jpivarski

from typing import Callable

import awkward as ak


class LazyMixin:
    def __init__(
        self,
        materialize: Callable[[], ak.contents.Content],
        length: ak._nplikes.shape.ShapeItem,

jpivarski

-- Question -----------------------------------------------

I am trying to rebin my hist objects from linear bins to log(x) bins. I see on the UHI read the docs on indexing (https://uhi.readthedocs.io/en/latest/indexing.html#) that it is possible to linear rebin with h == h[::rebin(2)], but this gives all bins the same width still. Is it possible to rebin with user defined bin edges? In this case, I wanted log(x) bins. If it is not possible, I am also thinking about making a new hist object that meets the log(x) bins by slicing the original histogram in a loop, each loop being each bin edges.

-- Answer ------------

Linear to log won't work because edges can't move. Ultimately variable space rebinning boils down to "map x bins into one" over and over, where x changes. The final set of bin edges will always be a subset of the original bin edges. You can make them sort-of log space with rounding to the nearest linear bin edge (but I'd highly recommend waiting for the next release of boost-histogram, as ful

jpivarski

jpivarski

jpivarski

jpivarski

jpivarski

import numpy as np

class System:
    G = 3
    min_distance = 0.1

    @classmethod
    def random(cls, num_particles, num_dimensions, mass_mean, mass_width, x_width, p_width):
        m = np.random.gamma(mass_mean / mass_width, mass_width, num_particles)
        x = np.random.normal(0, x_width, (num_particles, num_dimensions))

jpivarski

jpivarski

jpivarski