Elamur Groupantsmlab-numerics-reduce-folds
The groupant beside mlab/src/builtins/reduce.rs in the Mova repository: what the code computes, what it never does, how to go back from its outputs, its invariants, cost and checks.
Facts
Implementation material. A Mova code model describes this code at one source revision: partial models, finite checks and their stated scope. It is not a reviewed general exposition, and its labels are shown as written, never raised.
Components
| component | label | where |
|---|---|---|
| approximant | partial | approximant |
| ast | partial | ../syntax.json |
| cost | symbolic source derived | invariant |
| descriptant | present | descriptant |
| invariant | partial | invariant |
| tests | executed bounded checks | tests |
Obernants
The typed relations around the forward model: the behaviour complement (what the code never does), the reverse reconstruction (every input compatible with an output) and specialised views.
| id | type | label | where |
|---|---|---|---|
behavior_complement | behavior_complement | partial source derived model | behavior_complement |
reverse_reconstruction | reverse_reconstruction | partial source derived model | reverse_reconstruction |
column_fixture | column_fixture | partial source derived model | column_fixture |
identity | identity | partial source derived model | identity |
explicit_empty | explicit_empty | partial source derived model | explicit_empty |
real_scan | real_scan | partial source derived model | real_scan |
Model text
descriptant
mlab-numerics-reduce-folds
slices(dim1) lists column indices j*rows+i, dim2 row indices, else singleton. reduce_c folds (init,0) with complex operation and returns shaped result; absent dim on0x0 yields scalar init. cumulative uses same slices but real-only running accumulation.
Modeling epoch 2026-10-31
Prospective source reading preceded inline tests. This node records implementation behavior, not a complete MATLAB compatibility contract. Models retain limits discovered in source reading, including real-only projections and warning-only result gates. Synthetic CPU fixtures are independent of frozen suites. No external effects, teacher, corpora or benchmark main are executed.
approximant
Forward relation
slices(dim1) lists column indices j*rows+i, dim2 row indices, else singleton. reduce_c folds (init,0) with complex operation and returns shaped result; absent dim on0x0 yields scalar init. cumulative uses same slices but real-only running accumulation.
Executable Rust surrogate gap: no calibrated replacement is attached. Exact test calls are evidence for the reference, not a fast approximant. Numerical approximants would need a named domain, zero convention and metric at supported caps 1/2/5/10%; categorical errors/effects have no percent surrogate here.
invariant
Invariants and cost
Output shape dim1=1xcols,dim2=rowsx1,else original. Complex storage narrows via Mat::complex only for reduce_c.
Cost
O(N+S) fold, slice-index/vector-header allocation and output storage; cumulative O(N+S) plus cloned real cells. For0xM at dim1, N0 but S=M identity outputs.
Allocation failure, unchecked usize dimension arithmetic and IEEE nonfinite propagation are outside finite fixture guarantees. No measured throughput or global accuracy claim.
models
Typed source models: mlab-numerics-reduce-folds
behavior_complement
No default empty fold calls f. Singleton higher dimensions exclude cross-cell accumulation. cumulative excludes imaginary dependence.
reverse_reconstruction
Given sum result, compatible inputs have the requested shape and per-slice sum; product fibers depend on zeros and scaling pairs. Cumulative real sum can be differenced along same dimension absent rounding, but cannot recover imaginary input.
column_fixture
2x2 re[1,3,2,4] summed dim1→1x2[4,6].
identity
0x0 with no dim and init1 returns scalar1 before slicing.
explicit_empty
Explicit dim changes empty handling, producing shape derived from slices rather than scalar shortcut.
real_scan
cumulative output is new real Mat even for input with im; no complex prefix-product claim.
tests
Evidence and limits
Selected wave30_tests exercise this source; fragment-specific limits are stated in its operators.
Native command specifications, numeric exits, profile-paired logs, compiled controls and exact-restoration receipts are retained externally in the wave30 numerics run. Tests exercise actual source, not a copied model. The selected finite fixtures do not prove full convergence, backend accuracy on all matrices, resource exhaustion or side-effectful callback behavior.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/mlab/src/builtins/reduce.rs.groupant/reduce-folds.groupant/