Differentiating between visitor and bot in user agents

[etrader]

I am writing a simple counter, but it counts both visitors and bots. To distinguish between these visits, I use $_SERVER['HTTP_USER_AGENT']. By preg_match I can get which visit is by a bot. The common way in the internet is to define preg_match for each major bot (e.g. googlebot). As I explored, all bots have their official url in the uer_agent. Thus, preg_match by "http://" should consider all bots; as user_agent of a human visitor does not include a url. Right?

 

What's your idea to do so?

[LeadingWebDev]

<?php
echo $_SERVER['HTTP_USER_AGENT'] . "\n\n";
$browser = get_browser(null, true);
print_r($browser);
?>

should help you, but remember you should configure it on server and support him browser list.

Reference: http://php.net/manual/en/function.get-browser.php

[etrader]

<?php
echo $_SERVER['HTTP_USER_AGENT'] . "\n\n";
$browser = get_browser(null, true);
print_r($browser);
?>

should help you, but remember you should configure it on server and support him browser list.

Reference: http://php.net/manual/en/function.get-browser.php

 

Thanks LeadingWebDev,

Yes this is an alternative solution, as get_browser will provide [crawler] to be true or false. But I have read some reviews on the internet as the [crawler] is not very accurate, probably because get_browser is not very popular. On the other hand, it gives lots of information which are useless to me. Then, I though using $_SERVER['HTTP_USER_AGENT'] and performing preg_match is a lighter process. What do you think?

[ohdang888]

Well, really, you're limited to the information the client provides you. A.k.a - a bot could send a IE header, and you wouldn't know (using this method, at least).

 

but anyways, a simple search reveals a quick solution:

<?php
if (preg_match('/slurp|inktomisearch|[Gg]rub|[bb]ot|archiver|[ss]qworm/', $_SERVER['HTTP_USER_AGENT']))
     echo "Is not a human";
?>

[LeadingWebDev]

no, Regular Expressions loads server hard.

If u are building statistics that shouldn't catch crawlers (web search spiders) probably get_browser will help you, as every search bot have his own user agent.