Automatic download of images with Google images

Google images allows you to search the web for images instead of web pages. You can specify keywords but also image size, file extension, dominant color, …

Sometimes you just need a set of image to test your latest program. Instead of opening every web page and right-click/Save as image you might want to use the following script.

• Copy/paste the script at the end of the post in a file.
• Make it executable: $ chmod +x googleimage.sh
• Give it a try by searching red bananas image with a .jpg extension:
  $ cd /tmp
$ /path-of-the-script/google-image.sh -s bananas -c red -e jpg

• If you want more images use the “page” option:
  $ /path-of-the-script/google-image.sh -s bananas -c red -e jpg -p 0
$ /path-of-the-script/google-image.sh -s bananas -c red -e jpg -p 1
$ /path-of-the-script/google-image.sh -s bananas -c red -e jpg -p 2
$ /path-of-the-script/google-image.sh -s bananas -c red -e jpg -p 3
$ /path-of-the-script/google-image.sh -s bananas -c red -e jpg -p 4

The script use wget the retrieve the content of the page then it parses it to find images URL. The only tricky part is to set the user-agent (-U option of wget) to an actual web browser (here firefox), otherwise you won’t have any results. There is however some limitations: some images appears twice or more, you cannot use multiples keywords, …

#!/bin/bash

# TODO: command line takes arguments
# search, color (all), number of answers
#

# Dislay help message
function dispHelp()
{
echo "Google Image Retriever version $VERSION"
echo "Usage:  gir [options]"
echo "  -s     : specify keywords"
echo "  -e : file extension filter (gif, jpg, png, ...)"
echo "  -c      : dominant color: red,orange,yellow,green,teal,blue,purple,pink,white,grey,black,brown"
echo "  -p     : page result (20 images per pages)"
echo "  -h -help       : print usage information"
}

###############################################################################
### Begining/Initialisation ###
VERSION="0.1"  # Program version

SEARCH=""e
EXTENSION=""
COLOR=""
OFFSET=""

### Input option parsing ###
while getopts ":s:e:c:p:h" opt; do
case $opt in
s)
SEARCH=${OPTARG}
;;
e)
EXTENSION=${OPTARG}
;;
c)
COLOR=${OPTARG}
;;
p)
OFFSET=`echo ${OPTARG}*20 | bc`
;;
h|help)
if [ $# = 3 ]
then
echo "Too much arguments with -h option: use -h for help"
exit -1
fi
dispHelp
exit 0
;;
\?)
echo "Invalid option: -${OPTARG}" >&2
;;
esac
done

# download google page result
cmd="wget -U firefox -O - http://images.google.com/images?q=${SEARCH}\&imgcolor=${COLOR}\&start=${OFFSET}\&as_filetype=${EXTENSION}"
cmd="${cmd} | tr \"?&\" \"\\n\""
cmd="${cmd} | grep imgurl"
cmd="${cmd} | tr \"=\" \"\\n\""
cmd="${cmd} | grep http"
cmd="${cmd} | wget -U firefox -i -" # user agent: Firefox, read from pipeline
eval "${cmd}"

# remove identical files
rm *.?
rm *.??
rm *.html.?
rm *.html.??

echo "SEARCH=${SEARCH}"
echo "COLOR=${COLOR}"
echo "OFFSET=${OFFSET}"
echo "EXTENSION=${EXTENSION}"

About these ads