Table of Contents

• CLI settings
• Setup
  • symbolic links
  • Python and R workflow
• Config
  • Neovim
  • Fish shell
  • miniconda, and fixes
  • Tmux
  • Yabai
• Saving SSH settings
  • Add SSH host
  • Add private key
• Common commands

CLI settings

This repo stores the following CLI settings I use:

1. Neovim (text editor)
2. fish shell
3. Tmux
4. Yabai (window management for macOS)

Setup

symbolic links

For a new machine, all the configs done can be done by git clone and symbolic linking the files.

# git clone
git clone https://github.com/lingjie00/cli_settings .dotfiles

# symbolic link
./install

Python and R workflow

I use Neovim for script editing. Therefore, there is no specific configurations for script files (.py and .R)

However, there are occasions visualizations or experimentation are required. In these cases, a notebook style of editing is preferred.

For Python, I use Jupyter notebook. Jupyter notebook runs a local host at port (default) 8888, which I forward to local machine using SSH.

For R, if you install the GUI RStudio then it is very troublesome to forward the GUI using X11. The simple solution would be to install Rstudio server which will work similarity as Jupyter Lab.

Config

Neovim

I have switched from Vim to Neovim to display errors inline and to have some Neovim specific Plugins, including Neovim built-in LSPs.

I started switching to Vim and full terminal based workflow from May 2021. In Nov 2021 I have managed to switch to nearly 100% of my workflow to Vim/Neovim. I use Vimium for browsing. The only time I will still need my mouse is to navigate PDF documents.

For data science related work, I use Jupytext which sync between Jupyter notebook and Python script, balances the need to experiment with data in notebooks and having the best practice in software development.

Read more about Neovim from their official page

Fish shell

I have transit from Bash to Zsh to fish shell now. Fish shell is much more responsive than zsh.

Since most of my workflow is in Neovim and Tmux now, the most usage for terminal will be to prompt files and commands.

miniconda, and fixes

I use miniconda for my local Python environment handling. Miniconda is a light-weight version of the full conda installation.

Just beware, if you are upgrading from bash to zsh after installing conda, you might lose the settings. Luckily, the fix is easy thanks to this

/(your conda installation path)/bin/conda init zsh

Tmux

Tmux is a "terminal multiplexer". The biggest benefit of Tmux is the ability to launch multiple terminals within a single screen. I usually split my screen into 1) Neovim 2) Script Terminal (Python, R) 3) checking files (with ls) and launch another window within Tmux to launch jupyter server.

I use a home server to perform heavy-duty work. Using Tmux allows me to perform my workflow with a single SSH and allows me to continue my work after resuming server back from sleep.

Yabai

Unlike Linux which has a decent window management system. MacOS's default window management is confusing with windows popping everywhere. I use yabai to control my spaces.

Saving SSH settings

If you use SSH frequently, then perhaps you also face the problem of constantly needing to enter the IP address and password for the server. There are ways to simplify the workflow

Add SSH host

Add in ~/.ssh/config

Host example_host
    User name
    HostName 111.111.111.111
    IdentityFile ~/.ssh/host.pub
    Port 22

Add private key

1. Generate SSH key in host

ssh-keygen

2. Copy key to server

Remember to copy the SSH over with e.g. scp

ssh-copy-id -i ~./ssh/host.pub user@host

Common commands

There are some common commands in (n)Vim for different file types

Note:

• for bigger projects a MakeFile is recommended
• %:p display the file path
• %:p:h displays the folder path file is in

#########
# LaTex #
#########
# generate output
pdflatex -output-directory '%:p:h'  '%:p'
# clear temp files
rm '%:p:r.aux' '%:p:r.log' '%:p:r.out'

#####
# R #
#####
# R: run the current script
Rscript '%:p'
# Rmd: compile the Rmd file
Rscript -e "rmarkdown::render('%:p')"

############
# MarkDown #
############
# compile to pdf using pandoc
pandoc '%:p' -o '%:p:r.pdf'

##########
# Python #
##########
# run the current script
python3 '%:p'
# run jupyter ascending sync (a bit buggy now)
python3 -m jupyter_ascending.requests.sync --filename '%:p'

#######
# C++ #
#######
# compile
g++ '%' -o '%:r'

########
# Json #
########
# reformat json files with python json.tool
python -m json.tool

########
# Java #
########
# compile
javac %