A "DI", strictly speaking, is not required for this circuit. The output impedance of the speaker is low . . . so is its output, so I'd be tempted to use a higher ratio transformer than the 1:2 in his schematic.
Of course, when you start talking transformers for a mic that is designed to only pick up low frequencies, then you start talking about an expensive transformer. If this is just for a loud source like kick, doulos' simple wiring to XLR is best.
I don't get the point of all the switching options. Who cares? Switches add a lot of cost for not much value here. You don't need a lower impedance, so leave it at 8 ohms and take the extra signal.
Finally, I can't imagine why anyone would ever want to tie the XLR pin 1 to transformer primary, even across the cap/resistor arrangement. This box is not shielded anyway, unless we are talking about tying to a shield on the driver coil itself. If hum is a problem here, just shield the whole enclosure.
As an aside, many people have an erroneous notion that you need a large diaphragm like a speaker mic to pick up a lot of lows. That's actually very untrue; what these mics do is simply reject practically all highs above 4kHz, if not lower, and they have slow transient response to boot. You can use a regular mic and toss out highs with EQ and achieve much the same thing . . . add some saturation and get crazy with
a transient designer plug if you are fastidious . . .