histogram.py 文件源码

def __init__(self, mcmc_samples, bin_edges):
        """
        Histogram Inference a la Dan Foreman-Mackey

        Parameters:
        ===========
        - mcmc_samples:      numpy array of shape (Nobs, Nsamples)
                             MCMC samples for the thing you want to histogram

        - bin_edges:         numpy.ndarray array
                             The edges of the histogram bins to use.

        """
        self.mcmc_samples = mcmc_samples
        self.bin_edges = bin_edges
        self.bin_centers = (self.bin_edges[:-1] + self.bin_edges[1:]) / 2
        self.bin_widths = np.diff(self.bin_edges)
        self.Nbins = self.bin_widths.size
        self.Nobs = self.mcmc_samples.shape[0]

        # Find which bin each q falls in
        self.bin_idx = np.digitize(self.mcmc_samples, self.bin_edges) - 1

        # Determine the censoring function for each bin (used in the integral)
        self.censor_integrals = np.array([quad(func=self.censoring_fcn,
                                               a=left, b=right)[0] for (left, right) in
                                          zip(self.bin_edges[:-1], self.bin_edges[1:])])

        # Set values needed for multinest fitting
        self.n_params = self.Nbins
        self.param_names = [r'$\theta_{}$'.format(i) for i in range(self.Nbins)]