Elamur Groupantsworld-report-sampling
The groupant beside world/src/report.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 source model | forward |
| ast | partial | ../syntax.json |
| composition | partial source model | composition |
| cost | partial source model | cost |
| descriptant | partial source model | epoch-memory |
| invariant | partial source model | invariant |
| tests | partial source model | checks |
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 |
|---|---|---|---|
world-report-sampling-behavior-complement | behavior_complement | partial source model | behavior-complement |
world-report-sampling-reverse-reconstruction | reverse_reconstruction | partial source model | reverse-reconstruction |
world-report-sampling-sample-budget | resource_boundary | partial source model | sample-budget |
world-report-sampling-manual-join | constraint_relation | partial source model | manual-join |
Model text
Identity sampling, permissive verdict decoding and manual denominators.
forward
eval_tale obtains state answer, anchor fallback and baseline then scores; Agg sums supplied r_slm/r_anchor/r_base and separately sums state-source r_slm. report folds rows into fixed attribute order plus ordered maps for explicitness/story/first-space intent. sample ranks eligible identities by salted wrapping FNV/finalizer. read_verdicts skips malformed rows, overwrites duplicate keys; manual_md averages only matched recognized-type rows.
invariant
For Agg built from add, 0<=answered<=n; zero denominators use max(1). Only source determines answered membership; score fields are trusted and may be inconsistent with source. Manual scores are trusted rather than masked when Src::None. Unknown attributes contribute to all totals but lack a displayed attribute slice.
cost
N rows and total text bytes L: aggregation O(N log G + L), formatting linear in emitted bytes, sampling scans N once per seven attributes and sorts each slice O(N log N) overall with repeated key allocation. Manual lookup O(N log V), verdict decoding O(L + V log V). eval_tale cost includes external answer/baseline/scoring dependencies; no measured speed claim.
composition
Sample identities join verdicts on exact story and n. Duplicate sample rows may reuse one verdict and increase denominator; exclusions are identity-based.
epoch-memory
2026-10-30: first source-bound modeling epoch. Synthetic native fixtures only; no held-out tales, live parsers or annotators. CSV lost a quoted empty final record; a zero-length known section prevented decomposition progress. The repaired scanner retains quoted-field presence and decomposition ignores empty candidates. Report observations preserve existing permissive score/label semantics rather than silently enforcing a new measurement policy. Native finite evidence does not establish unrestricted correctness.
checks
Run cargo test -p world --lib report::wave9_models -- --test-threads=1. See the root sidecar evidence.md for repeated debug/release checks and compiled controls. Partial syntax excludes macro expansion and type/name resolution. No executable percentage approximant is claimed.
behavior-complement
Subject: current bound source and this node's scope. Domain and observation: With k=0 no rows are selected; excluded identity cannot be selected; malformed numeric verdict rows cannot replace a previous parsed key. These are inner absent transitions, not evidence of score validity.
Status: source-derived partial relation, supported only where named finite fixtures apply. Assume successful allocations, fixed dependencies and bounded native inputs; external filesystem/answer-generation effects remain unknown. The relation describes one terminating call, not all process traces.
reverse-reconstruction
Subject: current bound source and this node's scope. Domain and observation: A final verdict map admits arbitrary skipped lines and overwritten earlier rows. Manual percentages admit maps with different comments and omitted identities. A unique prior adjudication cannot be recovered.
Status: source-derived partial relation, supported only where named finite fixtures apply. Assume successful allocations, fixed dependencies and bounded native inputs; external filesystem/answer-generation effects remain unknown. The relation describes one terminating call, not all process traces.
sample-budget
Subject: current bound source and this node's scope. Domain and observation: Each recognized type contributes min(k,eligible row count); duplicates count as rows. Stable hash ties retain input order and forbid an unconditional permutation-invariance claim.
Status: source-derived partial relation, supported only where named finite fixtures apply. Assume successful allocations, fixed dependencies and bounded native inputs; external filesystem/answer-generation effects remain unknown. The relation describes one terminating call, not all process traces.
manual-join
Subject: current bound source and this node's scope. Domain and observation: Exact key join, then recognized type filter, then average matched occurrences. Three sample rows with two verdicts give N=2, independently of two state-source rows; the fixture separates both denominators.
Status: source-derived partial relation, supported only where named finite fixtures apply. Assume successful allocations, fixed dependencies and bounded native inputs; external filesystem/answer-generation effects remain unknown. The relation describes one terminating call, not all process traces.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/world/src/report.rs.groupant/sampling.groupant/