Skip to content
Python bindings for LibNormaliz
C++ Python M4 Other
Branch: master
Clone or download
Pull request Compare This branch is 68 commits behind Normaliz:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

PyNormaliz - An interface to Normaliz

What is PyNormaliz

PyNormaliz provides an interface to Normaliz ( via libNormaliz. It offers the complete functionality of Normaliz, and can be used interactively from python. For a first example, see this introduction by Richard Sieg.


The source packages of the Normaliz realeases contains PyNormaliz.


The PyNormaliz install script assumes that you have executed the

script. To install PyNormaliz and navigate to the Normaliz directory and type

python install --user

The script can be customized by some options. See Appendix E of the Normaliz manual.


The main command is Cone to create a cone, and the member functions of the cone class to compute properties. For a full list of input and output properties, see the Normaliz manual.

To create a cone, use

import PyNormaliz
C = PyNormaliz.Cone(cone = [[1,0],[0,1]])

All possible Normaliz input types can be given as keyword arguments.

To compute a property of the cone, use the provided getters, which correspond to Normaliz computation goals.


You can pass options to the compute functions

C.HilbertSeries(HSOP = True)

Low level commands

There is also a low-level API, directly using C functions:

To create a cone, use

C = NmzCone("cone", [[1,0],[0,1]])

or, equivalently,

C = NmzCone(["cone", [[1,0],[0,1]]])

NmzCone can take an arbitrary number of arguments, either as separated arguments or in a list. First is always a string, describing an input property for Normaliz, followed by a (possibly empty) matrix.

NmzCompute takes a cone as first argument, followed by arbitrary many strings, or a list of strings, describing Normaliz output types. NmzCompute lets Normaliz compute the necessary values, and returns true if everything was computed properly, false otherwise.

NmzCompute(C, "HilbertBasis")


NmzCompute(C, ["HilbertBasis"])

NmzIsComputed takes a cone and a string representing an output property, and returns true if the property is already computed for the cone, false otherwise. (In contrast to NmzCompute it does not start a computation.)

NmzIsComputed(C, "HilbertBasis")

NmzResult takes a cone and a string representing an output property, and returns the computed value of this property as a matrix, a list, or as a bool.

NmzResult(C, "HilbertBasis")
You can’t perform that action at this time.