Load Specific URLs from Google search?

[Anon]

Here's my goal. I'm a high school teacher looking for ways to detect plagiarism in student papers.

I want to be able to leech the top five links for a search topic, such as "Henry Kissinger" from google, and then load the contents of those urls into variables. I can then use the similar_text() function to determine the possibility of plagiarism for each document by comparing those results with the student's paper.

I know how to load specific urls into variables, and I've worked out a lot of the problems with similar_text() - I had to strip all of the tags out of the documents - but I'm not sure if there is a quick and dirty way to create five variables like this, based on search results from specifically google.com.

$link1 = "http://www.henrykissinger.com";
$link2 = "http://www.nobelprize.com/henry";
...etc.

Any help would be greatly appreciated.

[Anon]

Yes you can =)

do a HTTP Post to Google - listen for the results. You can parse the URL by
"<!--m-- >"
each result start with the comment out "!--m--" - parse each URL into an array. (do a view source on your search you will see it)

This way you don't even have to script out the results, just take out what you need.

[Anon]

Ok, you've intrigued me, but I'm not following how I go about doing this.... remember, you're dealing with a teacher here... "Those who CAN'T - Teach!".

Thanks!

Cong wrote:

Yes you can =)

do a HTTP Post to Google - listen for the results. You can parse the URL by
"<!--m-- >"
each result start with the comment out "!--m--" - parse each URL into an array. (do a view source on your search you will see it)

This way you don't even have to script out the results, just take out what you need.

[Anon]

First thing you want to do is write a php that'll do an HTTP POST/GET to Google, and then listen for the returned data. This file might look something like this:

// you'll have to find out all the parameter google will accept, in this case I am searching for "foo bar"
$path = "search?hl=en&lr=&safe=off&q=foo+bar";

$fp = fsockopen($host,80,&$errno,&$errstr,10000
);
fputs($fp, "GET $path HTTP/1.1\n");
while(!feof($fp)) {
$searchresults .= fgets($fp, 128);
}
fclose($fp);

next, what you want to do is parse $searchresults for the <!--m-->, everytime you see this marker setup a new array index and insert the URL into it.

now even more work, go through each of the element in the URLs array above, connect to it, grab the returned data and start using your similar_text thingie.