Issue #509: Remove remaining inner constructors that don't use new

EpiAware/src/EpiAwareUtils/EpiAwareUtils.jl

@@ -19,15 +19,14 @@ export HalfNormal, DirectSample, SafePoisson, SafeNegativeBinomial, SafeIntValue
        SafeDiscreteUnivariateDistribution
 
 #Export functions
-export scan, spread_draws, censored_cdf, censored_pmf, get_param_array, prefix_submodel, ∫F
+export spread_draws, censored_cdf, censored_pmf, get_param_array, prefix_submodel, ∫F
 
 # Export accumulate tools
 export get_state, accumulate_scan
 
 include("docstrings.jl")
 include("censored_pmf.jl")
 include("HalfNormal.jl")
-include("scan.jl")
 include("accumulate_scan.jl")
 include("prefix_submodel.jl")
 include("turing-methods.jl")

EpiAware/src/EpiAwareUtils/SafeNegativeBinomial.jl

@@ -84,7 +84,6 @@ struct SafeNegativeBinomial{T <: Real} <: SafeDiscreteUnivariateDistribution
     end
 end
 
-#Outer constructors make AD work
 function SafeNegativeBinomial(r::T, p::T) where {T <: Real}
     return SafeNegativeBinomial{T}(r, p)
 end

EpiAware/src/EpiInfModels/EpiData.jl

@@ -50,16 +50,16 @@ struct EpiData{T <: Real, F <: Function}
             length(gen_int),
             transformation)
     end
+end
 
-    function EpiData(; gen_distribution::ContinuousDistribution,
-            D_gen = nothing,
-            Δd = 1.0,
-            transformation::Function = exp)
-        gen_int = censored_pmf(gen_distribution, Δd = Δd, D = D_gen) |>
-                  p -> p[2:end] ./ sum(p[2:end])
+function EpiData(; gen_distribution::ContinuousDistribution,
+        D_gen = nothing,
+        Δd = 1.0,
+        transformation::Function = exp)
+    gen_int = censored_pmf(gen_distribution, Δd = Δd, D = D_gen) |>
+              p -> p[2:end] ./ sum(p[2:end])
 
-        return EpiData(gen_int, transformation)
-    end
+    return EpiData(gen_int, transformation)
 end
 
 @doc raw"

EpiAware/src/EpiInfModels/Renewal.jl

@@ -86,18 +86,6 @@ struct Renewal{E, S <: Sampleable, A} <:
     initialisation_prior::S
     recurrent_step::A
 
-    function Renewal(data::EpiData; initialisation_prior = Normal())
-        rev_gen_int = reverse(data.gen_int)
-        recurrent_step = ConstantRenewalStep(rev_gen_int)
-        return Renewal(data, initialisation_prior, recurrent_step)
-    end
-
-    function Renewal(; data::EpiData, initialisation_prior = Normal())
-        rev_gen_int = reverse(data.gen_int)
-        recurrent_step = ConstantRenewalStep(rev_gen_int)
-        return Renewal(data, initialisation_prior, recurrent_step)
-    end
-
     function Renewal(data::E,
             initialisation_prior::S,
             recurrent_step::A) where {
@@ -106,6 +94,18 @@ struct Renewal{E, S <: Sampleable, A} <:
     end
 end
 
+function Renewal(data::EpiData; initialisation_prior = Normal())
+    rev_gen_int = reverse(data.gen_int)
+    recurrent_step = ConstantRenewalStep(rev_gen_int)
+    return Renewal(data, initialisation_prior, recurrent_step)
+end
+
+function Renewal(; data::EpiData, initialisation_prior = Normal())
+    rev_gen_int = reverse(data.gen_int)
+    recurrent_step = ConstantRenewalStep(rev_gen_int)
+    return Renewal(data, initialisation_prior, recurrent_step)
+end
+
 """
 Create the initial state of the `Renewal` model.
 

EpiAware/src/EpiLatentModels/manipulators/broadcast/LatentModel.jl

@@ -27,12 +27,6 @@ struct BroadcastLatentModel{
     "The broadcast rule to be applied."
     broadcast_rule::B
 
-    function BroadcastLatentModel(model::M; period::Integer,
-            broadcast_rule::B) where {
-            M <: AbstractTuringLatentModel, B <: AbstractBroadcastRule}
-        BroadcastLatentModel(model, period, broadcast_rule)
-    end
-
     function BroadcastLatentModel(model::M, period::Integer,
             broadcast_rule::B) where {
             M <: AbstractTuringLatentModel, B <: AbstractBroadcastRule}
@@ -42,6 +36,12 @@ struct BroadcastLatentModel{
     end
 end
 
+function BroadcastLatentModel(model::M; period::Integer,
+        broadcast_rule::B) where {
+        M <: AbstractTuringLatentModel, B <: AbstractBroadcastRule}
+    BroadcastLatentModel(model, period, broadcast_rule)
+end
+
 @doc raw"
 Generates latent periods using the specified `model` and `n` number of samples.