Skip to content

Latest commit

 

History

History
108 lines (85 loc) · 3.36 KB

jupyter_lab_management_readme.md

File metadata and controls

108 lines (85 loc) · 3.36 KB

Jupyter Lab

If not already installed you will require Jupyter Lab, see conda documentation for management of your environment.

Creating a environment.yml file

Why python enviornments are used see doco here

A .yml file is used to list packages required by your application/project and is used to create the python environment.

The following is an example of a .yml which lists typical packages (some have versions applied eg python=3.6). You are also able to install using pip if a package is not avialable via the conda package repository

name: [Your Environment Name Here]
channels:
  - defaults
  - conda-forge 
dependencies: 
  - python=3.6
  - pandas=0.19.2
  - geopandas
  - requests
  - numpy
  - matplotlib
  - voila
  - jupyterlab 
  - ipywidgets 
  - jupyter
  - pint
  - autopep8
  - geojson
  - fiona
  - shapely
  - ipyleaflet
  - sidecar
  - pip
  - pip:
    - json

Setup

Prerequisites: your computer has conda installed and available from a command prompt (check with where conda on windows and which conda on Linux).

The following commands create an environment with a name that is listed in the environment.yml file.

cd c:\path\to\here [where the environment.yml is stored]
conda env create -f=./environment.yml

Alternatively

If you want another name you can override the default name:

set my_env_name=something_else 
conda env create -f=./environment.yml --name %my_env_name%

Activate the newly created environment, register it as a so-called kernel. There is a bit of legwork also to install jupyter-widgets.

set my_env_name=[Your name in the environment file]
conda activate %my_env_name%
jupyter-labextension install @jupyter-widgets/jupyterlab-manager
jupyter labextension install @jupyter-widgets/jupyterlab-sidecar
python -m ipykernel install --user --name %my_env_name% --display-name "Py3 BoM Water"

In the above steps you may get an error that node.js was not installed (or labextensions has an issue, reinstall nodejs to fix). Steps to install node.js are here https://anaconda.org/conda-forge/nodejs

If you need to customise your environment or more information see this conda documentation

Updating the environment

If one of the packages has changed you can executed the following

cd c:\path\to\here [where the environment.yml is stored]
set my_env_name=[Your name in the environment file]
conda activate %my_env_name%
conda env update --prefix ./env --file environment.yml  --prune

Getting started

The above steps only need to be followed initially or if the packages are updated. Once the conda env is active you can start jupyter lab with the following command

Using pyBRAT in Jupyter Lab: then to start jupyter run

jupyter lab

Alternatively

You can create a .bat file to activate your envirnment and start Jupyter Lab

set mypath=%cd%
@echo %mypath%
cd %mypath%
CALL activate [Your env name]
jupyter lab

If you have created widgets you can use Voila, which will start the notebook in a Voila server

voila ./notebooks/[Notebook name].ipynb