scotch implements algorithms for stochastic, continuous-time chains, or Markov processes. We are currently in alpha release; contributions are very welcome. For pull requests or bug reports, find us on Github. The core developers are Quentin Caudron and Ruthie Birger.
Contents
- Quick-Start : dive straight into describing stochastic processes and simulating them with scotch.
- Model Specifications : the full specs for scotch model files.
- Scotch API and Docs : docs for model building, simulating, sampling, and plotting scotch systems.
- Examples : a series of example scotch models that work out of the box.
Roadmap
Currently implemented :
- interactive, text-based "wizard" for full model specification
- Gillespie's (SSA) algorithm
- tau-leaping algorithm
- a quick-plot method for single realisations
- repeat trajectory sampling and bootstrapping confidence intervals
- complete event tracking
For the future :
- Gibson-Bruck algorithm
- adaptive timestepping in tau-leaping
- parameter inference
Bug Reports and Contributing
Please let us know about bugs as issues on Github. Pull requests are also absolutely welcome !
Dependencies
Required
- Python >= 2.7
- Numpy
- Scipy (for
model.sample()
andmodel.plotsamples()
) - Matplotlib (for
model.plot()
andmodel.plotsamples()
)
Optional
- Seaborn