Elamur Groupantskaggle-text-command
The groupant beside tools/kaggle/src/main.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 |
| composition | partial | composition |
| cost | symbolic source derived | invariant-and-cost |
| descriptant | present | descriptant |
| invariant | partial | invariant-and-cost |
| 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 |
|---|---|---|---|
kaggle-text-command-behavior_complement | behavior_complement | partial source derived model | behavior_complement |
kaggle-text-command-reverse_reconstruction | reverse_reconstruction | partial source derived model | reverse_reconstruction |
kaggle-text-command-feature_thresholds | feature_thresholds | partial source derived model | feature_thresholds |
kaggle-text-command-candidate_tie_order | candidate_tie_order | partial source derived model | candidate_tie_order |
kaggle-text-command-group_edge_relation | group_edge_relation | partial source derived model | group_edge_relation |
kaggle-text-command-fold_label_isolation | fold_label_isolation | partial source derived model | fold_label_isolation |
Parts
Model text
descriptant
kaggle-text-command
Epoch 2026-10-31
Full source read preceded changes. Prospective model predicted branch/order, resource and information-loss behavior. Tests exercise actual helpers on bounded synthetic inputs; command workflows remain unexecuted. Production source prefix is preserved; only inline tests added.
Load CSV records into P with optional parse defaults; feats yields eight categories; learn repeatedly picks best Laplace singleton/pair among remaining rows, removes covered rows and stores strict majority, retaining at most twelve rules. Because the rule-count check follows candidate construction, up to thirteen candidate-building iterations can execute. predict_list first matching condition; predict_wcg direct shared-key neighbors with train-known outcomes; cv id modulo partitions.
Executable percentage approximation is unsupported: categorical parsing/effects have no useful certified metric here. Partial syntax excludes macro expansion and name resolution.
approximant
Forward relation
Load CSV records into P with optional parse defaults; feats yields eight categories; learn repeatedly picks best Laplace singleton/pair among remaining rows, removes covered rows and stores strict majority, retaining at most twelve rules. Because the rule-count check follows candidate construction, up to thirteen candidate-building iterations can execute. predict_list first matching condition; predict_wcg direct shared-key neighbors with train-known outcomes; cv id modulo partitions.
Domain
Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used.
composition
Local composition
The scoped group-prediction fragment models its declared source interval. Parent relation: Load CSV records into P with optional parse defaults; feats yields eight categories; learn repeatedly picks best Laplace singleton/pair among remaining rows, removes covered rows and stores strict majority, retaining at most twelve rules. Because the rule-count check follows candidate construction, up to thirteen candidate-building iterations can execute. predict_list first matching condition; predict_wcg direct shared-key neighbors with train-known outcomes; cv id modulo partitions.
Shared representation/invariant: Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds. Child costs are included in parent costs, not added a second time. Other command paths are tagged alternatives; the child alone does not establish whole-file behavior.
models
Typed source models
All models are partial source-derived relations at this manifest snapshot. Scope is one call/trace under the stated domain; no universal semantic proof.
behavior_complement
First-match list cannot defer to a later rule. Female correction to false requires known dead woman/child and zero known living; unknown mates alone cannot flip gender.
Domain: Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used. Observation: Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds. Assumptions: bounded resources and source branch semantics. Oracle: selected actual-helper tests in tests.md constrain only their finite fixtures; remaining branches are source-derived.
reverse_reconstruction
Prediction/reason admits many passenger feature and training fibers; category bins and rounded fare keys discard numeric precision; boolean result cannot recover rule list or unique outcomes.
Domain: Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used. Observation: Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds. Assumptions: bounded resources and source branch semantics. Oracle: selected actual-helper tests in tests.md constrain only their finite fixtures; remaining branches are source-derived.
feature_thresholds
Age bins <13,<20,<50,else; fare <8,<15,<50,else; None separate. NaN falls into last bin, not unknown.
Domain: Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used. Observation: Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds. Assumptions: bounded resources and source branch semantics. Oracle: selected actual-helper tests in tests.md constrain only their finite fixtures; remaining branches are source-derived.
candidate_tie_order
Maximum Laplace accuracy breaks ties by fewer conditions then lexicographically smaller Cond; choose class p>n; retain rule only lap>=base_acc+.02.
Domain: Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used. Observation: Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds. Assumptions: bounded resources and source branch semantics. Oracle: selected actual-helper tests in tests.md constrain only their finite fixtures; remaining branches are source-derived.
group_edge_relation
Edge(a,b) iff distinct ids share surname+fare rounded2 or ticket. Union of direct neighbor sets is not transitive; empty ticket creates a group.
Domain: Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used. Observation: Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds. Assumptions: bounded resources and source branch semantics. Oracle: selected actual-helper tests in tests.md constrain only their finite fixtures; remaining branches are source-derived.
fold_label_isolation
cv test predicate id%k==fold; closure sees remaining labels while WCG all includes all identities/features. k>0 gives each input one test visit; empty denominator yields NaN.
Domain: Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used. Observation: Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds. Assumptions: bounded resources and source branch semantics. Oracle: selected actual-helper tests in tests.md constrain only their finite fixtures; remaining branches are source-derived.
tests
Native evidence — 2026-10-31
Package/bin: kaggle. Exact selected names:
wave33::features_groups_and_empty_learning
All 1 selected checks passed twice in debug and twice in release, one exact name per invocation with one test thread. Numeric exits are zero; sorted named outcomes match across all four runs. Added test blocks were formatted before these receipts; original source prefix (including earlier tests) remains byte-identical.
One compiled production mutant changed Some(a) if a < 13.0 to Some(a) if a <= 13.0. The new wave33 test compiled and failed its runtime assertion (exit 101). Exact original/restored SHA-256 equality and a restored positive exact test (exit 0) were verified. No assertion-only mutant is counted; existing negative assertions remain ordinary tests. No surviving compiled mutant is omitted.
The external wave33 text receipts retain native-commands.json, verified-outcomes.json, mutation-report.json, raw logs/statuses, paired syntax projections and prospective.md. No private filesystem path is part of this public model.
Scope: actual helpers on synthetic in-memory data only. Unique passenger ids for group interpretation; bounded u32 family sums and usize counts; finite ages/fares for meaningful numeric bins; missing labels treated false by learner. No competition files used. The scoped fragment reuses these tests where its helper is reached; other declared paths remain source-derived. Macro expansion, full command/effect execution, corpus accuracy and universal correctness are not established.
invariant-and-cost
Invariant and cost
Invariant
Feature order and BTreeMap candidate order are deterministic for bounded finite inputs. Label ties/default ties false. Group neighbors are direct key intersections, not connected components. Empty cv returns NaN, k=0 executes no folds.
Cost
Eight features produce 36 candidates per row; at most twelve rules are retained, with up to thirteen candidate-building iterations, giving O(1336N log(36N)) plus string comparison; groups worst-case O(N^3) dedup scans and is rebuilt per prediction; CV clones O(kN) records before learner costs.
Arithmetic and allocation premises are conditional. No timing or percentage surrogate claimed.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/tools/kaggle/src/main.rs.groupant/