fix: handle empty row or column colors in mixed mode sparse Jacobian

DifferentiationInterface/ext/DifferentiationInterfaceSparseMatrixColoringsExt/jacobian_mixed.jl

@@ -6,7 +6,10 @@ struct SMCMixedModeSparseJacobianPrep{
     BSr<:DI.BatchSizeSettings,
     P<:AbstractMatrix,
     C<:AbstractColoringResult{:nonsymmetric,:bidirectional},
-    M<:AbstractMatrix{<:Number},
+    Mf<:AbstractMatrix{<:Number},
+    Mr<:AbstractMatrix{<:Number},
+    Sfp<:NTuple,
+    Srp<:NTuple,
     Sf<:Vector{<:NTuple},
     Sr<:Vector{<:NTuple},
     Rf<:Vector{<:NTuple},
@@ -19,8 +22,10 @@ struct SMCMixedModeSparseJacobianPrep{
     batch_size_settings_reverse::BSr
     sparsity::P
     coloring_result::C
-    compressed_matrix_forward::M
-    compressed_matrix_reverse::M
+    compressed_matrix_forward::Mf
+    compressed_matrix_reverse::Mr
+    batched_seed_forward_prep::Sfp
+    batched_seed_reverse_prep::Srp
     batched_seeds_forward::Sf
     batched_seeds_reverse::Sr
     batched_results_forward::Rf
@@ -111,12 +116,24 @@ function _prepare_mixed_sparse_jacobian_aux_aux(
     groups_forward = column_groups(coloring_result)
     groups_reverse = row_groups(coloring_result)
 
+    seed_forward_prep = DI.multibasis(x, eachindex(x))
+    seed_reverse_prep = DI.multibasis(y, eachindex(y))
     seeds_forward = [DI.multibasis(x, eachindex(x)[group]) for group in groups_forward]
     seeds_reverse = [DI.multibasis(y, eachindex(y)[group]) for group in groups_reverse]
 
-    compressed_matrix_forward = stack(_ -> vec(similar(y)), groups_forward; dims=2)
-    compressed_matrix_reverse = stack(_ -> vec(similar(x)), groups_reverse; dims=1)
+    compressed_matrix_forward = if isempty(groups_forward)
+        similar(vec(y), length(y), 0)
+    else
+        stack(_ -> vec(similar(y)), groups_forward; dims=2)
+    end
+    compressed_matrix_reverse = if isempty(groups_reverse)
+        similar(vec(x), 0, length(x))
+    else
+        stack(_ -> vec(similar(x)), groups_reverse; dims=1)
+    end
 
+    batched_seed_forward_prep = ntuple(b -> copy(seed_forward_prep), Val(Bf))
+    batched_seed_reverse_prep = ntuple(b -> copy(seed_reverse_prep), Val(Br))
     batched_seeds_forward = [
         ntuple(b -> seeds_forward[1 + ((a - 1) * Bf + (b - 1)) % Nf], Val(Bf)) for a in 1:Af
     ]
@@ -136,15 +153,15 @@ function _prepare_mixed_sparse_jacobian_aux_aux(
         f_or_f!y...,
         DI.forward_backend(dense_backend),
         x,
-        batched_seeds_forward[1],
+        batched_seed_forward_prep,
         contexts...;
     )
     pullback_prep = DI.prepare_pullback_nokwarg(
         strict,
         f_or_f!y...,
         DI.reverse_backend(dense_backend),
         x,
-        batched_seeds_reverse[1],
+        batched_seed_reverse_prep,
         contexts...;
     )
 
@@ -156,6 +173,8 @@ function _prepare_mixed_sparse_jacobian_aux_aux(
         coloring_result,
         compressed_matrix_forward,
         compressed_matrix_reverse,
+        batched_seed_forward_prep,
+        batched_seed_reverse_prep,
         batched_seeds_forward,
         batched_seeds_reverse,
         batched_results_forward,
@@ -183,6 +202,8 @@ function _sparse_jacobian_aux!(
         coloring_result,
         compressed_matrix_forward,
         compressed_matrix_reverse,
+        batched_seed_forward_prep,
+        batched_seed_reverse_prep,
         batched_seeds_forward,
         batched_seeds_reverse,
         batched_results_forward,
@@ -200,15 +221,15 @@ function _sparse_jacobian_aux!(
         pushforward_prep,
         DI.forward_backend(dense_backend),
         x,
-        batched_seeds_forward[1],
+        batched_seed_forward_prep,
         contexts...,
     )
     pullback_prep_same = DI.prepare_pullback_same_point(
         f_or_f!y...,
         pullback_prep,
         DI.reverse_backend(dense_backend),
         x,
-        batched_seeds_reverse[1],
+        batched_seed_reverse_prep,
         contexts...,
     )
 

DifferentiationInterface/test/Core/SimpleFiniteDiff/test.jl

@@ -128,6 +128,11 @@ end
         @test only(row_groups(jac_rev_prep)) == 1:10
         @test only(column_groups(hess_prep)) == 1:10
     end
+
+    @testset "Empty colors for mixed mode" begin # issue 857
+        backend = MyAutoSparse(MixedMode(adaptive_backends[1], adaptive_backends[2]))
+        @test jacobian(copyto!, zeros(10), backend, ones(10)) isa AbstractMatrix
+    end
 end
 
 @testset "Misc" begin

DifferentiationInterface/test/testutils.jl

@@ -17,7 +17,7 @@ function MyAutoSparse(backend::AbstractADType)
     return AutoSparse(
         backend;
         sparsity_detector=TracerSparsityDetector(),
-        coloring_algorithm=GreedyColoringAlgorithm(),
+        coloring_algorithm=GreedyColoringAlgorithm(; postprocessing=true),
     )
 end