I think the best book is probably still Dr. Bernstein's Diabetes Solution. He doesn't advocate actually being in ketosis necessarily--the plan he describes I'm sure will have some people in a mild state of ketosis at times, at least.
Obviously extra problems for a more ketogenic diet for type I's. I think you'd probably need to test blood ketones and glucose regularly, that could get expensive. If one of Dr. Bernstein's patients measured a blood glucose of 60, he'd have them taking a couple of grams of glucose to correct the hypo. But occasional glucose that low is within normal variation for somebody on a high fat ratio ketogenic diet. A few years ago when I first tried a higher ratio diet, my blood glucose went into to mid-60's pretty regularly sometime in the afternoon. You can't tell from that blood glucose number alone whether you're in a true hypo.
A non-diabetic can probably figure it's fine if they feel fine, but type I's get resistant to the early signs of hypo, by the time it's more obvious, it can be too late.
I think I've seen a case study for a type I diabetic who also had epilepsy...
Successful treatment of type 1 diabetes and seizures with combined ketogenic diet and insulin.
Aguirre Castaneda RL1, Mack KJ, Lteif A.
Diabetic ketoacidosis (DKA) is a life-threatening condition and a major cause of morbidity and mortality in children with type 1 diabetes mellitus. The deficiency of insulin leads to metabolic decompensation, causing hyperglycemia and ketosis that resolves with the administration of insulin and fluids. However, an induced state of ketosis is the basis for the success of the ketogenic diet (KD), which is an effective therapy for children with intractable epilepsy. We report the case of a 2-year-old girl who presented to the emergency department with 1-week history of decreased activity, polyuria, and decreased oral intake. Her past medical history was remarkable for epilepsy, for which she was started on the KD with a significant improvement. Her laboratory evaluation was compatible with DKA, and fluids and insulin were given until correction. Because of concerns regarding recurrence of her seizures, the KD was resumed along with the simultaneous use of insulin glargine and insulin aspart. Urine ketones were kept in the moderate range to keep the effect of ketosis on seizure control. Under this combined therapy, the patient remained seizure-free with no new episodes of DKA.
Just the abstract... looks like type I probably developed when the child was already on the ketogenic diet, the ketogenic diet didn't cause ketoacidosis (it's even likely to have prevented it for a time, the extremely low insulin requirement of the ketogenic diet might have masked what would otherwise have shown as insulin insufficiency sooner on a diet with a higher insulin requirement).
Urine ketones... okay, I guess that makes sense. At mild levels of ketosis, they're not a perfect match for blood ketones. But at ketoacidosis levels--they are always going to show deep purple, the colour depends on blood levels of acetoacetone available to pass into the urine, and the kidney's ability to reabsorb those ketones from the urine, at ketoacidosis levels, the levels in the urine will be far past the kidney's ability to reabsorb enough ketones to make the strips a paler colour. Dark purple does not mean you're in ketoacidosis, but a moderate colour means you're probably not, at least on a ketogenic diet (although on a non-ketogenic diet it will likely mean that you are insulin insufficient).
Which all sounds like while you should get a good doctor to help with this if you're following Dr. Bernstein's not-quite-ketogenic plan, you might need an even better doctor if you want to tip over into ketosis a bit more.