Command-line primer

Tina Düren <t.duren@bath.ac.uk>, Lev Sarkisov <Lev.Sarkisov@ed.ac.uk> and Gaël Donval <g.donval@bath.ac.uk>

This guide quickly introduces you to working in a command-line Linux environment, with elements of data transfer, plotting and computer visualizations.

Objectives:

1. Getting familiar with the Linux environment for scientific computation;

2. Learning how to move/navigate within the file hierarchy in Linux; create, move, rename and delete files and directories; look inside and edit text files; execute commands;

3. Acquiring basic skills with graph plotting software and molecular visualization software.

Just like Windows, Linux is a powerful operating system (and in contrast to Windows, it is free). Linux is used for the majority of scientific computing tasks. However, most of what you are going to do is very text-based and takes a bit of getting used to.

Connecting to a remote Linux computer

Let us start.

You are logged in on a University Windows machine. All simulations will be performed on a remote Linux computer.

Start Xming and then use Kitty to connect to linux.bath.ac.uk and log on using your normal username and password.

You will now have opened a Linux terminal on a Windows machine and can work as if you were sitting at a Linux machine. The file store is the same and you can access all your files in your H: drive from Linux. This also means that you can access all files created in the Linux environment using Windows.

Note

If you want to use your own computer or want to go a bit further about this kind of remote access, have a look at How do you access Balena?.

The command prompt

The Linux terminal provides an environment in which you can run commands by typing them (remember most programs you are going to use are text-based):

$ <command>

In this tutorial $ is used as a visual clue to introduce commands you should type at the prompt. Here, <command> is just a placeholder though, not a real command. Once the command is typed, press Enter to execute it. Throughout your research projects, you will probably only need ten or so commands which allow you to move/navigate within the file hierarchy in Linux; create, move, rename and delete files and directories; look inside and edit text files; execute other commands.

Note

If you want a more in-depth explanation, have a look at More about the command prompt.

Getting familiar with basic commands

Click on the terminal, so that the command prompt is active. Type the following (keep in mind $ should not be typed, it is only a visual indication for you):

$ ls

The command ls stands for list: it allows you to list the contents of your current directory. This directory is currently your home directory (i.e. your H: drive). Many Linux commands can take options. For example, a more detailed list output can be generated with:

$ ls -ao

I would suggest that you keep all your simulation data in a separate directory. So let’s start by making a directory Research_Project (spaces are often problematic, so use an underscore instead):

$ mkdir Research_Project

Use the ls command to see if it is indeed there.

What we will do now is to change directory to that new one, make a subdirectory called test1 and create a text file in this subdirectory. You have to tell Linux that you want to change directory by using the cd command (and provide a destination):

$ cd Research_Project
$ mkdir test1
$ cd test1

How do you know that you are now in test1? Use the path to working directory command to show your current location:

$ pwd
/u/k/ggd21/Research_Project/test1

Note that this location involves a full hierarchy of folders and subfolders (directories), starting with / (a.k.a. the root directory) and separated by /. This hierarchy is called path. In the example given above /u/k/ggd21 is the path to my personal home directory (yours will obviously be different): this is where I started from. The second part, Research_Project/test1, is where I am with respect to that starting directory.

Note

/ really represents folders: foo/bar/baz just means that you want to access baz that is located in folder bar that is itself located in folder foo. There is nothing more to it.

Paths starting by / are called absolute paths or full paths because you can’t go up from that folder: it’s the root folder, the one that contains them all. For instance pwd always gives you absolute paths: they always start with /. Let us walk our way through the hierarchy:

$ cd /
$ ls

You are now at the root of the system: ls lists all the folders in there. Now have a look at the first element in the pwd output you did previously. In my case, it was simply u: it means that there should be a folder named u in the root. There is indeed as you can see with ls. We can now go inside and have a look:

$ cd u
$ ls

Now in my case (yours may differ), I need to look for k: it is indeed in there and I can cd my way into it. Proceed in the same way to your own home folder, Research_Project and finally test1.

Let us try to move a level up, back to Research_Project. You might be tempted to type something like:

$ cd Research_Project

This is not going to work: list files in your current directory; Research_Project is not in there! To go up in the hierarchy, type:

$ cd ..

Now check that you are indeed one level up. The command cd is no way limited to single folders: you can go two (or more) levels up trivially:

$ cd ../..

You can provide no path (in which case you get back to your home directory) or more paths; you can even provide absolute paths:

$ cd
$ cd Research_Project/test1
$ cd
$ cd /u/k/ggd21/Research_Project/test1

Absolute paths like the last one (that you have to adapt to your own home directory) will work wherever you happen to be in the hierarchy. Use pwd and ls at each step to be sure you understand where you are and where you are moving to.

Hint

Just like .. means one folder up, . means current folder and ~ means home directory. The commands cd and cd ~ are equivalent.

Now go back to test1 in Research_Project if you are not already in there. We are going to create a simple text file. For this we will use an inline text editor (in other words it does have any fancy interface such as Word) called Nano:

$ nano lines.txt

Within this editor you can type text, edit it, save files and so on, using a combination of Ctrl key and one of the letter options (see the lower tab of the screen for all the options).

Type in the following point for a simple line:

1.0  2.0
2.0  4.0
3.0  9.0
4.0  16.0

Save the file (Ctrl-o) and exit the editor (Ctrl-x). Use ls to double check that the file was indeed created. To check the content of that file use the command:

$ less lines.txt

Press q to get back to the prompt. This is also a good place in the tutorial to go back to the Nano editor and fool around a bit more to understand other options.

Imagine that lines.txt is the result of a calculation that we would like to plot (for example it could be an adsorption isotherm). There are several ways to do this:

• Use the program Gnuplot directly under Linux to plot the content of lines.txt:

  $ gnuplot -p -e "plot 'lines.txt'"
  

• As the Linux file space is the same as Windows file space you can also use Excel (or any other suitable program) to plot the lines.

  Note

  This is untrue on Balena, the supercomputer you will be using later to run your simulations: you will need to copy files to a special place called $BUCSHOME instead.

• Copy-pasting from less’ output is also perfectly possible.

You will often need to copy things around to avoid starting from scratch every time you do something new. You can copy files the following way:

$ cp <original file> <new file>

For instance:

$ cp lines.txt new_lines.txt

To copy a file into a different directory, use:

$ cp <original file> <directory>/<optional file name>

To copy a whole directory, you need to instruct cp to do the copy recursively (i.e. copy the folder and its content). You can do so by adding the -r flag:

$ cp -r <original directory> <new directory>

Likewise, to remove a file, use rm (and add -r for a directory):

$ rm new_lines.txt

Warning

You can easily delete everything you did in one click or with a single command. Nothing is going to ask you if you are absolutely certain you want to erase your whole work!

Exercises

Now it is time to practice these things.

1. Go into ~/Research_Project.

2. Copy test1 to test2.

3. Remove test2/lines.txt

4. Create test2/molecule.gro and edit it to contain:

This is Trappe TNT molecule
   16
    1TNT     C1    1   3.685   0.123   3.822
    1TNT     C2    2   3.671  -0.008   3.743
    1TNT     C3    3   3.599  -0.010   3.624
    1TNT     C4    4   3.586  -0.130   3.552
    1TNT     C5    5   3.645  -0.247   3.600
    1TNT     C6    6   3.717  -0.245   3.719
    1TNT     C7    7   3.730  -0.125   3.791
    1TNT     N8    8   3.536   0.114   3.572
    1TNT     O9    9   3.547   0.219   3.635
    1TNT    O10   10   3.473   0.113   3.467
    1TNT    N11   11   3.631  -0.374   3.523
    1TNT    O12   12   3.683  -0.477   3.565
    1TNT    O13   13   3.568  -0.376   3.418
    1TNT    N14   14   3.807  -0.123   3.919
    1TNT    O15   15   3.818  -0.018   3.982
    1TNT    O16   16   3.859  -0.226   3.961
  10.00000  10.00000  10.00000

That new file contains the coordinates of a tnt molecule. You are going to visualize it from Linux using a program called Vmd:

$ vmd

This will cause several screens to pop up. Drag the newly created molecule.gro into the main area of the main window (or click Open ‣ New Molecule and select that file in the hierarchy). Now familiarise yourself with Vmd; play with the molecule and visualisation options. Particularly, find out how to save what you see on screen to images that you may later want to use in your report.

Attention

We will be using Vmd quite extensively. At this stage, you might be told something like Error: Can't open display and nothing would be displayed.

If you are using the command line, be sure XQuartz is running on MacOS and don’t forget to use the ssh -X.

If you are using Putty or Kitty, be sure Xming is running, launch Kitty, load the linux.bath.ac.uk profile and on the left pannel click on Connection ‣ SSH ‣ X11 and click on Enable X11 forwarding. If still unsure, see this.

Going further

If you feel you need to know more about the command line, you could start with More about the command prompt which contains useful information if you are:

• unsure about what you are supposed to do or type,

• stuck with a command that does not stop,

• trying to find out how to know more about a command.

Do not be afraid to search on the web for advice on using commands; there are many sites devoted to question and answer forums for Linux or for scientific computing in general.