View source for BASH scripting

==Bourne-Again Shell (Bash)==

'''Bash''' is an acronym for "Bourne-Again Shell", the name of a code interpreter and a high-level programming language, and it is a must-know tool in Computational Chemistry and Biology.
You can use Bash scripting in Unix/Linux computers through a '''terminal'''.
When you initialize the '''shell''', i.e, the interpreter, your computer runs initialization files -- <code>~/.bash_profile</code>, <code>~/.bash_login</code>, and <code>~/.profile</code> (where <code>~/</code> points to your ''home'' directory) -- but we do not recommend changing these files unless you really know what you are doing.
In most cases, you can change the <code>~/.bashrc</code> file, which allows the user to customize the system according to their needs.

A '''bash script''' is a text file containing a series of instructions written in the bash language. You can create one by typing the following commands in the terminal:

  touch my_first_script.sh

which will generate a modifiable file that you can use to write the instructions to be executed by the shell. You can use the [[Vi]] text editor to write your code; just remember to add to the beginning of the file the following line:

  #!/bin/sh

This line tells the interpreter that this is a bash script. You can run your script by telling the interpreter:

  bash my_first_script.sh

or you can change the ''permissions'' of the file to make it an ''executable'' by typing:

  chmod +x my_first_script.sh

and then running:

  ./my_first_script.sh

Suppose your ''my_first_script.sh'' contains the following lines:

  #!/bin/sh
  
  number=6
  for ((i=0;i<${number};i++))
  do
      echo "Hello world ${i}"
  done

If you run <code>./my_first_script.sh</code>, the output will be:

  Hello world 0
  Hello world 1
  Hello world 2
  Hello world 3
  Hello world 4
  Hello world 5

For more on commands, see [[Unix]].

==Environment Variables==

===Declaring and accessing variables===

Bash allows the user to assign values to variables in the command line, but it is more common to set any variables inside your scripts or <code>~/.bashrc</code> file. In Bash, you define your variable using the following syntax:

  my_variable=''value''

Do not leave spaces between the variable name and its value. You can check the value of a variable by typing the following command in your terminal:

  echo $my_variable

The shell will show the following result in your screen:

  ''value''

Remember to tell the shell that you want to return the value of the <code>my_variable</code> by using the <code>$</code> sign, otherwise, you'll be shell to print the string <code>my_variable</code> on the screen.

In the previous section, the variable <code>number</code> contained the value <code>6</code>, and the variable <code>i</code> was an iteration counter that was called inside the ''for'' loop.
It is good practice to encapsulate the variable name with curly brackets <code>{}</code> to avoid ambiguities inside the code.


===Paths and Environment variables===

Every file has a [https://en.wikipedia.org/wiki/Path_(computing) path], i.e, a location within the file system. Paths can be specified in an absolute way -- with respect to the whole file system -- or in a relative way -- with respect to the working directory.
<code>HOME</code> and <code>PATH</code> are variables containing paths information, i.e., the "addresses" within the directory tree that allows the user to find their files and executables.

Variables such as <code>HOME</code> or <code>PATH</code> are inherited from the environment. 
They have short and easily memorizable names that characterize important '''environment''' features.
<code>HOME</code>, for instance, is the variable that identifies your ''home'' directory (e.g, <code>/gpfs/home/''your_username''</code> on Seawulf).
You can modify your home directory by changing <code>HOME</code> to whichever path you find more suitable, but it is advisable not to do it if you have scripts and programs that depend on a pre-existing value for the variable.

<code>PATH</code> is the variable that determines which directories the shell should look for the programs that the user might use.
If you use the Python programming language, you might need to create a <code>PYTHONPATH</code> variable to be able to include your own Python subroutines and classes. 
Similarly, if you use the Amber software, all Amber-related programs can be found in the path defined by <code>AMBERHOME</code>. 
We, DOCK6 developers and users, define <code>DOCKHOME</code> as the path to the most stable release or as we see it fit.

If you type the name of a file as if it were a command, the shell searches for this program in certain directories defined in the <code>PATH</code> variable.
<code>PATH</code> specifies the order in which these directories should be searched by the shell. 
You can add more directories to the variable by typing:

  export PATH=/new/path/to/directory:${PATH}

In the command above, you appended the path <code>/new/path/to/directory</code> to the beginning of the <code>PATH</code> variable.

==Basic commands==

Most basic commands can be found at [[Unix]]. 
Here are some tricks of the trade:

===Iterations===

If you have an iterative task to run, you can use a ''for'' loop. The simplest kind of for loop was already shown in [https://ringo.ams.stonybrook.edu/index.php/BASH_scripting#Bourne-Again_Shell_.28Bash.29 Bourne-Again Shell] section. 
Suppose, however, that each line of the file <code>systems.txt</code> contains the name of a system that you need to work on. 
The most straightforward way of doing it in bash is:

  for line in $(cat systems.txt)
  do 
       ## Commands
      echo ${line}
  done

''while'' loops can be used in a similar way:

  while IFS= read -r line
  do
      ## Commands
  done < system.txt

Every character after a <code>#</code> is a comment and will not be read by the interpreter. 
The exceptions to this rule are the <code>#!</code> that defines the shell and cluster management/job scheduling systems (see [[SLURM]]).
<code>IFS</code> stands for internal field separator and it is used by the shell to deal with word splitting.

===Conditionals===

Conditions are written as ''if...then'', ''if...then...else'' statements. 
You can nest ''if'' and ''else'' clauses as many times as necessary, but try to keep the decision-making process simple. 
Nested ''if'' clauses can be great sources of headaches.

  if  ''conditional_expression''
      then
      ##commands
  elif ''another_conditional_expression''
      then
      ##other commands
  else
      ##more commands
  fi

Bash does not care about code indentation, but it is useful to indent your code to increase its readability.

===Creating files during runtime===

You will frequently run bash scripts to prepare your files before a simulation.
This is easily done by:

  cat <<EOF > filename
  ''text''
  EOF

<code>EOF</code> means ''end of file'' and the sequence above means that ''text'' will be written in <code>filename</code>.