osaut / gsl-rb

Ruby GSL wrapper using FFI

gsl-rb wraps the GNU Scientific Library using FFI. See also Ruby/GSL for a c-based wrapper. GSL-rb uses FFI so it may be used from JRuby or any Ruby with FFI support.

Author

Patrick Mahoney ([email protected])

Copyright

Copyright © 2010 Patrick Mahoney

License

MIT

Usage¶ ↑

Objects such as vectors and matricies are fixed size objects; they do not automatically grow as new members are added.

The main numeric objects are GSL::Vector and GSL::Matrix. These classes have subclasses of several data typse such as GSL::Vector::Int. The default GSL::Vector.new simply calls GSL::Vector::Double.new, similar for the matrix classes.

At the moment, only type double should be expected to work.

Vector¶ ↑

A Vector is a one-dimensional array of numbers. The default Vector class uses machine ‘double’ values. Other types include Vector::Int and others that use machine ‘int’ values. Note these types refer to the underlying C-types and not to Ruby classes.

Matrix¶ ↑

A Matrix is a two-dimensional matrix. When indexing a matrix, [row, col] is the convention used.

Example: Indexing¶ ↑

require 'gsl'

v = GSL::Vector.new([1,2,3,4])
v[0] => 1
v[1] => 2

m = GSL::Matrix.new([[1,2,3],
                     [4,5,6],
                     [7,8,9]])

m[0,0] => 1
m[0,1] => 2
m[0,2] => 3
m[2,2] => 9

m = GSL::Matrix.new(3, 3, [1,2,3,4,5,6,7,8,9])
m.to_ary_rows => [[1,2,3],
                  [4,5,6],
                  [7,8,9]]

Example: Basic Linear Algebra Subprograms¶ ↑

require 'gsl'
require 'gsl/blas'

a = GSL::Vector.new([1,2,3])
b = GSL::Vector.new([2,2,2])

a.dot(b)    => 12
b.magnitude => 3.46...

m1 = GSL::Matrix.new([[ 1, 2, 3],
                      [ 4, 5, 6],
                      [ 7, 8, 9],
                      [10,11,12]])
m2 = GSL::Matrix.new([[-2],
                      [ 1],
                      [ 0]])
m1.mul(m2).to_ary_rows => [[ 0]
                           [-3],
                           [-6],
                           [-9]]

Example: Linear Algebra Solver¶ ↑

require 'gsl'
require 'gsl/linalg'

m = GSL::Matrix.new([[0.18, 0.60, 0.57, 0.96],
                     [0.41, 0.24, 0.99, 0.58],
                     [0.14, 0.30, 0.97, 0.66],
                     [0.51, 0.13, 0.19, 0.85]])

# Can also get the inverse after LU_decomp!  But as GSL docs state:
# "It is preferable to avoid direct use of the inverse whenever
# possible, as the linear solver functions can obtain the same result
# more efficiently and reliably"
m.inv