Change Log

Version 1.14.0

• Added deprecation notices and details of randomgen

Version 1.13.3

• Build Linux wheel using manylinux1
• Allow (randint()) to broadcast inputs
• Sync with upstream NumPy changes
• Add protection against negative inputs in (dirichlet())

Version 1.13.2

• Add Ziggurat generation for standard gamma (standard_gamma()) for both floats and doubles. The gamma generator uses a rejection sampler that depends on random double, normal and/or exponential values.

In [1]: import numpy as np

In [2]: import randomstate as rs

In [3]: rs.seed(23456)

In [4]: rs.standard_gamma(2, size=3, method='zig') # New method
Out[4]: array([2.72136867, 2.17323252, 2.16020074])

In [5]: rs.standard_gamma(2, size=3, method='inv') # Old method
Out[5]: array([4.36635148, 1.86539508, 2.32948412])

• Add Ziggurat generator for standard exponential (standard_exponential()) for both floats and doubles

In [6]: import numpy as np

In [7]: import randomstate as rs

In [8]: rs.seed(23456)

In [9]: rs.standard_exponential(3, method='zig') # New method
Out[9]: array([0.57465034, 0.7605474 , 2.57592118])

In [10]: rs.standard_exponential(3, method='inv') # Old method
Out[10]: array([0.37352726, 0.17668242, 0.45267949])

• Add SIMD-oriented Fast Mersenne Twister (SFMT) generator.
• Add complex normal (complex_normal())
• Added support for jumping the MT19937 generator
• Added support for jumping the SFMT generator

Version 1.13

• Synchronize with NumPy 1.13 release

Version 1.12

• Synchronize with NumPy 1.12 release
• Add out argument which allows filling existing arrays. This feature was added to facilitate multithreaded filling of arrays using parallel random number generators.
  • Uniforms (random_sample())
  • Normals (standard_normal())
  • Standard Exponentials (standard_exponential())
  • Standard Gammas (standard_gamma())

Version 1.11.4

• Fix for error in Ziggurat implementation of Normal

Version 1.11.3

• Extended 32-bit generation to

  • Uniforms (random_sample() and rand())
  • Normals (standard_normal() and randn())
  • Standard Gammas (standard_gamma())
  • Standard Exponentials (standard_exponential())

  using the dtype keyword.

• Removed random_uintegers since these are special cases of randint

• Release to include files required for install from PyPi

Version 1.11.2

• Added keyword argument dtype to random_sample which allows for single precision as well as double precision uniforms to be generated.

In [11]: import numpy as np

In [12]: import randomstate as rs

In [13]: rs.seed(23456)

In [14]: rs.random_sample(3, dtype=np.float64)
Out[14]: array([0.32180029, 0.32730047, 0.92742231])

In [15]: rs.seed(23456)

In [16]: rs.random_sample(3, dtype=np.float32)
Out[16]: array([0.32180023, 0.12272525, 0.32730043], dtype=float32)

Version 1.11.1

• Added xoroshiro128+ PRNG. This is an improved version of the xorshirt128+ PRNG and should be used instead. In the long run, xorshift128+ will likely be removed.
• Fixed DeprecationWarning when initializing a PRNG using a single element array.

Version 1.11

• Update to recent changes in NumPy’s RandomState
• Expose system entropy through randomstate.entropy.random_entropy()
• Add vector initialization for all PRNGs

Version 1.10.1

• Added support for jumping the MRG32K3A generator
• Added support for jumping the dSFMT generator
• Update to recent changes in NumPy’s RandomState

Version 1.10

• This is the initial release with compatibility with NumPy 1.10