15.1. Installation¶

Sections

System Requirements
- Python
- Windows
- Linux
Installation
Version Numbering
Version Upgrade and Downgrade
Repair Installation
Uninstall

15.1.1. System Requirements ¶

pSeven Core supports Windows 10 and newer (64-bit editions) and Linux x86_64, Python 3.6 or newer and Python 2.7.

15.1.1.1. Python ¶

Requirements to Python are:

Python 3.6 or newer.
NumPy 1.11.2 or newer.
Python 2.7 is also supported. Requires NumPy 1.6.0.

NumPy is not required during installation, though you will not be able to run pSeven Core until you install NumPy.

Additionally recommended:

pandas — data analysis and manipulation library, provides useful data structures.
SciPy — popular library for scientific programming.
Matplotlib — plotting library; compatible versions are Matplotlib 1.1 and newer.

While the above are not required, they are widely used in pSeven Core examples and guides.

Optional:

SHAP — implements a game theoretic approach to explain model output.

SHAP is required only by some pSeven Core approximation models and only if you are going to use the SHAP evaluation feature (the gtapprox.Model.shap_value() method) for that certain kind of models. Most models do not require SHAP to support shap_value().

15.1.1.2. Windows ¶

pSeven Core is tested on Windows 10, 64-bit desktop editions. Newer versions and corresponding server editions are also supported but not regularly tested.

15.1.1.3. Linux ¶

pSeven Core works on any x86_64 Linux with:

Linux kernel 2.6.18 or newer.
GNU C Library (glibc) 2.5 or newer.

15.1.2. Installation ¶

Install pSeven Core from PyPI with pip:

python -m pip install --upgrade psevencore

pip is the recommended installer for Python packages. For details and other installation options, see the Installing Packages tutorial in the Python Packaging User Guide, in particular Installing from PyPI.

Note

Register your license to be able to use pSeven Core. See section License Setup for details.

15.1.3. Version Numbering ¶

The pSeven Core package version numbering scheme is year.month.days without leading zeros.

Release series are assigned short version numbers vyear.month — for example, v2024.04 — with the leading zero in month. Short version numbers are used in various descriptions and documentation where the full package version number is not required. This simplified scheme is consistent with version numbering in pSeven Desktop and pSeven Enterprise.

In the full (package) version number, days is the number of days since the beginning of the month when this release series started. Note that days can be greater than 31. For example, the v2024.04 series might begin with the initial release 2024.4.18 in April, followed by the next release of this series 2024.4.40 in May, which would be an update or fix to the 2024.4.18 release.

15.1.4. Version Upgrade and Downgrade ¶

pip handles version upgrade and downgrade.

To upgrade to the latest version:

python -m pip install --upgrade psevencore

To downgrade or upgrade to any specific version, for example 2024.4.18 (see Version Numbering):

python -m pip install --force-reinstall psevencore==2024.4.18

Unless otherwise noted in section Version Compatibility Issues or the Changelog, new pSeven Core versions are backward compatible.

15.1.5. Repair Installation ¶

If you want to repair your installation, you can run pip with the --force-reinstall option.

To reinstall the latest version:

python -m pip install --force-reinstall psevencore

To replace the existing installation with a specific version, for example 2024.4.18 (see Version Numbering):

python -m pip install --force-reinstall psevencore==2024.4.18

15.1.6. Uninstall ¶

pip can uninstall pSeven Core versions 2024.03 and above:

python -m pip uninstall psevencore

Note

If you get the error message:

ERROR: Cannot uninstall ‘pSevenCore’. It is a distutils installed project and thus we cannot accurately determine which files belong to it which would lead to only a partial uninstall.

this means that you have an older pSeven Core version installed, which has to be removed manually. Locate its installation directory (by default, the Python site-packages directory) and delete the da directory and any pSevenCore-*.egg-info files or directories there. For full details, see Installation and Uninstallation in known issues.