Chad Scherrer
Chad Scherrer
> `ArrayInterface.size(ones(3)')` would lower to code like this at compile time: `Static.StaticInt(1), Base.arraysize(x, 1)`. This is a great example! It really helps to see what's computed at compile time. Could...
There are a couple of issues here involving using macros within `@resumable`. Usually, a quick fix would be to have `@resumable` call `macroexpand` on the expression before rewriting things. The...
In that case, maybe it would make sense to remove them from master and add a branch to prototype an alternative? I've got a sprint coming up for the JuliaCon...
Hi @baggepinnen , I was thinking some more a bout this today... I think we could do something like ```julia struct Weights{T,N} vals::Vector{T} members :: MutableLinkedList{WeightedParticles{T,N}} end struct WeightedParticles{T,N} x...
> The `members` field has a list of all `WeightedParticles` that are part of the same "multivariate" particles, i.e., `Vector{WeightedParticles}`? Each element in such a vector in turn has a...
For a moment I thought this might be entirely wrong, but I'm pretty sure this is correct. In particle filtering I've done (maybe there are variations?) each particle maintains a...
Ok here's an example. I think all of this checks out: ```julia using MonteCarloMeasurements using Distributions # Expected value of x log-weighted by ℓ function expect(x,ℓ) w = exp(ℓ -...
Hi @millerjoey , not sure if we've talked about this before, but lots of `Distribution`s have a `check_args` keyword argument. From a PPL perspective, it's a bit annoying because it's...
For now I just have ```julia function Base.getindex(a::AbstractArray{P}, i::Particles) where P
Thanks @baggepinnen , the "triangle of methods" idea is pretty clever, and seems like it might work