Doing a PhD is a gigantic task and especially when you are coming from Chemistry Honors/Majors background to computational chemistry. Many Indian universities except the IITs and IISERs do not provide necessary background in computer related tech stuffs which is daily required once you join the PhD. Here I am outlining some of the things which I needed to learn for surviving the PhD days.

Note

The following tutorial is made keeping in mind that the user is using Linuxas their primary operating system.

1 Git

One of the major things to first learn is using a version control. This can be thought as a time machine for your codes and text files or more generally non binary source files.

Note

What is a non binary file?

A non binary file is a simple text file and not any application like .exe or .docx or .pptx etc.

Git is the best tool for version control until today and it is used as an industry standard. There are websites called Github or Gitlab or Codeberg which are remote or online instances of the git program. So, git is an local tool. It can be installed through any linux package manager.

sudo apt install git

1.1 Configure Git

Once installed, one has to configure the git in their own machines. To check whether git is installed or not we shall test it with following command:

git --version

This would give you some number as an output, if not then install git in your machine.

Configure your git config file:

git config --global user.name "Raunak Farhaz"
git config --global user.email "rubbish@gmail.com"
git config --global core.editor nano
git config --global init.defaultBranch main

And now to check your git config type

git config --list

1.2 Initialize a git Branch

Want to create a git repository?

git init -b main

Always check the status of your git repository in every actions

git status

1.3 Adding stuff for git to track

There are following selection of commands that one can use for adding files to be tracked by git.

git add [filename]
git add .

The second command will add everything in the folder.

Note

if you added something before and then added to .gitignore file to ignore, then it would still be tracked. Now in this kind of scenario one should remove the gitcache completely and add everything without that particular file to get a clean untracking.

1.4 Committing stuff

Once added for tracking you have to commit to create a certain timestamp, use the following command

git commit -m "creative commit message"

If you use the git commit command without the -m flag you would be eventually redirected to the git-default text editor to input a commit message.

In that case even if you keep it empty and save the file it will be an empty commit message for that timestamp.

1.5 Checking the commits

Once committed you need to see the timestamps, every timestamp has a unique hash-code.

Note

first 7 characters of hash is enough for uniqueness

The command for checking the timestamps is git log, the following commands can be useful.

git log --stat
git log --oneline

The first one gives long info and the other one is self-explanatory.

To check the timestamps one can create an alias called git graph which I use often for my machine.

Creating an Alias

git config --global alias.creative "creative command"

To create the git graph alias in your machine type the following:

git config --global alias.graph "log --all --graph --decorate --oneline"

Caution

What not to track?
Try to avoid tracking binary files, for eg. .jpeg, .png, .docx etc. Text files are always preferable.

Example : In \(\LaTeX\) files, .tex has to be version controlled only but not the .pdf files.

1.6 Branching git repositories

Want to git branch a repo that you like to edit but do not want to mess with the original? Use the branching commands

$ git branch branch_name main
$ git branch branch_name master

Note

The name master to the original branch is an old convention, the modern one is main.

1.7 List all the branches

git branch

1.8 Switching branches

git switch branch_name

1.9 Merging branches

git merge branch_name

It branches with the branch node it was branched from.

1.10 Deleting branches

git branch -d branch_name

Caution

-D flag instead of -d deletes unmerged branches

Note

If you remove the .git folder, you lose everything tracked in your local machine.

2 Secure Shell Protocol (SSH)

The calculations in computational chemistry is not something that can be done in laptops. Yes, the fastness of the calculations have increased to much greater extent but still for larger molecules one needs to use high performance compute clusters or known as HPC. These machines do not have a graphical user interface (GUI) which means it cannot be used with a mouse pointer and click to function. Hence, one needs to use what is called command line interface (CLI) or using a terminal. The HPC has to be connected from a regular laptop or desktop in your office. The connection protocol is usually done via secure shell protocol or ssh. A usual ssh command is

ssh username@hostname.domain

keywords	meaning
username	username for the account in HPC
hostname	hostname of the node of HPC can be a single name or an IP address
domain	domain of the network connected to the HPC usually something `.com` or `.org`

After entering the command one has to enter their password associated to their accounts to access the remote server. To avoid this multiple password pressing one can use the following scheme.

2.1 Passwordless ssh login

First open terminal and check whether you have a ssh key in your local laptop/desktop by checking the existence of ~/.ssh folder. And then if not then follow the steps:

ssh-keygen -t rsa

this will generate two files id_pub and id_pub.rsa

Note

If you want the passphrase to be empty press two times Enter. It is definitely recommended to give a password to secure the ssh key in your own machine.

ssh-copy-id -i ~/.ssh/id_rsa.pub username@hostname

type the password for the hostmachine login and probably it should be done. Last but least check that the contents of id_pub.rsa has copied the same content in the ~/.ssh/authorized_keys file in your remote host.

2.2 Github or Gitlab ssh keys

For passwordless git push and pulls one can copy + paste the contents of id_pub.rsa from ~/.ssh/ directory into the github or gitlab authorization sections to push and pull without password

Tip

ssh cloning is much better and more stable than http