzero(X, axis=0)
    idf = np.log((N+1)/(df+1))+1
    return normalize(tf * idf, norm="l2")