All of the examples I've posted so far (and that I'm writing) are based on the Laplace equation: specify known voltages, and solve for no sources. This is a very useful technique.
Sometimes it is more useful to specify a charge distribution. To achieve this in #FEniCS we can make use of subdomains and/ or expressions. I'll be starting a post on solving for relatively arbitrary charge distributions. (Just got an example working myself).