- Don't install meta assets to
site-packages
folder - Pass GBM
rng
to underlying BrownianMotion - Update dependencies, and support Python 3.8+
- Removes zero args for dropping first sample vector value (breaking)
- Changes to diffusion process classes to align with common definitions (breaking)
- Refactor into processes and utils subpackages (breaking)
- Move base class checks into utils.validation and create abstract base classes for processes
- Provide RNG control and seeding functionality per instance and globally
- Add Dirichlet process
- Add generalized Diffusion process
- Fixed a bug with missing drift when sampling Brownian motion at specific times (thanks to MichaelHogervorst)
- Fixed implementation of fractional Brownian motion (thanks to Antony Lee)
- Fixed a bug with Bernoulli process success probability
- Added a MixedPoissonProcess (thanks to Gabinou)
- Introduced breaking changes that move the t argument of all processes to the end of the __init__ signature
- Added support for inverse Gaussian process
- Added support for colored noise processes (generalized power law, violet, blue, white, pink, red/Brownian)
- Added support for multifractional brownian motion
- Added more citations and bibliographical source page to docs
- First release.
- Support for multiple continuous-time, discrete-time, diffusion, and noise processes.