Working offline

Setup of python environments can be frustrating,
particularly if you're working with Windows.

There are many ways to setup python environments,
and which way is best often depends on the context.

If you're familiar with python and package managers,
you can use the list of requirements specified the yml files in
resources/ to create your env using any available approach.

For all other users, we either suggest the conda
or docker approach described below.

If you're on Windows, the approach should be the same.
However, on Windows, we suggest using
Windows Subsystem for Linux (WSL1).

Follow any guide to set up WSL1. One example
process is describes here.

(Mini)conda

We suggest following the Miniconda installer,
instead of the Anaconda distribution. A description for setup in WSL is available here.

If you do not work in WSL, we suggest installation of
Miniconda using Chocolatey Windows Package Manager
(elevated windows command line):

choco install miniconda3

Afterwards download or clone the repository,
and create a python environment from the file(s)
provided in the repo:

git clone https://gitlab.hrz.tu-chemnitz.de/tud-ifk/python_datascience_2022.git
cd python_datascience_2022
# not necessary, but recommended:
conda config --env --set channel_priority strict
conda env create -f resources/01_intro.yml

A shortcut may be to directly install using the exact requirements:

conda create \
    --name intro_env \
    --file resources/01_intro.txt

This will usually go faster, since conda does not have to "solve" the list of dependencies.

Since we are using a separate Jupyter Lab environment,
you must install those dependencies manually.

conda activate intro_env
conda install -c conda-forge ipywidgets \
    jupyter_contrib_nbextensions \
    jupyter_nbextensions_configurator \
    jupyterlab \
    jupytext \
    nbconvert

Note: On Windows, use ^ instead of \
as the multi-line-command character

Now, activate the environment and start jupyter lab:

jupyter lab

and open your webbrowser at the default URL (http://localhost:8888/)

IfK Jupyter Docker Container

We have prepared a Docker container that can be used
to serve Jupyter Lab with various environments.

First, setup Docker.
We also suggest using Chocolatey Windows Package Manager
(elevated windows command line):

choco install docker-desktop

See a complete list of instructions here.

You can use Windows Command Line, but we do suggest using
Windows Subsystem for Linux (WSL1)
for starting Docker Containers in Windows.

Afterwards, clone the IfK JupyterLab Docker Container.

git clone https://gitlab.vgiscience.de/lbsn/tools/jupyterlab.git
cd jupyterlab

Create a folder envs and copy 01_intro.yml to this folder.

mkdir envs
cd envs
wget https://gitlab.hrz.tu-chemnitz.de/tud-ifk/python_datascience_2022/-/raw/main/resources/01_intro.yml

Adjust the .env.example to specify that environment_default.yml should be used:

cd ..
cp .env.example .env

Edit the .env file and add:

ENVIRONMENT_FILE=envs/01_intro.yml
WORKER_ENV_NAME=intro_env
JUPYTER_NOTEBOOKS=/c/mypath/to/python_datascience_2022/

Build and startup the docker container:

docker-compose up -d

Go to http://localhost:8888/ and login with the password from .env.