Wiley Linux Command Line and Shell Scripting Bible 3rd (2015)

Part IV. Creating Practical Scripts

Chapter 26. Creating Fun Little Shell Scripts

In This Chapter

1. Sending a message

2. Getting inspiration

3. Sending a text

The primary reason for learning to write bash shell scripts is to be able to create your own Linux system utilities. Understanding how to write useful and practical script utilities is important. However, sometimes it helps to do something fun to learn a concept or skill. The scripts in this chapter are not necessarily practical, but they can be lots of fun! And they help solidify script-writing concepts.

Sending a Message

Messages can be sent in many ways in an office or a home environment — text message, e-mail, and even making a phone call. One method, not commonly used any more, is sending a message directly to a fellow system user's terminal. Because this technique is largely unknown, it can be fun to communicate with someone with this method.

This shell script utility helps you to quickly and easily send a message to someone who is logged onto your Linux system. It is a rather simple script, but it can be loads of fun!

Understanding the required functions

For this simple script, only a few functions are required. Several of the commands are common and have been covered in the book. However, a few of the commands have only been touched on, and you may not be familiar with the primary command needed. This section looks at the commands needed to put together this simple, but interesting script.

Determining who is on the system

The first utility needed is the who command. The who utility allows you to see all the users currently logged into the system:

$ who

christine tty2 2015-09-10 11:43

timothy tty3 2015-09-10 11:46

[...]

In this partial listing, all the information needed for sending messages is shown. By default, the who command gives you the short version of information available. The same information is provided, when who -s is issued:

· User name

· User's terminal

· Time the user logged into the system

To send a message, you only need the first two items. Both the user name and the user's current terminal are necessary.

Allowing messages

Users can disallow anyone to send them messages via the mesg utility. Therefore, before you start attempting to send messages, it's a good idea to check whether messages are allowed. For yourself, you can simply enter the mesg command as follows:

$ mesg

is n

The is n result shows that messaging is turned off. If the result showed is y, messages would be allowed.

Caution

Some distributions, such as Ubuntu, come with messaging turned off by default. Other distributions, such as CentOS, come with messaging turned on by default. Thus, you need to check your status and other user's message status before attempting to send a message.

To check everyone else's message status, you can use the who command again. Keep in mind that this checks the message status only for those who are currently logged into the system. You use the -T option to check their message status:

$ who -T

christine - tty2 2015-09-10 12:56

timothy - tty3 2015-09-10 11:46

[...]

The dash (-) after each user name indicates that messaging is turned off for those users. If it is turned on, you see a plus (+) sign.

To allow messages to be sent to you, if it is turned off, you need to use the message command with the y option:

$ whoami

christine

$ mesg y

$ mesg

is y

Messaging is turned on by the user christine, when the command mesg y is issued. The user's individual message status is checked, by issuing the mesg command. Sure enough, the command shows is y, which indicates messages are allowed to this user.

Using the who command, other users can see how the user christine has changed her message status. The message status is now set to a plus sign, which indicates the user is allowing messages to be sent to her.

$ who -T

christine + tty2 2015-09-10 12:56

timothy - tty3 2015-09-10 11:46

[...]

For two-way communication, you need to allow messaging and one or more users also need to allow messaging. In this example, the user timothy has also turned on his messaging:

$ who -T

christine + tty2 2015-09-10 12:56

timothy + tty3 2015-09-10 11:46

[...]

Now that messaging is allowed between you and at least one other user, you can try out the command to send messages. However, the who command is also still needed, because it provides the necessary information in order to send a message.

Sending a message to another user

The primary tool for this script is the write command. As long as messaging is allowed, the write command allows you to send a message to another logged-in user using his username and current terminal.

Note

The write command only allows you to successfully send messages to users logged onto a virtual console terminal (see Chapter 2). A user logged into the graphical environment will not be able to receive messages.

In this example, a message is sent from user christine to user timothy logged on the tty3 terminal. From christine's terminal, the session looks as follows:

$ who

christine tty2 2015-09-10 13:54

timothy tty3 2015-09-10 11:46

[...]

$ write timothy tty3

Hello Tim!

After the message is initiated by the write command, a blank line is shown for you to begin inputting the message text. It may be as many lines as you desire. When the Enter key is pressed, a new line is available for more message text. After you are finished entering message text, the whole message is sent by pressing the Ctrl+D key combination.

The receiver of the message sees something like the following:

Message from christine@server01 on tty2 at 14:11 ...

Hello Tim!

EOF

The receiver can see which user on which terminal sent the message. A time stamp is also included. Notice the EOF shown at the message's bottom. It indicates End Of File, which lets the message recipient know that the entire message is being displayed.

Tip

Often, a message recipient needs to press the Enter key in order to get a prompt to show again, after a message is received.

Now you can send messages! The next step is to use these commands to create the script.

Creating the script

Using a script to send messages helps overcome a few potential problems. First, if you have lots of users on the system, trying to find the one user you want to send a message to can be a pain! You must also determine whether that particular user has messaging turned on. In addition, a script speeds things up allowing you to quickly send a message to a particular user in one easy step.

Checking if user is logged on

The first issue is to let the script know to which user you want to send a message. This is easily done by sending a parameter (Chapter 14) along with the script's execution. For the script to determine whether that particular user is logged on the system, the whocommand is employed as shown in this bit of script code:

# Determine if user is logged on:

logged_on=$(who | grep -i -m 1 $1 | gawk '{print $1}')

In the preceding code, the results of the who command are piped into the grep command (Chapter 4). The grep command uses the -i option to ignore case, which allows the username to be entered using uppercase or lowercase letters. The -m 1 option is included on thegrep command, in case the user is logged into the system multiple times. The grep command produces either nothing, if the user is not logged on, or the username's first login information. This output is passed to the gawk command (Chapter 19). The gawk command returns only the first item, either nothing or the username. This final output from the gawk command is stored in the variable logged_on.

Tip

Some Linux distributions, such as Ubuntu, may not have the gawk command installed by default. To install it, type sudo apt-get install gawk. Also, you can find more information about installing software packages in Chapter 9.

When the variable, logged_on, contains either nothing (if the user is not logged on) or the username, it can be tested and acted upon:

if [ -z $logged_on ]

then

echo "$1 is not logged on."

echo "Exiting script..."

exit

Employing the use of an if statement and a test command (Chapter 12), the logged_on variable is tested to determine if it is a zero-length variable. If it is a zero-length variable, the script user is informed via echo commands that the user is not currently logged ontothe system, and the script is exited via the exit command. If the user is logged onto the system, the logged_on variable contains the user's username, and the script continues.

In the following example, a username, Charlie, is passed as a parameter to the shell script. This user is not currently logged onto the system:

$ ./mu.sh Charlie

Charlie is not logged on.

Exiting script...

The code worked perfectly! Now instead of you digging through the who command results to determine whether a user is logged onto the system, the message script does that for you.

Checking if user accepts messages

The next important item is to determine whether a logged on user accepts messages. This script portion operates very closely to the script section for determining whether a user is logged on:

# Determine if user allows messaging:

allowed=$(who -T | grep -i -m 1 $1 | gawk '{print $2}')

if [ $allowed != "+" ]

then

echo "$1 does not allowing messaging."

echo "Exiting script..."

exit

Notice that this time, the who -T command and option are used. This displays a + next to the username, if messaging is allowed. Otherwise, it displays a - next to the username, if messaging is not allowed. The results from the who command are then piped into grep andgawk to pull out only the messaging indicator. The messaging indicator is stored in the allowed variable. Finally, an if statement is employed to test for a messaging indicator not set to +. If the indicator is not set to +, the script user is informed and the script is exited. However, if the messaging indicator shows messaging is allowed, the script continues.

To test out this script's section, a user who is logged into the system with messaging disabled is tested. The user Samantha currently has messaging disabled:

$ ./mu.sh Samantha

Samantha does not allowing messaging.

Exiting script...

The test worked as expected. This script portion eliminates any need to manually check for messaging being enabled or disabled.

Checking if message was included

The message to be sent is also included as a script parameter. Therefore, another needed check is whether a message was included as a parameter to the mu.sh shell script. To test for the message parameter, an if statement, similar to those used earlier, must be included in the script's code:

# Determine if a message was included:

if [ -z $2 ]

then

echo "No message parameter included."

echo "Exiting script..."

exit

To test out this script portion, a message was not included for a user who is both logged into the system and allows messaging:

$ ./mu.sh Timothy

No message parameter included.

Exiting script...

This is exactly what is needed! Now that the script has performed these preliminary checks, the primary task of sending a message can be undertaken.

Transmitting a simple message

Before a message is sent, the user's current terminal must be identified and stored in a variable. The who, grep, and gawk commands are employed again:

# Send message to user:

uterminal=$(who | grep -i -m 1 $1 | gawk '{print $2}')

To transmit the message, both the echo and the write commands are used:

echo $2 | write $logged_on $uterminal

Because write is an interactive utility, it must have the message piped into it for the script to work properly. The echo command is used to send the message, $2, to STDOUT, which in turn is piped into the write command. The logged_on variable holds the username, and the uterminal variable holds the user's current terminal.

Now, you can test sending a simple message to a designated user via the script:

$ ./mu.sh Timothy test

The user Timothy receives the following message on his terminal:

Message from christine@server01 on tty2 at 10:23 ...

test

EOF

Success! You can now send simple one word messages to other users on your system via your script.

Transmitting a long message

Often, you want to send more than just a single word to another system user. Let's try a longer message using the current script:

$ ./mu.sh Timothy Boss is coming. Look busy.

The user Timothy receives the following message on his terminal:

Message from christine@server01 on tty2 at 10:24 ...

Boss

EOF

It didn't work. Only the first word of the message, Boss, was sent. This is due to the script using parameters (Chapter 14). Recall that the bash shell considers a space to differentiate between parameters. Thus, because there are spaces in the message, each word is treated as a different parameter. The script must be modified to fix this problem.

The shift command (Chapter 14) and a while loop (Chapter 13) help with this long message issue:

# Determine if there is more to the message:

shift

while [ -n "$1" ]

whole_message=$whole_message' '$1

shift

done

Recall that the shift command allows you to process the various provided script parameters without knowing the total number of parameters. The shift command simply moves the next parameter in line down to parameter $1. First, a primary shift must be issued before the while loop, because the message starts in parameter $2, instead of parameter $1.

After the while loop is initiated, it continues grabbing each message word, tacking the word onto the whole_message variable. The loop then shifts to the next parameter. After the final parameter is processed, the while loop exits and the whole_message variable contains the entire message to send.

One additional script modification is needed to fix this problem. Instead of just sending parameter $2 to the write utility, the script is modified to send the variable, whole_message:

# Send message to user:

uterminal=$(who | grep -i -m 1 $1 | gawk '{print $2}')

echo $whole_message | write $logged_on $uterminal

Now, again try to send that warning message about the boss coming his way to Timothy:

$ ./mu.sh Timothy Boss is coming

Usage: grep [OPTION]... PATTERN [FILE]...

Try 'grep --help' for more information.

Oops! That didn't work either. This is because when shift was used in the script, the $1 parameter contents were removed. Thus, when the script attempts to use $1 in the grep command, it generates an error. To fix this problem a variable, muser, is used to capture the$1 parameter's value:

# Save the username parameter

muser=$1

Now muser stores the username. The $1 parameter in the script's various grep and echo commands can be replaced by the muser variable:

# Determine if user is logged on:

logged_on=$(who | grep -i -m 1 $muser | gawk '{print $1}')

[...]

echo "$muser is not logged on."

[...]

# Determine if user allows messaging:

allowed=$(who -T | grep -i -m 1 $muser | gawk '{print $2}')

[...]

echo "$muser does not allowing messaging."

[...]

# Send message to user:

uterminal=$(who | grep -i -m 1 $muser | gawk '{print $2}')

[...]

To test out the script changes, a multi-word message is sent again. In addition, some emphasis is added to the message by tacking on exclamation points:

$ ./mu.sh Timothy The boss is coming! Look busy!

The user Timothy receives the following message on his terminal:

Message from christine@server01 on tty2 at 10:30 ...

The boss is coming! Look busy!

EOF

It worked! You can now employ the script to quickly send messages to other users on the system. Here's the final message script with all the needed checks and changes:

#!/bin/bash

#mu.sh - Send a Message to a particular user

#############################################

# Save the username parameter

muser=$1

# Determine if user is logged on:

logged_on=$(who | grep -i -m 1 $muser | gawk '{print $1}')

if [ -z $logged_on ]

then

echo "$muser is not logged on."

echo "Exiting script..."

exit

# Determine if user allows messaging:

allowed=$(who -T | grep -i -m 1 $muser | gawk '{print $2}')

if [ $allowed != "+" ]

then

echo "$muser does not allowing messaging."

echo "Exiting script..."

exit

# Determine if a message was included:

if [ -z $2 ]

then

echo "No message parameter included."

echo "Exiting script..."

exit

# Determine if there is more to the message:

shift

while [ -n "$1" ]

whole_message=$whole_message' '$1

shift

done

# Send message to user:

uterminal=$(who | grep -i -m 1 $muser | gawk '{print $2}')

echo $whole_message | write $logged_on $uterminal

exit

Because you have made it to the last chapter in this book, you should be ready for a script-writing challenge. Here are some suggested improvements for the message script that you can attempt on your own:

· Instead of passing the username and message as parameters, use options (see Chapter 14).

· If a user is logged into multiple terminals, allow a message to be sent to those multiple terminals. (Hint: Use multiple write commands.)

· If the message to be sent is for a user who is currently only logged into the GUI, produce a message for the script user and exit the script. (Remember the write command can only write to virtual console terminals.)

· Allow a long message stored in a file to be sent to a terminal. (Hint: Use the cat command output piped into the write utility, instead of the echo command.)

Not only does reading through the script help solidify the script-writing concepts you are learning, but so does modifying the script. Come up with your own creative modification. Have a little fun! It helps you learn.

Obtaining a Quote

Inspirational quotes have long been used in the business environment. You may have a few on your office wall right now. This fun little and interesting script helps you obtain a daily inspirational quote to use as you please.

This section takes you through how to create this script. Included is a new rich utility that has not been covered in the book yet. The script also uses some utilities that have been covered, such as sed and gawk.

Understanding the required functions

Several great websites allow you to obtain daily inspiration quotes. Just open your favorite search engine, and you can find many sites. After you find a site for your daily quote, you need a utility to download that quote. For this script, the wget utility is just what's needed.

Learning about the wget utility

The wget utility is a flexible tool that allows web pages to be downloaded to your local Linux system. From these pages, you can glean your daily inspirational quote.

Note

The wget command is an extremely rich utility. In this chapter, only a small portion of its power is used. Find out more about wget via the man pages.

To download a web page via wget, you just need the wget command and the website's address:

$ wget www.quotationspage.com/qotd.html

--2015-09-23 09:14:28-- http://www.quotationspage.com/qotd.html

Resolving www.quotationspage.com... 67.228.101.64

Connecting to www.quotationspage.com|67.228.101.64|:80. connected

HTTP request sent, awaiting response... 200 OK

Length: unspecified [text/html]

Saving to: "qotd.html"

[ <=> ] 13,806 --.-K/s in 0.1s

2015-09-23 09:14:28 (118 KB/s) - "qotd.html" saved [13806]

The website's information is stored in a file named after the web page. In this case, it's qotd.html. And as you might have guessed by now, the file is full of HTML code:

$ cat qotd.html

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">

<head>

<title type="main">Quotes of the Day - The Quotations Page</title>

[...]

Only a partial HTML code listing is shown here. For the script, the sed and gawk utilities help strip out the desired inspirational quote. But before tackling the script, you need a little more control over the wget utility's input and output.

You can use a variable to hold the web address (URL). Simply pass the variable to wget as a parameter. Just don't forget to use the $ along with the variable name:

$ url=www.quotationspage.com/qotd.html

$ wget $url

--2015-09-23 09:24:21-- http://www.quotationspage.com/qotd.html

Resolving www.quotationspage.com... 67.228.101.64

Connecting to www.quotationspage.com|67.228.101.64|:80 connected.

HTTP request sent, awaiting response... 200 OK

Length: unspecified [text/html]

Saving to: "qotd.html.3"

[ <=> ] 13,806 --.-K/s in 0.1s

2015-09-23 09:24:21 (98.6 KB/s) - "qotd.html.3" saved [13806]

The daily inspiration quote script is eventually to be run daily via cron (Chapter 16) or some other script automation utility. Thus, having the wget command's session output display to STDOUT is undesirable. To store the session output to a log file, use the option -o. This allows session output to be viewed at a later time:

$ url=www.quotationspage.com/qotd.html

$ wget -o quote.log $url

$ cat quote.log

--2015-09-23 09:41:46-- http://www.quotationspage.com/qotd.html

Resolving www.quotationspage.com... 67.228.101.64

Connecting to www.quotationspage.com|67.228.101.64|:80 connected.

HTTP request sent, awaiting response... 200 OK

Length: unspecified [text/html]

Saving to: "qotd.html.1"

0K .......... ... 81.7K=0.2s

2015-09-23 09:41:46 (81.7 KB/s) - "qotd.html.1" saved [13806]

The wget utility now stores its session output into the log file as it retrieves web page information. If desired, you can view the logged session output by using the cat command, as shown in the preceding code.

Note

For various reasons, you may decide that you do not want wget to produce a log file or display session output. In this case, just use the -q option, and the wget command quietly performs its directed duties.

To control where the web page information is stored, use the -O option on the wget command. Thus, instead of having the web address as the storage file name, you can use the filename of your choice:

$ url=www.quotationspage.com/qotd.html

$ wget -o quote.log -O Daily_Quote.html $url

$ cat Daily_Quote.html

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">

<head>

[...]

Using the -O option allows the web page data to be stored in the designated file, Daily_Quote.html. Now that the wget utility's output is controlled, the next required function, checking the web address's validity, can be explored.

Testing a web address

Web addresses change. Sometimes, it seems this happens daily. Therefore, it is important to test the address validity within the script. The wget utility gives the ability to conduct such a test with the --spider option:

$ url=www.quotationspage.com/qotd.html

$ wget --spider $url

Spider mode enabled. Check if remote file exists.

--2015-09-23 12:45:41-- http://www.quotationspage.com/qotd.html

Resolving www.quotationspage.com... 67.228.101.64

Connecting to www.quotationspage.com|67.228.101.64|:80 connected.

HTTP request sent, awaiting response... 200 OK

Length: unspecified [text/html]

Remote file exists and could contain further links,

but recursion is disabled -- not retrieving.

This output indicates that the URL is valid, but it's too much to read through. You can cut down on the output by adding the -nv option, which stands for non-verbose:

$ wget -nv --spider $url

2015-09-23 12:49:13

URL: http://www.quotationspage.com/qotd.html 200 OK

The -nv option allows just the web address's status to be displayed, making the output much easier to read. Contrary to what you may think, the OK at the non-verbose line's end does not indicate that the web address is valid. The indication is that the web address came back as it was sent. This concept is a little unclear, until you see an invalid web address.

To see an invalid web address indicator, the URL variable is changed to an incorrect web address. The wget command is reissued using this bad address:

$ url=www.quotationspage.com/BAD_URL.html

$ wget -nv --spider $url

2015-09-23 12:54:33

URL: http://www.quotationspage.com/error404.html 200 OK

Notice that the output still has an OK at its end. However, the web address ends in error404.html. This indicates the web address is invalid.

With the necessary wget command to grab the inspirational quote's web page information, and the ability to test the web page's address, it is time to start building the script. Your daily inspirational quote awaits retrieval.

Creating the script

To test the script as it is built, a parameter containing the website's URL is passed to the script. Within the script, the variable quote_url contains the passed parameter's value:

quote_url=$1

Checking the passed URL

It is always a good idea to have checks in place within your script. The first check is to ensure that the daily inspirational quote script website's URL is still valid.

As you would expect, the script checks the web address validity with wget and the --spider option. However, the resulting indicator must be saved so the indicator can be checked later with an if statement test. Thus, the resulting indicator must be saved to a variable. This is a little tricky with the wget command.

To save the indicator output, the standard $() syntax is used around the command. But in addition, STDERR and STDOUT redirection is needed. This is accomplished by tacking on 2>&1 to the end of the wget command:

check_url=$(wget -nv --spider $quote_url 2>&1)

Now the indicator status message is saved within the check_url variable. To carve out the error indicator, error404, from the check_url string, parameter expansion and the echo command can be used:

bad_url=$(echo ${check_url/*error404*/error404})

In this example, string parameter expansion allows the string stored in check_url to be searched. Think of string parameter expansion as a quick and easy sed alternative. Using wildcards around the search word, *error404* allows the entire string to be searched. If the search is successful, the echo command sends the string error404 to be stored into the bad_url variable. If the search is not successful, the bad_url variable contains the check_url variable's contents.

Now an if statement (Chapter 12) is employed to check the bad_url variable's string. If the string error404 is found, a message is displayed and the script exits:

if [ "$bad_url" = "error404" ]

then

echo "Bad web address"

echo "$quote_url invalid"

echo "Exiting script..."

exit

An easier and shorter method can be used. This method removes the need for string parameter expansion and the bad_url variable altogether. A double bracket if statement allows a search to be conducted of the check_url variable:

if [[ $check_url == *error404* ]]

then

echo "Bad web address"

echo "$quote_url invalid"

echo "Exiting script..."

exit

The test statement within the if structure searches the check_url variable's string. If the string error404 is found anywhere within the variable string, a message is displayed and the script exits. If the indicator string does not contain the error message, the script continues. This statement saves time and effort. No need for any string parameter expansion or even the bad_url variable.

Now that the check is in place, the script can be tested with an invalid web address. The url variable is set to an incorrect URL and passed to the get_quote.sh script:

$ url=www.quotationspage.com/BAD_URL.html

$ ./get_quote.sh $url

Bad web address

www.quotationspage.com/BAD_URL.html invalid

Exiting script...

That works great. Just to make sure that all is well, now a valid web address is tested:

$ url=www.quotationspage.com/qotd.html

$ ./get_quote.sh $url

No error message received. The script works perfectly so far! This is the only check needed, so the next item to be added to the script is obtaining the web page's data.

Obtaining web page information

Grabbing the inspiration daily quote's web page data is simple. The wget command shown earlier in the chapter is used in the script. The only needed change is to store the log file and the HTML file, which contains the web page information, in the /tmp directory:

wget -o /tmp/quote.log -O /tmp/quote.html $quote_url

Before moving on to the rest of the script, this code section should be tested using a valid web address:

$ url=www.quotationspage.com/qotd.html

$ ./get_quote.sh $url

$ ls /tmp/quote.*

/tmp/quote.log /tmp/quote.html

$ cat /tmp/quote.html

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">

<head>

[...]

</body>

</html>

The script still works well! The log file, /tmp/quote.log, and the html file, /tmp/quote.html, were properly created.

Tip

If you do not want cookies to be involved when obtaining website information, you can add the --no-cookies option to the wget command. By default, storing cookies is turned off.

The next task is to dig the daily inspirational quote out of the HTML code within the downloaded web page HTML file. This task requires both the sed and the gawk utilities.

Parsing out the desired information

In order to pull out the actual inspirational quote, some processing must take place. This part of the script uses sed and gawk to parse out the desired information.

Note

This section is where the most variety is introduced when you modify this script for your own use. The sed and gawk utilities are used to search for keywords that are specific for this particular quote website's data. You may need to use different keywords as well as different sed and gawk commands to extract the data you desire.

The script first needs to remove all the HTML tags from the downloaded web page's information stored in the /tmp/quote.html file. The sed utility can provide this capability:

sed 's/<[∧>]*//g' /tmp/quote.html

The preceding code should look very familiar. It was covered in Chapter 21 in the “Removing HTML tags” section.

After the HTML tags are removed, the output looks like the following:

$ url=www.quotationspage.com/qotd.html

$ ./get_quote.sh $url

[...]

>Quotes of the Day - The Quotations Page>

[...]

>>Selected from Michael Moncur's Collection of Quotations

- September 23, 2015>>

>>>Horse sense is the thing a horse has which keeps

[...]

This snipped listing shows that there is still too much unnecessary data in this file. Therefore, some additional parsing must be done. Fortunately, the quote text needed is situated right next to the current date. Therefore, the script can use the current date as a search term!

The grep command, the $() format, and the date command can help here. The output from the sed command is piped into the grep command. The grep command uses the current date formatted to match the date used on the quotation's web page. After the date text line is found, two additional text lines are pulled with the -A2 parameter:

sed 's/<[∧>]*//g' /tmp/quote.html |

grep "$(date +%B' '%-d,' '%Y)" -A2

Now the script's output looks similar to the following:

$ ./get_quote.sh $url

>>Selected from Michael Moncur's Collection of Quotations

- September 23, 2015>>

>>>Horse sense is the thing a horse has which keeps it from

betting on people.> >>>>>>>>>>>>>>>>>>W. C. Fields> (1880 -

1946)> >>>

>>Newspapermen learn to call a murderer 'an alleged murderer'

and the King of England 'the alleged King of England' to

avoid libel suits.> >>>>>>>>>>>>>>>>>>Stephen Leacock> (1869

- 1944)> >>> - More quotations on: [>Journalism>] >

Tip

If your Linux system's date is set differently than the quote of the day page's date, you get a blank line instead of a quote. The preceding grep command assumes your system date is the same as the web page's date.

Although the output is greatly reduced, there is still too much clutter in the text. The extra > symbols can easily be removed with the sed utility. In the script, the output from the grep command is piped into the sed utility, which strips off the > symbols:

sed 's/<[∧>]*//g' /tmp/quote.html |

grep "$(date +%B' '%-d,' '%Y)" -A2 |

sed 's/>//g'

With the new script line, the output is now a little clearer:

$ ./get_quote.sh $url

Selected from Michael Moncur's Collection of Quotations

- September 23, 2015

Horse sense is the thing a horse has which keeps it from

betting on people. W. C. Fields (1880 - 1946)

Newspapermen learn to call a murderer 'an alleged murderer'

and the King of England 'the alleged King of England' to

avoid libel suits. Stephen Leacock (1869 - 1944) -

More quotations on: [Journalism]

Now we're getting somewhere! However, we can still remove a little more clutter from the quotation.

You may have noticed that two quotations are listed in the output instead of one. This happens occasionally with this particular website. Some days, it may be one quote, and other days, it may be two. Therefore, the script needs a way to pull out only the first quote.

The sed utility can help again with this problem. Using the sed utility's next and delete commands (Chapter 21), the string is located. After it's found, sed moves to the next line of the data and deletes it:

sed 's/<[∧>]*//g' /tmp/quote.html |

grep "$(date +%B' '%-d,' '%Y)" -A2 |

sed 's/>//g' |

sed '/ /{n ; d}'

Now the script can be tested to see if the new sed addition fixes the multiple quotation problem:

$ ./get_quote.sh $url

Selected from Michael Moncur's Collection of Quotations

- September 23, 2015

Horse sense is the thing a horse has which keeps it from

betting on people. W. C. Fields (1880 - 1946)

The extra quotation is removed! One item remains for the quotation cleanup. At the quotation's end, the string is still hanging around. The script could use another sed command to remove this pesky item, but just for variety, the gawk command is used:

sed 's/<[∧>]*//g' /tmp/quote.html |

grep "$(date +%B' '%-d,' '%Y)" -A2 |

sed 's/>//g' |

sed '/ /{n ; d}' |

gawk 'BEGIN{FS=" "} {print $1}'

In the preceding code, the input field separator variable, FS, is used with the gawk command (Chapter 22). The string is set as a field separator, which causes gawk to drop it from the output:

$ ./get_quote.sh $url

Selected from Michael Moncur's Collection of Quotations

- September 23, 2015

Horse sense is the thing a horse has which keeps it from

betting on people. W. C. Fields (1880 - 1946)

One last needed script action is to save this quotation text to a file. Here the tee command (Chapter 15) helps. Now the entire quote extraction process looks as follows:

sed 's/<[∧>]*//g' /tmp/quote.html |

grep "$(date +%B' '%-d,' '%Y)" -A2 |

sed 's/>//g' |

sed '/ /{n ; d}' |

gawk 'BEGIN{FS=" "} {print $1}' |

tee /tmp/daily_quote.txt > /dev/null

The extracted quote is saved to /tmp/daily_quote.txt, and any output produced by the gawk command is redirected to /dev/null (see Chapter 15). To make the script a little more self-directed, the URL is hard-coded into the script:

quote_url=www.quotationspage.com/qotd.html

Now these two new changes to the daily inspirational quote script can be tested:

$ ./get_quote.sh

$ cat /tmp/daily_quote.txt

Selected from Michael Moncur's Collection of Quotations

- September 23, 2015

Horse sense is the thing a horse has which keeps it from

betting on people. W. C. Fields (1880 - 1946)

That works perfectly! The daily inspiration quote was extracted from the website's data and stored in a text file. You may have noticed by now that this quotation is less a traditional inspirational quote and more a humorous quote. Just know that some people find humor inspirational!

For your review, here's the final daily inspirational quote script with all the needed checks and changes:

#!/bin/bash

# Get a Daily Inspirational Quote

#####################################

# Script Variables ####

quote_url=www.quotationspage.com/qotd.html

# Check url validity ###

check_url=$(wget -nv --spider $quote_url 2>&1)

if [[ $check_url == *error404* ]]

then

echo "Bad web address"

echo "$quote_url invalid"

echo "Exiting script..."

exit

# Download Web Site's Information

wget -o /tmp/quote.log -O /tmp/quote.html $quote_url

# Extract the Desired Data

sed 's/<[∧>]*//g' /tmp/quote.html |

grep "$(date +%B' '%-d,' '%Y)" -A2 |

sed 's/>//g' |

sed '/ /{n ; d}' |

gawk 'BEGIN{FS=" "} {print $1}' |

tee /tmp/daily_quote.txt > /dev/null

exit

This script is an excellent opportunity to try out some of your newly learned script writing and command line skills. The following are a few suggested changes for the daily inspirational quote script that you can attempt on your own:

· Change the website to your favorite quotation or sayings website, and make the necessary changes to the quote extraction commands.

· Try different sed and gawk commands for extracting the daily quotation.

· Set up the script to run daily on an automated basis via cron (see Chapter 16).

· Add a command to display the quote text file at certain times, such as when you first log in for the day.

Reading your daily quotes can inspire you. They may just inspire you to get out of that next business meeting. The next chapter section helps you write a script that does just that.

Generating an Excuse

You've been there. That endless staff meeting that is full of unimportant information. You would really rather be working on that fascinating bash shell script project back at your desk. Here's a little fun script you can use to get out of the next staff meeting.

Short Message Service (SMS) allows text messages to be sent between cell phones. However, you can also use SMS to send text messages directly from e-mail or the command line. The script in this section allows you to construct a text message to be sent at a specified time directly to your phone. Receiving a “critical” message from your Linux system is the perfect excuse for leaving a staff meeting early.

Understanding the required functions

You can send an SMS message from the command line in several ways. One way is via your system's e-mail using your phone carrier's SMS service. Another way is using the curl utility.

Learning about curl

Similar to wget, the curl utility allows you to transfer data from a particular web server. Unlike wget, it also allows you to transfer data to a web server. Transferring data to a particular web server is exactly what is needed here.

Tip

Some Linux distributions, such as Ubuntu, may not have the curl command installed by default. To install it, type sudo apt-get install curl. Also, you can find more information about installing software packages in Chapter 9.

Besides the curl utility, you need a website that provides free SMS message transfer. The one used here for this script is http://textbelt.com/text. This website allows you to send up to 75 text messages per day for free. You need it only for one text message, so it should be no problem.

Tip

If your company already uses an SMS provider, such as http://sendhub.com or http://eztexting.com, you can use those sites in your script instead. Be aware that the syntax needs to change depending upon those SMS provider's requirements.

To use curl and http://textbelt.com/text to send yourself a text message, you need to use the following syntax:

$ curl http://textbelt.com/text \

-d number=YourPhoneNumber \

-d "message=Your Text Message"

The -d option tells curl to send specified data to the website. In this case, the website needs particular data sent in order to send a text message. This data includes YourPhoneNumber, which is your cell phone number starting with the area code. And it also includes Your Text Message, which is the text message you desire to send.

Tip

The curl utility can handle much more than simply transferring data to and from a web server. It can handle many other network protocols, such as FTP, without any human intervention as well. Look at the man pages for curl to discover its rich power.

When the message is sent, the website provides a success message, ”success”: true, if no problems occurred:

$ curl http://textbelt.com/text \

> -d number=3173334444 \

> -d "message=Test from curl"

{

"success": true

Or it provides a fail message, ”success”: false, if data, such as the phone number, is incorrect:

$ curl http://textbelt.com/text \

-d number=317AAABBBB \

-d "message=Test from curl"

{

"success": false,

"message": "Invalid phone number."

Note

If your cell phone carrier is not in the United States of America, it is likely that http://textbelt.com/text will not work for you. You can try http://textbelt.com/Canada if your cell phone carrier is in Canada. If your cell phone carrier is located elsewhere, try http://textbell.com/intl instead. For additional help, see http://textbelt.com.

The success/fail messages are very helpful, but they are unwanted for the script. To remove these messages, simply redirect STDOUT to /dev/null (see Chapter 15). Unfortunately, now curl supplies undesired output:

$ curl http://textbelt.com/text \

> -d number=3173334444 \

> -d "message=Test from curl" > /dev/null

% Total % Received % Xferd Average Speed...

Dload Upload...

0 21 0 21 0 45 27 58 ...

The preceding snipped listing shows various statistics, which may be helpful when debugging your curl command. However, for the script, this information must be suppressed. Fortunately, the curl command has a -s option, which makes it silent:

$ curl -s http://textbelt.com/text \

> -d number=3173334444 \

> -d "message=Test from curl" > /dev/null

That is much better. The curl command is ready to be put into a script. However, before looking at the script, one more topic needs to be addressed: sending text messages via e-mail.

Choosing to use e-mail

If you choose not to use the text message relay service provided by http://textbelt.com/text or if for some reason it doesn't work for you, you can always substitute sending a text message via e-mail. This section briefly covers how to accomplish this substitution.

Caution

If your cell phone carrier is not in the United States of America, it is likely that this web service will not work for you. Also, your cell phone carrier may block SMS messages from this site. In this case, you must attempt to use e-mail instead.

Whether or not e-mail works as a substitute depends upon your cell phone carrier. If your cell phone carrier has an SMS gateway, you are in luck. Contact your cell phone carrier and find out the name of the gateway. Often, it is something similar to txt.att.net orvtext.com.

Tip

You can often find out your cell phone carrier's SMS gateway on your own via the Internet. One great site listing various SMS gateways, along with usage tips, is http://martinfitzpatrick.name/list-of-email-to-sms-gateways/. If you cannot find your carrier there, use your favorite search engine to locate it.

The basic syntax for sending a text message via e-mail is as follows:

mail -s "your text message" your_phone_number@your_sms_gateway

Note

If the mail command does not work on your Linux system, you need to install the mailutils package. See Chapter 9 for a review of installing software packages.

Unfortunately, after you enter the syntax, you must type your message and press Ctrl+D to send the text message. This is similar to sending a regular e-mail (see Chapter 24). Using this method doesn't work well in a script. Instead, you can store your e-mail message in a file and use it to send a text message. The basic idea for this method is as follows:

$ echo "This is a test" > message.txt

$ mail -s "Test from email" \

3173334444@vtext.com < message.txt

Now the e-mail syntax is more compatible with a script. However, be aware that many problems may exist with this approach. First, you must have a mail server running on your system (see Chapter 24). Secondly, your phone service provider may block SMS messages coming from your system via e-mail. This is often true, if you are attempting this method from your home.

Tip

If your phone service provider blocks SMS messages coming from your system, you can use a cloud-based e-mail provider as an SMS relay. Use your favorite Internet browser and search for the words SMS relay your_favorite_cloud_email and see what sites come up.

Although sending a text message via e-mail is a potential alternative, it can be fraught with problems. If you can, it is much easier to use a free SMS relay website and the curl utility. The script in the next section uses curl to send a text message to the phone of your choice.

Creating the script

After you have the required functions, creating the script to send a text message is fairly simple. You just need a few variables and the curl command.

You need three variables for the script. Setting up these particular data items as variables makes it easier if any of this information changes. The variables are shown here:

phone="3173334444"

SMSrelay_url=http://textbelt.com/text

text_message="System Code Red"

The only other needed item is the curl utility. Thus, here is the entire send a text message script:

#!/bin/bash

# Send a Text Message

################################

# Script Variables ####

phone="3173334444"

SMSrelay_url=http://textbelt.com/text

text_message="System Code Red"

# Send text ###########

curl -s $SMSrelay_url -d \

number=$phone \

-d "message=$text_message" > /dev/null

exit

If you see this script as simple and easy, you are right! Even more important, that means you have learned a great deal about shell script writing. Even easy scripts need to be tested, so be sure to test this script using your cell phone number in the phone variable before continuing.

Tip

While you are testing your script, be aware that this website, http://textbelt.com/text, does not allow you to send more than three text messages to the same phone number in less than three minutes.

To have a text message sent to you at a desired time, you must employ the at command. If you need a reminder, the at command was covered in Chapter 16.

First, you can test the use of the at command with your new script. Have the at utility execute the script by using the -f option along with the script's file name, send_text.sh, in this case. Have the script run immediately using the Now option:

$ at -f send_text.sh Now

job 22 at 2015-09-24 10:22

The script runs instantly. However, it may be a minute or two before you receive the text message on your phone.

To have the script run at another time, you simply use other at command options (see Chapter 16). In the following example, the script is run 25 minutes from the current time.

$ at -f send_text.sh Now + 25 minutes

job 23 at 2015-09-24 10:48

Note in the example, the at command provides an informational message when the script is submitted. The date and time listed in that message is when the script will execute.

What fun! Now you have a script utility that will be of assistance when you need an excuse to get out of that staff meeting. Better yet, you could modify the script to send yourself truly serious system messages that need to be addressed.

Summary

This chapter showed how to put some of the shell-scripting information presented in the book to use for fun little shell scripts. Each script reinforced material covered in the chapters along with a few new commands and ideas.

The chapter demonstrated how to send a message to another user on the Linux system. The script checked to see whether the user was logged on to the system and whether the user allowed messaging. After those checks were made, the passed message was sent using the write command. Included were some suggestions for modifying this script, which improve your shell-scripting abilities.

The next section walked you through obtaining website information using the wget utility. The created script pulled a quote from the web. After retrieval, the script used several utilities to pull out the actual quote text. These now familiar commands included sed, grep,gawk, and the tee command. For this script, suggestions were made for how the script could be modified. These are well worth pursuing to solidify and improve your new skills.

The chapter ended with a very fun and simple script for sending yourself a text message. We explored the curl utility, along with SMS concepts. Although this is a fun script, it can be modified and used for more serious purposes.

Thanks for joining us on this journey through the Linux command line and shell scripting. We hope you've enjoyed the journey and have learned how to get around on the command line and how to create shell scripts to save time. But don't stop your command line education here. There's always something new being developed in the open source world, whether it's a new command line utility or a full-blown shell. Stay in touch with the Linux community and follow along with the new advances and features.