pub struct CompareReport {Show 14 fields
pub old_file: String,
pub new_file: String,
pub dev_delta_overall: f64,
pub validation_delta_overall: f64,
pub generalization_efficiency: f64,
pub dev_deltas: BTreeMap<String, f64>,
pub validation_deltas: BTreeMap<String, f64>,
pub max_section_regression: f64,
pub max_regression_section: String,
pub gate_ge: f64,
pub guard_section_delta: f64,
pub ge_passed: bool,
pub guard_passed: bool,
pub failures: Vec<String>,
}Expand description
Wave-11 gate report over two saved holdout result files (JSON
HoldoutComparisons): the GE gate (--gate-ge MIN, default 0.0 — fails
when GE <= MIN; semantic waves must generalize) and the per-section
validation regression guard (--guard-section-delta MAX, default 0.05 —
fails when ANY startup-required section regresses by more than MAX
between the two runs, in development or validation).
Fields§
§old_file: String§new_file: String§dev_delta_overall: f64new.dev.mean_overall - old.dev.mean_overall.
validation_delta_overall: f64new.holdout.mean_overall - old.holdout.mean_overall.
generalization_efficiency: f64validation_delta_overall / dev_delta_overall (guarded).
dev_deltas: BTreeMap<String, f64>Per-section delta = new - old, development corpus.
validation_deltas: BTreeMap<String, f64>Per-section delta = new - old, validation corpus.
max_section_regression: f64Largest startup-required section drop (new - old) across both corpora; 0.0 when nothing regressed.
max_regression_section: StringThe section with the largest drop (section@corpus), e.g.
contracts@validation; deterministic (first at the max in
STARTUP_SECTIONS order, development before validation).
gate_ge: f64§guard_section_delta: f64§ge_passed: bool§guard_passed: bool§failures: Vec<String>Human-readable failure reasons; empty when the report passes.