Skip to content

Commit

Permalink
Remove unneccessary nsteps argument
Browse files Browse the repository at this point in the history
  • Loading branch information
@brenhinkeller
brenhinkeller committed Mar 2, 2025
1 parent 7c2df18 commit f489c80
Showing 1 changed file with 14 additions and 15 deletions.
29 changes: 14 additions & 15 deletions src/metropolis.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -151,22 +151,22 @@ function metropolis_min(nsteps::Integer, dist::Collection, mu::Collection, sigma
# Allocate ouput array
tmindist = Array{float(eltype(mu))}(undef,nsteps)
# Run Metropolis sampler
return metropolis_min!(tmindist, nsteps, dist, mu, sigma; kwargs...)
return metropolis_min!(tmindist, dist, mu, sigma; kwargs...)
end
function metropolis_min(nsteps::Integer, dist::Collection{T}, analyses::Collection{UPbAnalysis{T}}; kwargs...) where {T}
# Allocate ouput arrays
tmindist = Array{T}(undef,nsteps)
t0dist = Array{T}(undef,nsteps)
# Run Metropolis sampler
metropolis_min!(tmindist, t0dist, nsteps, dist, analyses; kwargs...)
metropolis_min!(tmindist, t0dist, dist, analyses; kwargs...)
return tmindist, t0dist
end


"""
```julia
metropolis_min!(tmindist::DenseArray, nsteps::Integer, dist::Collection, mu::AbstractArray, sigma::AbstractArray; burnin::Integer=0)
metropolis_min!(tmindist::DenseArray, t0dist::DenseArray, nsteps::Integer, dist::Collection, analyses::Collection{<:UPbAnalysis}; burnin::Integer=0) where {T}
metropolis_min!(tmindist::DenseArray, dist::Collection, mu::AbstractArray, sigma::AbstractArray; burnin::Integer=0)
metropolis_min!(tmindist::DenseArray, t0dist::DenseArray, dist::Collection, analyses::Collection{<:UPbAnalysis}; burnin::Integer=0) where {T}
```
In-place (non-allocating) version of `metropolis_min`, fills existing array `tmindist`.
Expand All @@ -179,8 +179,7 @@ zircon eruption ages from a distribution of zircon crystallization ages.
metropolis_min!(tmindist, 2*10^5, MeltsVolcanicZirconDistribution, mu, sigma, burnin=10^5)
```
"""
function metropolis_min!(tmindist::DenseArray, nsteps::Integer, dist::Collection, mu::AbstractArray, sigma::AbstractArray; burnin::Integer=0)
@assert length(tmindist) == nsteps
function metropolis_min!(tmindist::DenseArray{<:Number}, dist::Collection{<:Number}, mu::AbstractArray{<:Number}, sigma::AbstractArray{<:Number}; burnin::Integer=0)

# standard deviation of the proposal function is stepfactor * last step; this is tuned to optimize accetance probability at 50%
stepfactor = 2.9
Expand Down Expand Up @@ -256,8 +255,8 @@ function metropolis_min!(tmindist::DenseArray, nsteps::Integer, dist::Collection
end
return tmindist
end
function metropolis_min!(tmindist::DenseArray{T}, t0dist::DenseArray{T}, nsteps::Integer, dist::Collection{T}, analyses::Collection{UPbAnalysis{T}}; burnin::Integer=0, t0prior=Uniform(0,minimum(age68.(analyses))), lossprior=Uniform(0,100)) where {T}
@assert length(tmindist) == length(t0dist) == nsteps
function metropolis_min!(tmindist::DenseArray{T}, t0dist::DenseArray{T}, dist::Collection{T}, analyses::Collection{UPbAnalysis{T}}; burnin::Integer=0, t0prior=Uniform(0,minimum(age68.(analyses))), lossprior=Uniform(0,100)) where {T<:AbstractFloat}
@assert eachindex(tmindist) == eachindex(t0dist)

# standard deviation of the proposal function is stepfactor * last step; this is tuned to optimize accetance probability at 50%
stepfactor = 2.9
Expand Down Expand Up @@ -396,17 +395,17 @@ function metropolis_minmax(nsteps::Integer, dist::Collection, mu::AbstractArray,
tmaxdist = Array{float(eltype(mu))}(undef,nsteps)
tmindist = Array{float(eltype(mu))}(undef,nsteps)
# Run metropolis sampler
return metropolis_minmax!(tmindist, tmaxdist, lldist, acceptancedist, nsteps, dist, mu, sigma; kwargs...)
return metropolis_minmax!(tmindist, tmaxdist, lldist, acceptancedist, dist, mu, sigma; kwargs...)
end
function metropolis_minmax(nsteps::Integer, dist::Collection{T}, analyses::Collection{<:UPbAnalysis{T}}; kwargs...) where T
function metropolis_minmax(nsteps::Integer, dist::Collection{T}, analyses::Collection{<:UPbAnalysis{T}}; kwargs...) where {T<:AbstractFloat}
# Allocate ouput arrays
acceptancedist = falses(nsteps)
lldist = Array{T}(undef,nsteps)
t0dist = Array{T}(undef,nsteps)
tmaxdist = Array{T}(undef,nsteps)
tmindist = Array{T}(undef,nsteps)
# Run metropolis sampler
return metropolis_minmax!(tmindist, tmaxdist, t0dist, lldist, acceptancedist, nsteps, dist, analyses; kwargs...)
return metropolis_minmax!(tmindist, tmaxdist, t0dist, lldist, acceptancedist, dist, analyses; kwargs...)
end

"""
Expand All @@ -426,8 +425,8 @@ crystallization ages.
metropolis_minmax!(tmindist, tmaxdist, lldist, acceptancedist, 2*10^5, MeltsVolcanicZirconDistribution, mu, sigma, burnin=10^5)
```
"""
function metropolis_minmax!(tmindist::DenseArray, tmaxdist::DenseArray, lldist::DenseArray, acceptancedist::BitVector, nsteps::Integer, dist::Collection, mu::AbstractArray, sigma::AbstractArray; burnin::Integer=0)
@assert length(tmindist) == length(tmaxdist) == length(lldist) == length(acceptancedist ) == nsteps
function metropolis_minmax!(tmindist::DenseArray, tmaxdist::DenseArray, lldist::DenseArray, acceptancedist::BitVector, dist::Collection{<:Number}, mu::AbstractArray{<:Number}, sigma::AbstractArray{<:Number}; burnin::Integer=0)
@assert eachindex(tmindist) == eachindex(tmaxdist) == eachindex(lldist) == eachindex(acceptancedist)

# standard deviation of the proposal function is stepfactor * last step; this is tuned to optimize accetance probability at 50%
stepfactor = 2.9
Expand Down Expand Up @@ -506,8 +505,8 @@ function metropolis_minmax!(tmindist::DenseArray, tmaxdist::DenseArray, lldist::
end
return tmindist, tmaxdist, lldist, acceptancedist
end
function metropolis_minmax!(tmindist::DenseArray{T}, tmaxdist::DenseArray{T}, t0dist::DenseArray{T}, lldist::DenseArray{T}, acceptancedist::BitVector, nsteps::Integer, dist::Collection{T}, analyses::Collection{UPbAnalysis{T}}; burnin::Integer=0) where {T}
@assert length(tmindist) == length(tmaxdist) == length(t0dist) == length(lldist) == length(acceptancedist ) == nsteps
function metropolis_minmax!(tmindist::DenseArray{T}, tmaxdist::DenseArray{T}, t0dist::DenseArray{T}, lldist::DenseArray{T}, acceptancedist::BitVector, dist::Collection{T}, analyses::Collection{UPbAnalysis{T}}; burnin::Integer=0) where {T <: AbstractFloat}
@assert eachindex(tmindist) == eachindex(tmaxdist) == eachindex(t0dist) == eachindex(lldist) == eachindex(acceptancedist)

# standard deviation of the proposal function is stepfactor * last step; this is tuned to optimize accetance probability at 50%
stepfactor = 2.9
Expand Down

0 comments on commit f489c80

Please sign in to comment.