Source code for tardis.energy_input.gamma_ray_packet_source

import numpy as np
import pandas as pd

from tardis.energy_input.energy_source import (
    positronium_continuum,
)
from tardis.energy_input.GXPacket import (
    GXPacketCollection,
)
from tardis.energy_input.samplers import sample_energy
from tardis.energy_input.util import (
    H_CGS_KEV,
    doppler_factor_3d,
    get_index,
    get_random_unit_vector,
)
from tardis.transport.montecarlo.packet_source import BasePacketSource


[docs] class RadioactivePacketSource(BasePacketSource): def __init__( self, packet_energy, gamma_ray_lines, positronium_fraction, inner_velocities, outer_velocities, inv_volume_time, times, energy_df_rows, effective_times, taus, parents, average_positron_energies, average_power_per_mass, **kwargs, ): self.packet_energy = packet_energy self.gamma_ray_lines = gamma_ray_lines self.positronium_fraction = positronium_fraction self.inner_velocities = inner_velocities self.outer_velocities = outer_velocities self.inv_volume_time = inv_volume_time self.times = times self.energy_df_rows = energy_df_rows self.effective_times = effective_times self.taus = taus self.parents = parents self.average_positron_energies = average_positron_energies self.average_power_per_mass = average_power_per_mass self.energy_plot_positron_rows = np.empty(0) super().__init__(**kwargs)
[docs] def create_packet_mus(self, no_of_packets, *args, **kwargs): return super().create_packet_mus(no_of_packets, *args, **kwargs)
[docs] def create_packet_radii( self, no_of_packets, inner_velocity, outer_velocity ): """Initialize the random radii of packets in a shell Parameters ---------- packet_count : int Number of packets in the shell inner_velocity : float Inner velocity of the shell outer_velocity : float Outer velocity of the shell Returns ------- array Array of length packet_count of random locations in the shell """ z = np.random.random(no_of_packets) initial_radii = ( z * inner_velocity**3.0 + (1.0 - z) * outer_velocity**3.0 ) ** (1.0 / 3.0) return initial_radii
[docs] def create_packet_nus( self, no_of_packets, energy, intensity, positronium_fraction, positronium_energy, positronium_intensity, ): """Create an array of packet frequency-energies (i.e. E = h * nu) Parameters ---------- no_of_packets : int Number of packets to produce frequency-energies for energy : One-dimensional Numpy Array, dtype float Array of frequency-energies to sample intensity : One-dimensional Numpy Array, dtype float Array of intensities to sample positronium_fraction : float The fraction of positrons that form positronium positronium_energy : array Array of positronium frequency-energies to sample positronium_intensity : array Array of positronium intensities to sample Returns ------- array Array of sampled frequency-energies array Positron creation mask """ nu_energies = np.zeros(no_of_packets) positrons = np.zeros(no_of_packets) zs = np.random.random(no_of_packets) for i in range(no_of_packets): nu_energies[i] = sample_energy(energy, intensity) # positron if nu_energies[i] == 511: # positronium formation 25% of the time if fraction is 1 if zs[i] < positronium_fraction and np.random.random() < 0.25: nu_energies[i] = sample_energy( positronium_energy, positronium_intensity ) positrons[i] = 1 return nu_energies, positrons
[docs] def create_packet_directions(self, no_of_packets): """Create an array of random directions Parameters ---------- no_of_packets : int Number of packets to produce directions for Returns ------- array Array of direction vectors """ directions = np.zeros((3, no_of_packets)) for i in range(no_of_packets): directions[:, i] = get_random_unit_vector() return directions
[docs] def create_packet_energies(self, no_of_packets, energy): """Create the uniform packet energy for a number of packets Parameters ---------- no_of_packets : int Number of packets energy : float The packet energy Returns ------- array Array of packet energies """ return np.ones(no_of_packets) * energy
[docs] def create_packet_times_uniform_time(self, no_of_packets, start, end): """Samples decay time uniformly (needs non-uniform packet energies) Parameters ---------- no_of_packets : int Number of packets start : float Start time end : float End time Returns ------- array Array of packet decay times """ z = np.random.random(no_of_packets) decay_times = z * start + (1 - z) * end return decay_times
[docs] def create_packet_times_uniform_energy( self, no_of_packets, start_tau, end_tau=0.0, decay_time_min=0.0, decay_time_max=0.0, ): """Samples the decay time from the mean lifetime of the isotopes Parameters ---------- no_of_packets : int Number of packets start_tau : float Initial isotope mean lifetime end_tau : float, optional Ending mean lifetime, by default 0.0 for single decays decay_time_min : float, optional Minimum time to decay, by default 0.0 decay_time_max : float, optional Maximum time to decay, by default 0.0 Returns ------- array Array of decay times """ decay_times = np.ones(no_of_packets) * decay_time_min for i in range(no_of_packets): # rejection sampling while (decay_times[i] <= decay_time_min) or ( decay_times[i] >= decay_time_max ): decay_times[i] = -start_tau * np.log( np.random.random() ) - end_tau * np.log(np.random.random()) return decay_times
[docs] def calculate_energy_factors(self, no_of_packets, start_time, decay_times): """Calculates the factors that adjust the energy of packets emitted before the first time step and moves those packets to the earliest possible time Parameters ---------- no_of_packets : int Number of packets start_time : float First time step decay_times : array Packet decay times Returns ------- array Energy factors array Adjusted decay times """ energy_factors = np.ones(no_of_packets) for i in range(no_of_packets): if decay_times[i] < start_time: energy_factors[i] = decay_times[i] / start_time decay_times[i] = start_time return energy_factors, decay_times
[docs] def create_packets(self, decays_per_isotope, *args, **kwargs): """Initialize a collection of GXPacket objects for the simulation to operate on. Parameters ---------- decays_per_isotope : array int64 Number of decays per simulation shell per isotope Returns ------- list List of GXPacket objects array Array of main output dataframe rows array Array of plotting output dataframe rows array Array of positron output dataframe rows """ number_of_packets = decays_per_isotope.sum().sum() decays_per_shell = decays_per_isotope.sum().values locations = np.zeros((3, number_of_packets)) directions = np.zeros((3, number_of_packets)) packet_energies_rf = np.zeros(number_of_packets) packet_energies_cmf = np.zeros(number_of_packets) nus_rf = np.zeros(number_of_packets) nus_cmf = np.zeros(number_of_packets) shells = np.zeros(number_of_packets) times = np.zeros(number_of_packets) # set packets to IN_PROCESS status statuses = np.ones(number_of_packets, dtype=np.int64) * 3 positronium_energy, positronium_intensity = positronium_continuum() self.energy_plot_positron_rows = np.zeros((number_of_packets, 4)) packet_index = 0 # go through each shell for shell_number, pkts in enumerate(decays_per_shell): isotope_packet_count_df = decays_per_isotope.T.iloc[shell_number] for isotope_name, isotope_packet_count in zip( self.gamma_ray_lines.keys(), isotope_packet_count_df.values ): isotope_energy = self.gamma_ray_lines[isotope_name][0, :] isotope_intensity = self.gamma_ray_lines[isotope_name][1, :] isotope_positron_fraction = self.calculate_positron_fraction( self.average_positron_energies[isotope_name], isotope_energy, isotope_intensity, ) tau_start = self.taus[isotope_name] if isotope_name in self.parents: tau_end = self.taus[self.parents[isotope_name]] else: tau_end = 0 # sample radii at time = 0 initial_radii = self.create_packet_radii( isotope_packet_count, self.inner_velocities[shell_number], self.outer_velocities[shell_number], ) # sample directions (valid at all times) initial_directions = self.create_packet_directions( isotope_packet_count ) # packet decay time initial_times = self.create_packet_times_uniform_energy( isotope_packet_count, tau_start, tau_end, decay_time_min=0, decay_time_max=self.times[-1], ) # get the time step index of the packets initial_time_indexes = np.array( [ get_index(decay_time, self.times) for decay_time in initial_times ] ) # get the time of the middle of the step for each packet packet_effective_times = np.array( [self.effective_times[i] for i in initial_time_indexes] ) # scale radius by packet decay time. This could be replaced with # Geometry object calculations. Note that this also adds a random # unit vector multiplication for 3D. May not be needed. initial_locations = ( initial_radii * packet_effective_times * self.create_packet_directions(isotope_packet_count) ) # get the packet shell index initial_shells = np.ones(isotope_packet_count) * shell_number # the individual gamma-ray energies that make up a packet # co-moving frame, including positronium formation initial_nu_energies_cmf, positron_mask = self.create_packet_nus( isotope_packet_count, isotope_energy, isotope_intensity, self.positronium_fraction, positronium_energy, positronium_intensity, ) # equivalent frequencies initial_nus_cmf = initial_nu_energies_cmf / H_CGS_KEV # compute scaling factor for packets emitted before start time # and move packets to start at that time # probably not necessary- we have rejection sampling in the # create_packet_times_uniform_energy method energy_factors, initial_times = self.calculate_energy_factors( isotope_packet_count, self.times[0], initial_times ) # the CMF energy of a packet scaled by the "early energy factor" initial_packet_energies_cmf = ( self.create_packet_energies( isotope_packet_count, self.packet_energy ) * energy_factors ) # rest frame gamma-ray energy and frequency # this probably works fine without the loop initial_packet_energies_rf = np.zeros(isotope_packet_count) initial_nus_rf = np.zeros(isotope_packet_count) for i in range(isotope_packet_count): doppler_factor = doppler_factor_3d( initial_directions[:, i], initial_locations[:, i], initial_times[i], ) initial_packet_energies_rf[i] = ( initial_packet_energies_cmf[i] / doppler_factor ) initial_nus_rf[i] = initial_nus_cmf[i] / doppler_factor self.energy_plot_positron_rows[i] = np.array( [ packet_index, isotope_positron_fraction * self.packet_energy, # * inv_volume_time[packet.shell, decay_time_index], initial_radii[i], initial_times[i], ] ) packet_index += 1 # deposit positron energy for time in initial_time_indexes: self.energy_df_rows[shell_number, time] += ( isotope_positron_fraction * self.packet_energy ) # collect packet properties locations[ :, packet_index - isotope_packet_count : packet_index ] = initial_locations directions[ :, packet_index - isotope_packet_count : packet_index ] = initial_directions packet_energies_rf[ packet_index - isotope_packet_count : packet_index ] = initial_packet_energies_rf packet_energies_cmf[ packet_index - isotope_packet_count : packet_index ] = initial_packet_energies_cmf nus_rf[ packet_index - isotope_packet_count : packet_index ] = initial_nus_rf nus_cmf[ packet_index - isotope_packet_count : packet_index ] = initial_nus_cmf shells[ packet_index - isotope_packet_count : packet_index ] = initial_shells times[ packet_index - isotope_packet_count : packet_index ] = initial_times return GXPacketCollection( locations, directions, packet_energies_rf, packet_energies_cmf, nus_rf, nus_cmf, statuses, shells, times, )
[docs] def calculate_positron_fraction( self, positron_energy, isotope_energy, isotope_intensity ): """Calculate the fraction of energy that an isotope releases as positron kinetic energy Parameters ---------- positron_energy : float Average kinetic energy of positrons from decay isotope_energy : numpy array Photon energies released by the isotope isotope_intensity : numpy array Intensity of photon energy release Returns ------- float Fraction of energy released as positron kinetic energy """ return positron_energy / np.sum(isotope_energy * isotope_intensity)
[docs] class GammaRayPacketSource(BasePacketSource): def __init__( self, packet_energy, gamma_ray_lines, positronium_fraction, inner_velocities, outer_velocities, inv_volume_time, times, energy_df_rows, effective_times, taus, parents, average_positron_energies, average_power_per_mass, **kwargs, ): self.packet_energy = packet_energy self.gamma_ray_lines = gamma_ray_lines self.positronium_fraction = positronium_fraction self.inner_velocities = inner_velocities self.outer_velocities = outer_velocities self.inv_volume_time = inv_volume_time self.times = times self.energy_df_rows = energy_df_rows self.effective_times = effective_times self.taus = taus self.parents = parents self.average_positron_energies = average_positron_energies self.average_power_per_mass = average_power_per_mass self.energy_plot_positron_rows = np.empty(0) super().__init__(**kwargs)
[docs] def create_packet_mus(self, no_of_packets, *args, **kwargs): return super().create_packet_mus(no_of_packets, *args, **kwargs)
[docs] def create_packet_radii(self, sampled_packets_df): """Initialize the random radii of packets in a shell Parameters ---------- packet_count : int Number of packets in the shell sampled_packets_df : pd.DataFrame Dataframe where each row is a packet Returns ------- array Array of length packet_count of random locations in the shell """ z = np.random.random(len(sampled_packets_df)) initial_radii = ( z * sampled_packets_df["inner_velocity"] ** 3.0 + (1.0 - z) * sampled_packets_df["outer_velocity"] ** 3.0 ) ** (1.0 / 3.0) return initial_radii
[docs] def create_packet_nus( self, no_of_packets, packets, positronium_fraction, positronium_energy, positronium_intensity, ): """Create an array of packet frequency-energies (i.e. E = h * nu) Parameters ---------- no_of_packets : int Number of packets to produce frequency-energies for packets : pd.DataFrame DataFrame of packets positronium_fraction : float The fraction of positrons that form positronium positronium_energy : array Array of positronium frequency-energies to sample positronium_intensity : array Array of positronium intensities to sample Returns ------- array Array of sampled frequency-energies """ energy_array = np.zeros(no_of_packets) zs = np.random.random(no_of_packets) for i in range(no_of_packets): # positron if packets.iloc[i]["decay_type"] == "bp": # positronium formation 75% of the time if fraction is 1 if zs[i] < positronium_fraction and np.random.random() < 0.75: energy_array[i] = sample_energy( positronium_energy, positronium_intensity ) else: energy_array[i] = 511 else: energy_array[i] = packets.iloc[i]["radiation_energy_kev"] return energy_array
[docs] def create_packet_directions(self, no_of_packets): """Create an array of random directions Parameters ---------- no_of_packets : int Number of packets to produce directions for Returns ------- array Array of direction vectors """ directions = np.zeros((3, no_of_packets)) for i in range(no_of_packets): directions[:, i] = get_random_unit_vector() return directions
[docs] def create_packet_energies(self, no_of_packets, energy): """Create the uniform packet energy for a number of packets Parameters ---------- no_of_packets : int Number of packets energy : float The packet energy Returns ------- array Array of packet energies """ return np.ones(no_of_packets) * energy
[docs] def create_packet_times_uniform_time(self, no_of_packets, start, end): """Samples decay time uniformly (needs non-uniform packet energies) Parameters ---------- no_of_packets : int Number of packets start : float Start time end : float End time Returns ------- array Array of packet decay times """ z = np.random.random(no_of_packets) decay_times = z * start + (1 - z) * end return decay_times
[docs] def create_packet_times_uniform_energy( self, no_of_packets, isotopes, decay_time ): """Samples the decay time from the mean lifetime of the isotopes Parameters ---------- no_of_packets : int Number of packets isotopes : pd.Series Series of packet parent isotopes decay_time : array Series of packet decay time index Returns ------- array Array of decay times """ decay_times = np.zeros(len(no_of_packets)) for i, isotope in enumerate(isotopes.to_numpy()): decay_time_min = self.times[decay_time[i]] if decay_time_min == self.times[-1]: decay_time_max = self.effective_times[-1] else: decay_time_max = self.times[decay_time[i] + 1] # rejection sampling while (decay_times[i] <= decay_time_min) or ( decay_times[i] >= decay_time_max ): decay_times[i] = -self.taus[isotope] * np.log( np.random.random() ) return decay_times
[docs] def create_packets( self, decays_per_isotope, number_of_packets, *args, **kwargs ): """Initialize a collection of GXPacket objects for the simulation to operate on. Parameters ---------- decays_per_isotope : array int64 Probability of decays per simulation shell per isotope per time step number_of_packets : int Number of packets to create Returns ------- GXPacketCollection """ # initialize arrays for most packet properties locations = np.zeros((3, number_of_packets)) directions = np.zeros((3, number_of_packets)) packet_energies_rf = np.zeros(number_of_packets) packet_energies_cmf = np.zeros(number_of_packets) nus_rf = np.zeros(number_of_packets) nus_cmf = np.zeros(number_of_packets) times = np.zeros(number_of_packets) # set packets to IN_PROCESS status statuses = np.ones(number_of_packets, dtype=np.int64) * 3 self.energy_plot_positron_rows = np.zeros((number_of_packets, 4)) # compute positronium continuum positronium_energy, positronium_intensity = positronium_continuum() # sample packets from dataframe, returning a dataframe where each row is # a sampled packet sampled_packets_df = decays_per_isotope.sample( n=number_of_packets, weights="decay_energy_erg", replace=True, random_state=np.random.RandomState(self.base_seed), ) # get unique isotopes that have produced packets isotopes = pd.unique(sampled_packets_df.index.get_level_values(2)) # compute the positron fraction for unique isotopes isotope_positron_fraction = self.calculate_positron_fraction(isotopes) # get the packet shell index shells = sampled_packets_df.index.get_level_values(1) # get the inner and outer velocity boundaries for each packet to compute # the initial radii sampled_packets_df["inner_velocity"] = self.inner_velocities[shells] sampled_packets_df["outer_velocity"] = self.outer_velocities[shells] # sample radii at time = 0 initial_radii = self.create_packet_radii(sampled_packets_df) # get the time step index of the packets initial_time_indexes = sampled_packets_df.index.get_level_values(0) # get the time of the middle of the step for each packet packet_effective_times = np.array( [self.effective_times[i] for i in initial_time_indexes] ) # packet decay time times = self.create_packet_times_uniform_energy( number_of_packets, sampled_packets_df.index.get_level_values(2), packet_effective_times, ) # scale radius by packet decay time. This could be replaced with # Geometry object calculations. Note that this also adds a random # unit vector multiplication for 3D. May not be needed. locations = ( initial_radii * packet_effective_times * self.create_packet_directions(number_of_packets) ) # sample directions (valid at all times), non-relativistic directions = self.create_packet_directions(number_of_packets) # the individual gamma-ray energy that makes up a packet # co-moving frame, including positronium formation nu_energies_cmf = self.create_packet_nus( number_of_packets, sampled_packets_df, self.positronium_fraction, positronium_energy, positronium_intensity, ) # equivalent frequencies nus_cmf = nu_energies_cmf / H_CGS_KEV # per packet co-moving frame total energy packet_energies_cmf = self.create_packet_energies( number_of_packets, self.packet_energy ) # rest frame gamma-ray energy and frequency # this probably works fine without the loop # non-relativistic packet_energies_rf = np.zeros(number_of_packets) nus_rf = np.zeros(number_of_packets) for i in range(number_of_packets): doppler_factor = doppler_factor_3d( directions[:, i], locations[:, i], times[i], ) packet_energies_rf[i] = packet_energies_cmf[i] / doppler_factor nus_rf[i] = nus_cmf[i] / doppler_factor # deposit positron energy in both output arrays # this is an average across all packets that are created # it could be changed to be only for packets that are from positrons self.energy_plot_positron_rows[i] = np.array( [ i, isotope_positron_fraction[sampled_packets_df["isotopes"][i]] * packet_energies_cmf[i], # this needs to be sqrt(sum of squares) to get radius np.linalg.norm(locations[i]), times[i], ] ) # this is an average across all packets that are created # it could be changed to be only for packets that are from positrons self.energy_df_rows[shells[i], times[i]] += ( isotope_positron_fraction[sampled_packets_df["isotopes"][i]] * packet_energies_cmf[i] ) return GXPacketCollection( locations, directions, packet_energies_rf, packet_energies_cmf, nus_rf, nus_cmf, statuses, shells, times, )
[docs] def calculate_positron_fraction(self, isotopes): """Calculate the fraction of energy that an isotope releases as positron kinetic energy Parameters ---------- isotopes : array Array of isotope names as strings Returns ------- dict Fraction of energy released as positron kinetic energy per isotope """ positron_fraction = {} for isotope in isotopes: isotope_energy = self.gamma_ray_lines[isotope][0, :] isotope_intensity = self.gamma_ray_lines[isotope][1, :] positron_fraction[isotope] = self.average_positron_energies[ isotope ] / np.sum(isotope_energy * isotope_intensity) return positron_fraction