To use 'tick' and 'Joblib' dependencies, I had to roll back to Python 3.6. Then had to roll back to scikit-learn==0.21.2 for Joblib dependency.
Below are the steps I did in iTerm2 to create a conda enviroment to run all three notebooks in your example folder (I'm running JupyterLab 2.2.6 and Spyder 4.1.5 via Anaconda on Mohave OSX):
Create a new conda environment
conda create --name py362 python=3.6.2
conda activate py362
########################################################################
This section is only if you want to install the variable explorer extension in JupyterLab on Python 3.6.2:
First..Completely install conda nodes.js:
conda install -c conda-forge nodejs
conda install -c conda-forge/label/gcc7 nodejs
conda install -c conda-forge/label/cf201901 nodejs
conda install -c conda-forge/label/cf202003 nodejs
Now install extension:
jupyter labextension install @lckr/jupyterlab_variableinspector
########################################################################
Run the following commands before running CHIP .ipynb examples:
conda install numpy
conda install matplotlib
conda install networkx
conda install joblib
pip install tick
conda install scikit-learn==0.21.2
conda install dill
Good point to backup:
conda create --name py362clone --clone py362
For Spyder
If everything runs good in JupyterLab, now install Spyder 4.1.5
conda install -c anaconda spyder
will need to convert .ipynb to .py for Spyder IDE usage
For plots to show in JupyterLab or Spyder:
Need to put this command in every cell before first plt. command:
%matplotlib inline