Repository navigation
Drop the BlockSparseArrays dependency and adopt TensorAlgebra v0.10 - #180
Merged
Merged
Conversation
GradedArrays implements the block-sparse interface on its own graded array and axis types, so it no longer needs to depend on BlockSparseArrays. The interface functions it was extending are now declared and owned by GradedArrays in `src/blocksparseinterface.jl`, reached downstream by qualified import, and `blockstoredlength` is computed directly on top of the SparseArraysBase block storage. Display no longer routes through a `BlockSparseArray`. `print_array` renders through a BlockArrays block array whose unstored blocks are FillArrays `Zeros`, reproducing the same block grid and dotted unstored entries. Matricization is unchanged, served by TensorAlgebra's BlockArrays extension.
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #180 +/- ##
==========================================
- Coverage 83.02% 83.00% -0.02%
==========================================
Files 23 23
Lines 1832 1830 -2
==========================================
- Hits 1521 1519 -2
Misses 311 311
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
mtfishman
changed the title
mtfishman
marked this pull request as draft
June 23, 2026 20:20
Bump compat to TensorAlgebra 0.10 and SparseArraysBase 0.10 and adapt to their interfaces: - Use TensorAlgebra's `zero!` and `scale!` directly, dropping the FunctionImplementations dependency and the local generic `scale!` fallback. - Express the fermionic contraction twist through TensorAlgebra v0.10's per-position fusion styles. `default_contract_algorithm` puts a `TwistedSectorFusion` on the right factor, whose `matricizeop` inserts the twist, replacing the `contractopadd!` override. - Permute axis tuples with `blockpermute` instead of the removed tuple-by-bipermutation `getindex`. - Drop the NamedDimsArrays extension. NamedDimsArrays is deprecated, and ITensorBase already forwards `conj` through its named ranges, which covers the only live use because `conj` on a graded axis is defined as `dual`.
mtfishman
changed the title
mtfishman
marked this pull request as ready for review
June 23, 2026 21:44
mtfishman
enabled auto-merge (squash)
June 23, 2026 21:44
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
GradedArrays implements the block-sparse interface on its own graded array and axis types, so it no longer depends on BlockSparseArrays. The interface functions it was extending are now declared and owned by GradedArrays in
src/blocksparseinterface.jl, reached downstream by qualified import, andblockstoredlengthis computed directly on the SparseArraysBase block storage. Display no longer routes through aBlockSparseArray:print_arrayrenders through a BlockArrays block array whose unstored blocks are FillArraysZeros, reproducing the same block grid and dotted unstored entries.This release also moves the package onto TensorAlgebra v0.10 and SparseArraysBase v0.10.
scale!andzero!now come from TensorAlgebra, dropping the FunctionImplementations dependency, and the fermionic contraction twist is expressed with TensorAlgebra v0.10's per-position fusion styles in place of the oldcontractopadd!override. The deprecated NamedDimsArrays extension is removed, now that ITensorBase plumbsconjthrough named axes generically.This is a breaking release (
0.10.0): downstream code that reached the block-sparse interface functions through BlockSparseArrays now needs to reach them through GradedArrays instead.