Remove the matricize axis-fusion sub-interface#188
Merged
Conversation
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #188 +/- ##
==========================================
+ Coverage 79.52% 82.39% +2.86%
==========================================
Files 22 22
Lines 669 619 -50
==========================================
- Hits 532 510 -22
+ Misses 137 109 -28
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:
|
The dense `matricize` (`ReshapeFusion`) now computes its fused row and column axes inline from the bipartitioned axes, instead of routing through the generic `matricize_axes`/`fused_axis`/`tensor_product_axis`/`trivial_axis` sub-interface. That interface existed only to let fusion compose across array types, threaded with a `Val{:codomain}`/`Val{:domain}` side argument, and it no longer has a live consumer: the graded path was migrated to compute its axes directly in ITensor/GradedArrays.jl#188.
This removes `matricize_axes`, `fused_axis`, `tensor_product_axis`, `trivial_axis`, and `tensor_product_fusionstyle`. `trivialrange` stays, as a separate type-level identity range unrelated to the matricize fold. The removal is breaking: older GradedArrays releases extended the sub-interface, so this goes out as a minor release.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Read the codomain and domain extents straight off `size(a)` and reshape to their products, rather than folding lengths over the bipartitioned axes. `prod` of an empty group is 1, so the scalar codomain or domain case falls out without a guard. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The `test`, `docs`, and `examples` subprojects pin the in-repo `TensorAlgebra` through a path source, so their `[compat]` has to admit the new version. Move it from `0.12` to `0.13`. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`BlockSparseArrays`, `DiagonalArrays`, `FusionTensors`, `KroneckerArrays`, and `NamedDimsArrays` are deprecated and will not track this breaking change, so testing against them no longer adds signal. Keep the live consumers `GradedArrays`, `ITensorBase`, `ITensorNetworksNext`, and `SparseArraysBase`. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mtfishman
force-pushed
the
mf/matricize-inline-axes
branch
from
cf5f492 to
77afc92
Compare
mtfishman
added a commit
to ITensor/GradedArrays.jl
that referenced
this pull request
Jun 29, 2026
Widens the `TensorAlgebra` compat to admit 0.13. GradedArrays computes its matricize axes inline and no longer uses the axis-fusion sub-interface that 0.13 removes (ITensor/TensorAlgebra.jl#188), so it works with both 0.12 and 0.13. Also drops the deprecated `FusionTensors` from the integration test matrix. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mtfishman
added a commit
to ITensor/ITensorBase.jl
that referenced
this pull request
Jun 29, 2026
Widens the `TensorAlgebra` compat to admit 0.13. ITensorBase does not use the axis-fusion sub-interface that 0.13 removes (ITensor/TensorAlgebra.jl#188), so it works with both 0.12 and 0.13. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mtfishman
added a commit
to ITensor/SparseArraysBase.jl
that referenced
this pull request
Jun 29, 2026
## Summary Widens the `TensorAlgebra` compat to admit 0.13. The `SparseArraysBaseTensorAlgebraExt` does not use the axis-fusion sub-interface that 0.13 removes (ITensor/TensorAlgebra.jl#188), so admitting it is a compat-only change. Also drops the deprecated consumers `BlockSparseArrays`, `DiagonalArrays`, and `FusionTensors` from the integration test matrix, keeping `GradedArrays`. --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mtfishman
added a commit
to ITensor/ITensorBase.jl
that referenced
this pull request
Jun 29, 2026
## Summary Widens the `TensorAlgebra` compat to admit 0.13. ITensorBase does not use the axis-fusion sub-interface that 0.13 removes (ITensor/TensorAlgebra.jl#188), so admitting it is a compat-only change. --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
mtfishman
added a commit
to ITensor/GradedArrays.jl
that referenced
this pull request
Jun 29, 2026
## Summary Widens the `TensorAlgebra` compat to admit 0.13. GradedArrays computes its matricize axes inline and does not use the axis-fusion sub-interface that 0.13 removes (ITensor/TensorAlgebra.jl#188), so admitting it is a compat-only change. Also drops the deprecated `FusionTensors` from the integration test matrix. --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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
1 participant
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
The dense
matricize(ReshapeFusion) now reshapes to the products of the codomain and domain dimension sizes inline, instead of routing through the genericmatricize_axes/fused_axis/tensor_product_axis/trivial_axissub-interface. That interface existed only to let fusion compose across array types, threaded with aVal{:codomain}/Val{:domain}side argument, and it no longer has a live consumer: the graded path was migrated to compute its axes directly in ITensor/GradedArrays.jl#188.This removes
matricize_axes,fused_axis,tensor_product_axis,trivial_axis, andtensor_product_fusionstyle.trivialrangestays, as a separate type-level identity range unrelated to the matricize fold. The removal is breaking: older GradedArrays releases extended the sub-interface, so this goes out as a minor release.Also drops the deprecated consumers
BlockSparseArrays,DiagonalArrays,FusionTensors,KroneckerArrays, andNamedDimsArraysfrom the integration test matrix, since they will not track this breaking change. The live consumersGradedArrays,ITensorBase,ITensorNetworksNext, andSparseArraysBasestay.