Updates to get dense working

NDTensors/src/abstractarray/fill.jl

@@ -2,16 +2,18 @@ using .TypeParameterAccessors: set_ndims, specify_default_parameters, unwrap_arr
 
 ## TODO I modified this to accept any type and just match the output to the number of dims.
 # for example generic_zeros(Vector, 2,3) = Matrix{Float64}[0,0,0;0,0,0;]
-function generic_randn(
-  arraytype::Type{<:AbstractArray}, dims...; rng=Random.default_rng()
-)
-  arraytype_specified = set_ndims(specify_default_parameters(unwrap_array_type(arraytype)), length(dims));
+function generic_randn(arraytype::Type{<:AbstractArray}, dims...; rng=Random.default_rng())
+  arraytype_specified = set_ndims(
+    specify_default_parameters(unwrap_array_type(arraytype)), length(dims)
+  )
   data = similar(arraytype_specified, dims)
   return randn!(rng, data)
 end
 
 function generic_zeros(arraytype::Type{<:AbstractArray}, dims...)
-  arraytype_specified = set_ndims(specify_default_parameters(unwrap_array_type(arraytype)), length(dims));
+  arraytype_specified = set_ndims(
+    specify_default_parameters(unwrap_array_type(arraytype)), length(dims)
+  )
   ElT = eltype(arraytype_specified)
   return fill!(similar(arraytype_specified, dims), zero(ElT))
 end

NDTensors/src/abstractarray/ndims.jl

@@ -1,5 +1,6 @@
 ## NDTensors.ndims (not imported from Base)
-using .TypeParameterAccessors: TypeParameterAccessors, Self, type_parameter, set_ndims, set_parameter
+using .TypeParameterAccessors:
+  TypeParameterAccessors, Self, type_parameter, set_ndims, set_parameter
 
 ndims((array::AbstractArray)) = ndims(typeof(array))
 ndims(arraytype::Type{<:AbstractArray}) = type_parameter(arraytype, Base.ndims)

NDTensors/src/abstractarray/similar.jl

@@ -1,5 +1,6 @@
 ## todo working on this still
-using .TypeParameterAccessors: IsWrappedArray, set_eltype, specify_default_parameters, unwrap_array_type
+using .TypeParameterAccessors:
+  IsWrappedArray, set_eltype, specify_default_parameters, unwrap_array_type
 
 ## Custom `NDTensors.similar` implementation.
 ## More extensive than `Base.similar`.
@@ -107,9 +108,7 @@ function similartype(arraytype::Type{<:AbstractArray}, eltype::Type, dims::Tuple
 end
 
 ## Set eltype captures WrappedArray types as long as `position(::Type, ::typeof(parenttype))` is defined
-function similartype(
-  arraytype::Type{<:AbstractArray}, eltype::Type
-)
+function similartype(arraytype::Type{<:AbstractArray}, eltype::Type)
   return set_eltype(arraytype, eltype)
 end
 

NDTensors/src/dense/typeparameteraccessors.jl

@@ -1,5 +1,11 @@
 using .TypeParameterAccessors:
-  TypeParameterAccessors, default_type_parameter, default_type_parameters, parenttype, position, unwrap_array_type, set_parenttype
+  TypeParameterAccessors,
+  default_type_parameter,
+  default_type_parameters,
+  parenttype,
+  position,
+  unwrap_array_type,
+  set_parenttype
 
 ## Dense
 datatype(storetype::Type{<:Dense}) = parenttype(storetype)
@@ -29,4 +35,16 @@ function TypeParameterAccessors.position(::Type{<:Dense}, ::typeof(parenttype))
   return TypeParameterAccessors.Position(2)
 end
 
-TypeParameterAccessors.default_type_parameters(::Type{<:Dense}) = (default_type_parameter(Vector, eltype), Vector{default_type_parameter(Vector, eltype)})
+function TypeParameterAccessors.default_type_parameters(::Type{<:Dense})
+  return (
+    default_type_parameter(Vector, eltype), Vector{default_type_parameter(Vector, eltype)}
+  )
+end
+
+function TypeParameterAccessors.position(::Type{<:DenseTensor}, ::typeof(Base.ndims))
+  return TypeParameterAccessors.Position(2)
+end
+
+function TypeParameterAccessors.set_ndims(type::Type{<:DenseTensor}, N)
+  return set_parameter(type, Base.ndims, N)
+end

NDTensors/src/diag/diagtensor.jl

@@ -1,4 +1,5 @@
 using .DiagonalArrays: diaglength
+using .TypeParameterAccessors: unwrap_array_type
 
 const DiagTensor{ElT,N,StoreT,IndsT} = Tensor{ElT,N,StoreT,IndsT} where {StoreT<:Diag}
 const NonuniformDiagTensor{ElT,N,StoreT,IndsT} =
@@ -109,7 +110,7 @@ end
 
 # convert to Dense
 function dense(T::DiagTensor)
-  return dense(unwrap_type(T), T)
+  return dense(unwrap_array_type(T), T)
 end
 
 # CPU version
@@ -124,7 +125,7 @@ end
 # GPU version
 function dense(::Type{<:AbstractArray}, T::DiagTensor)
   D_cpu = dense(Array, cpu(T))
-  return adapt(unwrap_type(T), D_cpu)
+  return adapt(unwrap_array_type(T), D_cpu)
 end
 
 # UniformDiag version

NDTensors/src/tensorstorage/default_storage.jl

@@ -1,11 +1,12 @@
+using .TypeParameterAccessors: specify_default_parameters
 ## This is a fil which specifies the default storage type provided some set of parameters
 ## The parameters are the element type and storage type
 default_datatype(eltype::Type=default_eltype()) = Vector{eltype}
 default_eltype() = Float64
 
 ## TODO use multiple dispace to make this pick between dense and blocksparse
 function default_storagetype(datatype::Type{<:AbstractArray}, inds::Tuple)
-  datatype = specify_parameters(datatype)
+  datatype = specify_default_parameters(datatype)
   return Dense{eltype(datatype),datatype}
 end
 

NDTensors/src/tensorstorage/set_types.jl

@@ -1,6 +1,5 @@
-using .SetParameters:
-  SetParameters, Position, get_parameters, specify_parameters, unspecify_parameters
-function SetParameters.set_ndims(arraytype::Type{<:TensorStorage}, ndims::Int)
+using .TypeParameterAccessors: TypeParameterAccessors
+function TypeParameterAccessors.set_ndims(arraytype::Type{<:TensorStorage}, ndims::Int)
   # TODO: Change to this once `TensorStorage` types support wrapping
   # non-AbstractVector types.
   # return set_datatype(arraytype, set_ndims(datatype(arraytype), ndims))