Add scaling_type serialisation

frame-consulting · Jun 24, 2024 · 93dc264 · 93dc264
1 parent 91ed77d
commit 93dc264
Show file tree

Hide file tree

Showing 6 changed files with 120 additions and 3 deletions.
diff --git a/src/models/rates/GaussianHjmModel.jl b/src/models/rates/GaussianHjmModel.jl
@@ -39,6 +39,7 @@ end
         factor_alias::AbstractVector
         correlation_holder::Union{CorrelationHolder, Nothing}
         quanto_model::Union{AssetModel, Nothing}
+        scaling_type::BenchmarkTimesScaling
     end
 
 A Gaussian HJM model with piece-wise constant benchmark rate volatility and
@@ -54,6 +55,7 @@ struct GaussianHjmModel <: SeparableHjmModel
     factor_alias::AbstractVector
     correlation_holder::Union{CorrelationHolder, Nothing}
     quanto_model::Union{AssetModel, Nothing}
+    scaling_type::BenchmarkTimesScaling
 end
 
 """
@@ -76,7 +78,7 @@ function gaussian_hjm_model(
     sigma_f::BackwardFlatVolatility,
     correlation_holder::Union{CorrelationHolder, Nothing},
     quanto_model::Union{AssetModel, Nothing},
-    scaling_type::BenchmarkTimesScaling = ForwardRateScaling,
+    scaling_type::BenchmarkTimesScaling = _default_benchmark_time_scaling,
     )
     # Check inputs
     @assert length(delta()) > 0
@@ -117,7 +119,7 @@ function gaussian_hjm_model(
         y0 = y[:,:,k]
     end
     return GaussianHjmModel(alias, delta, chi, sigma_T, y,
-        state_alias, factor_alias, correlation_holder, quanto_model)
+        state_alias, factor_alias, correlation_holder, quanto_model, scaling_type)
 end
 
 """

diff --git a/src/models/rates/SeparableHjmModel.jl b/src/models/rates/SeparableHjmModel.jl
@@ -118,6 +118,14 @@ https://ssrn.com/abstract=4638188 for details.
 )
 
 
+"""
+    const _default_benchmark_time_scaling = ForwardRateScaling
+
+We specify a default value to allow for backward compatibility in
+seialisation.
+"""
+const _default_benchmark_time_scaling = ForwardRateScaling
+
 """
     benchmark_times_scaling_forward_rate(chi::AbstractVector, delta::AbstractVector)
 

diff --git a/src/serialisation/Models.jl b/src/serialisation/Models.jl
@@ -7,6 +7,20 @@ Serialise model Context.
 serialise(o::Context) = serialise_struct(o)
 
 
+"""
+    serialise(o::BenchmarkTimesScaling)
+
+Serialise a BenchmarkTimesScaling enumeration object.
+"""
+function serialise(o::BenchmarkTimesScaling)
+    d = OrderedDict{String, Any}()
+    d["typename"]    = string(typeof(o))
+    d["constructor"] = "BenchmarkTimesScaling"
+    d["enumeration"] = Integer(o)
+    return d
+end
+
+
 """
     serialise(o::GaussianHjmModel)
 
@@ -30,6 +44,9 @@ function serialise(o::GaussianHjmModel)
     else
         d["quanto_model"] = serialise_key(o.quanto_model.alias)
     end
+    if o.scaling_type != _default_benchmark_time_scaling
+        d["scaling_type"] = serialise(o.scaling_type)
+    end
     return d
 end
 

diff --git a/src/serialisation/RebuildModels.jl b/src/serialisation/RebuildModels.jl
@@ -29,6 +29,7 @@ function model_parameters(m::GaussianHjmModel)
     else
         d["quanto_model"] = m.quanto_model.alias
     end
+    d["scaling_type"] = m.scaling_type
     # we add another dict layer to allow combining models and ts.
     return Dict(m.alias => d)
 end
@@ -143,6 +144,7 @@ function build_model(
             m_dict["sigma_f"],
             ch,
             quanto_model,
+            m_dict["scaling_type"],
         )
     end
     if m_dict["type"] == LognormalAssetModel

diff --git a/test/unittests/serialisation/models.jl b/test/unittests/serialisation/models.jl
@@ -133,6 +133,94 @@ using YAML
         end
     end
 
+    @testset "GaussianHjmModel with BenchmarkTimesScaling (de-)serialisation." begin
+        models = setup_models(ch_full)
+        ref_dict = Dict(
+            "EUR-USD" => models[2],
+            "One" => ch_one,
+            "Full" => ch_full
+        )
+        #
+        d = OrderedDict(
+            "typename" => "DiffFusion.GaussianHjmModel",
+            "constructor" => "gaussian_hjm_model",
+            "alias" => "USD",
+            "delta" => OrderedDict{String, Any}(
+                "typename" => "DiffFusion.BackwardFlatParameter",
+                "constructor" => "BackwardFlatParameter",
+                "alias" => "",
+                "times" => [0.0],
+                "values" => [[1.0], [7.0], [15.0]],
+                ),
+            "chi" => OrderedDict{String, Any}(
+                "typename" => "DiffFusion.BackwardFlatParameter",
+                "constructor" => "BackwardFlatParameter",
+                "alias" => "",
+                "times" => [0.0],
+                "values" => [[0.01], [0.1], [0.3]],
+                ),
+            "sigma_f" => OrderedDict{String, Any}(
+                "typename" => "DiffFusion.BackwardFlatVolatility",
+                "constructor" => "BackwardFlatVolatility",
+                "alias" => "USD",
+                "times" => [0.0],
+                "values" => [[0.005], [0.006], [0.007]]
+                ),
+            "correlation_holder" => "{Full}",
+            "quanto_model" => "nothing",
+            "scaling_type" => OrderedDict{String, Any}(
+                "typename"    => "DiffFusion.BenchmarkTimesScaling",
+                "constructor" => "BenchmarkTimesScaling",
+                "enumeration" => 1,
+            ),
+        )
+        o = DiffFusion.deserialise(d, ref_dict)
+        s = DiffFusion.serialise(o)
+        if VERSION >= v"1.7" # equality tests fail with Julia 1.6
+            @test s == d
+        end
+        #
+        d = OrderedDict(
+            "typename" => "DiffFusion.GaussianHjmModel",
+            "constructor" => "gaussian_hjm_model",
+            "alias" => "EUR",
+            "delta" => OrderedDict{String, Any}(
+                "typename" => "DiffFusion.BackwardFlatParameter",
+                "constructor" => "BackwardFlatParameter",
+                "alias" => "",
+                "times" => [0.0],
+                "values" => [[1.0], [10.0]],
+                ),
+            "chi" => OrderedDict{String, Any}(
+                "typename" => "DiffFusion.BackwardFlatParameter",
+                "constructor" => "BackwardFlatParameter",
+                "alias" => "",
+                "times" => [0.0],
+                "values" => [[0.01], [0.15]],
+                ),
+            "sigma_f" => OrderedDict{String, Any}(
+                "typename" => "DiffFusion.BackwardFlatVolatility",
+                "constructor" => "BackwardFlatVolatility",
+                "alias" => "EUR",
+                "times" => [0.0],
+                "values" => [[0.008], [0.009000000000000001]],
+                ),
+            "correlation_holder" => "{Full}",
+            "quanto_model" => "{EUR-USD}",
+            "quanto_model" => "nothing",
+            "scaling_type" => OrderedDict{String, Any}(
+                "typename"    => "DiffFusion.BenchmarkTimesScaling",
+                "constructor" => "BenchmarkTimesScaling",
+                "enumeration" => 2,
+            ),
+        )
+        o = DiffFusion.deserialise(d, ref_dict)
+        s = DiffFusion.serialise(o)
+        if VERSION >= v"1.7" # equality tests fail with Julia 1.6
+            @test s == d
+        end
+    end
+
     @testset "LognormalAssetModel (de-)serialisation." begin
         models = setup_models(ch_full)
         ref_dict = Dict(

diff --git a/test/unittests/serialisation/rebuild_models.jl b/test/unittests/serialisation/rebuild_models.jl
@@ -61,7 +61,7 @@ using Test
         @test length(d) == 1
         @test haskey(d, DiffFusion.alias(usd_model))
         d = d[DiffFusion.alias(usd_model)]
-        for (a, b) in zip(keys(d), ["correlation_holder", "quanto_model", "chi", "sigma_f", "type", "delta", "alias"])
+        for (a, b) in zip(keys(d), ["correlation_holder", "quanto_model", "chi", "sigma_f", "scaling_type", "type", "delta", "alias"])
             @test a == b
         end
         #