v0.9.5

BlockSparseArrays v0.9.5

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• More general truncation and slicing (#159) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])
• Generalize block sparse matricize (#163) (@mtfishman)
• [pre-commit.ci] pre-commit autoupdate (#165) (@pre-commit-ci[bot])
• Blockdiagonal factorizations refactor (#166) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#167) (@pre-commit-ci[bot])
• Use DerivableInterfaces.permuteddims (#168) (@mtfishman)
• More robust blockreshape (#169) (@mtfishman)
• Better check when blockreshape is type unstable (#170) (@mtfishman)
• Clean up extraneous comments (#171) (@mtfishman)
• CompatHelper: bump compat for MatrixAlgebraKit to 0.3, (keep existing compat) (#172) (@github-actions[bot])
• Fix 0-dim broadcasting bug (#174) (@mtfishman)
• Improve BlockUnitRange indexing (#175) (@mtfishman)
• Update to TensorAlgebra 0.4 (#176) (@mtfishman)
• Backport "Fix 0-dim broadcasting bug" to 0.9 (#177) (@mtfishman)

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)
• [BUG] SVD of rectangular block structure (#162)

v0.9.4

BlockSparseArrays v0.9.4

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• More general truncation and slicing (#159) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])
• Generalize block sparse matricize (#163) (@mtfishman)
• [pre-commit.ci] pre-commit autoupdate (#165) (@pre-commit-ci[bot])
• [pre-commit.ci] pre-commit autoupdate (#167) (@pre-commit-ci[bot])
• Use DerivableInterfaces.permuteddims (#168) (@mtfishman)
• More robust blockreshape (#169) (@mtfishman)
• Better check when blockreshape is type unstable (#170) (@mtfishman)
• Clean up extraneous comments (#171) (@mtfishman)

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.9.3

BlockSparseArrays v0.9.3

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• More general truncation and slicing (#159) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])
• Generalize block sparse matricize (#163) (@mtfishman)
• [pre-commit.ci] pre-commit autoupdate (#165) (@pre-commit-ci[bot])
• [pre-commit.ci] pre-commit autoupdate (#167) (@pre-commit-ci[bot])
• Use DerivableInterfaces.permuteddims (#168) (@mtfishman)
• More robust blockreshape (#169) (@mtfishman)
• Better check when blockreshape is type unstable (#170) (@mtfishman)

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.9.2

BlockSparseArrays v0.9.2

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• More general truncation and slicing (#159) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])
• Generalize block sparse matricize (#163) (@mtfishman)
• [pre-commit.ci] pre-commit autoupdate (#165) (@pre-commit-ci[bot])
• [pre-commit.ci] pre-commit autoupdate (#167) (@pre-commit-ci[bot])
• Use DerivableInterfaces.permuteddims (#168) (@mtfishman)
• More robust blockreshape (#169) (@mtfishman)

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.9.1

BlockSparseArrays v0.9.1

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• More general truncation and slicing (#159) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])
• Generalize block sparse matricize (#163) (@mtfishman)
• [pre-commit.ci] pre-commit autoupdate (#165) (@pre-commit-ci[bot])
• [pre-commit.ci] pre-commit autoupdate (#167) (@pre-commit-ci[bot])
• Use DerivableInterfaces.permuteddims (#168) (@mtfishman)

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.9.0

BlockSparseArrays v0.9.0

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• More general truncation and slicing (#159) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])
• Generalize block sparse matricize (#163) (@mtfishman)
• [pre-commit.ci] pre-commit autoupdate (#165) (@pre-commit-ci[bot])
• [pre-commit.ci] pre-commit autoupdate (#167) (@pre-commit-ci[bot])

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.8.1

BlockSparseArrays v0.8.1

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• More general truncation and slicing (#159) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.8.0

BlockSparseArrays v0.8.0

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.6, (keep existing compat) (#160) (@github-actions[bot])

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.7.22

BlockSparseArrays v0.7.22

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)
• Towards more general truncation and slicing (#158) (@mtfishman)

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)

v0.7.21

BlockSparseArrays v0.7.21

Merged pull requests:

• Package setup (#5) (@mtfishman)
• CompatHelper: bump compat for GPUArraysCore to 0.2, (keep existing compat) (#7) (@github-actions[bot])
• Update for latest Derive, SparseArraysBase (#10) (@mtfishman)
• Skeleton update (#11) (@mtfishman)
• Set up package extensions, improve permutedims (#14) (@mtfishman)
• Rename BlockZero to GetUnstoredBlock (#15) (@mtfishman)
• [Enhancement] Add singular value decomposition (#16) (@lkdvos)
• Bump versions (#18) (@mtfishman)
• Use MapBroadcast.jl and DerivableInterfaces.jl (#20) (@mtfishman)
• Improve showing unstored values (#22) (@mtfishman)
• Simplify show methods (#23) (@mtfishman)
• Start adding exports (#25) (@mtfishman)
• Fix setting zero dimensional block (#28) (@mtfishman)
• Fix zero-dimensional conversion (#30) (@mtfishman)
• Fix tests (#31) (@mtfishman)
• Fix broken SVD tests (#33) (@lkdvos)
• Construct BlockSparseArray when slicing with graded unit ranges (#36) (@mtfishman)
• Fix ambiguities with BlockArrays.jl (#39) (@mtfishman)
• Fix a bug in blocktype (#40) (@mtfishman)
• More compact show for block sparse arrays (#42) (@mtfishman)
• CompatHelper: bump compat for TypeParameterAccessors to 0.3, (keep existing compat) (#43) (@github-actions[bot])
• Fix similar with eltype change (#44) (@ogauthe)
• Simplify show implementation (#46) (@lkdvos)
• Update skeleton (#47) (@lkdvos)
• Make similar with graded axes more generic (#48) (@mtfishman)
• Add IntegrationTest workflow (#49) (@mtfishman)
• Update Registrator workflow (#51) (@mtfishman)
• Fix splitdims(::BlockSparseArrays) (#52) (@ogauthe)
• adapt to TensorAlgebra.blockedperm changes (#56) (@ogauthe)
• Fix ambiguity error in similar of graded vector (#58) (@mtfishman)
• Better overloading of similar of block sparse broadcasting expressions (#61) (@mtfishman)
• Delete unused file from BlockSparseArraysTensorAlgebraExt (#62) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.3, (keep existing compat) (#64) (@github-actions[bot])
• Fix generic show for block sparse arrays (#65) (@mtfishman)
• Delete unneeded show helper function (#66) (@mtfishman)
• Update IntegrationTest github action (#67) (@lkdvos)
• Require undef in BlockSparseArray constructors (#68) (@mtfishman)
• Add missing compat entries (#71) (@mtfishman)
• Fix printing higher dimensional block sparse arrays (#74) (@mtfishman)
• CompatHelper: bump compat for SparseArraysBase to 0.4 for package test, (keep existing compat) (#76) (@github-actions[bot])
• Fix vector show (#79) (@mtfishman)
• CompatHelper: bump compat for DiagonalArrays to 0.3, (keep existing compat) (#80) (@github-actions[bot])
• Fix adapting and printing subarrays (#82) (@mtfishman)
• Upgrade to DerivableInterfaces v0.4 (#84) (@mtfishman)
• Update license, acknowledgements, docs logos (#86) (@mtfishman)
• Update logos (#87) (@mtfishman)
• Adapt to TensorProducts.jl (#89) (@ogauthe)
• Update Registrator workflow (#92) (@mtfishman)
• Remove usage of GradedUnitRanges (#93) (@mtfishman)
• More downstream tests (#94) (@mtfishman)
• define TensorAlgebra.matricize (#95) (@ogauthe)
• Fix ambiguities in similar and blockreshape (#99) (@mtfishman)
• Update .github/workflows/FormatPullRequest.yml (#100) (@mtfishman)
• Update .pre-commit-config.yaml (#103) (@mtfishman)
• Automatic JuliaFormatter.jl run (#104) (@github-actions[bot])
• [pre-commit.ci] pre-commit autoupdate (#105) (@pre-commit-ci[bot])
• Delete sub_axis (#106) (@mtfishman)
• [Fix] Stackoverflow in BlockSparseArray constructor with wrong number of axes (#109) (@lkdvos)
• [pre-commit.ci] pre-commit autoupdate (#110) (@pre-commit-ci[bot])
• Add MatrixAlgebraKit svd (#111) (@lkdvos)
• Add truncation functionality for SVD (#113) (@lkdvos)
• Make SVD more general to accommodate graded arrays (#114) (@mtfishman)
• Noncontiguous slicing (#116) (@mtfishman)
• Add support for block sparse QR decomposition (#117) (@mtfishman)
• Define simpler codepath for blockwise map (#118) (@mtfishman)
• Bump to MatrixAlgebraKit v0.2 (#119) (@mtfishman)
• Bump subdirs to v0.6 (#120) (@mtfishman)
• Make SVD a bit more customizable (#121) (@mtfishman)
• Implement left_orth/right_orth (#122) (@mtfishman)
• Implement blockwise LQ (#124) (@mtfishman)
• Implement blockwise polar decomposition (#125) (@mtfishman)
• More customizability (#126) (@mtfishman)
• Better definitions of norm and tr (#127) (@mtfishman)
• Begin implementation of null (#128) (@mtfishman)
• Block sparse eig (#129) (@mtfishman)
• Add support for logical indexing that preserves block sparsity (#131) (@mtfishman)
• Simplified truncated SVD with logical indexing (#132) (@mtfishman)
• Simplify default_algorithm logic based on MatrixAlgebraKit v0.2.2 (#133) (@mtfishman)
• CompatHelper: add new compat entry for StableRNGs at version 1 for package test, (keep existing compat) (#134) (@github-actions[bot])
• Start adding support for matrix functions (#135) (@mtfishman)
• Update for DerivableInterfaces.jl v0.5 (#137) (@mtfishman)
• Update for TypeParameterAccessors.jl v0.4 (#138) (@mtfishman)
• Convenient constructors like blocksparse and blocksparsezeros (#141) (@mtfishman)
• More specific interfaces (#142) (@mtfishman)
• Enable specialization when dividing by scalars (#143) (@mtfishman)
• More generalizations for generic block types (#144) (@mtfishman)
• More general block types in broadcast style (#145) (@mtfishman)
• blocksparsezeros with Float64 default (#146) (@mtfishman)
• Make block sparse SVD more generic (#147) (@mtfishman)
• Fix zero-dim view! (#149) (@mtfishman)
• Start handling abstract block types (#150) (@mtfishman)
• Fix test for matmul with abstract block types (#151) (@mtfishman)
• Make map/broadcast a bit more customizable (#152) (@mtfishman)
• Add KroneckerArrays.jl as a downstream test (#153) (@mtfishman)
• Fix matrix factorizations of block sparse arrays with abstract block type (#154) (@mtfishman)
• Matrix functions with abstract block type (#155) (@mtfishman)
• Generalize BlockIndexVector (#156) (@mtfishman)
• Generalize blockwise slicing (#157) (@mtfishman)

Closed issues:

• [ENHANCEMENT] Blockwise matrix factorizations (#3)
• Simplify zero dimensional BlockIndex constructors (#4)
• Set up package extensions properly (#6)
• [ENHANCEMENT] Remove NDTensors as test dependency (#12)
• [ENHANCEMENT] Redesign and rename BlockZero (#13)
• [BUG] Matrix multiplication with lazy transpose/adjoint fills all blocks (#24)
• [ENHANCEMENT] Define SparseArraysBase.isstored for block sparse arrays (#26)
• [BUG] zerodim BlockSparseArray errors on setindex (#27)
• [BUG] Array conversion fails on zero dim (#29)
• [BUG] Failures in svd caused by diagind(::AbstractMatrix, ::IndexStyle) (#32)
• [BUG] NamedDimsArrays broadcasting broken (#37)
• [BUG] similar calls on block sparse arrays with graded axes that change the element type fail (#38)
• [BUG] Error when printing adjointed BlockSparseVector (#41)
• [BUG] splitdims(::BlockSparseArrays, ::Tuple{AbstractGradedUnitRanges) result is wrong (#50)
• [ENHANCEMENT] splitdims to use views (#53)
• [ENHANCEMENT] splitdims and fusedims to use BlockedTuple (#54)
• [BUG] BlockArraysExtensions breaks TensorAlgebra.contract (#57)
• [BUG] fusedims on BlockSparseArray with gradedrange axes and trailing singleton broken (#59)
• [ENHANCEMENT] Remove BlockSparseArray{Float64}([2, 3], [2, 3]) constructor (#69)
• [ENHANCEMENT] Fix show for AbstractBlockSparseArray with unconvential type parameter orderings (#63)
• [BUG] display of 3D BlockSparseArray errors (#73)
• [BUG] BlockSparseVector displays sparse zero as 0 instead of . (#77)
• [BUG] display(view(::BlockSparseArray)) errors for full view (#78)
• [BUG] fusedims sometimes initializes all blocks (#88)
• [BUG] call(::BlockSparseArrayInterface, ::typeof(similar), ::BlockSparseArray is ambiguous (#98)
• [BUG] Stackoverflow for BlockSparseArray constructor with wrong number of axis. (#108)
• [BUG] Some SVD truncations don't work (#115)
• [BUG] a'a fills in all blocks (#123)
• [BUG] view!(::BlockSparseArray{Float64, 0}) is broken (#148)