general purpose database cleanup tools (design patterns or Python)?

[Skip Montanaro]

A side effect of people using the Mojam and Musi-Cal web sites is that a
fairly complex MySQL database gets built.  There are several tables in the
database (venues, concerts, cities, etc) that can accumulate errors fairly
rapidly because the system allows anyone with concerts to publicize to add
them to our database.  Because of the size of the tables and their
interdependencies, manually correcting all but the most straightforward
errors is impossible.

I'm looking for pointers to design patterns or tools (preferably in Python)
that might be applicable to the task.  I think such tools would separate the
description of the database's structure from the rules about deleting and
merging rows, which would, in turn, be separate from the engine that
actually manipulates the database.

We have a tool now that works reasonably well but is very difficult to
maintain and enhance because none of the three functions are separate.
Before I launch into a complete redesign, I thought I'd see what already
exists.  This can't be a problem unique to our environment.

Thx,

-- 
Skip Montanaro (skip at mojam.com)
Support Mojam & Musi-Cal: http://www.musi-cal.com/sponsor.shtml
(847)971-7098