I've written a small pythonic apache access log parser framework which will hopefully make it easier for us to do more aggregate data analysis in the same vein of what we have tried to do with Fedora Maps.

Attached you'll find three python files

parser.py : which holds the logic for parsing log files.

handlers.py: it has an example of an analysis handler, which looks for substring matches in the request field and determines unique ips and the number of visits in the time window.

weekly-log-parse.py: which sketches an example of how to use the example unique ip handler to for logs over the past week. I've tested it with my local apache logs, and added logic based on

Fedora Map making calculations and other future analysis could be added in as additional handlers as interest warrants.

ref: https://www.redhat.com/archives/fedora-devel-list/2009-January/msg01456.html