dimlpfidex 
Discretized Interpretable Multi Layer Perceptron (DIMLP) and related algorithms
Installation
pip install dimlpfidex
Contribution
Get the project code
To get the latest source code, install git and clone the repository:
$ git clone https://github.com/HES-XPLAIN/dimlpfidex.git
To download the required dependencies on your system, run:
$ git submodule init
$ git submodule update
Install C++ toolchain
Linux, macOS, Windows/WSL
Install with your package manager:
- a C++ compiler (gcc/g++)
- the ninja build system (ninja or ninja-build)
- cmake
Windows
[!TIP]
Consider using Windows/WSL instead.
winget install Microsoft.VisualStudio.2022.Community
The MSVC compiler is necessary for proper packaging of the bindings.
winget install Kitware.CMake
Ensure cmake.exe
is accessible in the $PATH
environment variable.
cmake.exe --version
[!NOTE]
You will have to restart your computer for the changes to take effect.
Install Python
Install Python, version 3.9 or newer (3.11 is recommended):
- Linux, macOS, Windows/WSL: Use your package manager to install
python3
and python3-dev
- Windows:
winget install Python.Python.3.11
[!WARNING]
On Windows, avoid installing Python through the Microsoft Store as the package has additional permission restrictions.
Install Python dependencies
Using pip
python -m venv .venv
source .venv/bin/activate
pip install .
[!NOTE]
On Windows, use .venv\Scripts\activate
instead.
Work with virtualenv
To activate the virtualenv, use the standard methods:
- Unix:
source .venv/bin/activate
- Windows:
.venv\Scripts\activate
To leave the virtualenv, use deactivate
.
Add dependencies
To add new dependencies to the project, add them to the pyproject.toml
file.
To add them to the virtualenv, use:
pip install .
Install Pre-commit hooks
Git hooks are used to ensure quality checks are run by all developers every time
before a commit.
Install with pip install pre-commit
.
To enable pre-commit:
pre-commit install
Pre-commit hooks can be run manually with:
pre-commit run --all-files
Compile
Compile:
mkdir build && cd build
cmake ..
cmake --build .
[!NOTE]
On Windows, you may have to use cmake -DCMAKE_PREFIX_PATH="C:\<absolute\path\to>\.venv" ..
instead.
[!TIP]
To speed up the compilation process, you can also add -j X
with X
being your number of CPU cores.
[!WARNING]
If you need to rebuild the project, you must erase the content of the build/
directory.
Package
Create archives for distribution, from the root of the project:
python -m build
Documentation
Install Doxygen:
- Linux, macOS, Windows/WSL: Use your package manager to install
doxygen
- Windows:
winget install DimitriVanHeesch.Doxygen
Create the documentation:
mkdir build && cd build
cmake -DBUILD_DOCUMENTATION=ON ..
cmake --build .
The generated HTML documentation will be found in build/docs/sphinx
.
Release
To publish the package on PyPI, refer to RELEASE.
Credits
Our test suite is using Obesity or CVD risk dataset from AravindPCoder (under CC BY-SA 4.0 license)