knox-mi-graph 1.1.0

Knowledge graph builder

knox-d

MI-graph (Machine learning graph)

machine learning experts, do things

Usage

run the following commands

• pip install --extra-index-url https://repos.knox.cs.aau.dk knox-mi-graph
• mi_graph input.json

You can try it out with our example, and put a -v for visualisation

mi_graph examples/grundfos_test.json

Run under development

at root level of the repository run:

python mi_graph file.json

Setup: virtualenv environment

Make sure pip is up-to-date:

python -m pip install --upgrade pip

Install virtualenv:

python -m pip install virtualenv

Activate environment

To create an environment run:

python -m venv env

to create an environment with the name "env". Now to activate the environment run the activate script based on your OS in "env/Scripts".

On windows 10 you can run:

env/Scripts/Activate.ps1

in your powershell, to activate the environment.

If running scripts is denied - run the following command to remove the restriction. Stackoverflow

set-executionpolicy remotesigned

Install dependencies

To install dependencies run:

pip install -r .\requirements.txt

If you encounter problems see this.

Setup: conda environment

Install conda from their website. (We recommend the mini version / miniconda)

Initialize the environment with:

• conda create --name knox-env python=3.8
• activate knox-env
• pip install -r requirements.txt And you are good to go.

Test dependencies

For developers, you also need to install the test requirements:

• pip install -r tests/requirements.txt

pylint

Before you make a pull request to master, you should run branch though pylint.

you can use pylint_runner to run all folders in the solution, or with pylint like so Run:

pip install pylint

And then

pylint folder/

Build module

You cannot be in a virtual environment, when building

• python3 setup sdist bdist_wheel
• python3 -m twine upload dist/*

TODO

• upload to knox repo instead of pypi