While I can see benefits to a rectangular setup with LEDs I'm not sure how snaking back and forth helps. As you gain length (time) you decrease cross section area and therefore flow speed increases proportionally and you end up with the same contact time per flow rate, but with smaller passages that drive higher frictional losses in the line. A line source on the axis of a tube is pretty efficient as far as avoiding self-shading goes. There's a reason spiral CFLs aren't used on places that don't meet their particular space constraints.
It would be interesting if you could do a shell with two different concentric passages. One close to the UV and one farther away. The section at the higher intensity for parasites, and the section at lower intensity to have higher throughput for algae reduction. Of course the surface fouling on the inside would inhibit this, so I'm not sure how well it would work in practice.
