import math


class TfidfConverter:
    def __init__(self, freq_vectors):
        self.freq_vectors = freq_vectors
        self.doc_size = len(freq_vectors)
        self.term_freq = {}
        self.doc_freq = {}
        self.idf = {}
        self.tfidf_vectors = {}

    def generate_tfidf_vectors(self):
        # Calculate TF and DF
        self.populate_term_doc_freq()

        # Calculate IDF
        self.calculate_idf()

        # Calculate TF-IDF
        for cluster in self.freq_vectors.keys():
            tf = self.term_freq[cluster]

            tfidf_vector = {}
            for term, freq in tf.items():
                tfidf = freq * self.idf[term]

                tfidf_vector[term] = tfidf

            self.tfidf_vectors[cluster] = tfidf_vector

    def calculate_idf(self):
        for term, doc_freq in self.doc_freq.items():
            self.idf[term] = math.log(self.doc_size / doc_freq)

    def populate_term_doc_freq(self):
        for cluster, freq_vector in self.freq_vectors.items():
            total = float(sum(freq for freq in freq_vector.values()))

            freq_dict = {}
            for term, count in freq_vector.items():
                freq_dict[term] = count / total

                if term in self.doc_freq:
                    self.doc_freq[term] = self.doc_freq[term] + 1.0
                else:
                    self.doc_freq[term] = 1.0

            self.term_freq[cluster] = freq_dict