Custom Functionality

Useful custom scripts I use in my setups. There are various functionalities included here, and can be accessed in the appropriate sections:

Setup
Requirements
Tunneling & Mounting
VIM

Initial Setup

My recommendation for quick setup is to clone the repo as hidden in the home directory. E.g.,

git clone https://github.com/zhampel/custom_scripts.git ~/.custom_scripts

Then, one can add in sourcing lines in ~/.bashrc (~/.bash_profile on a MacOS) to have access to the functions on the cmd line. For me these lines are:

source ${HOME}/.custom_scripts/bash/ipython_tunnel.sh
source ${HOME}/.custom_scripts/bash/mnt_machines.sh /Users/${USER}/MNTMachines

If you only want specific functionalities, relevant lines are listed in respective sections below.

Requirements

For SSH tunneling functions, one must have bash>3. Current Linux machines should have at least bash>=4, but for some reason MacOS still holds out at 3.X. The output of bash --version on my machine gives 3.2.57. Thus, to upgrade bash on the terminal on a mac (via homebrew):

brew update && install bash
sudo bash -c 'echo /usr/local/bin/bash >> /etc/shells'
chsh -s /usr/local/bin/bash

Tunneling & Mounting

Setup

Lines in .bashrc

To setup the SSH and mounting functions, the following lines must be included in ~/.bashrc (~/.bash_profile on MacOS):

source ${HOME}/.custom_scripts/bash/ipython_tunnel.sh
source ${HOME}/.custom_scripts/bash/mnt_machines.sh /Users/${USER}/MNTMachines

assuming that this repo was cloned to ~/.custom_scripts. This simply gives cmd line access to functions in those two shell scripts. The argument to mnt_machines.sh is the local directory where I access mounted remote machines. You can change that to anywhere appropriate, but if you keep that line as I wrote it, run mkdir /Users/${USER}/MNTMachines to avoid a directory dne (does not exist) error.

SSH Configuration

The SSH and drive mounting capabilities also require the use of a SSH configuration file, i.e. ~/.ssh/config. This file contains the necessary information for logging into remote machines. Here's an example with parameters defined which should facilitate smooth operation:

Host ex-machina
    HostName 10.33.0.97
    ForwardAgent yes
    ForwardX11 yes
    IdentityFile /Users/aiava/.ssh/id_rsa
    User ava

The first line Host defines a simple name, an alias, for the computer with IP address identified by the HostName keyword below. The following two lines assist in running graphics applications remotely, and are not necessary for remote cmd line work. The IdentityFile keyword points to the local SSH identification file (i.e. the one without the .pub extension) used for secure access. This file can be generated locally via ssh-keygen or provided via some other service. Finally, the User parameter is the username that you'll use on the remote machine, i.e. ex-machina.

Usage

Mounting a Machine

To mount a remote machine and thus have access to its files on your local desktop, mount it like so, using ex-machina as the example remote computer above: mymnt ex-machina

It will request user ava's password for the SSH key file. Simple as that. You should now find a directory /Users/${USER}/ex-machina that points to the home folder on that machine!

To unmount: unmntme ex-machina

Nota bene:

• On the terminal, you must not be inside the mounting directory (/Users/${USER}/ex-machina in this case), as it can cause major hang ups in MacOS Finder and terminal. Wait till successfully mounting before navigating to the directory via the Finder.

In summary:

1. To mount: mymnt ex-machina
2. Access files in MNTMachines/ex-machina directory
3. To unmount: unmtme ex-machina

Tunneling to a Machine

If you have a remote machine running a jupyter notebook that you want to access locally, you have to first run some lines in both locations. E.g. on ex-machina, run jupyter notebook --no-browser --port=8889, which says to start a jupyter session without spinning up a browser, and allow us to listen on remote port 8889. Then we can access that session locally by running ipytunnel ex-machina 8889, requiring password as usual. This will spit out a local port number between 8888-8988, something like this: Connected to lambda via port id 8890 with process id 36197. Open up a browser and go to localhost:8890. You should now be able to play with the jupyter notebook locally! To kill the session, run ipykill 36197 or if you can't recall the process id, it can be found via ipyfind ex-machina.

Nota bene:

• Allowable port numbers (both local and remote) are limited to the range [8888, 8988]. So only 100 remote notebooks can be opened at once....
• If there are multiple sessions with ex-machina opened, then ipyfind ex-machina will show them all, or if you remember the local port 8890 ipyfind 8890 will print the relevant process id.
• Keep in mind that to have multiple notebooks running on the same remote server, they each require different ports, hence the explicit inclusion in the ipytunnel command. Thus, to have two notebooks running on ex-machina I would do jupyter notebook --no-browser --port=8889 and jupyter notebook --no-browser --port=8890, for example.

In summary:

On remote machine ex-machina

1. jupyter notebook --no-browser --port=ZZZZ

On local machine
2. ipytunnel ex-machina ZZZZ
3. Get local port XXXX from output of (2), open localhost:XXXX in browser.
4. Work
5. ipyfind ex-machina or ipyfind XXXX
6. ipykill YYYY

VIM

I almost exclusively use VIM as my text editor for reasons.

Run Control File: vimrc

There are a number of nice default that I've found or stolen over the years, which are included in my .vimrc file so that they are default. To use them yourself, softlinks would be fine or simply copy the file into your home directory:

cp ~/.custom_scripts/vim/vimrc ~/.vimrc

Don't forget that period in ~/.vimrc, otherwise VIM won't use this run-control file.

Colors

Additionally, I prefer using the distinguished colorscheme, as it's easy on the eyes and it's already specified in the .vimrc included here for your enjoyment. Simply copy/paste the file distinguished.vim into the vim colors directory:

cp ~/.custom_scripts/vim/colors/distinguished.vim ~/.vim/colors/

If you desire not to use the distinguished colorscheme and thus don't copy the file, no worries, VIM will grab the default scheme.