On Friday, February 25, 2005 3:36 PM [GMT],
Hugh Morris <[EMAIL PROTECTED]> wrote:

> Could someone tell me if it is possible to get the search query from
> requests referred from images.google.com (and the other image search
> engines).
>
> The request strings are a bit untidy, like:
>
>
http://images.google.com/imgres?imgurl=http://www.washto2004.org/net/external/images/sunday/BNSF%2520PHOTOS/MOUNTAIN%2520PEAK.JPG&imgrefurl=http://www.washto2004.org/net/external/images/sunday/BNSF%2520PHOTOS/&h=1200&w=1600&sz=696&tbnid=A3gVV6ed4DkJ:&tbnh=112&tbnw=149&start=1&prev=/images%3Fq%3Dmountain%2Bpeak%26hl%3Den%26lr%3D%26sa%3DG

Your search string is in the last part of that referring URL -
prev=/images%3Fq%3Dmountain%2Bpeak%26hl%3Den%26lr%3D%26sa%3DG

That's an escaped version of the original quesrystring -
prev=/images?q=mountain+peak&hl=en&lr=&sa=G

If I add
SEARCHENGINE http://images.google.com/* prev
to my analog.cfg file, then I get this entry in my Search Query Report:
1  /images?q=mountain peak&hl=en&lr=&sa=g

And I get these 2 entries in my Search Word Report:
1  peak&hl=en&lr=&sa=g
1  /images?q=mountain

I don't know if you could clean the results up any more than that
without preprocessing.

Aengus

+------------------------------------------------------------------------
|  TO UNSUBSCRIBE from this list:
|    http://lists.meer.net/mailman/listinfo/analog-help
|
|  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
|  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
+------------------------------------------------------------------------

Reply via email to