I'm trying to build a filtering proxy for personal use but I can't seem to find a solution that would be doing filtering by web page contents. All I find is just filtering by URL. Any ideas?
Get Squid working first - then compile/install DG via whatever method (I always compiled it)
It has great documentation and there's lots of examples.
You might also look at ClearOS (formerly Clark Connect) - I believe they use DG in their web filter, along with Squid. It makes a nice package with web-based management (Based on CentOS as the underlying OS).