Rule #1: This is a photon, and all we are sure about is that a photon is an electromagnetic phenomenon. This fact, though, carries an implication: that the effect, being electromagnetic, can be directed - and possibly shaped - by a magnetic field.
Rule #2: Electromagnetic effects propogate at the speed of light. So sue me: we know, too, its speed of propogation. But, conveniently, this second fact is often ignored or misrepresented. It, too, carries an implication. Above, one of the answers (Andrew's) suggests that the effect can be described in language identical to the description of a toroidal magnetic field.
The speed of propogation implies, of course, that if the photon is a single point source, by travelling so fast it could appear (to our eyes) as being a toroidal shaped shell, since if it is trapped in a tiny magnetic "bottle" it would (at that speed) be capable of being in so many places over the course of one twenty-fourth of a second that it would appear to be at all points on the surface of that minute toroidal shell simultaneously.
In my opinion, Heisenberg's uncertainty principle is understandable on a consideration of what is happening here: to our unaided senses, the photon seems to be present at all points on the surface of our (hypothetical) toroidal-shaped magnetic "bottle", simultaneously. When physicists speak of being "uncertain" where the photon is, I try to bear in mind the speed at which it propogates, and the tiny distances involved at the quantum level: in a 24th of a second (the exposure duration of each individual frame of a movie) it can potentially pass through every single point on the theorised toroidal surface at least a hundred times. That, it seems to me, is the very essence of the nature of uncertainty.
I don't claim that it is a point-like particle. I'm not convinced by the arguments that suggest it is. But if we theorise that it might be, we get some insight into why uncertainty over its position and direction can exist, if both are changing too rapidly for us to measure, maybe too rapidly for any possible measurement to be meaningful.
Does this mean the photon "looks like" a point-like particle? Or does it mean it "looks like" a toroidal-shaped shell totally surrounding and obscuring the magnetic dipole? I suppose, if you freeze-frame, and so discount its motion entirely, it could look like the former; but if the universe of spacetime includes time, then over any meaningful time-period it could look like the latter.
Uncertainty: do we create this ourselves?
Any possible measurement we make in so minute an environment is certain to fatally disrupt the effect we are trying to measure, because the effect being measured is so ephemeral. When we "measure" it, we inject so much additional energy that we totally disrupt the tiny system we are measuring.
We only know how to measure by employing some type of bombardment: we are trying to do the equivalent of measure the position and momentum of a soap bubble, by the subtle procedure of hurling house-bricks at it!
This, I think, is what physicists mean when they say that particles (in general, not just photons) are uncertain in either position or momentum: the energy imparted to the "particle" in measuring either of those properties alters both its position and its momentum, rendering a further measurement (i.e. of the other property) meaningless.
It is most difficult to get a clear picture of what an object looks like by smashing it to pieces, with all the subtlety of an atom-smashing technique governed by the delicate scientific approach termed (in the scientific community yet!) as "brute force and ignorance".