Conda environment files

The "yml" files found here are the environment files used to create specialized conda environments for use with python in the University of Idaho Glacier Dynamics lab. Each yml file can be used to construct an environment of internally consistent python packages for use in your research.

These environments are already available to all users on the lab server, kennicott, and are also what Tim uses on his personal computer. For use on your personal computer, you can download these environments as well, to create identical environments for computing while not connected to kennicott. To download, git clone this repository, click on the green "Code" button and download the whole lot by selecting "download zip", or right-click on each yml file individually and save the file to your computer.

The spatialenv environments in particular have many components that take conda a long while (more than ~30 min) to solve. Be patient.

By building each environment from scratch, through these environment (".yml") files, environments are easier to maintain up-to-date and conflict free than if new modules are added incrementally with occasional conda install commands. Read more about managing conda environments here, at the conda website, to create new environments on your local computer from these .yml files. On a PC computer running Windows, management of these environments happens through the "Anaconda prompt."

Tim expects to build new environments approximately every 6 months (Jan and Jul), and name each environment by ending it with the last two digits of the year, and then "a" or "b," for the first and second environments of the year.

Launching and working with Jupyter lab

After Jupyter Lab is all set up, Jupyter Lab can be launched by typing jupyter lab into bash, Anaconda Prompt, or terminal.

Within Jupyter Lab, to have interactive plotting, include the following line near the top of your python notebook file:

%matplotlib widget

This is an example of a Jupyter "magic", and controls iPython operation. Other magics also control plotting, or work within the operating system.

Working with Jupyter Lab

Elements of the Jupyter Lab interface are still somewhat immature and under development. The following steps were necessary to get jupyter lab running with different environments and interactive plots in early 2021, but no longer appear necessary.

In order for Jupyter Lab to work correctly and have interactive plots where you can zoom around, you'll have to take several actions. Each of these actions should be run from the bash/Anaconda Prompt/terminal prompt. Enter these after conda activateing each of the applicable environments that have been built, such as base21a, spatialenv21a, etc.

1. Delete a line from a jupyter configuration file. Within ~/anaconda3/envs/ENV_NAME/etc/jupyter/jupyter_config.json, delete the line "kernel_spec_manager_class": "nb_conda_kernels.CondaKernelSpecManager" This can be accomplished by typing, for example,

vi ~/anaconda3/envs/base21a/etc/jupyter/jupyter_config.json

then press a to make the file editable, delete the line above, press esc to get out of editing mode, and finally press wq to save the file and quit the program vi.

2. Run the following lines:

jupyter labextension install @jupyter-widgets/jupyterlab-manager
jupyter labextension install jupyter-matplotlib
jupyter nbextension enable --py widgetsnbextension

python -m ipykernel install --user --name NAME_OF_ENVIRO

The purpose of the last line is to allow the specific environment you're setting up to appear as an optional environment within jupyter lab. This command was identified here and here.