ff4a1b3d4a36fd539b918c5d41bce8715bf65714
11 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Brummel
|
ff4a1b3d4a |
feat(0092): run-from-blueprint infrastructure — restructure + topology_hash + shared seam (Tasks 1-3)
Cycle-1 infrastructure for #165 (World/C21). Behaviour-preserving + workspace green; the CLI arm + tests (Tasks 4-5) follow. Task 1 — restructured stage1_r_graph into stage1_signal() (the serializable signal leg: SMA-cross -> Bias, a `price` input-role + a `bias` output) + wrap_stage1r(signal, ...) (broker/sinks/exec/cost around a nested signal), with stage1_r_graph() = wrap_stage1r(stage1_signal(...)). DEVIATION from the plan's "callers untouched" claim, verified behaviour-preserving: nesting the signal prefixes its param-space names (stage1_signal.fast.length), which broke the sweep's bare .axis("fast.length"); fixed exactly as the #137 stop-knob machinery does — FAST_LENGTH_SUFFIX/SLOW_LENGTH_SUFFIX suffix-resolution + stage1_r_friendly_name arms mapping the prefixed names back to the bare manifest names. All observable behaviour (manifest names, member keys, metrics, traces) byte-identical; the flat graph and every metric unchanged (full suite green, incl. sweep/walkforward/MC). Task 2 — RunManifest.topology_hash: Option<String> (Tier-1 optional, serde default/skip — old manifests byte-identical, #156), threaded None into all 24 literal sites across 7 crates. Also fixed the aura-registry RunManifestRead read-mirror to carry topology_hash (it was hardcoding None on load) — else a stored hash would silently drop on the registry round-trip once Task 3 stores a real one. Task 3 — the shared run_signal_stage1r seam (hash the signal, wrap, compile with params, bootstrap, run, manifest with params from param_space + topology_hash) + the sha256_hex topology_hash helper (sha2 in aura-cli, off the frozen engine, invariant 8). run_stage1_r now carries its topology_hash too (every run becomes self-identifying, #158). The seam is unwired until Task 4 (transient #[allow(dead_code)], retired there); RED-first seam tests added. Verified: cargo test --workspace green (51 suites, 0 failures); cargo clippy --workspace --all-targets -D warnings clean. refs #165 |
||
|
Brummel
|
0a06afed29 |
feat(0078): cross-instrument generalization as a validation step
Last piece of the Inferential-validation milestone (#146): a new `aura generalize` subcommand grades how consistently a single candidate holds across a set of instruments — the across-instrument axis of the same anti-false-discovery concern as trials-deflation (#144, within one family) and plateau-over-peak (#145, within one instrument's surface). Aggregator/validator, not a third selector (user-ratified): it runs a brought candidate (a single-cell stage1-r grid — --fast/--slow/--stop- length/--stop-k pinned to single values) across an instrument list (--real SYM1,SYM2,...), collects per-instrument R-metrics, and reduces them to a worst-case floor (min over instruments) + a sign-agreement count + the per-instrument breakdown. R-only (C10 — R is the only account/instrument-agnostic unit); the worst-case min is the cross- instrument sibling of PlateauMode::Worst and satisfies the non- domination constraint by construction (a strong instrument cannot lift a min). Rejected: a cross-instrument t-stat mean/std*sqrt(M), which reintroduces the pooled mean the milestone warns against. - RunManifest gains a first-class `instrument` lineage field (serde- widened, C14/C18 — legacy lines load as None, existing manifest bytes unchanged); the compat read-mirror threads it in lockstep. - FamilyKind::CrossInstrument (C12 comparison axis); the M per-instrument runs persist as a CrossInstrument family, each member self-identifying via its stamped instrument. The aggregate is recomputable from the members (McAggregate precedent), printed live. - The generalization reduction lives registry-side beside optimize_ deflated/optimize_plateau (C9), reusing resolve_metric/metric_value/ is_r_metric; check_r_metric is the data-free pre-check the CLI runs before evaluating any instrument. - Additive: existing sweep/walkforward/mc/standalone paths stay byte- identical (C23 — instrument omitted when None via skip_serializing_if). Verified by the orchestrator: cargo test --workspace green (0 failed — the new generalization_*, check_r_metric, parse_generalize_*, and the gated cross-instrument E2E all pass; the E2E ran the real GER40/USDJPY path on this host, exit 0); cargo clippy --workspace --all-targets -D warnings clean. closes #146 |
||
|
Brummel
|
a2959050a4 |
feat(0076): deflate the sweep winner's metric for the number of trials
Adds optimize_deflated beside optimize in aura-registry: the in-sample walk-forward selection now records, on each OOS winner's manifest, how much the winner's metric is inflated by the size of the search it won — a deflated score and (R arm) an overfit probability — without changing which member wins (additive, C23; a regression test pins optimize_deflated's winner == optimize's). - R arm (sqn_normalized/expectancy_r/...): a centred moving-block reality-check. Each member's per-trade R series is mean-subtracted to impose the no-edge null, resampled via the r_bootstrap kernel (extracted to a shared resample_block, byte-identical), and the best-of-K null-max gives overfit_probability = (count(null >= raw)+1)/(n+1) and deflated_score = raw - p95(null). Deterministic given the seed (C1). - total_pips arm: a closed-form expected-max-of-K dispersion floor (inv_norm_cdf + expected_max_of_normals), no probability. The record (FamilySelection on RunManifest) rides the proven serde(default, skip_serializing_if) widening + a compat.rs mirror, so legacy runs.jsonl/families.jsonl lines load unchanged and unstamped sweep/mc/standalone manifests stay byte-identical (C14/C18). The SelectionMode enum reserves a Plateau slot for #145. CLI: walk-forward stamps both arms; runs family <id> rank surfaces a human-readable deflated/overfit line. Quality review (re-dispatched after the implement-loop's quality gate exhausted) caught a real defect: on the R arm with positive resamples but no member carrying trade_rs (a family loaded from disk — trade_rs is serde-skipped), the null was all-NEG_INFINITY, turning deflated_score into +inf. Fixed by collapsing "no usable null" (zero resamples OR no trade_rs) into one degenerate floor (deflated == raw, overfit == 1.0); pinned by optimize_deflated_no_trade_rs_floors_instead_of_infinity. Verified: cargo test --workspace and cargo clippy --workspace --all-targets -D warnings both green. Frictionless Stage-1 R; n_eff advisory and reload-time recompute deferred (spec 0076 Out of scope). closes #144 |
||
|
Brummel
|
c192dfd344 |
audit(0075): cycle close — drift-clean; doc-link + trade_rs tripwire tidy; drop ephemera
Architect drift review (c99e7b8..2e77ab1) against the design ledger +
CLAUDE.md invariants: drift-clean, debt only — no contract or spec
violation. Verified intact: C18 wire shape (trade_rs is #[serde(skip)] +
excluded from the hand-written PartialEq; runs.jsonl/families.jsonl bytes
and digests unmoved; mc_r_bootstrap is a printed line, never a record),
C1 determinism (seeded r_bootstrap + disjoint walkforward windows, both
pinned by byte-identical-rerun e2e tests), C7 frictionless Stage-1
(round_trip_cost stays 0.0), C8/C12 streaming reduction (IS sweep folds
O(trades)/member, OOS run non-reduce). The verbatim summarize_r /
r_metrics_from_rs float duplication is guarded by the cross-reducer
equality test — sound.
Tidy applied this close:
- doc-link drift: the new E[R] notation in r_bootstrap / RBootstrap /
run_mc_r_bootstrap doc-comments was parsed by rustdoc as an intra-doc
link `[R]` (4 warnings). Wrapped in backticks; doc build now clean.
- coverage micro-gap [low, architect]: added a tripwire asserting a
populated trade_rs is absent from the serialized JSON (the serde(skip)
C18-preservation property had no direct test — only Vec::new() was
serialized before).
Carry-on (benign debt, not fixed): the mc usage literal carries a `usage:`
prefix the sibling sweep/walkforward usages omit — required because the
new `["mc", rest @ ..]` arm now handles the `--real` rejection and the
preserved golden asserts stderr.contains("usage"); churning the golden to
drop the prefix is not worth it.
Removed the cycle's ephemeral spec + plans (docs/specs/0075,
docs/plans/0075, docs/plans/0076) per the lifecycle rule.
Verified: cargo test --workspace 610 passed / 0 failed, clippy -D warnings
clean, cargo doc --no-deps clean.
closes #139
|
||
|
Brummel
|
2e77ab1dda |
feat(0075): Monte-Carlo R-bootstrap for OOS validation (iter 2)
Add `aura mc --strategy stage1-r [--real <SYM>]`: run the stage1-r
walk-forward, pool its OOS per-trade R series, and report a moving-block
bootstrap E[R] distribution / CI (one `mc_r_bootstrap` line with an `e_r`
quantile block + P(E[R]<=0)). Completes spec 0075's Monte-Carlo half on top
of iter 1's walk-forward R-reporting.
Engine: `r_bootstrap(rs, n_resamples, block_len, seed) -> RBootstrap` — a
deterministic moving-block bootstrap (non-circular, final block truncated;
block_len=1 = i.i.d. trade-shuffle; clamped to [1,n]; empty -> all-zero),
reusing the existing MetricStats/quantile and SplitMix64.
CLI: parse_mc_args/McArgs split — bare/`--name`/`--trace` route to the
unchanged synthetic seed-resweep `run_mc` (byte-for-byte preserved); the R
path accepts ONLY `--strategy stage1-r`, so a bare `--real` stays a usage
error (mc_rejects_real_flag_with_usage_exit_2 preserved, doc-comment
refreshed). run_mc_r_bootstrap reuses walkforward_family + RMetrics.trade_rs.
Corrective fixes over the plan snippets, both verified: the mc usage string
carries a `usage:` token (the new `["mc", rest @ ..]` arm now handles the
`--real` rejection, and the preserved golden asserts `stderr.contains
("usage")`); `#[allow(clippy::large_enum_variant)]` on McArgs keeps the
unboxed Stage1RGrid field layout.
Scope: single `--real <SYM>` per invocation; cross-symbol pooling is a
deferred follow-on (needs a multi-symbol grammar). Verified: cargo build
clean, full `cargo test --workspace` green (609 passed), clippy -D warnings
clean; bare `aura mc`, SMA walkforward, stage1-r single-run, and C18
round-trip goldens all preserved.
refs #139
|
||
|
Brummel
|
c6c1d3bb10 |
feat(stage1-r): rename the exposure_sign_flips metric to bias_sign_flips (#126)
Complete the typed-field half of the exposure->bias rename - the one the cycle-0065
close audit flagged as aging worst (a serde-stable on-disk key).
- RunMetrics.exposure_sign_flips -> bias_sign_flips with
#[serde(alias = "exposure_sign_flips")], so legacy runs.jsonl still deserialises
(pinned by the legacy-read tests, which keep the old key); new output serialises
bias_sign_flips (the byte-pin tests updated to the new key).
- The registry rank metric: Metric::ExposureSignFlips -> BiasSignFlips, the rank
string accepts BOTH "bias_sign_flips" and "exposure_sign_flips" (CLI back-compat),
the error/listing updated; the rank tests exercise both names.
- The MC aggregate field renamed too (McAggregate is hand-rendered to JSON, not
serde-derived - no alias needed).
Scope held to the typed metric. The strategy-output PARAM NAMESPACE - the
.named("exposure") Bias instance, its exposure.scale knob path (Composite::param_space
derives the knob from the node name), and the exposure_scale manifest label - is a
coherent ~30-site, sweep-pinned surface (the implement loop correctly flagged renaming
the knob as an unscoped expansion touching walkforward + the cli_run byte-pins). Kept
uniformly as "exposure" and deferred to a follow-on so the cluster renames as one
consistent unit. SimBroker's pre-reframe exposure concept and the on-disk "exposure"
tap label stay by the #117 fork decision.
cargo test --workspace green; clippy --workspace --all-targets -D warnings clean; a
live `aura run --harness stage1-r` emits "bias_sign_flips".
closes #126
refs #117
|
||
|
Brummel
|
b4e84335c4 |
feat(stage1-r): Sizer + RiskExecutor + R-metric enrichment (iter 2)
Iteration 2 of cycle 0065 (Stage-1 R signal quality) — the node + metric layer. Builds on iter-1's PositionManagement dense R-record + core summarize_r. What ships: - PositionManagement gains a 4th `size` input (slot 3 -> col 10), fed by the Sizer. R is computed size-INVARIANTLY (pure stop-distance ratio), so size never touches realized_r — pinned by a RED test (scaling size leaves every realized_r unchanged) at the node and again end-to-end through the executor. - Sizer (aura-std): `size = risk_budget / stop_distance` — flat-1R (risk_budget 1.0 => one risk unit per trade, size inversely proportional to the stop, not a constant). Reads bias (firing/presence) + stop_distance; the Stage-2 fixed-fractional sizer slots in unchanged (swap risk_budget for risk_fraction*equity, same node shape). - summarize_r enrichment: SQN (sqrt(n)*mean_R/sample-stdev_R; n<2 or zero-variance -> 0), conviction_terciles_r (E[R] by |bias_at_entry| tercile), net_expectancy_r (gross minus one round-trip cost per trade, charged in R via the latched_dist recovered from the entry_price/stop_price columns). summarize_r gains a `round_trip_cost` param (price units). The net-of-cost recovery is tested through the real producer->consumer seam, not just hand-built rows. - RunMetrics.r: Option<RMetrics> with #[serde(default, skip_serializing_if = "Option::is_none")] — legacy runs.jsonl (no `r` key) deserialise to None and a pip-only run's on-disk shape stays byte-unchanged (C14/C18 back-compat). Every RunMetrics literal threaded (report.rs, aura-registry, aura-engine/mc). - RiskExecutor (aura-engine integration-test fixture, sibling of vol_stop_composite): FixedStop -> Sizer -> PositionManagement behind open bias+price input roles, price fanned to both the stop and PM, the Sizer's size into PM. Bias is produced in-graph from the single price source (a second bias *source* would k-way-merge into separate cycles and mark stale prices). The Veto is a DOCUMENTED SEAM, not a runtime node (a pass-through identity is what C19/C23 DCE deletes). Tests: the composite bootstraps + runs + folds to the documented hand value; R invariant under risk_budget while the size column scales; a live-folded RMetrics survives the RunMetrics serde round-trip. The dense-record size column (10) is brought under the cross-crate layout guard (stage1_r_e2e r_col_indices_match_producer_field_layout) so the executor fixture's size-invariance read is drift-protected like the others. Scope: the CLI/recording surface (#129) is sub-split into a separate iteration 3 and is NOT in this commit. Verified: cargo build --workspace clean; cargo test --workspace 500 passed, 0 failed; cargo clippy --workspace --all-targets -D warnings clean. refs #117 #127 #128 #129 |
||
|
Brummel
|
d858caf67b |
chore: scrub dangling references to deleted specs/plans from sources
docs/specs and docs/plans were retired (prior commit); the source/test
comments that cited them ("spec 0050 §4.1", "spec §Testing N", "per spec",
"the spec's ...") now point at nothing. Strip every such pointer while
preserving the technical substance, the design-ledger contract refs
(C1/C11/C20/C34/C12.1/...), and the Gitea issue refs (#41).
Comment/doc edits only across 20 files — no logic change; full workspace
suite green, clippy clean.
|
||
|
Brummel
|
cd3d1ca9ed |
refactor(aura-core): split Scalar into a tag-free Cell + ScalarKind
Motivation
----------
`Scalar` was a tagged enum (I64/F64/Bool/Ts), so every scalar value
physically carried its own kind tag. But the kind is already known from
the schema/port/column the value flows through (C7: the type is a
property of the column, not of the value — the hot path is already
columnar `Column<T>`, and `AnyColumn::get` *reconstructs* the tag from
the column on the way out). The per-value tag was therefore redundant
with the kind the surrounding context already holds.
That redundancy had three costs:
* It baked an implicit `match` (a branch) into every function that read
a Scalar payload — even where the caller statically knew the type.
The tag could never be exploited away.
* Size: a tagged enum is tag + payload = 16 bytes (f64/i64 alignment),
twice the 8 bytes the value needs. A `Column<Scalar>` would be double
the memory and half the cache utilisation.
* It is the shared root of several downstream papercuts we keep hitting
— the lossy f64 manifest field, the `unreachable!` panic on a
non-numeric param, the serde-tag question — all symptoms of "the type
is baked into the value".
Change
------
Introduce `Cell`: a type-erased 64-bit word (`struct Cell(u64)`) that is
not readable without external type context. It is constructed per base
type (`from_i64/from_f64/from_bool/from_ts`) and read only by naming the
type at the call site (`i64()/f64()/bool()/ts()`) — each a branch-free
bit-cast. The hot path resolves the kind once at the boundary (from the
schema) and then reads natively, with no per-value branch. `Cell` knows
nothing of `Scalar` or `ScalarKind`; the dependency is strictly one-way,
and it lives in its own `cell.rs` (more is planned on top of it).
`Scalar` becomes `struct { kind: ScalarKind, cell: Cell }` — the
self-describing form for the dynamic boundaries (builder binding,
serialization, rendering), built on top of `Cell`. Its `as_*` accessors
now `debug_assert` the kind and return the native value (free in
release); calling the wrong accessor is a caller bug, not a checked
`Option`. The variant constructors `Scalar::I64(..)` become associated
fns `Scalar::i64(..)`.
`PartialEq` is hand-written (not derived) to preserve the former enum's
value semantics: kinds must match, then native payloads compare, so f64
keeps IEEE-754 behaviour (`NaN != NaN`, `+0.0 == -0.0`) and a kind
mismatch is never equal even when the raw words coincide. A fixture
(`scalar_eq_is_value_not_bitwise`) pins exactly the cases where bit- and
value-equality diverge, so it can't silently regress. `Cell`'s own
`Eq`/`Hash` stay bitwise — correct for a raw word.
The change is behaviour-preserving: Scalar's observable behaviour is
identical to the pre-Cell enum (the value-equality fixture proves it);
only the internal representation changed. The ~440 call sites across the
workspace are a mechanical constructor rename plus ~12 destructuring
sites (match-arms / `let`-patterns) rewritten to `kind()` + `as_*`.
Verified: cargo build --workspace --all-targets, cargo clippy --workspace
--all-targets -- -D warnings, cargo test --workspace — all green.
|
||
|
Brummel
|
4764656062 |
feat(aura-engine): walk-forward family — WindowRoller + walk_forward (C12 axis 3)
The third C12 orchestration axis: walk-forward varies the data window. A
WindowRoller is a pure iterator of WindowBounds { is, oos } — bounds only, never
tick data (#71 firewall). walk_forward runs a caller closure per split disjointly
over the shared run_indexed core (C1), and stitches the OOS pip-equity segments
into one continuous curve (each segment offset by the running sum of prior
segments' finals; an empty segment contributes 0.0, mirroring summarize). C2
no-look-ahead is structural: the roller only ever emits oos.0 == is.1 + 1, pinned
by a zero-tick bounds test. The in-sample optimize (axis 2) is closure-supplied,
not called by the engine: aura-engine has no aura-registry dependency (C9), and
C12 forbids baking search policy into the primitive — the CLI bridges both crates.
Param-stability is on-demand, not stored (the R2 decision, recorded with
provenance on #69): WalkForwardResult stores only the raw per-window outcomes +
the stitched curve, and param_stability(&result) -> Vec<MetricStats> is a public
reduction over the retained per-window chosen params. A stored summary would be
recomputable from the windows (redundant) and would force one statistic canonical
when "stability" admits several; this mirrors SweepFamily (raw points + on-demand
optimize), not McFamily (stored aggregate). MetricStats::from_values is extracted
from McAggregate::from_draws (behaviour-preserving — the 7 mc tests stay green)
and gains serde so the CLI summary renders it and #70 lineage can persist it.
`aura walkforward` runs a built-in rolling walk-forward over a synthetic windowed
source: per window it sweeps the built-in grid in-sample, optimizes by total_pips,
runs the chosen params out-of-sample, persists each OOS RunReport (C18), and prints
a stitched summary line.
Two plan deviations, both compiler-forced and consistent with siblings: added
#[derive(Debug)] to WindowRoller (the RED tests' unwrap_err needs Ok: Debug; the
sibling types already derive Debug); removed a now-redundant test-module
SyntheticSpec import after Step 2 hoisted it to the top-level use.
Verification: cargo test --workspace green (aura-engine 152 incl. walkforward 9 +
mc 9, aura-cli 14); clippy --workspace --all-targets -D warnings exit 0; cargo doc
--workspace --no-deps clean.
refs #69
|
||
|
Brummel
|
e2c550f2f9 |
feat(aura-engine): monte_carlo family — McFamily over a seed set (C12 axis 4)
`monte_carlo(base_point, seeds, run_one) -> McFamily`, the Monte-Carlo
orchestration family: a fixed base point run over a seed set, each seed a
disjoint C1 realization. C12 axis 4 — Monte-Carlo IS a sweep over seeds, so it
reuses the disjoint-parallel executor rather than forking a run loop.
The reuse is realized by extracting the index-parallel core out of
`sweep_with_threads` into a `pub(crate) run_indexed<T>` (generic over the
per-job result, over a 0..n job range); `sweep_with_threads` becomes a thin
adapter over it. Behaviour-preserving (C1/C11/C23): the three pre-existing sweep
tests stay green — they ARE the proof the refactor preserved behaviour. A small
incidental win: the adapter consumes `points` via zip, dropping the old
per-point clone.
mc.rs adds McDraw{seed,report} / McFamily{draws,aggregate} (draws in seed-INPUT
order, independent of thread completion) / McAggregate (mean + p5/p25/p50/p75/p95
of all three run metrics — the V1 decision: covers every metric, not a "chosen"
one) / MetricStats, a private type-7 linear-interpolation `quantile`, and
`McAggregate::from_draws` (a pure post-run reduction recomputable from the
retained raw draws). Eager-agnostic firewall (#71): the API takes seeds + a
per-draw closure `Fn(u64,&[Scalar])->RunReport`, never a materialized stream Vec;
seed->Source construction stays a closure-body concern. An empty seed set is a
debug-asserted precondition, preserving the no-Result `-> McFamily` signature.
7 RED-first mc::tests pin: one draw per seed in input order; determinism across
1 vs 8 threads (C1); draw == an independent seeded run; distinct seeds perturb
the family; aggregate == from_draws(draws); type-7 quantile/mean math on a known
fixture; quantile endpoints + singleton. Workspace green, clippy -D warnings
clean, cargo doc clean.
refs #68
|