porn filtering, blacklists, and squid log file analysis

["Dave" <dmehler26@xxxxxxxxxx>]

Hello,
   I'm trying to implement porn filtering. I'm trying a variety of setups 
to see which will give me the best results. First i'm using squid (2.6 port 
on FreeBSD), as a transparent proxy in all setups. Setup1 is using 
squidGuard, and the Mesd blacklist. When i dropped in mesd to the picture 
the situation improved, a lot of previously accessible sites were now 
blocked. My volunteer has a test machine for this and was able to google and 
to either pull up images, nothing with pornographic-like names, but that 
kind of images, and sites that weren't on the list. I update the blacklist 
every night, but i need to write a script that goes through the access.log, 
finds machine accesses and where they go, and then sets up a list of sites. 
It then goes through said list, eliminating all duplicate entries, and sees 
which domains still work, those that do are automatically added to a custom 
squidguard blacklist and squidguard is reconfigured, squid reloaded.
   After that explanation i use grep on the access.log to find only the 
accesses from the machine i want my test box, put that in another file. I 
then use cut to take out i think it's field 10 or 11 it's the url of the 
page, drop that in another file. The problem is i have a file containing 
9500 entries, manually going through this isn't an option. If anyone can 
help with this i can put the file somewhere where it can be downloaded.
       On the subject of blacklists aside from the mesd list, is there 
anymore lists for squid/squidguard, that are free or free for noncommercial 
purposes?
   My second setup involves dansguardian. My issue with this is first the 
last time i tried this yes it worked though i never stress-tested this to 
the extent i'm going for now, and second it seemed to slow the internet down 
very noticeably to the point where everyone was telling me. I've got squid 
as a transparent proxy using pf and i'd like to keep that arrangement, last 
time i had to change this if there's an alternative i'm open to suggestions.
Thanks.
Dave.