Spider Watcher

Spider Watcher
Author: Mikel Beck (mikel.beck@elite-computing.net)

This hack keeps track of the spiders (Search Engine robots) that visit your fourm. Every time a guest visits a page, the guest's IP address, user agent and the page they visited are logged to the database.

When somebody views the spider statistics page, this data is "rolled up", meaning the raw data is collated, the spider's name is determined by comparing the user agent to data contained in the spiders_bulletin.xml file, and the number of pages and visits is summarized and writted back to the database. In addition, and data from non-bots is removed.

The data is then displayed in a easy to read format for your viewing pleasure.

If the user viewing the report has permissions to view IP addresses, these are displayed as well.

A live version of the report from one of my sites can be seen here: http://www.happyhourpub.com/spiders.php

Also see the attached screenshot for an exmaple.


Revision History:
1.0.0 Beta 1 - 01/05/2006
- Initial Release

1.0.0 Beta 2 - 01/06/2006
- Included templates for spiders.php
- Removed text from templates, added them as phrases

1.0.0 Beta 3 - 01/07/2006
- Split up the display of "known" and "unknown" spiders

1.0.0 Beta 4 - 01/25/2006
- Corrected potentional SQL injection issue in plug-in
- Reduced the number of SQL queries required to display statistics
- Corrected date/time display issue

1.0.0 Beta 5 - 02/01/2006
- Reduced the number of SQL queries required to display statistics

1.0.0 Beta 6 - 02/08/2006
- No release

1.0.0 Beta 7 - 02/11/2006
- Corrected issue with "unknown" spiders not being displayed properly.
- Added tracking of the type of spider (searchspider, link checker, etc)

1.0.0 Beta 8 - 02/19/2006
- Change the display of IP addresses to be a pop-up so they're all not displayed on the main page.
- Combined the spiders that have the same name but different user agents.

1.0.0 Beta 9 - 03/10/2006
- Changed the display to group similar spiders together (search spiders, http check spiders, etc)

1.0.0 Beta 10 - 08/08/2006
- Changed how the rollup functions. Instead of rolling up every time somebody views the spider page, it rolls up once per hour.
- Corrected a few bugs here and there, mostly related to removing entries from the database.

Installation Instructions
1. Upload spiders.php to the root of your forum.
2. Upload spiders_rollup.php to the includes/cron directory.
3. Import the file product-spiderwatcher.xml using the Manage Products module.
4. Add a link to spiders.php on your navbar or footer.
5. Add a cron job with the following information:
Title: Spider Watcher Rollup
Day of the Week: *
Day of the Month: *
Hour: *
Minute: 0 - - -
Log entries: Yes
Filename: ./includes/cron/spiders_rollup.php

Upgrade Instructions
1. Upload (and overwrite) spiders to the root of your forum.
2. Upload spiders_rollup.php to the includes/cron directory.
3. Import the file product-spiderwatcher.xml using the Manage Products module. Make sure the "Allow Overwrite" option is set to "Yes".
4. Add a link to spiders.php on your navbar or footer.
5. Add a cron job with the following information:
Title: Spider Watcher Rollup
Day of the Week: *
Day of the Month: *
Hour: *
Minute: 0 - - -
Log entries: Yes
Filename: ./includes/cron/spiders_rollup.php

***UPGRADE NOTE***
When you upgrade from version 1.0.0 Beta 7 to 1.0.0 Beta 8 your existing spider data will be lost!

To make sure that you can decode the maximum amount of spiders, you should grab the latest spiderlist.xml and replace the spiders_vbulletin.xml file in your forumhome/includes/xml/ directory with the one from this thread: http://www.vbulletin.com/forum/showthread.php?t=76662

[Detomah]

Originally Posted by Megareus Rex

You might want to consider adding an IP display limit with a "View All" link to see all the IPs at your request, rather than displaying all on the main page itself.

I too would love to see something like this implemented, as Yahoo Slurp is awefully prolific and seems to be using a different IP address every single visit and at 17,000+ visits, it's getting pretty damn long at the moment.

[Megareus Rex]

Just another query update: Its now up to anywhere from 2200 to 4000 queries >.<

[StarBuG]

Any solution in sight for larger forums?

[mikelbeck]

Yes, I'm close to having a new version ready... One that should fix the heavy query issue.

[ChuanSE]

which is the hack you use on your forum to state winxp like the users didn't register yet mikelbeck? ya know, the yellow bar on top of the browserscreen in yellow

[Zia]

hello

Its My Spider's Page Listing Status

Page generated in 2.55627 seconds with 394 queries [Server Loads: 1.61 2.18 : 2.45]
[Page Compression: 42.06 k/45.78 k (8.13%)]

michelbeck..
can u add a modification in ur next release..

that now its devided in tow part

like the ip-address showing all..& it making the page too long..too long..just like Yahoo!Slurp...hundreds of IP...
it will show Ip Range..least this shorten page..

& the 2 part known & unknow...here can u devide few more..like
Search/Forum/Feed & Links/
its just my idea..

Thankx

[Brandon Sheley]

Originally Posted by mikelbeck

Yes, I'm close to having a new version ready... One that should fix the heavy query issue.

can I make a request, that shouldn't be hard.. at least I wouldn't think it should..

I'm doing and working on a lot of SEO changes for my site, and something like this could help me see whats working and whats not...



.could u add something in vboptions that could "reset" the count ?

. show a more detailed page of where they are looking ?

showthread= ?

have this an options to show too.. cuz I'm sure larger sites wouldn't want all this information logged.. maybe this would just involve more work then u want to do..


but the reset would be great for me, or maybe if you could post a SQL query that I can run that would do this ?

thank you

[redspider]

Hello, could someone please tell me why i am getting this error?

Warning: array_multisort(): Argument #1 is expected to be an array or a sort flag in /spiders.php on line 223

Also, i have no Spiders showing and there are several on my site as we speak...Thank you for any help in advance -

[redspider]

Ok, i dont know what happened, but my Sider Watch just mysteriously started working...No more error, and the Spiders are showing just fine...Maybe it just takes some time to kick in lol who knows, who cares? it works now...AWESOME HACK! Thank you!

[mikelbeck]

Ok, here's where I'm at right now.

I've got version 1.0.0 Beta 5 running on one of my sites. I just ran the spider page for the first time today, it took 0.21342 seconds with 57 queries. That's quite a bit better. ;-)

The problem is that the plug-in is a little heavy, I'd rather not do it this way. It checks to make sure that the wol_spiders is available (that's what contains the spider user agents) and then updates the database with the correct information at that point, instead of when the spiders page is run. I'm not sure if the code that loads up wol_spiders runs on EVERY page load or not, if it does that's bad. If it doesn't it should be OK. I'm still working on it...

The way the data is written to the database has changed in this new version, so the data that's currently in the database will be wiped out when this version is installed. Sorry, there's no way around that.

I'm still working on it, I should be able to put this version out in another couple of days.