uqsa 0.8.0

• Most output can be suppressed with a global option: options(uqsa.verbose=FALSE).
• Objects of class errors seem to cause slow downs on windows when doing arithmetic, we now more carefully drop errors when a function doesn’t need them.
• Fixed some bugs related to extensions of shared libraries on windows (dll rather than so).

uqsa 0.7.8

• A new function abc_mcmc was introduced, it is a rewrite of the older ABCMCMC function, with different arguments.
• New Kullback-Leibler divergence function: KLD which can serve as fallback if other packages are unavailable for KLD calculations.
• made some examples smaller, and tests smaller so that they run faster.
• fixed several small bugs.
• removed most printing of messages, unless they are part of a warning
• interactive calls to functions that do heavy work now print a progress bar.
• examples, vignettes, and articles all restore par and option settings on completion.

uqsa 0.7.1

Windows compatibility: all checks pass.

uqsa 0.6.3

The way that we generate C code and how shared libraries are made has changed. In the past we were aiming for a workflow, where a user would have their model in a repository. This repository would have the generated C code for their model as well, so that there was no need to keep regenerating it. The same repository could have a copy of the model in various formats such as mod, sbml, vf, etc.

However, these derived formats make it quite difficult to keep everything synchronized and up to date. Therefore we consider it the more common case that we frequently change the model TSV files and then update the C code.

The model, once converted to an ODE or CME (for chemical master equation type models) now has a field for the C file storage location and the shlib function uses this field to compile the model to a shared library (same directory). The path to the so file is also stored in the ODE or CME object.

Many other changes have been made to improve convenience, as well es to fix bugs.