mschauer

using Random, LinearAlgebra, Distributions, StatsBase
# Linear Algebra of Markov chain

# Example: Weather Markov chain of Oz
S = [:R, :S, :C]
P = [0.5 0.25 0.25
     0.5  0    0.5
     0.25 0.25 0.5]

sum(P[2, :])

mschauer

using Observables, BenchmarkTools, Test

struct StaticObservable{T, Listeners} <: Observables.AbstractObservable{T}
    listeners::Listeners
    val::Base.RefValue{T}
end

StaticObservable(obs::Observable) = StaticObservable(tuple(obs.listeners...), Ref(obs[]))
function StaticObservable(val::Ref, listeners...) 
    _notify(val, listeners)

mschauer

n = 1000
x = range(0, 1, length=n)
ς = 1.5 # noise level

μ = 3*x.*sin.(2pi*x) # periodic signal in time domain
#μ = 6*sqrt.(abs.(x .- 0.5)) # this one is difficult to estimate

# Model: Signal distorted by white noise
y = μ + ς*randn(n)  

mschauer

using Mitosis
using MitosisStochasticDiffEq
import MitosisStochasticDiffEq as MSDE
using StaticArrays, LinearAlgebra
using OrdinaryDiffEq

# Match with B and sigma
B(θ) = [-0.1 0.2θ; -0.2θ -0.1]
beta(θ) = [0.,0.]
Σ(θ) = 0.15*I(2)

mschauer

using AdvancedMH
using Distributions
using Random
using MCMCChains
using StructArrays
struct MvWiener{Tu0,Ttrange} <: ContinuousMatrixDistribution
    u0::Tu0
    trange::Ttrange
end

mschauer

using LinearAlgebra, Test, Random, Statistics
outer(x) = x*x'
lchol(x) = cholesky(Symmetric(x)).L
Random.seed!(1)

# dimensions
n = 10 # observed
p = 5 # latent

# parameters

mschauer

using StaticArrays
struct EulerMaruyama!
end


"""
    tangent!(du, u, dz, P)


"""

mschauer

  using GitHub, JSON
  
  #myauth = GitHub.authenticate(ENV["GITHUB_AUTH"])
  
  gist_json(fn; description="$fn", public=true) = JSON.parse(
  """
  {
    "description": "$description",
    "public": $public,
    "files": {

mschauer

function mappairreduce(r, f, itr, dec)
    ϕ  = iterate(itr)
    if ϕ === nothing
        return dec
    end
    xlast, state = ϕ
    while true
        ϕ = iterate(itr, state)
        if ϕ === nothing 
            return dec

mschauer

using Makie
using Colors
n = 100
f(x) = 0.5 .*reim(((x[1] + x[2]*im)/33)^2)
function f2(x) 
    0.5.*abs(2*(x[1] + x[2]*im)/1089)
end
#f(x) = (x[1],x[2])
#f2(x) = 1.0