Convert scalar data to 0-dimensional arrays (#36)

* Add as_array

* Update namedtuple_of_arrays

* Test with non-array inputs

* Generate 0-dimensional data

* Ensure all variables are arrays.

* Increment version numbers

* Don't continue if errored

* Handle non-arrays in namedtuple_to_dataset

* Document conversion to a 0-dimensional array

* Bump DimensionalData compat

* Use Pkg API instead of REPL mode

* Use Julia shell

* Add missing close paren

.github/workflows/Subpackage_CI.yml

@@ -50,11 +50,13 @@ jobs:
         shell: julia --project=monorepo {0}
         run: |
           using Pkg;
-          pkg"dev . ./lib/${{ matrix.subpackage }}"
+          Pkg.develop([Pkg.PackageSpec(; path="."),
+                       Pkg.PackageSpec(; path="./lib/${{ matrix.subpackage }}")])
       - name: Run the tests
-        continue-on-error: true
-        run: >
-          julia --color=yes --project=monorepo -e 'using Pkg; Pkg.test("${{ matrix.subpackage }}", coverage=true)'
+        run: |
+          using Pkg
+          Pkg.test("${{ matrix.subpackage }}"; coverage=true)
+        shell: julia --color=yes --project=monorepo {0}
       - uses: julia-actions/julia-processcoverage@v1
         with:
           directories: lib/${{ matrix.subpackage }}/src

Project.toml

@@ -1,7 +1,7 @@
 name = "InferenceObjects"
 uuid = "b5cf5a8d-e756-4ee3-b014-01d49d192c00"
 authors = ["Seth Axen <[email protected]> and contributors"]
-version = "0.2.5"
+version = "0.2.6"
 
 [deps]
 Compat = "34da2185-b29b-5c13-b0c7-acf172513d20"
@@ -10,7 +10,7 @@ DimensionalData = "0703355e-b756-11e9-17c0-8b28908087d0"
 
 [compat]
 Compat = "3.46.0, 4.2.0"
-DimensionalData = "0.20, 0.21, 0.22, 0.23"
+DimensionalData = "0.23.1"
 OffsetArrays = "1"
 OrderedCollections = "1"
 julia = "1.6"

lib/InferenceObjectsNetCDF/Project.toml

@@ -1,7 +1,7 @@
 name = "InferenceObjectsNetCDF"
 uuid = "7cb6d088-77df-42c3-8f05-5ca8d42599d1"
 authors = ["Seth Axen <[email protected]>"]
-version = "0.2.3"
+version = "0.2.4"
 
 [deps]
 DimensionalData = "0703355e-b756-11e9-17c0-8b28908087d0"
@@ -10,7 +10,7 @@ NCDatasets = "85f8d34a-cbdd-5861-8df4-14fed0d494ab"
 Reexport = "189a3867-3050-52da-a836-e630ba90ab69"
 
 [compat]
-DimensionalData = "0.20, 0.21, 0.22, 0.23"
+DimensionalData = "0.23.1"
 InferenceObjects = "0.2"
 NCDatasets = "0.12"
 Reexport = "1"

lib/InferenceObjectsNetCDF/src/InferenceObjectsNetCDF.jl

@@ -108,7 +108,7 @@ end
 
 _var_to_array(var, load_mode) = var
 function _var_to_array(var, load_mode::Val{:eager})
-    arr = Array(var)
+    arr = as_array(Array(var))
     attr = var.attrib
     try
         arr_nomissing = NCDatasets.nomissing(arr)
@@ -187,4 +187,7 @@ function to_netcdf(data, ds::NCDatasets.NCDataset; group::Symbol=:posterior)
     return ds
 end
 
+as_array(x) = fill(x)
+as_array(x::AbstractArray) = x
+
 end

src/dataset.jl

@@ -52,6 +52,8 @@ end
 
 Convert `NamedTuple` mapping variable names to arrays to a [`Dataset`](@ref).
 
+Any non-array values will be converted to a 0-dimensional array.
+
 # Keywords
 
   - `attrs::AbstractDict{<:AbstractString}`: a collection of metadata to attach to the
@@ -79,7 +81,7 @@ function namedtuple_to_dataset(
     default_dims=DEFAULT_SAMPLE_DIMS,
 )
     dim_arrays = map(keys(data)) do var_name
-        var_data = data[var_name]
+        var_data = as_array(data[var_name])
         var_dims = get(dims, var_name, ())
         return array_to_dimarray(var_data, var_name; dims=var_dims, coords, default_dims)
     end

src/from_namedtuple.jl

@@ -54,6 +54,11 @@ whose first dimensions correspond to the dimensions of the containers.
 
   - `InferenceData`: The data with groups corresponding to the provided data
 
+!!! note
+    If a `NamedTuple` is provided for `observed_data`, `constant_data`, or
+    predictions_constant_data`, any non-array values (e.g. integers) are converted to
+    0-dimensional arrays.
+
 # Examples
 
 ```@example

src/utils.jl

@@ -7,6 +7,9 @@ ordered with dimensions of the innermost container first and outermost last.
 recursive_stack(x) = x
 recursive_stack(x::AbstractArray{<:AbstractArray}) = recursive_stack(stack(x))
 
+as_array(x) = fill(x)
+as_array(x::AbstractArray) = x
+
 """
     namedtuple_of_arrays(x::NamedTuple) -> NamedTuple
     namedtuple_of_arrays(x::AbstractArray{NamedTuple}) -> NamedTuple
@@ -25,7 +28,7 @@ ntarray = InferenceObjects.namedtuple_of_arrays(data);
 ```
 """
 function namedtuple_of_arrays end
-namedtuple_of_arrays(x::NamedTuple) = map(recursive_stack, x)
+namedtuple_of_arrays(x::NamedTuple) = map(as_array ∘ recursive_stack, x)
 namedtuple_of_arrays(x::AbstractArray) = namedtuple_of_arrays(namedtuple_of_arrays.(x))
 function namedtuple_of_arrays(x::AbstractArray{<:NamedTuple{K}}) where {K}
     return NamedTuple{K}(recursive_stack(getproperty.(x, k)) for k in K)

test/dataset.jl

@@ -152,5 +152,12 @@ using InferenceObjects, DimensionalData, Test
         @test metadata["inference_library"] == "MyLib"
         @test !haskey(metadata, "inference_library_version")
         @test metadata["mykey"] == 5
+
+        ds2 = namedtuple_to_dataset((x=1, y=randn(10)); default_dims=())
+        @test ds2 isa Dataset
+        @test ds2.x isa DimensionalData.DimArray{<:Any,0}
+        @test DimensionalData.dims(ds2.x) == ()
+        @test ds2.y isa DimensionalData.DimArray{<:Any,1}
+        @test DimensionalData.dims(ds2.y) == (Dim{:y_dim_1}(1:10),)
     end
 end

test/from_dict.jl

@@ -53,18 +53,22 @@ using InferenceObjects, OrderedCollections, Test
         library = "MyLib"
         dims = (; w=[:wx])
         coords = (; wx=1:2)
-        idata1 = from_dict(dict; group => Dict(:w => [1.0, 2.0]), dims, coords, library)
+        idata1 = from_dict(
+            dict; group => Dict(:w => [1.0, 2.0], :v => 2.5), dims, coords, library
+        )
         test_idata_group_correct(idata1, :posterior, keys(sizes); library, dims, coords)
         test_idata_group_correct(
-            idata1, group, (:w,); library, dims, coords, default_dims=()
+            idata1, group, (:w, :v); library, dims, coords, default_dims=()
         )
 
         # ensure that dims are matched to named tuple keys
         # https://github.com/arviz-devs/ArviZ.jl/issues/96
-        idata2 = from_dict(dict; group => Dict(:w => [1.0, 2.0]), dims, coords, library)
+        idata2 = from_dict(
+            dict; group => Dict(:w => [1.0, 2.0], :v => 2.5), dims, coords, library
+        )
         test_idata_group_correct(idata2, :posterior, keys(sizes); library, dims, coords)
         test_idata_group_correct(
-            idata2, group, (:w,); library, dims, coords, default_dims=()
+            idata2, group, (:w, :v); library, dims, coords, default_dims=()
         )
     end
 end

test/from_namedtuple.jl

@@ -55,18 +55,18 @@ using InferenceObjects, Test
         library = "MyLib"
         dims = (; w=[:wx])
         coords = (; wx=1:2)
-        idata1 = from_namedtuple(nt; group => (w=[1.0, 2.0],), dims, coords, library)
+        idata1 = from_namedtuple(nt; group => (w=[1.0, 2.0], v=2.5), dims, coords, library)
         test_idata_group_correct(idata1, :posterior, keys(sizes); library, dims, coords)
         test_idata_group_correct(
-            idata1, group, (:w,); library, dims, coords, default_dims=()
+            idata1, group, (:w, :v); library, dims, coords, default_dims=()
         )
 
         # ensure that dims are matched to named tuple keys
         # https://github.com/arviz-devs/ArviZ.jl/issues/96
-        idata2 = from_namedtuple(nt; group => (w=[1.0, 2.0],), dims, coords, library)
+        idata2 = from_namedtuple(nt; group => (w=[1.0, 2.0], v=2.5), dims, coords, library)
         test_idata_group_correct(idata2, :posterior, keys(sizes); library, dims, coords)
         test_idata_group_correct(
-            idata2, group, (:w,); library, dims, coords, default_dims=()
+            idata2, group, (:w, :v); library, dims, coords, default_dims=()
         )
     end
 end

test/test_helpers.jl

@@ -23,13 +23,13 @@ random_dataset(args...) = Dataset(random_dim_stack(args...))
 
 function random_data()
     var_names = (:a, :b)
-    data_names = (:y,)
+    data_names = (:y, :v)
     stats_names = (:diverging, :energy, :n_steps)
     stats_eltypes = (diverging=Bool, n_steps=Int)
     coords = (
         chain=1:4, draw=1:100, shared=["s1", "s2", "s3"], dima=1:4, dimb=2:6, dimy=1:5
     )
-    dims = (a=(:shared, :dima), b=(:shared, :dimb), y=(:shared, :dimy))
+    dims = (a=(:shared, :dima), b=(:shared, :dimb), y=(:shared, :dimy), v=())
     metadata = Dict{String,Any}("inference_library" => "PPL")
     posterior = random_dataset(var_names, dims, coords, metadata, (;))
     posterior_predictive = random_dataset(data_names, dims, coords, metadata, (;))

test/utils.jl

@@ -11,8 +11,20 @@ module TestSubModule end
         @test InferenceObjects.recursive_stack(1:5) == 1:5
     end
 
+    @testset "as_array" begin
+        @test InferenceObjects.as_array(3) == fill(3)
+        @test InferenceObjects.as_array(2.5) == fill(2.5)
+        @test InferenceObjects.as_array("var") == fill("var")
+        x = randn(3)
+        @test InferenceObjects.as_array(x) == x
+        x = randn(2, 3)
+        @test InferenceObjects.as_array(x) == x
+        x = randn(2, 3, 4)
+        @test InferenceObjects.as_array(x) == x
+    end
+
     @testset "namedtuple_of_arrays" begin
-        @test InferenceObjects.namedtuple_of_arrays((x=3, y=4)) === (x=3, y=4)
+        @test InferenceObjects.namedtuple_of_arrays((x=3, y=4)) == (x=fill(3), y=fill(4))
         @test InferenceObjects.namedtuple_of_arrays([(x=3, y=4), (x=5, y=6)]) ==
             (x=[3, 5], y=[4, 6])
         @test InferenceObjects.namedtuple_of_arrays([