ISU Bootcamp Shell Tutorial

The Shell

Original Material by Milad Fatenejad, Sasha Wood, and Radhika Khetani

Modified by Adina Howe

A good cheatsheet but there are many out there

http://chembytes.wdfiles.com/local--files/cheatsheets/linux4beginner.pdf

Some good resources (for those without man pages on git bash):

Man Pages: http://unixhelp.ed.ac.uk/CGI/man-cgi Shell Literacy: http://explainshell.com/explain?cmd=ls

What is the shell how do I access the shell?

The shell is a program that presents a command line interface which allows you to control your computer using commands entered with a keyboard instead of controlling graphical user interfaces (GUIs) with a mouse/keyboard combination.

Use a browser to open the tutorial on github, located at:

https://github.com/adina/2014-01-08-iastate

Click on the directory named day1.shell.

A terminal is a program you run that gives you access to the shell. There are many different terminal programs that vary across operating systems.

There are many reasons to learn about the shell. In our opinion, the most important reasons are that:

It is very common to encounter the shell and command-line-interfaces in scientific computing, so you will probably have to learn it eventually
The shell is a really powerful way of interacting with your computer. GUIs and the shell are complementary - by knowing both you will greatly expand the range of tasks you can accomplish with your computer. You will also be able to perform many tasks more efficiently.

The shell is just a program and there are many different shell programs that have been developed. The most common shell (and the one we will use) is called the Bourne-Again SHell (bash). Even if bash is not the default shell, it is usually installed on most systems and can be started by typing bash in the terminal. Many commands, especially a lot of the basic ones, work across the various shells but many things are different. I recommend sticking with bash and learning it well. (Here is a link for more information)

To open a terminal, just single click on the "Terminal" icon on the Desktop.

The Example: Manipulating Experimental Data Files

We will spend most of our time learning about the basics of the shell by manipulating some experimental data from a hearing test. To get the data for this test, you will need internet access. Just enter the command:

git clone -b R-workshop https://github.com/adina/2014-01-08-iastate

Followed by:

cd day1.shell

This command downloads all the data that will be needed for this tutorial.

Let's get started

One very basic command is echo. This command just prints text to the terminal. Try the command:

echo Hello, World

Then press enter. You should see the text "Hello, World" printed back to you. The echo command is useful for printing from a shell script, for displaying variables, and for generating known values to pass to other programs.

Moving around the file system

Let's learn how to move around the file system using command line programs. This is really easy to do using a GUI (just click on things). Once you learn the basic commands, you'll see that it is really easy to do in the shell too.

First we have to know where we are. The program pwd (print working directory) tells you where you are sitting in the directory tree. The command ls will list the files in files in the current directory. Directories are often called "folders" because of how they are represented in GUIs. Directories are just listings of files. They can contain other files or directories.

Whenever you start up a terminal, you will start in a special directory called the home directory. Every user has their own home directory where they have full access to do whatever they want.

File Types

When you enter the ls command lists the contents of the current directory. There are several items in the home directory, notice that some of them are files and some of them are directories.

Lets create an empty file using the touch command. Enter the command:

touch testfile

Then list the contents of the directory again. You should see that a new entry, called testfile, exists. The touch command just creates an empty file.

Some terminals will color the directory entries in this very convenient way. Another way you can display direcories is by using ls -F instead of ls. The -F argument modifies the results so that a slash is placed at the end of directories. If the file is executable meaning that it can be run like a program, then a star will be placed at the end of of the file name.

You can also use the command ls -l to see whether items in a directory are files or directories. ls -l gives a lot more information too, such as the size of the file and information about the owner. If the entry is a directory, then the first letter will be a "d". The fifth column shows you the size of the entries in bytes. Notice that testfile has a size of zero.

Now, let's get rid of testfile. To remove a file, just enter the command:

rm testfile

The rm command can be used to remove files. If you enter ls again, you will see that testfile is gone.

Changing Directories

Now, let's move to a different directory. Right now, we are in the "day1.shell" directory which is in the "2014-01-08-iastate" directory. You can see this with the command pwd. The command cd (change directory) is used to move around. Let's move into the ~/2014-01-08-iastate directory. Enter the following command:

cd ~/2014-01-08-iastate

Now use the ls command to see what is inside this directory. This directory contains all of the material for this boot camp. Now move to the directory containing the data for the shell tutorial:

cd day1.shell

If you enter the cd command by itself, you will return to the home directory. Try this, and then navigate back to the shell directory.

Arguments

Most programs take additional arguments that control their exact behavior. For example, -F and -l are arguments to ls. The ls program, like many programs, take a lot of arguments. But how do we know what the options are to particular commands?

Most commonly used shell programs have a manual. You can access the manual using the man program. Try entering:

man ls

This will open the manual page for ls. Use the space key to go forward and b to go backwards. When you are done reading, just hit q to exit.

Programs that are run from the shell can get extremely complicated. To see an example, open up the manual page for the find program, which we will use later this session. No one can possibly learn all of these arguments, of course. So you will probably find yourself referring back to the manual page frequently.

Examining the contents of other directories

By default, the ls commands lists the contents of the working directory (i.e. the directory you are in). You can always find the directory you are in using the pwd command. However, you can also give ls the names of other directories to view. Navigate to the home directory if you are not already there. Then enter the command:

ls /home/2014-01-08-iastate

Saving time with shortcuts, wild cards, and tab completion

Shortcuts

There are some shortcuts which you should know about. Dealing with the home directory is very common. So, in the shell the tilde character, ~, is a shortcut for your home directory. Navigate to the shell directory, then enter the command:

ls ~

This prints the contents of your home directory, without you having to type the full path. The shortcut .. always refers to the directory above your current directory. Thus:

ls ..

prints the contents of the /home/swc/boot-camps. You can chain these together, so:

ls ../../

prints the contents of /home/swc which is your home directory. Finally, the special directory . always refers to your current directory. So, ls, ls ., and ls ././././. all do the same thing, they print the contents of the current directory. This may seem like a useless shortcut right now, but we'll see when it is needed in a little while.

To summarize, the commands ls ~, ls ~/., ls ../../, and ls /home/swc all do exactly the same thing. These shortcuts are not necessary, they are provided for your convenience.

Our data set: Cochlear Implants

A cochlear implant is a small electronic device that is surgically implanted in the inner ear to give deaf people a sense of hearing. More than a quarter of a million people have them, but there is still no widely-accepted benchmark to measure their effectiveness. In order to establish a baseline for such a benchmark, our supervisor got teenagers with CIs to listen to audio files on their computer and report:

the quietest sound they could hear
the lowest and highest tones they could hear
the narrowest range of frequencies they could discriminate

To participate, subjects attended our laboratory and one of our lab techs played an audio sample, and recorded their data - when they first heard the sound, or first heard a difference in the sound. Each set of test results were written out to a text file, one set per file. Each participant has a unique subject ID, and a made-up subject name. Each experiment has a unique experiment ID. The experiment has collected 351 files so far.

The data is a bit of a mess! There are inconsistent file names, there are extraneous "NOTES" files that we'd like to get rid of, and the data is spread across many directories. We are going to use shell commands to get this data into shape. By the end we would like to:

Put all of the data into one directory called "alldata"
Have all of the data files in there, and ensure that every file has a ".txt" extension
Get rid of the extraneous "NOTES" files

If we can get through this example in the available time, we will move onto more advanced shell topics...

Wild cards

Navigate to the ~/2014-01-08-iastate/day1.shell/data/THOMAS directory. This directory contains our hearing test data for THOMAS. If we type ls, we will see that there are a bunch of files which are just four digit numbers. By default, ls lists all of the files in a given directory. The * character is a shortcut for "everything". Thus, if you enter ls *, you will see all of the contents of a given directory. Now try this command:

ls *1

And this command

ls *4*1

lists every file in the current directory which contains the number 4, and ends with the number 1. There are four such files: 0241, 0341, 0431, and 0481.

So how does this actually work? Well...when the shell (bash) sees a word that contains the * character, it automatically looks for files that match the given pattern. In this case, it identified four such files. Then, it replaced the *4*1 with the list of files, separated by spaces. In other the two commands:

ls *4*1
ls 0241 0341 0431 0481

are exactly identical. The ls command cannot tell the difference between these two things.

Short Exercise

Do each of the following using a single ls command without navigating to a different directory.

List all of the files in the data directory gerdal that contain the number 2
List all of the files in that contain the number 2, followed by the number (in any position) 9

Tab Completion

Navigate to the home directory (.../2014-01-08-iastate). Typing out directory names can waste a lot of time. When you start typing out the name of a directory, then hit the tab key, the shell will try to fill in the rest of the directory name. For example, enter:

cd d<tab>

The shell will fill in the rest of the directory name for day1.shell. Now enter:

ls ~/2014-01-08-iastate/day1.shell/e<tab><tab>

When you hit the first tab, nothing happens. The reason is that there are multiple directories in the home directory which start with e. Thus, the shell does not know which one to fill in. When you hit tab again, the shell will list the possible choices.

Tab completion can also fill in the names of programs. For example, enter e<tab><tab>. You will see the name of every program that starts with an e. One of those is the directory exercises. If you enter exe<tab> you will see that tab completion works.

Command History

You can easily access previous commands. Hit the up arrow.
Hit it again. You can step backwards through your command history. The down arrow takes your forwards in the command history.

^-C will cancel the command you are writing, and give you a fresh prompt.

^-R will do a reverse-search through your command history. This is very useful.

Which program?

Commands like ls, rm, echo, and cd are just ordinary programs on the computer. A program is just a file that you can execute. The program which tells you the location of a particular program. For example:

which ls

Will return "/bin/ls". Thus, we can see that ls is a program that sits inside of the /bin directory. Now enter:

which find

You will see that find is a program that sits inside of the /usr/bin directory.

So ... when we enter a program name, like ls, and hit enter, how does the shell know where to look for that program? How does it know to run /bin/ls when we enter ls. The answer is that when we enter a program name and hit enter, there are a few standard places that the shell automatically looks. If it can't find the program in any of those places, it will print an error saying "command not found". Enter the command:

echo $PATH

This will print out the value of the PATH environment variable. More on environment variables later. Notice that a list of directories, separated by colon characters, is listed. These are the places the shell looks for programs to run. If your program is not in this list, then an error is printed. The shell ONLY checks in the places listed in the PATH environment variable.

Navigate to the day1.shell directory and list the contents. You will notice that there is a program (executable file) called hello in this directory. Now, try to run the program by entering:

hello

You should get an error saying that hello cannot be found. That is because the directory .../2014-01-08-iastate/day1.shell is not in the PATH. You can run the hello program by entering:

./hello

Remember that . is a shortcut for the current working directory. This tells the shell to run the hello program which is located right here. So, you can run any program by entering the path to that program. You can run hello equally well by specifying:

/home/swc/2014-01-08-iastate/day1.shell/hello

Or by entering:

../day1.shell/hello

When there are no / characters, the shell assumes you want to look in one of the default places for the program.

Examining Files

We now know how to switch directories, run programs, and look at the contents of directories, but how do we look at the contents of files?

The easiest way to examine a file is to just print out all of the contents using the program cat. Enter the following command:

cat ex_data.txt

This prints out the contents of the ex_data.txt file. If you enter:

cat ex_data.txt ex_data.txt

It will print out the contents of ex_data.txt twice. cat just takes a list of file names and writes them out one after another (this is where the name comes from, cat is short for concatenate).

Short Exercises

Print out the contents of the ~/2014-01-08-iastate/day1.shell/dictionary.txt file. What does this file contain?
Without changing directories, (you should still be in shell), use one short command to print the contents of all of the files in the /home/swc/2014-01-08-iastate/day1.shell/data/THOMAS directory.

cat is a terrific program, but when the file is really big, it can be annoying to use. The program, less, is useful for this case. Enter the following command:

less ~/2014-01-08-iastate/day1.shell/dictionary.txt

less opens the file, and lets you navigate through it. The commands are identical to the man program. Use "space" to go forward and hit the "b" key to go backwards. The "g" key goes to the beginning of the file and "G" goes to the end. Finally, hit "q" to quit.

less also gives you a way of searching through files. Just hit the "/" key to begin a search. Enter the name of the word you would like to search for and hit enter. It will jump to the next location where that word is found. Try searching the dictionary.txt file for the word "cat". If you hit "/" then "enter", less will just repeat the previous search. less searches from the current location and works its way forward. If you are at the end of the file and search for the word "cat", less will not find it. You need to go to the beginning of the file and search.

Remember, the man program uses the same commands, so you can search documentation using "/" as well!

Redirection

Let's turn to the experimental data from the hearing tests that we began with. This data is located in the ~/2014-01-08-iastate/day1.shell/data directory. Each subdirectory corresponds to a particular participant in the study. Navigate to the Bert subdirectory in data. There are a bunch of text files which contain experimental data results. Lets print them all:

cat au*

Now enter the following command:

cat au* > ../all_data

This tells the shell to take the output from the cat au* command and dump it into a new file called ../all_data. To verify that this worked, examine the all_data file. If all_data had already existed, we would overwritten it. So the > character tells the shell to take the output from what ever is on the left and dump it into the file on the right. The >> characters do almost the same thing, except that they will append the output to the file if it already exists.

Short Exercise

Use >>, to append the contents of all of the files which contain the number 4 in the directory:

~/2014-01-08-iastate/day1.shell/data/gerdal

to the existing all_data file. Thus, when you are done all_data should contain all of the experiment data from Bert and any experimental data file from gerdal that contains the number 4.

Creating, moving, copying, and removing

We've created a file called all_data using the redirection operator >. This file is critical - it's our analysis results - so we want to make copies so that the data is backed up. Lets copy the file using the cp command. The cp command backs up the file. Navigate to the data directory and enter:

cp all_data all_data_backup

Now all_data_backup has been created as a copy of all_data. We can move files around using the command mv. Enter this command:

mv all_data_backup /tmp/

This moves all_data_backup into the directory /tmp. The directory /tmp is a special directory that all users can write to. It is a temporary place for storing files. Data stored in /tmp is automatically deleted when the computer shuts down.

The mv command is also how you rename files. Since this file is so important, let's rename it:

mv all_data all_data_IMPORTANT

Now the file name has been changed to all_data_IMPORTANT. Let's delete the backup file now:

rm /tmp/all_data_backup

The mkdir command is used to create a directory. Just enter mkdir followed by a space, then the directory name.

Short Exercise

Do the following:

Rename the all_data_IMPORTANT file to all_data.
Create a directory in the data directory called foo
Then, copy the all_data file into foo

By default, rm, will NOT delete directories. You can tell rm to delete a directory using the -r option. Enter the following command:

rm -r foo

Count the words

The wc program (word count) counts the number of lines, words, and characters in one or more files. Make sure you are in the data directory, then enter the following command:

wc Bert/* gerdal/*4*

For each of the files indicated, wc has printed a line with three numbers. The first is the number of lines in that file. The second is the number of words. Finally, the total number of characters is indicated. The final line contains this information summed over all of the files. Thus, there were 10445 characters in total.

Remember that the Bert/* and gerdal/*4* files were merged into the all_data file. So, we should see that all_data contains the same number of characters:

wc all_data

Every character in the file takes up one byte of disk space. Thus, the size of the file in bytes should also be 10445. Let's confirm this:

ls -l all_data

Remember that ls -l prints out detailed information about a file and that the fifth column is the size of the file in bytes.

The awesome power of the Pipe

Suppose I wanted to only see the total number of character, words, and lines across the files Bert/* and gerdal/*4*. I don't want to see the individual counts, just the total. Of course, I could just do:

wc all_data

Since this file is a concatenation of the smaller files. Sure, this works, but I had to create the all_data file to do this. Thus, I have wasted a precious 7062 bytes of hard disk space. We can do this without creating a temporary file, but first I have to show you two more commands: head and tail. These commands print the first few, or last few, lines of a file, respectively. Try them out on all_data:

head all_data
tail all_data

The -n option to either of these commands can be used to print the first or last n lines of a file. To print the first/last line of the file use:

head -n 1 all_data
tail -n 1 all_data

Let's turn back to the problem of printing only the total number of lines in a set of files without creating any temporary files. To do this, we want to tell the shell to take the output of the wc Bert/* gerdal/*4* and send it into the tail -n 1 command. The | character (called pipe) is used for this purpose. Enter the following command:

wc Bert/* gerdal/Data0559 | tail -n 1

This will print only the total number of lines, characters, and words across all of these files.

The philosophy behind these command line programs is that none of them really do anything all that impressive. BUT when you start chaining them together, you can do some really powerful things really efficiently. If you want to be proficient at using the shell, you must learn to become proficient with the pipe and redirection operators: |, >, >>.

A sorting example

Let's create a file with some words to sort for the next example. We want to create a file which contains the following names:

Bob
Alice
Diane
Charles

To do this, we need a program which allows us to create text files. There are many such programs, the easiest one which is installed on almost all systems is called nano. Navigate to /tmp and enter the following command:

nano toBeSorted

Now enter the four names as shown above. When you are done, press CONTROL+O to write out the file. Press enter to use the file name toBeSorted. Then press CONTROL+x to exit nano.

When you are back to the command line, enter the command:

sort toBeSorted

Notice that the names are now printed in alphabetical order.

Short Exercise

Use the echo command and the append operator, >>, to append your name to the file, then sort it and make a new file called Sorted.

Let's navigate back to ~/2014-01-08-iastate/day1.shell/data. Enter the following command:

wc Bert/* | sort -k 3 -n

We are already familiar with what the first of these two commands does: it creates a list containing the number of characters, words, and lines in each file in the Bert directory. This list is then piped into the sort command, so that it can be sorted. Notice there are two options given to sort:

-k 3: Sort based on the third column
-n: Sort in numerical order as opposed to alphabetical order

Notice that the files are sorted by the number of characters.

Short Exercise

Combine the wc, sort, head and tail commands so that only the wc information for the largest file is listed

Hint: To print the smallest file, use:

wc Bert/* | sort -k 3 -n | head -n 1

Printing the smallest file seems pretty useful. We don't want to type out that long command often. Let's create a simple script, a simple program, to run this command. The program will look at all of the files in the current directory and print the information about the smallest one. Let's call the script smallest. We'll use nano to create this file. Navigate to the data directory, then:

nano smallest

Then enter the following text:

#!/bin/bash
wc * | sort -k 3 -n | head -n 1

Now, cd into the Bert directory and enter the command ../smallest. Notice that it says permission denied. This happens because we haven't told the shell that this is an executable file. If you do ls -l ../smallest, it will show you the permissions on the left of the listing.

Enter the following commands:

chmod a+x ../smallest
../smallest

The chmod command is used to modify the permissions of a file. This particular command modifies the file ../smallest by giving all users (notice the a) permission to execute (notice the x) the file. If you enter:

ls -l ../smallest

You will see that the file name is green and the permissions have changed. Congratulations, you just created your first shell script!

Searching files

You can search the contents of a file using the command grep. The grep program is very powerful and useful especially when combined with other commands by using the pipe. Navigate to the Bert directory. Every data file in this directory has a line which says "Range". The range represents the smallest frequency range that can be discriminated. Lets list all of the ranges from the tests that Bert conducted:

grep Range *

Short Exercise

Create an executable script called smallestrange in the data directory, that is similar to the smallest script, but prints the file containing the file with the smallest Range. Use the commands grep, sort, and tail to do this.

Finding files

The find program can be used to find files based on arbitrary criteria. Navigate to the data directory and enter the following command:

find . -print

This prints the name of every file or directory, recursively, starting from the current directory. Let's exclude all of the directories:

find . -type f -print

This tells find to locate only files. Now try these commands:

find . -type f -name "*1*"
find . -type f -name "*1*" -or -name "*2*" -print
find . -type f -name "*1*" -and -name "*2*" -print

The find command can acquire a list of files and perform some operation on each file. Try this command out:

find . -type f -exec grep Volume {} \;

This command finds every file starting from .. Then it searches each file for a line which contains the word "Volume". The {} refers to the name of each file. The trailing \; is used to terminate the command. This command is slow, because it is calling a new instance of grep for each item the find returns.

A faster way to do this is to use the xargs command:

find . -type f -print | xargs grep Volume

Or if you want to do it like the "-exec" example before:

find . -type f -print | xargs -I {} grep Volume {}

find generates a list of all the files we are interested in, then we pipe them to xargs. xargs takes the items given to it and passes them as arguments to grep. xargs generally only creates a single instance of grep (or whatever program it is running).

Short Exercise

Navigate to the data directory. Use one find command to perform each of the operations listed below (except number 2, which does not require a find command):

Find any file whose name is "NOTES" within data and delete it
Create a new directory called cleaneddata
Move all of the files within data to the cleaneddata directory
Rename all of the files to ensure that they end in .txt (note: it is ok for the file name to end in .txt.txt

Hint: If you make a mistake and need to start over just do the following:

Navigate to the day1.shell directory
Delete the data directory
Enter the command: git checkout -- data You should see that the data directory has reappeared in its original state

BONUS

Redo exercise 4, except rename only the files which do not already end in .txt. You will have to use the man command to figure out how to search for files which do not match a certain name.

Bonus:

backtick, xargs: Example find all files with certain text

alias -> rm -i

variables -> use a path example

.bashrc

ssh and scp

Regular Expressions

Permissions

Chaining commands together