Download Panoramio pictures via user id

panoramio.sh

#!/bin/sh

delay=0
timeout=10
retry=2

usage()
{
	echo "Usage: `basename $0` userid"
	exit 1
}

[ $# -ne 1 ] && usage

global_index=1
page=1
while true ; do
	echo "[INFO] Processing page $page"
	wget -q -O .tmp http://www.panoramio.com/user/$1?photo_page=$page
	if [ $? -ne 0 ] ; then
		echo "[WARNING] Can not get content of page $page"
		break
	fi

	if [ -z "$total" ] ; then
		total=`grep -iE '<a href="\/user\/.*\/stats' .tmp | head -n 1 | sed 's/.*stats">//;s/<.*//'`
	fi
	imgs=`grep -ioE '<a href="\/photo\/[0-9]*"$' .tmp | sed 's/.*\///;s/"//'`
	alts=`grep -ioE '^              >..*<\/a>$' .tmp | sed 's/<.*//;s/.*>//'`

	index=1
	for img in $imgs ; do
		url=http://static.panoramio.com/photos/original/$img.jpg
		alt=`{ for alt in $alts ; do echo $alt ; done } | head -n $index | tail -n 1`
		file=`echo $global_index | awk '{printf "%03d", $0}'`_$alt.jpg
		index=$((index+1))
		global_index=$((global_index+1))

		if [ -f $file ] ; then
			continue
		fi

		echo "[INFO] Downloading $url"
		wget -q --tries=$retry --timeout=$timeout -O "$file" "$url"
		if [ $? -ne 0 ] ; then
			rm -f "$file"
			echo "[WARNING] Failed to download $url to $file"
		fi
		sleep $delay
	done

	if [ $global_index -gt $total ] ; then
		break
	fi	

	page=$((page+1))
done

rm -f .tmp

exit 0

I'd love to fix the filename to add a time and date but also the full alt as a file name.

My alts and filenames are mixing up.

Back story, I'm trying to save someone's history after his funeral and this is my first example of history disappearing. (Will Flickr disappear also at some point?!)

Could the filename be something like 2016-01-01T00:00:00Z-went_into_the_wild.jpg? Converting the alt text into file friendly would be amazing and getting the full alt into the filename.

How can I help?

Personally, this was an improvement to the filenames immediately.

//alts=grep -ioE '^ >..*<\/a>$' .tmp | sed 's/<.*//;s/.*>//'
alts=grep -ioE '^ >..*<\/a>$' .tmp | sed -e 's/[^A-Za-z0-9._-]/_/g'

Still plenty of room for improvement.

Nick

Thanks.. Very useful right now that they have disabled the API.
I'm using this on Win 10 Ubuntu Bash, and everything works.

The problem is though, the script can't properly name the files

@elyobelyob I tried this line

alts=grep -ioE '^ >..*<\/a>$' .tmp | sed -e 's/[^A-Za-z0-9._-]/_/g'

But instead of the original naming, it names files just with numbers and no text