A couple of short responses, corrections, etc.
Rook is entirely correct, except for substituting pounds for tons when discussing how much uranium ore is necessary. He failed to mention two things. First, uranium hexaflouride, the uranium gas used in the gaseous diffusion process, is highly caustic. This requires both special process piping and special diffusion filters. During the Manhattan Project, these caused no end of trouble. They would cause just as much trouble today as they are not exactly off the shelf items and attempting to purchase them directly would have raised all kinds of red flags.
Second, there are other ways to seperate uranium. The most common is to use a magnetic field to seperate the isotopes by weight. This is pretty low yield and requires ungodly amounts of electricity.
I concur with Rook's assesment that it's easier to a) steal weapons grade fissionables, b) buy them from Russia or China, or c) buy a whole bomb from a source like that. Making it from scratch is a huge industrial endeavour.