mandatory bluff follow through?

Isn't villain's continuing range also going to have diamonds on a ton? We don't expect him to peel dry two pair do we?

Yes which is why I'm a bit torn. He could peel dry two pairs, but most hands are at least pair+fd.

I would just delayed cbet turn. If we had nut fd or at least A♦: x-r could be nice. Then we have also kind of easier time to pull the trigger on river like this with nut blocker. Plus of course more fe if we block nfd from villain.

I were also considering just x-c turn because we should be beating his bluffs though some KTT type might bluff sometimes.

Regarding river sizing I would go with pot or check approach. We can't value bet straight so if we bet it should be polarised into nuts (maybe 2nd nuts) or bluff. Weaker flushes work better as x-c theory wise at least imo.

Usually I would x-c turn with this hand +nfd no reason to raise. I would x-r with weaker fds due to visibility issues.
Delayed c-bet is interesting. I usually think of it when I'm IP not OOP. Yeah I agree delayed c-bet with A♦: is very nice.

The idea with x-r nut fd or A♦: blocker would be to have more equity for the bluff. So if river is blank we can still follow though with straight blocker but if flush complete we can either pot or check as trap. Issue with that option is that people might 3bet too much with straights on turn and force us to fold good amount of equity.

But yeah I would anyway aim to have more equity or more blocker value to continuing range if x-r turn. Might be wrong approach though.

In turn for our straights I would in general delayed cbet more of weaker straights and look for x-r or sometimes x-c with strong ones. Reasoning being if we have straight with strong support we actually block villain chance to continue versus delayed cbet and we would prefer to give him chance to bluff. Also if we have dry straight we can obviously value bet and call versus raise but it's not necessarily good enough to x-r. Some of weakest straights can also work nice as x-c though.