This repo is meant to facilitate quickly setting up a new WSL working environment for data science on a Windows OS. Note that with the advent of WSL2, the process here might change
- On Windows install
- git for windows
- meld to C:\meld (no longer, just use VSCode for git merges and diffs)
- notepad++ with compare, markdown, and json plugins and Navajo theme selected
- VS2019 with extensions: HotKeys, HotCommands, Scope Studio, Productivity Power Tools (only if using Visual Studio)
- Windows Terminal
- VS Code (and log into github account and sync settings)
- Docker
- LinqPad
- Install WSL
- Move the following to home
- .bash_profile
- .bashrc
- .dircolors
- .gitconfig
- .vimrc
- Install conda and then environments
- Install base environment from environment_base.yml.
- Install sql environment from environment_sql.yml.
- Install environments and nbextensions in Jupyter Notebook
- Install nbstripout
- Replace WindowsTerminal settings.json file with WindowsTerminal/settings.json
- Sudo apt-get install
- gedit and sublime (optional, now I just use VSCode)
- msft sql odbc drivers for python/sql
Notes Steps to get jupyter working in WSL: https://gist.github.com/kauffmanes/5e74916617f9993bc3479f401dfec7da#gistcomment-3121935
So that jupyter notebook starts automatically in your windows browser
- Add this line to your ~/.jupyter/jupyter_notebook_config.py
c.NotebookApp.use_redirect_file = False - Set the BROWSER env var in your .bashrc Eg: export BROWSER='/mnt/c/Program Files (x86)/Microsoft/Edge/Application/msedge.exe'
2026: WSL mostly works out of the box today. Oh my posh needs some finagling, check bashrc. Also added profile.ps1 for powershell setup and .claude for claude hooks and settings.json.