Installation#

Requirements#

Module

Version

Python

>= 3.9

Other versions of Python (>3.6) might be supported. Python version 3.10 and above are now supported.

Numpy

>= 1.21.0

Older versions of numpy may be compatible.

Scipy

>= 1.8.0

MKL

>= 2022.0.1

Intel® Math Kernel Library (installed with mkl-service) provides the BLAS libraries. It is recommended that mkl-service is installed through conda

Aesara

>= 2.4.0

Latest version of Aesara can be downloaded from https://github.com/aesara-devs/aesara

Installing through Conda (miniconda)#

Download and install Miniconda.

Run the following command to install Conda and the required Conda dependencies:

You will need Miniconda (Full Anaconda works fine, but miniconda is recommmended for a minimal installation). Ensure that Conda is using at least Python 3.9.

Once Conda is installed, install the required dependencies from conda via:

Windows

$ conda install mkl-service conda-forge::cxx-compiler conda-forge::m2w64-toolchain -y

Mac OS X

$ conda install mkl-service Clang

Linux

$ conda install mkl-service conda-forge::cxx-compiler

Stable release installation#

To install the package, run this command in your terminal to download and install the latest branch of PyCMTensor from PyPI after you have installed the Conda dependencies:

pip install pycmtensor -U

This is the preferred method to install PyCMTensor, as it will always install the most recent stable release.

Optional: Alternatively, if you want the development version from the Github repository:

pip install git+https://github.com/mwong009/pycmtensor.git@develop -U

Development#

The sources for PyCMTensor can be downloaded from the Github repo.

git clone git://github.com/mwong009/pycmtensor

To set up PyCMTensor in a local development environment, you need to set up a virtual environment and install the project requirements. Follow the instructions to install Conda (miniconda), then start a new virtual environment with the provided environment_<your OS>.yml file.

For example in windows:

conda env create -f environment_windows.yml

Next, activate the virtual environment and install poetry via pip.

conda activate pycmtensor-dev
pip install poetry

Lastly, install the project and development dependencies

poetry install -E dev