v0.8.0
CADET-Process v0.8.0 is the culmination of the last 6 months of development.
All users are encouraged to upgrade to this release.
This release requires Python 3.9+
Highlights of this release
- Overhaul of parameter descriptors: Cleaner structure, new validation methods and improved testing.
- Improved indexing for multidimensional parameters for events and optimization variables.
- Improved parallelization: Selector for backend allows more flexibility when running optimization on multiple cores.
- Overhaul of
reactions
module: Individual reaction parameters can now be accessed using the standardparameters
interface (required for optimization). - Improved handling of variable transforms for optimization problems with linear constraints.
Breaking changes
Adsorption rates and reference concentrations
The parameter behavior for adsorption rates in CADET-Core and CADET-Process has been a point of discussion.
This parameter can represent either the "real"
To clarify this, CADET-Process initially introduced the adsorption_rate
parameter to always signify the "real"
An additional internal property, adsorption_rate_transformed
, was introduced to handle the transformation automatically before passing the value to CADET-Core.
This aimed to simplify the transfer of values to different reference concentrations or adsorption models.
Despite these efforts, this inconsistency between CADET-Core and CADET-Process lead to some confusion among users.
Starting from CADET-Process v0.8.0, a more harmonized approach has been adopted.
Now, adsorption_rate
and desorption_rate
in the SMA model (and similar models) directly map to
For users who still need access to the "real" parameter values, for example, when transferring parameters between different systems, the attributes adsorption_rate_untransformed
and desorption_rate_untransformed
have been introduced into CADET-Process binding model classes.
Pull requests for 0.8.0
- 38: Fix parallelization error
Cadet
object has no attribute_is_file_class
- 40: Add selector for parallelization backend
- 46: Fix linear constraints error in Scipy
- 54: Fix indexing for Events and OptimizationVariables
- 55: Rework interface regarding reference concentrations
- 59: Fix/change addict version