FJsample

Jackknife & Bootstrap resampling in Fortran with python bindings.

Current functionality includes 1st, 2nd, 3rd order jackknife sub ensemble generation, jackknife error estimation (not covariance!), a calculation of the mean, bootstrap resampling and a standard deviation calculation. For details see the example program(s) in app.

Compiler support

do concurrent locality specifier is required by default. This is in the bootstrap code in src/boot.f90. gfortran15.1+, intel IFX 2024+, NVidia NVHPC all have this support.

Build with fortran-lang/fpm

Fortran Package Manager (fpm) is a great package manager and build system for Fortran. You can build using provided fpm.toml:

fpm build

To use FJsample within your fpm project, add the following to your fpm.toml file:

[dependencies]
FJsample = { git="https://github.com/RJabi/FJsample.git" }

Python

Build this library for python via

cd python
pip install -r build-requirements.txt
pip install -e . --no-build-isolation

This will give the python module FJsample

Alternatively it may be used by including it as a submodule in your repository and then using a conda environment file which looks something like

name: G2L_CSRun
channels:
  - conda-forge
dependencies:
  - conda-forge::fpm            # Fortran package manager
  - conda-forge::meson-python   # For building fortran for use in python
  - conda-forge::ninja
  - conda-forge::pip
  - conda-forge::numpy
  - pip:
       - ../../libs/FJsample/python

If you have difficulty due to the do-concurrent locality specifier, check the python/meson.build to specify a different compiler. Note that you should also export that compiler to your (shell) environment using i.e.

export FC=gfortran15.1
export FPM_FC=gfortran15.1

as it is difficult to set the compiler manually in the meson build system. Currently the meson.build file is hardcoded to use a compiler called gfortran15.1.

---

Generated with cookiecutter template:

https://github.com/SalvadorBrandolin/fortran_meson_py