Pyconstruct is a Python library for declarative, constrained, structured-output prediction. When using Pyconstruct, the problem specification can be encoded in MiniZinc, a high-level constraint programming language. This means that domain knowledge can be declaratively included in the inference procedure as constraints over the optimization variables.
Check out the Quick Start guide to learn how to solve your first problem with Pyconstruct.
Have a look at the docs and the reference manual too, to learn more about it!
Pyconstruct can be installed through pip
:
pip install pyconstruct
Or by downloading the code from Github and running the following from the downloaded directory:
python setup.py install
Before using Pyconstruct you will need to install MiniZinc as well. Download the latest release of MiniZincIDE and follow the instructions.
Check out the Installation guide for more details.
This project is developed at the SML research group at the University of Trento, Italy and DTAI at KU Leuven, Belgium. The main developers and maintainers are:
- Paolo Dragone (SML)
- Stefano Teso (DTAI)
- Andrea Passerini (SML)