Elamur Groupantsmlab-ext-boost-metric
The groupant beside mlab/tests/ext_boost.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 | partial bounded parent evidence | 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 |
zero_metric | zero_metric | partial source derived model | zero_metric |
fault_witness | fault_witness | partial source derived model | fault_witness |
branch_relation | branch_relation | partial source derived model | branch_relation |
resource_cost | resource_cost | partial source derived model | resource_cost |
Model text
descriptant
mlab-ext-boost-metric
measure retains normal reference values only; e=abs((got-want)/want)/EPSILON with NaN got=>infinity, strict greater updates first worst. Existing CASES bound4eps; synthetic tests do not certify those datasets.
Epoch 2026-10-31: isolated source-bound fragment over bytes [2204,4687); includes no claim beyond this relation. Source hash 7fa4b847808f0a485411988bb342bd0b0540d350c50e6166f08224c1dbac9dfc.
approximant
metric: forward relation
measure retains normal reference values only; e=abs((got-want)/want)/EPSILON with NaN got=>infinity, strict greater updates first worst. Existing CASES bound4eps; synthetic tests do not certify those datasets.
Domain: Synthetic bounded rectangular rows with valid col index; dataset suites deliberately not executed.
Executable surrogate gap: no measured cheaper implementation with a useful scoped metric is supplied. Categorical/error/effect relations are exact obligations, not percent losses. Future numeric caps only1/2/5/10%.
invariant
metric: invariant and cost
Relative metric is undefined at0 and deliberately excluded; this is not a zero-reference accuracy claim.
erf fault adds1 to first output; synthetic erf0 confirms hook, without running frozen Boost points.
Scope-specific cost: O(rows) filtering and discrepancy scan plus interpreter vector execution; O(rows) references/results, zero references removed before call.
models
behavior_complement
Zero/subnormal/NaN/infinite reference values do not contribute. max_eps0 leaves worst_xNaN; ties do not update worst.
reverse_reconstruction
Stat admits all row sets with same normal count and maximum relative discrepancy, including arbitrary excluded rows; worst_x alone not full error curve.
zero_metric
Relative metric is undefined at0 and deliberately excluded; this is not a zero-reference accuracy claim.
fault_witness
erf fault adds1 to first output; synthetic erf0 confirms hook, without running frozen Boost points.
branch_relation
measure retains normal reference values only; e=abs((got-want)/want)/EPSILON with NaN got=>infinity, strict greater updates first worst. Existing CASES bound4eps; synthetic tests do not certify those datasets.
resource_cost
O(rows) filtering and discrepancy scan plus interpreter vector execution; O(rows) references/results, zero references removed before call.
tests
Evidence boundary
See ../tests.md for the actual selected source test and mutant. These tests sample the file, not every fragment branch. Zero/subnormal/NaN/infinite reference values do not contribute. max_eps0 leaves worst_xNaN; ties do not update worst. Other branches here are source-derived and unexecuted unless the parent test explicitly calls them.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/mlab/tests/ext_boost.rs.groupant/metric.groupant/