More Debt Watching from the Command Line

Those of you who watch our Tech Tip Videos may have seen my video on how to fetch the US National Debt at the command line. The script contained here uses the idea I developed there and expands it to fetch the debt twice, with an optional pause in between, and then display the amount that the debt has increased during the pause.

You can use the script in the following fashion:

bash ndebt.sh 10
bash ndebt.sh 3:10
bash ndebt.sh 1:15:10

The argument to the script is the pause time, which can be given in seconds, or minutes:seconds, or hours:minutes:seconds. The script fetches the debt at the start, pauses for the requested amount of time, then fetches the debt again and prints out the amount the debt has increased.

For example, here we run the script and specify a time of 20 seconds:

$ bash ndebt.sh 20
During the last 20 seconds the US National Debt has increased by 901128.13

The script itself follows:

#!/bin/bash


# Check for pause time.
pause_time=0

if [[ "$1" ]]; then
	if   [[ $1 =~ ^([0-9]+):([0-9]+):([0-9]+)$ ]]; then
		pause_time=$(((${BASH_REMATCH[1]}*60*60) + (${BASH_REMATCH[2]}*60) + (${BASH_REMATCH[3]})))
	elif [[ $1 =~ ^([0-9]+):([0-9]+)$ ]]; then
		pause_time=$(((${BASH_REMATCH[1]}*60) + (${BASH_REMATCH[2]})))
	elif [[ $1 =~ ^([0-9]+)$ ]]; then
		pause_time=$1
	else
		echo "Bad pause time: $1" >&2
		exit 1
	fi
fi


# Get national debt.
function get_debt()
{
	local t=$(wget --quiet http://brillig.com/debt_clock -O - | grep debtiv.gif)

	t=$(sed -e 's/.*ALT="\$//' -e 's/".*//' -e 's/[ ,]//g' <<<$t)
	echo $t
}

# Print time item.
# Pass time value and one of 'hour', 'minute', 'ssecond'.
function fmt_time()
{
	local t=$1
	local p=$2
	if [[ $t -gt 0 ]]; then
		if [[ $t -eq 1 ]]; then
			printf '%d %s' $t $p
		else
			printf '%d %ss' $t $p
		fi
	fi
}

# Print the elapsed time between the first and second argument.
# Times given in seconds.
# Earliest time is first arugment.
function elapsed_time()
{
	local st=$1
	local et=$2
	local dt=$((et - st))
	local ds=$((dt % 60))
	local dm=$(((dt / 60) % 60))
	local dh=$((dt / 3600))
	
	echo $(fmt_time $dh 'hour') $(fmt_time $dm 'minute') $(fmt_time $ds 'second')
}


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

stime=$(date +%s)
sdebt=$(get_debt)

if [[ $pause_time -gt 0 ]]; then sleep $pause_time; fi

etime=$(date +%s)
edebt=$(get_debt)

t=$(elapsed_time $stime $etime)
d=$(bc <<<"scale=2; $edebt - $sdebt")

printf "During the last %s the US National Debt has increased by %s\n" "$t" "$d"


## vim: tabstop=4: shiftwidth=4: noexpandtab:
## kate: tab-width 4; indent-width 4; replace-tabs false;

The first part of the script uses bash's regular expression comparison operator (=~) to see if the passed argument is a valid time. Then come a few functions.

The first function, get_debt(), fetches the debt page from brillig.com and then extracts and cleans up the debt number. The debt number is extracted from the alt attribute of the image at the top of the page. The string there contains spaces and commas and a dollar sign, those are removed to give us a usable number.

The second function fmt_time() formats a time component for our output. For example, fmt_time 2 second prints "2 seconds", fmt_time 0 second prints "" (an empty string).

The third function, elapsed_time() takes two time values in seconds and calculates the difference between the two. It then formats the elapsed time in a readable fashion. For example elapsed_time 0 3659 prints "1 hour 59 seconds".

After that comes the main code. It records the start time, fetches the starting debt and then pauses. After the pause it records the end time and fetches the ending debt. Then bc is used to calculate the difference and the result is printed. Note that the time printed is the actual elapsed time and so it may be slightly different than the elapsed time you request on the command line, but usually only by a second or two.