remove numerical pading as an option and hard code

EpiAware/src/EpiObsModels/ObservationErrorModels/NegativeBinomialError.jl

@@ -3,8 +3,8 @@
 The `NegativeBinomialError` struct represents an observation model for negative binomial errors. It is a subtype of `AbstractTuringObservationModel`.
 
 ## Constructors
-- `NegativeBinomialError(; cluster_factor_prior::Distribution = HalfNormal(0.1), pos_shift::AbstractFloat = 1e-6)`: Constructs a `NegativeBinomialError` object with default values for the cluster factor prior and positive shift.
-- `NegativeBinomialError(cluster_factor_prior::Distribution; pos_shift::AbstractFloat = 1e-6)`: Constructs a `NegativeBinomialError` object with a specified cluster factor prior and default value for the positive shift.
+- `NegativeBinomialError(; cluster_factor_prior::Distribution = HalfNormal(0.1))`: Constructs a `NegativeBinomialError` object with default values for the cluster factor prior.
+- `NegativeBinomialError(cluster_factor_prior::Distribution)`: Constructs a `NegativeBinomialError` object with a specified cluster factor prior.
 
 ## Examples
 ```julia
@@ -14,25 +14,10 @@ nb_model = generate_observations(nb, missing, fill(10, 10))
 rand(nb_model)
 ```
 "
-struct NegativeBinomialError{S <: Sampleable, T <: AbstractFloat} <:
-       AbstractTuringObservationErrorModel
+@kwdef struct NegativeBinomialError{S <: Sampleable, T <: AbstractFloat} <:
+              AbstractTuringObservationErrorModel
     "The prior distribution for the cluster factor."
-    cluster_factor_prior::S
-    "The positive shift value."
-    pos_shift::T
-
-    function NegativeBinomialError(;
-            cluster_factor_prior::Distribution = HalfNormal(0.01),
-            pos_shift::AbstractFloat = 1e-6)
-        new{typeof(cluster_factor_prior), typeof(pos_shift)}(
-            cluster_factor_prior, pos_shift)
-    end
-
-    function NegativeBinomialError(cluster_factor_prior::Distribution;
-            pos_shift::AbstractFloat = 1e-6)
-        new{typeof(cluster_factor_prior), typeof(pos_shift)}(
-            cluster_factor_prior, pos_shift)
-    end
+    cluster_factor_prior::S = HalfNormal(0.01)
 end
 
 @doc raw"
@@ -49,6 +34,6 @@ end
 This function generates the observation error model based on the negative binomial error model with a positive shift. It dispatches to the `NegativeBinomialMeanClust` distribution.
 "
 function observation_error(obs_model::NegativeBinomialError, Y_t, sq_cluster_factor)
-    return NegativeBinomialMeanClust(Y_t + obs_model.pos_shift,
+    return NegativeBinomialMeanClust(Y_t,
         sq_cluster_factor)
 end

EpiAware/src/EpiObsModels/ObservationErrorModels/PoissonError.jl

@@ -3,8 +3,7 @@ The `PoissonError` struct represents an observation model for Poisson errors. It
 is a subtype of `AbstractTuringObservationErrorModel`.
 
 ## Constructors
-- `PoissonError(; pos_shift::AbstractFloat = 0.)`: Constructs a `PoissonError`
-object with default values for the cluster factor prior and positive shift.
+- `PoissonError()`: Constructs a `PoissonError` object.
 
 ## Examples
 ```julia
@@ -15,19 +14,12 @@ rand(poi_model)
 ```
 "
 struct PoissonError{T <: AbstractFloat} <: AbstractTuringObservationErrorModel
-    "The positive shift value."
-    pos_shift::T
-
-    function PoissonError(; pos_shift::AbstractFloat = 0.0)
-        @assert pos_shift>=0.0 "The positive shift value must be non-negative."
-        new{typeof(pos_shift)}(pos_shift)
-    end
 end
 
 @doc raw"
 The observation error model for Poisson errors. This function generates the
-observation error model based on the Poisson error model with a positive shift.
+observation error model based on the Poisson error model.
 "
 function observation_error(obs_model::PoissonError, Y_t)
-    return Poisson(Y_t + obs_model.pos_shift)
+    return Poisson(Y_t)
 end

EpiAware/src/EpiObsModels/ObservationErrorModels/methods.jl

@@ -1,5 +1,7 @@
 @doc raw"
-Generates observations from an observation error model. It provides support for missing values in observations (`y_t`), and missing values at the beginning of the expected observations (`Y_t`). It dispatches to the `observation_error` function to generate the observation error distribution which uses priors generated by `generate_observation_error_priors` submodel. For most observation error models specific implementations of `observation_error` and `generate_observation_error_priors` are required but a specific implementation of `generate_observations` is not required.
+Generates observations from an observation error model. It provides support for missing values in observations (`y_t`), and missing values at the beginning of the expected observations (`Y_t`). It also pads the expected observations with a small value (1e-6) to mitigate potential numerical issues.
+
+It dispatches to the `observation_error` function to generate the observation error distribution which uses priors generated by `generate_observation_error_priors` submodel. For most observation error models specific implementations of `observation_error` and `generate_observation_error_priors` are required but a specific implementation of `generate_observations` is not required.
 "
 @model function EpiAwareBase.generate_observations(
         obs_model::AbstractTuringObservationErrorModel,
@@ -13,8 +15,10 @@ Generates observations from an observation error model. It provides support for
         @assert length(y_t)==length(Y_t) "The observation vector and expected observation vector must have the same length."
     end
 
+    pad_Y_t = Y_t + 1e-6
+
     for i in findfirst(!ismissing, Y_t):length(Y_t)
-        y_t[i] ~ observation_error(obs_model, Y_t[i], priors...)
+        y_t[i] ~ observation_error(obs_model, pad_Y_t[i], priors...)
     end
 
     return y_t, priors

EpiAware/test/EpiAwareUtils/generate_epiware.jl

@@ -1,10 +1,8 @@
-
 @testitem "`generate_epiaware` with direct infections and RW latent process runs" begin
     using Distributions, Turing, DynamicPPL
     # Define test inputs
     y_t = missing # Data will be generated from the model
     data = EpiData([0.2, 0.3, 0.5], exp)
-    pos_shift = 1e-6
     time_horizon = 100
 
     #Define the epi_model
@@ -42,7 +40,6 @@ end
     # Define test inputs
     y_t = missing# rand(1:10, 365) # Data will be generated from the model
     data = EpiData([0.2, 0.3, 0.5], exp)
-    pos_shift = 1e-6
 
     #Define the epi_model
     epi_model = ExpGrowthRate(data, Normal())
@@ -56,7 +53,7 @@ end
     #Define the observation model - no delay model
     time_horizon = 5
     obs_model = NegativeBinomialError(
-        truncated(Gamma(5, 0.05 / 5), 1e-3, 1.0); pos_shift
+        truncated(Gamma(5, 0.05 / 5), 1e-3, 1.0)
     )
 
     # Create full epi model and sample from it
@@ -80,7 +77,6 @@ end
     # Define test inputs
     y_t = missing# rand(1:10, 365) # Data will be generated from the model
     data = EpiData([0.2, 0.3, 0.5], exp)
-    pos_shift = 1e-6
 
     #Define the epi_model
     epi_model = Renewal(data, Normal())
@@ -94,8 +90,7 @@ end
     #Define the observation model - no delay model
     time_horizon = 5
     obs_model = NegativeBinomialError(
-        truncated(Gamma(5, 0.05 / 5), 1e-3, 1.0);
-        pos_shift
+        truncated(Gamma(5, 0.05 / 5), 1e-3, 1.0)
     )
 
     # Create full epi model and sample from it

EpiAware/test/EpiObsModels/ObservationErrorModels/NegativeBinomialError.jl

@@ -2,20 +2,19 @@
     using Distributions
     # Test default constructor
     nb = NegativeBinomialError()
+    @test typeof(nb) <: NegativeBinomialError
+    @test typeof(nb) <: TuringObservationErrorModel
     @test all(rand(nb.cluster_factor_prior, 100) .>= 0.0)
     @test isapprox(mean(nb.cluster_factor_prior), 0.01)
-    @test nb.pos_shift ≈ 1e-6
 
     # Test constructor with custom prior
     prior = Gamma(2.0, 1.0)
     nb = NegativeBinomialError(prior)
     @test nb.cluster_factor_prior == prior
-    @test nb.pos_shift ≈ 1e-6
 
-    # Test constructor with custom prior and pos_shift
-    nb = NegativeBinomialError(prior; pos_shift = 1e-3)
+    # Test constructor with custom prior
+    nb = NegativeBinomialError(prior)
     @test nb.cluster_factor_prior == prior
-    @test nb.pos_shift ≈ 1e-3
 end
 
 @testitem "Testing NegativeBinomialError against theoretical properties" begin
@@ -27,7 +26,7 @@ end
     α = 0.2  # Cluster factor (dispersion parameter)
 
     # Define the observation model
-    nb_obs_model = NegativeBinomialError(pos_shift = 0.0)
+    nb_obs_model = NegativeBinomialError()
 
     # Generate observations from the model
     Y_t = fill(μ, n)  # True values

EpiAware/test/EpiObsModels/ObservationErrorModels/PoissonError.jl

@@ -1,14 +1,9 @@
 @testitem "PoissonErrorConstructor" begin
     using Distributions
     # Test default constructor
-    poi = PoissonError()
-    @test poi.pos_shift ≈ zero(Float64)
-    poi_float = PoissonError(; pos_shift = 0.0f0)
-    @test poi_float.pos_shift ≈ zero(Float32)
-
-    # Test constructor with pos_shift
-    poi2 = PoissonError(; pos_shift = 1e-3)
-    @test poi2.pos_shift ≈ 1e-3
+    poi = PoissonError
+    @test typeof(poi) <: PoissonError
+    @test typeof(poi) <: TuringObservationErrorModel
 end
 
 @testitem "Testing PoissonError against theoretical properties" begin
@@ -19,7 +14,7 @@ end
     μ = 10.0  # Mean of the poisson distribution
 
     # Define the observation model
-    poi_obs_model = PoissonError(pos_shift = 0.0)
+    poi_obs_model = PoissonError()
 
     # Generate observations from the model
     Y_t = fill(μ, n)  # True values