Handy Bash Functions for ML & Data Science

Once I figure out a complicated command, I make a bash function or alias for it. Here are some that I find useful:

# set up python environment.
# usage: $ makeenv <env_name>
# Note: the pip update is b/c sometimes I've gotten stuck with an ancient pip where nothing works!
makeenv() { python3 -m venv ~/envs/"$1"; source ~/envs/"$1"/bin/activate; python3 -m pip install pip -Uqq; }

# loads python environment.  
# usage: $ loadenv <env_name>
loadenv() { source ~/envs/"$1"/bin/activate; }
# same thing different name
gimme() { source ~/envs/"$1"/bin/activate; }

# downloads colab notebook.  [Writes to colab.ipynb]
# usage: $ grabcolab <sharing_URL>    
grabcolab() { fileid=$( echo "$1" | sed -E 's/.*drive\/(.*)\?.*/\1/' ); wget -O colab.ipynb 'https://docs.google.com/uc?export=download&id='$fileid; }

# runs a notebook on command line, as if it were a python script.  
# usage: $ nbrun colab.ipynb  
# [Note for student nbs, run this from an access-reduced account!]
# On my personal page I remove the pip install but put it in to make it handy for students
nbrun() { pip install -Uqq jupytext; jupytext --to py "$1";  mv  "${1%.*}".py run_this.ipy; ipython run_this.ipy;}

# creates .tgz file from directory, runs in parallel
# $ usage: $ partar <dir_name>  [no / on end.   requires pigz: $ sudo apt install pigz
partar() { tar -I pigz -cf "$1.tgz" "$1";  }

# uncompresses .tgz file into directory, runs in parallel.
# usage $ paruntar <.tgz file>
paruntar() { tar -I pigz -xf "$1";  }  

# long-list sorted by timestamp, with more-style pagination.  $@ is all arguments.
# usage: ltt *.ipynb
llt() { ls --color=auto -p -lth --color=always "$@" | less -R -X --quit-if-one-screen; }

# which alias is this?  includes functions too.  
# example: $ wa partar
wa() { type "$1";}

…these are all in my ~/.bash_aliases file, which is invoked at the bottom of my ~/.bashrc file in the line: source ~/.bash_aliases.
Note that "$1" just refers to the first argument of the command.

I used to use Anaconda & conda but found it “bulky” and “invasive”, so now I just use python venv.

Image via VQGAN+CLIP, prompt=‘computer code | bash shell scripting | data science | by james gurney | trending on artstation’