scriptygoddess

1 | Richy C.

January 18th, 2003 at 7:26 pm

I take it you are aware that MT RefSearch does a similar thing and copes with around 240 search engines? Include ones which can have 'multiple parameters' in the search field?

The code is reasonably documented – so if you want to take a peek – be my guest!

2 | Jennifer

January 18th, 2003 at 8:02 pm

Actually – I have MT RefSearch working here too – hadn't even thought to look to see if it had seperated out a function that would do that… Changing the code above.
To download the full code to MT RefSearch – go here.

3 | Jason

May 25th, 2003 at 4:31 pm

I have ben looking everywhere to find a function to strip out the keywords from the search engines refering url and looked at the code you have here on this site but I seem to get lost at the end. See, what I need it to do is just…instead of writing to a flat file or lof file, I want to store the keywords in a mysql database so that I can include it with a stats program I have been working on. Any tips? Anything would be appreciated!

4 | RaulentRoi

July 11th, 2003 at 3:55 pm

Thank you goddess for this code. It works just fine and I couldn't have done it without you.

I did have a couple of problems with it having to do with French accents. And also I chose to clean the keywords when some unwanted characters were in it, like dots or commas. I also eliminated all the one letter words. This is what it turned out like (This is a rough looking code, but I 'm just a beginner):

//*First I query the referrals URLs in my data base, which I name URLJustvisited

$URLJustVisited = mysql_result($result,$i,"URLJustVisited");
$keywords = ExtractKeywords($URLJustVisited);
foreach ($keywords as $value) {
if ($keywords != "") {

//*Now I clean up the unwanted characters – it may seem strange to replace a character by itself, but it does have an effect on the outcome
$value1 = str_replace("+", " ", $value);
$value2 = str_replace("-", "", $value1);
$value3a = str_replace("qu'", "", $value2);
$value3b = str_replace("d'", "", $value3a);
$value3c = str_replace("l'", "", $value3b);
$value3d = str_replace("n'", "", $value3c);
$value3e = str_replace("m'", "", $value3d);
$value3f = str_replace("j'", "", $value3e);
$value3g = str_replace("t'", "", $value3f);
$value3h = str_replace("'", "", $value3g);
$value4 = str_replace("é", "é", $value3h);
$value5 = str_replace("é", "é", $value4);
$value6 = str_replace("â", "â", $value5);
$value7 = str_replace("ç", "ç", $value6);
$value8 = str_replace("ç", "ç", $value7);
$value9 = str_replace("è", "è", $value8);
$value10 = str_replace("è", "è", $value9);
$value11a = str_replace(":", "", $value10);
$value11b = str_replace(".", "", $value11a);
$value11c = str_replace(";", "", $value11b);
$value11d = str_replace(",", "", $value11c);
$value12 = str_replace("\"", "", $value11d);
$value13 = str_replace(" ", " ", $value12);
print "$value13<br>";

//*Now I clear all one letter results since they are not keywords

if (strlen($value13) > 1) {

//********************************
//*Now I proceed with the insert – update database with $value13

I had a problem while inserting into DataBase: A "duplicate entry warning" (on die), despite the fact that I querried to make sure that $value did not already exist prior to attempt to insert it in the database.

Why this happens is still a mystery to me. I went around it by disabling the "OR die" which is not a big problem for my usage since it concerns a very few words that are very unlikely to require analysis.

Finally, I added a delete command to get rid of a bunch of words that are not keywords (the, this, at, for ….).

Thanks again.