nupyprop 0.1.7.post125

A simulator for neutrino propagation through the earth.

Propagate neutrinos through the earth.

A python package and command line utility, including fortran for performance with openMP.

Documentation (WIP): https://nupyprop.readthedocs.io/en/latest/

Note: While the documentation is currently WIP, users and developers should consult the nuPyProp tutorial repository for visualizing output from the code and creating user-defined models.

Installation

with pip

python3 -m pip install nupyprop

with conda

We recommend installing nupyprop into a conda environment like so. In this example the name of the environment is “nupyprop”

conda create -n nupyprop -c conda-forge -c nuspacesim nupyprop
conda activate nupyprop

Usage

nupyprop --help

Example for running tau propagation for 10⁷ GeV neutrinos at 10 degrees with 10⁷ neutrinos injected with stochastic energy loss & with all other parameters as defaults:

nupyprop -e 7 -a 10 -t stochastic -s 1e7

Run parameters are defined in run.py. Different switches are described as follows:

1. -e or --energy: incoming neutrino energy in log_10(GeV). Works for single energy or multiple energies. For multiple energies, separate energies with commas eg. 6,7,8,9,10,11. Default energies are 10⁶,10^6.25,10^6.5,…10¹¹ GeV.

2. -a or --angle: slant Earth angles in degrees. Works for single angle or multiple angles. For multiple angles, separate angles with commas eg. 1,3,5,7,10. Default angles are 1->42 degrees, in steps of 1 degree.

3. -i or --idepth: depth of ice/water in km. Default value is 4 km.

4. -cl or --charged_lepton: flavor of charged lepton used to propagate. Can be either muon or tau. Default is tau.

5. -n or --nu_type: type of neutrino matter. Can be either neutrino or anti-neutrino. Default is neutrino.

6. -t or --energy_loss: energy loss type for lepton - can be stochastic or continuous. Default is stochastic.

7. -x or --xc_model: neutrino/anti-neutrino cross-section model used. Can be from the pre-defined set of models (see xc-table) or custom. Default is ct18nlo.

8. -p or --pn_model: photonuclear interaction energy loss model used. Can be from the pre-defined set of models (see pn-table) or custom. Default is allm.

9. -f or --fac_nu: rescaling factor for BSM cross-sections. Default is 1.

10. -s or --stats: statistics or no. of injected neutrinos. Default is 1e7 neutrinos.

11. -htc or --htc_mode: High throughput computing mode. If set to yes, the code will be optimized to run in high throughput computing mode. Default is no.

Note: This program uses OpenMP for propagating the huge number of neutrinos injected. For this purpose, the code will use all the threads available in your processor by default. To control the number of threads used for running the code, use export OMP_NUM_THREADS=x, where x is the number of threads you want the code to run with.

Viewing output results: output_*.h5 will contain the results of the code after it has finished running. In the terminal, run vitables (optional dependency) and open the output_*.h5 file to view the output results.

output_*.h5 naming convention is as follows: output_A_B_Ckm_D_E_F_G, where

A = Neutrino type: nu is for neutrino & anu is for anti-neutrino.

B = Charged lepton type: tau is for tau leptons & muon is for muons.

C = idepth: depth of water layer (in km).

D = Neutrino (or anti-neutrino) cross-section model.

E = Charged lepton photonuclear energy loss model.

F = Energy loss type: can be stochastic or continuous.

G = Statistics (ie. no. of neutrinos/anti-neutrinos injected).

Model Tables

Neutrino/Anti-Neutrino Cross-Section Model	Reference
Abramowicz, Levin, Levy, Maor (ALLM)	hep-ph/9712415, Phys. Rev. D 96, 043003
Block, Durand, Ha, McKay (BDHM)	Phys. Rev. D 89, 094027, Phys. Rev. D 96, 043003
CTEQ18-NLO	Phys. Rev. D 103, 014013, Phys. Rev. D 81, 114012
Connolly, Thorne, Waters (CTW)	Phys. Rev. D 83, 113009
nCTEQ15	Phys. Rev. D 93, 085037, Phys. Rev. D 81, 114012
User Defined	See nuPyProp tutorial repository

Charged Lepton Photonuclear Energy Loss Model	Reference
Abramowicz, Levin, Levy, Maor (ALLM)	hep-ph/9712415, Phys. Rev. D 63, 094020
Bezrukov-Bugaev (BB)	Yad. Fiz. 33, 1195, Phys. Rev. D 63, 094020
Block, Durand, Ha, McKay (BDHM)	Phys. Rev. D 89, 094027, Phys. Rev. D 63, 094020
Capella, Kaidalov, Merino, Tran (CKMT)	Eur. Phys. J. C 10, 153 Phys. Rev. D 63, 094020
User Defined	See nuPyProp tutorial repository

Code Execution Timing Tables

Charged Lepton	Energy Loss Type	E|nu| [GeV]	Angles	N|nu|;;in	Time (hrs)
τ	Stochastic	107	1-35	108	1.07*, 0.26***
τ	Continuous	107	1-35	108	0.88*
τ	Stochastic	108	1-35	108	6.18*, 1.53***
τ	Continuous	108	1-35	108	5.51*
τ	Stochastic	109	1-35	108	27.96*, 5.08***
τ	Continuous	109	1-35	108	19.11*
τ	Stochastic	1010	1-35	108	49.80*, 12.43***
τ	Continuous	1010	1-35	108	35.59*
τ	Stochastic	1011	1-35	108	12.73***
τ	Continuous	1011	1-35	108

Charged Lepton	Energy Loss Type	E|nu| [GeV]	Angles	N|nu|;;in	Time (hrs)
μ	Stochastic	106	1,2,3,5,7,10,12,15,17,20,25,30,35	108
μ	Continuous	106	1,2,3,5,7,10,12,15,17,20,25,30,35	108	0.95*
μ	Stochastic	107	1,2,3,5,7,10,12,15,17,20,25,30,35	108
μ	Continuous	107	1,2,3,5,7,10,12,15,17,20,25,30,35	108	3.19*
μ	Stochastic	108	1,2,3,5,7,10,12,15,17,20,25,30,35	108
μ	Continuous	108	1,2,3,5,7,10,12,15,17,20,25,30,35	108	5.17*
μ	Stochastic	109	1,2,3,5,7,10,12,15,17,20,25,30,35	108	111.77**
μ	Continuous	109	1,2,3,5,7,10,12,15,17,20,25,30,35	108	7.42*
μ	Stochastic	1010	1,2,3,5,7,10,12,15,17,20,25,30,35	108	98.17*
μ	Continuous	1010	1,2,3,5,7,10,12,15,17,20,25,30,35	108	9.76*
μ	Stochastic	1011	1,2,3,5,7,10,12,15,17,20,25,30,35	108
μ	Continuous	1011	1,2,3,5,7,10,12,15,17,20,25,30,35	108

* - Intel Core i7-8750H; 6 cores & 12 threads. ** - Intel Core i5-10210; 4 cores & 8 threads. *** - UIowa Argon cluster; 56 cores.

For debugging/development: The correct order to look at the code is in the following order:

1. data.py: contains functions for reading/writing from/to hdf5 files.

2. geometry.py: contains the Earth geometry modules (including PREM) for use with python/fortran.

3. models.py: contains neutrino cross-section & charged lepton energy loss model templates.

4. propagate.f90: heart of the code; contains fortran modules to interpolate between geometry variables, cross-sections, energy loss parameters & propagate neutrinos and charged leptons through the Earth.

5. main.py: forms the main skeleton of the code; propagates the neutrinos and charged leptons, and calculates the p_exit and collects outgoing lepton energies.

6. run.py: contains all the run parameters and variables needed for all the other .py files.

Developing the code on Ubuntu

These notes should help developers of this code build and install the package locally using a pep518 compliant build system (pip).

1. Install the non-pypi required dependencies as described for users above.

2. Install a fortran compiler. ex: sudo apt-get install gfortran

3. git clone the source code: git clone [e-mail]:NuSpaceSim/nupyprop.git

4. cd nupyprop

5. build and install the package in ‘editable’ mode python3 -m pip install -e .

Developing the code on MacOS

These notes should help developers of this code build and install the package locally using a pep518 compliant build system (pip). Currently we do not support the default system python3 on MacOS which is out of date and missing critical functionality. Use the homebrew python instead, or a virtualenv, or a conda environment.

1. Install the non-pypi required dependencies as described for users above.

2. Install a fortran compiler. ex: brew install gcc

3. git clone the source code: git clone [e-mail]:NuSpaceSim/nupyprop.git

4. cd nupyprop

5. build and install the package in ‘editable’ mode python3 -m pip install -e .