def _calc_zipf_counts()

in data_measurements/zipf/zipf.py [0:0]

• 17 lines of code
• 3 McCabe index (conditional complexity)

    def _calc_zipf_counts(self):
        """
        The fit is based on an optimal xmin (minimum rank)
        Let's use this to make count estimates for the zipf fit,
        by multiplying the fitted pmf value by the sum of counts above xmin.
        :return: array of count values following the fitted pmf.
        """
        logs.info("Getting predicted counts.")
        if not self.alpha:
            logs.warning("Have not yet fit -- need the alpha value.")
            logs.warning("Fitting now...")
            self.calc_fit()
        logs.info(self.word_counts_unique)
        logs.info(self.xmin)
        logs.info(self.xmax)
        # The subset of words that fit
        word_counts_fit_unique = self.word_counts_unique[
                                 self.xmin + 1: self.xmax]
        pmf_mass = float(sum(word_counts_fit_unique))
        zipf_counts = np.array(
            [self._estimate_count(rank, pmf_mass) for rank in
             self.word_ranks_unique]
        )
        return zipf_counts