Google Hacks 2Nd Edition [Electronic resources] نسخه متنی

Hack 51. Browse the World Wide Photo Album

Take a random stroll through the
world's photo album using some clever Google Image
searches (and, optionally, a smidge of programming
know-how) .

The proliferation of digital cameras and growing popularity of camera
phones are turning the Web into a worldwide photo album.
It's not only the holiday snaps of your Aunt Minnie
or minutiae of your moblogging friend's day that are
available to you. You can actually take a stroll through the publicly
accessible albums of perfect strangers if you know where to look.
Happily, Google has copies, and a couple of hacks know just where to
look.

3.5.1. Random Personal Picture Finder

Digital photo files have relatively standard filenames (e.g.,
DSC01018.JPG) by default and are usually
uploaded to the Web without being renamed. The

Random Personal
Picture Finder (http://www.diddly.com/random) sports a clever
little snippet of JavaScript code that simply generates one of these
filenames at random and queries Google Images for it.

The result, shown in Figure 3-3, is something like
looking through the world's photo album: people
eating, working, posing, and snapping photos of their cats,
furniture, or toes. And since it's a normal Google
Images search, you can click on any photo to see the story behind it,
and the other photos nearby.

Neat, huh?

Note that people snap pictures of not just their toes (or the toes of others). While an informal series of Shift-Reloads in my browser turned up only a couple of questionable bits of photographic work, you should assume the results are not workplace- or child-safe.

The code behind the scenes, as I mentioned, is really very
simplea swatch of JavaScript (view the source of http://www.diddly.com/random/randoml in
your browser to see the JavaScript bits for yourself) and list of
camera types and their respective filename structures (http://www.diddly.com/random/aboutl).
You're simply redirected to Google Images with
generated search query in tow.

A smidge of Python illustrates just how simple it is to generate a
link to some random collection of photos shot with a Canon digital
camera:

$ python
Python 2.3 (#1, Sep 13 2003, 00:49:11)
[GCC 3.3 20030304 (Apple Computer, Inc. build 1495)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> from random import randint
>>> linkform = 'http://images.google.com/images?q=IMG_%s.jpg'
>>> print linkform % str(randint(1, 9999)).zfill(4)
http://images.google.com/images?q=IMG_7931.jpg You can easily use this as the basis of a CGI script that acts in the
same manner as the Random Personal Picture Finder does.

3.5.2. WebCollage

And if you think the Random Personal Picture Finder is fun,
you'll love Jamie Zawinski's

WebCollage (http://www.jwz.org/webcollage/), a Perl
script that finds random pages, strips them of their images, and puts
these images together to create a collage. The collage is remixed and
added to once a minute, your browser reloading a fresh version every
so often (or just hit Shift-Reload on your browser yourself to speed
up the process). Figure 3-4 shows a WebCollage
snapshot in time.

Click on any one element and you'll be taken to the
page that the image is from.

You can grab the Perl source (http://www.jwz.org/webcollage/webcollage) and
generate a page similar to Jamie's like so:

$ webcollage -size '800x600' -imagemap ~/public_html/collage where 800x600 is the size and
~/public_html/collage is the path to and
file you'd like to save the image map as.

WebCollage is also included in Zawinski's

XScreenSaver (http://www.jwz.org/xscreensaver), a
screensaver system for computers running the X Windows System
(usually some brand of Unix).

As Jamie notes, he's disabled the Google search in WebCollage (you can just uncomment it in your own downloaded version) because it's not a Google API-based application and so is not in keeping with Google's Terms of Service ["A Note on Spidering and Scraping" in Chapter 9]. You can, however, very easily write a Google-specific API-friendly version.

Aaron Swartz