Energy modeling by @jmuhlich in #553
This adds support for BNG's energy modeling formalism. Models can now contain EnergyPatterns as well as energy-based Rules with the Arrhenius rate law parameterized by activation energy (E₀) and rate distribution (φ). Energy-based models can be simulated directly in PySB or exported …
Remove pure-python exporter (#543). The pure Python exporter was a legacy feature to export a PySB model for use without PySB itself. We recommend using a virtual environment or container as a better alternative (e.g. Ananconda, Docker). Users requiring this feature who cannot use alternatives could copy …
The only change in this release is that Python version requirements are now specified in setup.py metadata (Python 2.7, or >=3.5). This is anticipated to be the final PySB release to support Python 2.7.
Local functions allow PySB expressions to accept a tagged MonomerPattern or ComplexPattern as an argument. For example, in the following snippet, A synthesizes C with rate dependent on bound B:
BngSimulator, an interface to all simulators provided by BioNetGen: ODE, partitioned-leaping, SSA (stochastic) and NFsim (network-free stochastic). See pysb.simulator.bng.BngSimulator (#272).
General improvements
Faster and more memory-efficient construction of ODEs from rules (#293).
ScipyOdeSimulator is faster in "pure Python" mode (i.e. without weave and a …
StochKitSimulator, an interface to the stochastic simulator StochKit. See pysb.simulator.stochkit.StochKitSimulator.
InitialsSensitivity, a tool to perform pairwise sensitivity analysis of initial conditions. See pysb.tools.sensitivity_analysis.InitialSensitivity. EXPERIMENTAL.
Import models directly from BioModels from just their ID (BIOMDxxx... or MODELxxx...) with pysb.importers.sbml.model_from_biomodels().
CupSodaSimulator, a wrapper for the cupSODA GPU-based ODE integrator. No special Python libraries are necessary since cupSODA comes as a standalone binary, but if PyCUDA is installed then PySB can auto-detect the configuration details of your GPU. See the documentation in pysb.simulator.CupSodaSimulator for more.
*Brand new Simulator classes and SimulationResult classes in the
pysb.simulator package. The old pysb.integrate API is unchanged but has been
reimplemented using the Simulator class.
* Model import from other types of models in the pysb.importers package.
Currently supports BNGL and SBML model import.
PySB has officially adopted semantic versioning. This means v1.1.0 includes new features as compared to v1.0.x, but is still backwards compatible so you won't have to change any of your models to upgrade to it.
New features
Full Python 3 compatibility. All tests pass under Python …