GSL interface for Ruby with Numo::NArray

GitHub | RubyGems

• GSL - GNU Scientific Library - version >=2.0 and <=2.3.
• Pre-alpha version under development.
• Call for help! - writing tests and examples.

Numo::GSL API document

Implemented Modules:

• Numo::GSL – Mathematical Functions
  • (Modules/Classes below are defined in Numo::GSL module, e.g., Const => Numo::GSL::Const)
• Const – Constants
• Poly – Polynomials
• Sf – Special Functions
• Rng – Random Number Generation
  • Ran – Random Number Distriutions
  • Pdf – Probability Density Functions
  • Cdf – Cumulative Distribution Functions
• Stats – Statistics
  • Rstat – Runnning Statistics
• Histogram – Histograms
  • Histogram2D – 2D Histograms
• Spline – Interpolarion
  • Spline2D – 2D Interpolarion
• Wavelet – Wavelet Transforms
  • Wavelet2D – 2D Wavelet Transforms
• Fit – Linear regression
  • Multifit – Multi-parameter regression
• SpMatrix – Sparse Matrices
  • SpBlas – Sparse BLAS
  • SpLinalg – Sparse Linear Algebra

More modules will be implemented.

Naming convention

[C] GSL function/constant => [Ruby] Numo::GSL function/constant * Constants M_2_PI => Numo::GSL::M_2_PI GSL_CONST_MKSA_ANGSTROM => Numo::GSL::Const::MKSA_ANGSTROM * Module function gsl_acosh() => Numo::GSL.acosh() gsl_sf_bessel_J0() => Numo::GSL::Sf.bessel_J0() * Class method gsl_rng_alloc() => Numo::GSL::Rng.new gsl_rng_get() => Numo::GSL::Rng#get * Subclass gsl_rng_type *gsl_rng_mt19937; => Numo::GSL::Rng::Mt19937 < Numo::GSL::Rng * Exception gsl_ran_gaussian_pdf() => Numo::GSL::Pdf.gaussian gsl_ran_gaussian() => Numo::GSL::Rng#gaussian (Rng includes Numo::GSL::Ran)

Installation

• Install Numo::NArray

• Install GSL - GNU Scientific Library version between 2.0 and 2.3.

• Install Numo::GSL

shell $ gem install numo-gsl

Quick start

If you’re familiar with Docker, the following commands should work in most cases:

git clone https://github.com/ruby-numo/numo-gsl cd gsl docker build -t numogsl . docker run -d -p 8888:8888 numogsl start-notebook.sh --NotebookApp.token=''

and open a web browser to http://localhost:8888 .

Our Docker image is based on Minimal Jupyter Notebook Stack. See https://github.com/jupyter/docker-stacks/tree/master/minimal-notebook for more details on the Docker command options.