Iteration stderr-markers-2, tasks 3-4 of the stderr-class-markers plan.
aura-runner's four continuing-run diagnostics join the class grammar as
hand-written literals (the #295 boundary keeps the macro owner in the
presentation crate): the stale-dylib warning gains the `aura: ` namespace,
and the three tap/no-nominee notices become `aura: note: ` — the latter
two extracted into pure note-builder fns with unit pins. The two e2e
pins on the cost-model tap string moved with the retag
(research_docs.rs). A new project_nodes e2e exercises the real stale-
dylib load path: fresh build emits nothing, a source newer than the
dylib emits exactly one class-marked warning and still succeeds.
The #313 notice ships on both walk-forward paths: `aura: note: all N
walk-forward windows recorded zero trades` before the summary JSON,
exit 0 untouched (a null result is a valid research result, #198).
Wording and condition live in one shared diag helper taking the
per-window trade counts (orchestrator tidy of the reviewer's
duplicated-predicate residue); the call sites keep only the
path-specific field extraction. Three e2es pin it: the equal-length
degenerate on the synthetic path (exactly one note, no warnings, exit
0, summary unchanged), a traded negative twin, and the real/sugar-path
degenerate gated on the GER40 archive.
refs #278, refs #313
Iteration stderr-markers-1, tasks 1-2 of the stderr-class-markers plan
(spec signed via grounding-check PASS, decisions logged on #278).
The two stderr diagnostic classes become machine-separable (refs #278):
a new crate-internal diag module owns the grammar as note!/warning!
macros (`aura: note: <text>` benign continuing-run diagnostic,
`aura: warning: <text>` recorded fault the run survives); the four
undifferentiated sites are retagged — gate-emptied cell -> note,
failed cell / walkforward cell / mc cell -> warning — and the four
scaffold literals migrate onto the shared macro (output bytes
unchanged). The always-on record summary line gains the bare `aura: `
namespace it was missing; bare-prefix stays the vocabulary for error
lines accompanying a non-zero exit (C14 partition) and plain info
lines. Exit codes unchanged throughout.
Verification: RED->GREEN warning pin on the per-cell-fault e2e; new
gate-emptied e2e (impossible n_trades threshold) pins the note marker,
exit 0, and the summary-line prefix; scaffold e2es and the full
workspace suite green; clippy -D warnings clean on aura-cli. The C28
shell-module roster test gained the diag.rs entry.
Known residue for the next slices: aura-runner literals (stale-dylib
warning, three tap/no-nominee notes) and the #313 zero-trade notice
follow in tasks 3-5. The verb_sugar warning call sites remain e2e-
uncovered (a hostless walkforward fault fixture is structurally
unreachable — single-instrument validation refuses before the cell
loop); their grammar is pinned via the shared macro and the campaign
warning e2e.
refs #278
The fieldtest's spec_gap closed per the #316 decision (tighten, not
ratify): doc_gate normalizes the doc against both the full entry name
and its bare display segment after the last :: -- an extension node
ns::Scale with doc "Scale" now refuses (RestatesName), matching the
register seam's effective bare-name behaviour. Unnamespaced entries
are unchanged (the checks coincide); genuine meaning lines containing
the name as a substring stay accepted (equality, not containment).
The C29 contract sentence carries the precision.
Gates: workspace green (98 suites, 0 failed -- incl. the C29 coverage
walk over every shipped doc under the tightened gate), clippy clean.
refs #316
Executable spec for the fieldtest spec_gap (decision on #316: tighten,
not ratify): doc_gate("lab316_nodes::Scale", "Scale") must be
Err(RestatesName) -- a doc normalizing to the final ::-segment of a
namespaced entry name restates the name. Unnamespaced entries and
genuine meaning lines containing the name stay Ok (guard assertions
green). Currently red: the shipped check compares the full name only.
refs #316
Three of the four immediate fieldtest findings for the self-description
cycle:
- authoring-guide S0: the worked Scale literal gains the required
NodeSchema.doc line (+ the C29 comment) -- the shipped example was a
hard E0063 for anyone pasting it (the exact #314 teaching-failure
class, caught from the consumer seat).
- The op table becomes "The nine ops": the C29 doc op is documented
with its JSON shape, the register requirement, the at-most-once rule
and the gate; the README op-kinds list names it too. The fieldtester
had to guess the shape from commit bodies -- now it is on the public
surface.
- The store-gate refusal's fix hint is surface-neutral: builder
.doc(...), blueprint "doc" member, or op-script {"op":"doc"} --
previously it prescribed the Rust builder to a JSON author.
Gates: registry, graph_construct (63) and project_load (14) suites
green; refusal-prose pins unchanged (pinned substrings untouched).
refs #316
Consumer-seat run over all four C29 seams (load, register, op-script doc,
document description) from the public interface only. The three working
findings verify the #314 failure class removed end to end; the findings
route: guide sync + refusal-prose neutrality + doc_gate short-name parity
this run, discoverability surfaces to the tracker.
refs #316
Cycle-close audit for the self-description cycle. Architect: drift_found
(mild), core holds -- the gate is pure string shape in core imported by
every seam (C28), the engine never evaluates description text (C17/inv
10), determinism pins untouched (C1); the store boundary is gated by
construction with the raw write private; Op::Doc is a closed typed
construct (C25). Regression: aura-bench green (exit 0) after one
cycle-caused infra fix and one ratified re-pin.
Infra fix: the bench campaign seeds (BP_SMA_A/B) predate C29 and were
refused by the store gate (exit 2, infrastructure-first rule) -- both
thread the authored sma_signal meaning line.
Ratify: cli_fixed_cost's run_line_fnv moved because bench_sma_a.json now
carries its doc -- the blueprint's content id shifts exactly as the spec
anticipates for fixtures gaining docs; campaign result fingerprints are
unchanged (descriptions stay out of execution, C29). Baselines re-pinned
via aura-bench pin (throughput baselines re-pinned in the same sweep,
all deltas within tolerance at pin time).
Drift disposition:
- [high] C29 guarantee overclaim vs emit/binding-column vocabularies ->
contract now carries an explicit Scope paragraph naming the residue,
tracked as #321 (also covers the [medium] non-roster schema walk gap).
- [low] Delay's meaning line cited the stale C5 for the delay register;
corrected to C9.
refs #316, refs #321
The cycle's design principle enters the ledger: every closed-vocabulary
entry the binary exposes carries a one-line meaning behind one shared
deterministic shape gate (doc_gate) at three seams -- compile/unit for
engine-shipped entries, load for native node crates, register for the
content-addressed store (documents: an additive-optional gated
description). Forbids the engine evaluating description text (C17 /
invariant 10), any influence on execution/identity/determinism (C1),
retroactive invalidation of registered artifacts, and machine-invented
meaning lines. Why: the audience is headless LLM agents (#319); field
evidence #314 showed schema knowledge being recovered by CAS forensics
from the release binary -- the removed failure class.
All spec acceptance criteria are now met across the four iterations of
this cycle (core carrier + std texts; domain threading; load seam;
register seam + op-script doc slot + document description).
closes#316
Iteration 3 completion (tasks 5-6): the construction vocabulary gains
the slot its own #125 scope-cut comment anticipated. Op::Doc { text }
is the op-script twin of the builder's .doc(...) -- a closed, typed
metadata construct (C25, no logic content); at most one per script, a
second refuses via OpError::DuplicateDoc (the existing duplicate-fault
family), rendered "a doc op may appear at most once". GraphSession
threads the text through Composite::with_doc at finish, so op-built
composites pass the C29 store gate like builder-built ones.
The op-script register test runs on the described variant
(SIGNAL_DOC_DESCRIBED) for both shapes; the doc-less script now pins
the C29 refusal, and the E2E phase added the RestatesName op-script
case. Full gates: cargo test --workspace green (98 suites, 0 failed);
clippy -D warnings clean.
refs #316
Iteration 3 task 4: register_refuses_undescribed_composites_end_to_end
pins the store gate's verb surface -- a doc-less root refuses naming
`sma_signal` and the rule, a doc-less named nested composite refuses
naming `inner`, the described shipped example registers (exit codes and
stderr through the C14 contract). The pre-existing twin pin
graph_doc_field_moves_content_id_but_not_identity_id stays green.
Checkpoint note: the loop escalated this task spec-ambiguous because
its plan gate demanded a fully green graph_construct suite while the
op-script register test is known-open until the doc-op task lands (a
plan-sequencing defect, orchestrator's authoring error); the task's own
work was verified green and is committed here, the doc-op task follows.
refs #316
Iteration 3 of the self-description plan, tasks 1-3 (checkpoint commit
mid-iteration; the register e2e, the op-script doc slot, and the full
gates follow):
- aura-research: DocFault::BadDescription { subject, fault }; a present
document description passes doc_gate in validate_process /
validate_campaign (absent is never a fault and stays byte-identical);
prose arm in doc_fault_prose; tests incl. content-id participation.
- aura-registry: the store boundary gates admission -- the public
put_blueprint parses the canonical bytes and walks root + named
nested composite docs (None refuses like ""), refusing via
UndescribedComposite { name, fault } / MalformedBlueprint; the raw
write survives as private put_blueprint_unchecked (store-mechanics
tests + the never-retroactive witness: doc-less pre-C29 entries stay
readable). Registry-internal seeders triaged junk-vs-real.
- authored root docs: the 4 example blueprints + their 4 open-fixture
twins, the scaffold starter template, the scaled_signal and hl_signal
inline fixtures, and the three programmatic member.rs builders
(breakout, meanrev, hl_channel) via GraphBuilder::doc -- mechanical
propagation of the authored lines.
Recovery note: the implement run's task 3 discarded tasks 1+2's edits
in three files via a file-level checkout (the issue-#23 guard shape);
the loop's boundary snapshot caught it and the content was restored
from recovery snapshot 11a6ae2, then re-verified (research 81 green,
registry 51 green, cli 58/59).
Known-open at this checkpoint: the op-script register test is red --
op-built composites have no doc surface (#125 scope cut); the op
vocabulary grows the anticipated doc slot in the next task of this
iteration (decision logged on the issue).
refs #316
Iteration 2 of the self-description plan: field presence is compile-
enforced since iteration 1, but an extension crate can still ship the
empty-string alibi (or a name-restating doc) the compiler cannot see.
load_crate now walks every charter-checked type id, resolves its
builder, and runs aura_core::doc_gate over schema.doc; a fault refuses
the load via ProjectError::UndescribedVocabularyEntry { type_id, fault }
(exit 1 through the CLI's Display-driven surface, C14).
The variant carries the DocGateFault beyond the spec's minimal sketch
because the spec's error-handling section requires the refusal to name
entry + field + rule -- with only the type id the prose could not
distinguish the two faults (decision logged on #316).
Two gate-failing fixture crates drive the e2e, modelled on
badcharter-project: undescribed-project (und::Opaque, doc: "") for the
Empty arm, restated-project (restated::Echo, doc: "Restated Echo") for
the RestatesName arm -- the second contributed by the E2E phase to
close the only-unit-pinned gap on that arm. A runner unit test pins
both rendered Display arms directly. The described twins (demo-project,
nested-nodes-project) stay green and are now load-bearing for the gate.
Gates: cargo test -p aura-cli --test project_load (14 green, incl. both
refusal e2es); cargo test --workspace green; clippy -D warnings clean.
refs #316
C29 compile/unit seam, tasks 6-10 -- iteration 1 of the self-description
plan complete: every engine-shipped vocabulary entry now carries a
gate-clean one-line meaning.
- aura-runner: the internal tap recording sink threads its doc (the
last unthreaded NodeSchema site) -- first full workspace build since
the field became required.
- aura-cli scaffold: the sample node's template carries a doc line plus
a maintenance comment, so every scaffolded extension crate starts
gate-clean; both extension-vocabulary fixture exemplars thread real
meaning lines (the load seam walks these in iteration 2). New E2E:
the scaffold's own doc must pass its own gate -- an alibi doc in the
template would teach the wrong pattern to every new project, and the
compiler checks only the field's presence, never its shape.
- aura-research: metric_vocabulary()/tap_vocabulary() promote from bare
name arrays to MetricSchema/TapSchema carriers (id + doc, 17 + 4
authored lines); callers adapt to .id with byte-identical output
(introspection goldens unchanged). The #190 guard keeps its triple --
the doc column adds no fourth roster site.
- Coverage: tests/self_description.rs walks all five vocabularies
(blocks, metrics, taps, folds, std nodes) through doc_gate -- the
guard that keeps any future edit from blanking a doc.
Adjudicated review residue: the 17 metric doc strings are surfaced by
no CLI path yet -- deliberate; task 8 pins callers byte-identical, the
surfacing belongs to the CLI self-description work (#315). Plan-cited
coordinates lib.rs:1003/:1283 resolved to the real caller sites
(:1019/:1299); the plan's literal test code was reconciled to the real
accessors (FoldRegistry::core().roster(), Env::std()).
Gates: cargo test --workspace green (0 failed, incl. the new coverage
walk + scaffold E2E); cargo clippy --workspace --all-targets --
-D warnings clean.
refs #316
C29 compile/unit seam, tasks 4+5 of the self-description plan.
aura-engine (task 4): derive_signature stamps doc: "" with the stance
recorded in-code -- a derived composite signature is graph wiring, not
a vocabulary entry; no seam walks its doc, the described surface is the
composite's own doc at the register seam. All in-crate test literals
thread doc: "test-only schema".
Domain crates (task 5): the 12 production builder sites in
aura-market / aura-strategy / aura-backtest carry authored meaning
lines; test sites in aura-backtest / aura-composites / aura-ingest
thread the test-only doc.
Three texts were corrected against the actual node semantics after
quality review rather than kept from the first authoring pass:
SimBroker is the frictionless integrator of held exposure times price
return into cumulative pip equity (no fills/stops/lifecycle -- that is
PositionManagement's domain), the shared cost-node line names the
PM-geometry inputs and both charge modes (AtClose / PerHeldCycle)
instead of a "per-trade gross R" mapping, and LongOnly's line
conditions on its enabled param and speaks port-term "exposure".
Gates: cargo test green for all six touched crates (engine, market,
strategy, backtest, composites, ingest); the workspace-wide gate
follows task 6 (aura-runner is the one remaining unthreaded site,
E0063 by design until then).
refs #316
C29 compile/unit seam, task 3 of the self-description plan: every
aura-std NodeSchema literal threads its one-line doc. Four texts were
corrected against the actual eval/finalize semantics rather than taken
from the plan table verbatim: Delay is a lag-N register (not one-step),
GatedRecorder flushes an ungated final row at finalize, Latch is a
level-sensitive set/reset register (captures no input), SeriesReducer
emits its single summary row at finalize (not per cycle).
Gate: cargo build -p aura-std --lib clean; full std test run follows at
the all-crates gate once strategy/backtest/engine thread their sites
(the earlier per-crate test gate was unsatisfiable in isolation --
std's dev-deps pull crates whose sites belong to later tasks).
refs #316
C29 compile/unit seam, core carrier (iteration 1 of the self-description
spec): DocGateFault { Empty, RestatesName } + doc_gate() — deterministic
string-shape check, never content judgement — and NodeSchema gains the
required pub doc: &'static str, so every construction site must declare
a meaning line (compiler-enumerated). Core's own 10 test literals thread
doc: "test-only schema"; doc_gate/DocGateFault re-exported at crate root.
Downstream crates are E0063-broken by design until their sites thread
the field (tasks 3-7 of the plan); gates here are crate-scoped:
cargo test -p aura-core (59 green) + clippy -p aura-core (clean).
Interim subset commit of the implement run after adjudicating a
cross-task discard: task 2's file checkout had destroyed task 1's
node.rs/lib.rs work; restored from the loop's recovery snapshot
53c55c3 and merged with task 2's field threading.
refs #316
Cycle-close fieldtest for #283 + #77 (tap subscribers), exercising the new
tap surface as a downstream consumer over the public interface only (design
ledger, glossary, cargo doc, example corpus — never the crate sources).
Five examples, one per axis, all built and run against worktree HEAD bd0c557:
- tap_1_record_spread.ops.json — record → runs/traces/graph/spread.json
- tap_4_measure_ic.ops.json — two-tap run → aura measure ic (clean IC)
- tap_consumer/src/tap_2_fold.rs — Named{mean}/Named{count} via run_signal_r
- tap_consumer/src/tap_3_live.rs — Live(closure); bit-identical to fold + record
- tap_consumer/examples/refusal_probe.rs — unknown-label refusal (exit 1, no write)
The consumer crate is a detached workspace (own empty [workspace], path-deps on
aura-core/-engine/-runner), exactly as a real World program; it is not a
workspace member and does not build under `cargo build --workspace`. Verified
here with a fresh `cargo build --bins --examples` (Finished, 0 errors).
No merge-blocking defect: record/fold/live/measure over the shared pair are
correct and deterministic. Follow-ups filed at cycle close:
- #309 (feature) — a plain `aura run` echoes no recorded trace handle (F1+SG1)
- #310 (idea) — Named fold has no data/CLI reach; ratify Rust-only vs selector (SG2)
- #311 (bug) — same-name re-run orphans still-chartable tap files (SG3, pre-existing)
- #297 (comment) — library refusal = process::exit, not Result (F2, already tracked)
Cycle-close audit for the #283/#77 tap-subscribers cycle
(09994b8..6e3f394). Architect review: code contract-clean; what
holds — C1/C7/C8 (empty-output caller-built consumers, (Timestamp,
Cell) zero-heap payload, lifecycle hooks inside the sequential loop,
byte-identical twin-run and streamed-vs-legacy byte-equality pins),
C28 (assembly-crate placement of the writer-holding consumer,
c28_layering guard green), C25 (registry growth = new Rust entry,
data-expressible Named form, roster-enumerating refusals). All drift
items were doc lifts, resolved in this commit:
- C8: the lifecycle is now a symmetric pair — initialize() as
finalize's start-of-stream mirror (infallible by signature;
acquire-in-initialize, degrade-to-inert, surface-once-at-finalize);
history sidecar records the refinement and the #77 resolution the
old Deferred paragraph still carried as open.
- C27: Guarantee and Current state now describe the shipped model —
tap plan (Named{label, params} | Live) over the layered fold
registry, BOTH entry points (run_signal_r AND run_measurement) on
the one shared wiring pair, record streaming at constant memory.
- C28: aura-runner now defines one graph node of its own (the
in-graph record consumer — exactly the graph-meets-persistence
assembly position); the #297 refusal-site count grows ~20 → ~24
(the four tap-plan refusals, already typed before the exit).
- glossary §tap: the subscription/fold-registry sense added; Avoid
line steers probe/monitor/scope away from the observation slot.
Bench (report-only, loadavg 5.8 > 2 on the box — read with care),
all fingerprints OK, no regression:
engine_throughput bars_per_s 14005262 wall_s 0.714s Δ bars_per_s +3.0% wall_s -3.0%
ingest_throughput bars_per_s 12970670 wall_s 0.387s Δ bars_per_s +0.5% wall_s -0.5%
campaign_sweep cpu_percent 2202.0 peak_rss_mb 92.2 wall_s 1.396s Δ cpu_percent +1.5% peak_rss_mb -6.1% wall_s -1.5%
campaign_heavy cpu_percent 2155.0 peak_rss_mb 94.8 wall_s 5.585s Δ cpu_percent +1.8% peak_rss_mb -3.5% wall_s -0.4%
cli_fixed_cost help_ms 1.5ms run_ms 3.6ms Δ help_ms -1.8% run_ms -2.6%
No baseline update (report-only surface; no ratify due).
Correction on 6e3f394's body: "no Probe symbol is introduced
anywhere" — introduced is the operative word; the pre-existing
aura-core test fixture `struct Probe` stays, as the cycle's design
record specified.
Follow-ups on the tracker: #308 (legacy Recorder surface migration),
#307 (binary trace format). Cycle spec/plan working files discarded.
refs #283 refs #77
Completes the #283 registry cycle. run_measurement drops its
duplicated Recorder-bind + post-run try_iter drain for the shared
bind_tap_plan/BoundTaps pair — the mirror of run_signal_r is now the
same code, so the two declared-tap entry points cannot drift; the
measure CLI arm passes TapPlan::record_all() (semantics unchanged,
constant memory). Measure-side library tests pin a fold plan's
one-row trace (streamed mean bit-equal to the slice mean over the
synthetic prices) and the tap-free write-nothing guarantee through
the shared pair.
Run-side integration tests pin the #283 acceptance surface: fold and
live plans agree exactly with the recorded series of the same
blueprint (C1, three runs, three drain policies); a live-only plan
persists nothing; plan refusals (unknown tap, roster-enumerating
unknown label) fire typed before any store I/O; two identical
record-all runs produce byte-identical tap files. The previously
untested `aura: writing tap traces failed` register gains end-to-end
coverage on both arms: store root occupied by a file (pre-run
refusal) and a read-only run dir (deferred initialize-open failure —
the run completes, the failure surfaces terminally, no index.json is
written, the store's treat-as-absent crash shape).
Two forced deviations from the plan's scripted literals, both
verified equivalent: an explicit match replaces expect_err (BoundTaps
is deliberately not Debug — it carries a TraceStreamer), and the test
helper takes &Path, not &PathBuf (clippy::ptr_arg under -D warnings).
Closes#77 as specced: zero per-cycle heap on the tap path
((Timestamp, Cell) payload, committed earlier this branch) and the
Recorder→Probe rename retired (no Probe symbol anywhere; decision
recorded on #77, 2026-07-21).
Verification: cargo test --workspace — 1517 passed, 0 failed;
cargo clippy --workspace --all-targets -- -D warnings clean;
tap_recording 5/5 (byte-compat anchors + both register arms).
refs #283 refs #77
The #283 registry model's assembly seam: tap_plan.rs carries
TapSubscription (Named{label, params} — fully data-expressible — |
Live(closure), the single non-data variant), TapPlan (name →
subscription + default; the plan travels with the FoldRegistry it
resolves against, the Env-at-resolve pattern, so layered hosts inject
vocabulary as data), and the registry itself: eight core entries
(record, count, sum, mean, min, max, first, last), each with a doc
line (help + roster-enumerating refusals) and a scalar-typed param
schema (all v1 entries param-less; the seam ships now because it sits
in every entry's build signature). Typed TapPlanError refusals fire
pre-bootstrap, before any store I/O; the terminal Store variant keeps
the established "writing tap traces failed" register bytes.
bind_tap_plan/BoundTaps is the shared wiring pair both declared-tap
entry points call; run_signal_r is re-wired onto it here (Recorder
bind + post-run try_iter drain replaced — no O(cycles) retention, no
per-cycle heap), with all six callers threaded (both CLI run-arm and
the four example-identity test calls pass TapPlan::record_all()).
run_measurement follows in the next commit.
Registry growth stays a new Rust entry (C25); a layer-registered
entry is pinned by the injection test, param validation by a
schema-carrying test entry (missing/mismatched/unknown all refused
with named messages).
Verification: cargo build --workspace clean; aura-runner 85+ tests
green incl. both example-identity anchors; tap_recording byte-compat
anchors green (recorded traces byte-identical, tap-free writes
nothing, duplicate-tap refusal unchanged); clippy -D warnings clean.
refs #283 refs #77
Two #283 groundwork moves for the fold-registry model:
- TraceStreamer::register_tap returns a TapWriterOpener — everything
needed to open one tap's incremental writer later, from inside a
consumer node's initialize, without holding the streamer. tap_writer
is re-expressed as register_tap(..).open(), so there is exactly one
open path; pinned by a byte-equality test between the two.
- TapRecorder leaves aura-std (which must not know aura-registry, C28
layering) and is reworked in aura-runner as the record entry's
consumer: it holds the TapTraceWriter in-graph — initialize opens
(deferred; an open error is stored and the node degrades to inert),
eval appends per fired warm cycle (zero heap, no thread, no
channel), finalize finishes the writer and reports exactly one
terminal outcome. The SyncSender-based aura-std variant and its
planned writer-thread protocol are retired; newest_cell becomes pub
for the assembly crate.
Verification: cargo clippy --workspace --all-targets -D warnings
clean; aura-core/engine/registry/std/runner suites green (new tests:
opener equivalence, happy-path capture + single Ok outcome,
failed-open inert + single Err outcome).
refs #283 refs #77
Adds the second lifecycle hook the #283 design revision pairs with
finalize: Node::initialize (default no-op) runs once per node, in
topological order, before the first source value — inside the
deterministic sequence, no within-sim concurrency (C1). Consumers
acquire run resources there (the record consumer opens its streaming
writer); the hook is infallible by signature, so an acquiring node
stores its error, degrades to inert, and surfaces the failure once,
terminally, at finalize.
Every existing node keeps the default. Pinned by the mirror test
run_initializes_every_node_once_before_the_stream (init0,init1
strictly before any eval, exactly once), the shape of the existing
finalize test. RED was observed by temporarily reverting the flush
(test fails with evals-before-inits), then re-applied.
Ledger note for cycle close: C8's lifecycle section gains initialize
as finalize's mirror.
refs #283
Adds the incremental write path the tap-subscriber cycle streams
record subscriptions through: TraceStore::begin_run opens a run
directory and returns a TraceStreamer; per tap, TapTraceWriter
appends (ts, cell) points into two temp column streams and finish()
assembles the canonical ColumnarTrace JSON by streaming
concatenation (io::copy), removing the temps. index.json is written
last by TraceStreamer::finish, preserving the store's crash
discipline (no index -> NotFound, treat-as-absent).
Compatibility is pinned as an equality property, not a frozen byte
assumption: a new test asserts the streamed output byte-identical to
TraceStore::write's for the same rows (including empty and bool
columns), so the two write paths cannot drift and every reader
(read, read_family, chart) stays untouched. TraceStore::write is
refactored onto shared write_tap_file/write_index helpers used by
both paths. Temp streams open read+write (OpenOptions), since
finish() seeks back and re-reads them.
Verification: cargo test -p aura-registry — 78 passed (3 new:
byte-equality incl. empty+bool, crash shape NotFound, caller-order
index).
refs #283
The three C8 sinks realizing the per-tap subscription model (#283): TapFold
(closed FoldKind vocabulary Count|Sum|Mean|Min|Max|First|Last, in-graph
accumulation, exactly one finalize row), TapRecorder (bounded
SyncSender<(Timestamp, Cell)>, lossless backpressure — a full channel blocks
the sim, wall time never state), and TapLive (consumer-owned closure, inline
on the sim thread). All three move (Timestamp, Cell) — 16 B, Copy — instead
of (Timestamp, Vec<Scalar>): zero per-cycle heap on the tap path (#77 part
2). The shared kind-dispatched newest-cell read lives in tap_cell.rs.
Known-good subset commit of iter tap-subscribers (tasks 1-3 of 7, each gated
implementer -> spec-compliance -> quality DONE; the loop blocked on a task-4
plan defect, repaired separately): the trace-store streaming writer, the
TapPlan wiring seam, and both entry-point rewires follow in this same cycle.
Verified directly: cargo test -p aura-std — 109 passed, 0 failed.
refs #283, refs #77
The single-file ledger had grown to 2968 lines / ~42k tokens, mixing
current design law with accreted history: 59 cycle-stamped realization
blocks, 18 [HISTORY] passages, 22 supersession markers, and the C10 /
C22 / C24 reframe sagas layered several supersessions deep. A
code-grounding audit (31 agents, adversarially verified) confirmed 11
defects stated as current truth: stale crate homes from the C28 #288
roster split (cost nodes, PositionManagement, PositionEvent, Session),
the renamed InputSpec->PortSpec, the pre-#241 project model in C16 and
the open-threads section, a stale HarnessKind retirement deferral in
C24, and three C28-internal inconsistencies.
New shape, per the ailang precedent:
- INDEX.md stays the sole addressable entry point: foundation, external
components, a C-id-keyed contract map (one line per contract), and
only the genuinely open architectural threads.
- contracts/cNN-<slug>.md carries each contract's current truth only:
Guarantee / Forbids / Why with ratified refinements integrated, plus
a code-anchored Current state. All confirmed defects are fixed here;
crate anchors were re-verified against the tree.
- contracts/cNN-<slug>.history.md (18 sidecars) and INDEX.history.md
preserve every superseded block verbatim, stamps and issue refs
intact, under a frozen-record banner. Nothing was deleted: superseded
design intent remains an addressable working-tree artifact, off the
per-cycle audit walk.
- Ledger discipline is now stated in INDEX.md: live files are edited in
place at cycle close, superseded text moves verbatim to the sidecar,
and a supersession marker in a live file is itself an audit finding.
Every contract file was verified against its old text by an independent
zero-loss pass (statement-by-statement) plus a code-accuracy spot check;
C-ids and contract titles are unchanged, so existing C-id citations in
code, tests, and issues resolve as before.
aura {process|campaign} show now accepts the 8-hex prefix every surface
prints: an exact match is tried first, then the prefix is resolved
against the store's candidate ids — a unique match prints the identical
stored bytes as the full id (#164 byte discipline), an ambiguous prefix
refuses listing every sharing candidate, and a genuinely-unknown id
keeps its established refusal. The resolution mirrors the semantics
reproduce shipped with #298, closing the asymmetry the #300 fieldtest
named (df_4). Registry gains list_process_ids/list_campaign_ids
(missing store dir reads as empty, matching the blueprint-scan
discipline); campaign_not_found_prose becomes the single source for
show_campaign's two not-found arms. The ambiguity arm is pinned by a
unit test at the resolution seam — hash-derived ids make a
shared-prefix E2E fixture impractical.
aura {process|campaign} show is exact-match-only while every surface
prints 8-hex prefixes and reproduce has resolved prefixes since #298;
the fieldtest's prefix paste (df_4) hits the same flat not-found prose
as a genuinely-unknown id. The E2E test pins the wanted headline: a
unique 8-hex prefix of a registered id prints byte-identically what the
full-id show prints (#164 byte discipline), both verbs. The ambiguity
refusal is deliberately left to a unit test at the resolution seam —
hash-derived ids make a shared-prefix E2E fixture impractical.
An unknown select rule in a process document now refuses with the
accepted vocabulary beside the quoted offending value:
block std::sweep: unknown select rule "plateau" (accepted: argmax |
plateau:mean | plateau:worst)
mirroring the roster-naming metric-refusal shape. The roster is derived
from slot_kind_label(SlotKind::SelectRule) — CLI describe output and
the document refusal stay in sync by construction — and a companion
test pins the "select rule: " prefix that derivation trims, so a
label-wording drift breaks a test instead of silently degrading the
prose. No alias-hint sentence: the enumeration alone rescues the
bare-plateau paste case (fieldtest df_3), and the document schema stays
strict per #300 F7.
A process document carrying an unknown select rule refuses correctly
but names no accepted vocabulary; the fieldtest's bare-plateau paste
(df_3) hits an opaque wall. The unit test pins the wanted property at
the site that builds the prose (select_from): the refusal enumerates
argmax, plateau:mean, plateau:worst beside the quoted offending value.
No alias-hint sentence — the enumeration alone rescues the paste case
and the project's refusal prose stays sparse; the document schema stays
strict per #300 F7.
register_process and register_campaign now carry the same
env.provenance() no-project guard as their show/runs siblings, refusing
with exit 1 and the missing-Aura.toml prose instead of letting
Env::runs_root()'s relative fallback materialise ./runs/ in an innocent
cwd. Five pre-existing register tests ran from a bare temp cwd and were
reaching their asserted behaviour only through this bug; they now run
inside the fresh_project() fixture, preserving their original property
(content addressing, intrinsic-validation refusals) in a real project.
Verified: the RED E2E test (register_outside_a_project_refuses_and_
leaves_no_store, both verbs, refusal + no side-effect store) observed
GREEN by the loop's independent verify alongside the full suite.
register_process/register_campaign call env.registry() without the
no-project guard their show/runs siblings carry, so Env::runs_root()'s
relative fallback silently creates ./runs/ in an innocent cwd and
registers the document there (exit 0). The E2E test pins both verbs and
both halves of the wanted refusal: exit 1 naming the missing Aura.toml,
and no store directory created as a side effect.
show_process/show_campaign now guard env.provenance().is_none() before
touching the registry — the guard their siblings (campaign_runs,
validate_campaign_file) already carry — refusing with the honest
degradation prose ('needs a project ... no Aura.toml found up from
{cwd}') instead of blaming the id with 'not found in the project
store' against a store that does not exist. Refusal exit 1 unchanged.
The register-verb sibling gap (silently creating a stray ./runs/) is
deliberately NOT touched here — tracked as #305.
RED-first: the in-process pin landed at 803c970; independent verify
re-ran it green with the suite.
closes#304
Outside any project, show_process/show_campaign print 'not found in the
project store' — asserting a store that does not exist and blaming the
id, where the sibling verbs (campaign_runs, validate) degrade honestly
naming the missing Aura.toml. The in-process RED pins: the refusal
stays Err, its prose names 'no Aura.toml found up from', and does NOT
claim 'not found in the project store'. Diagnosed cause: both show fns
lack the env.provenance().is_none() guard their siblings use, so the
no-project fallback store's Ok(None) masquerades as an unregistered id.
refs #304
Downstream-author fieldtest of the #300 surface, public interface only,
real GER40 archive. All four axes hold: the closed
generate -> show -> hand-extend -> re-register -> run loop (2-cell
extended campaign), the typed stop grammar's clap rejections, bare
plateau's content-id identity with plateau:mean, and the show read-back
(byte-identical, no framing) with a clean unknown-id refusal.
Findings routed: 1 bug (show outside a project blames the id instead of
the missing Aura.toml — RED-first fix follows), 2 friction items filed
to the tracker (opaque process-document select refusal; no prefix-id
hint on show), 4 working (carry-on). The fieldtest spec survives
git-ignored as the next planning cycle's reference input.
refs #300
Architect review (cycle 757e3ac): core clean — show is pure shell
dispatch over Env::registry() (C28); both ledger amendments match the
F1-F8 decision log exactly; document schema untouched; the #164
no-framing pin is stricter than the spec snippet asked. Four peripheral
drift items, resolved:
- fix: docs/authoring-guide.md verb rosters now name show (process) and
show|runs (campaign) — the guide teaches exactly the
generate->retrieve->extend loop this cycle closed.
- fix: docs/glossary.md campaign/process document entries carry the
settled verb sets (record-reality glossary write).
- fix: show_process's not-found refusal now sources its prose from
ref_fault_prose(ProcessNotFound) instead of a hand-copied format!
(byte-identical, test-pinned; the campaign-side prose has no RefFault
sibling and stays the single campaign instance).
- ratify (log-only): the shipped show helpers use print! (no framing,
#164) where the spec snippet sketched println! — the code is the
stricter, correct form; recorded on the reference issue's decision
log. The spec is a git-ignored working file swept at this close.
Regression: aura-bench report-only run, all fingerprints OK, deltas in
noise (host load 9.6 noted; engine/ingest/campaign/cli lanes all within
±5%). Full aura-cli gates re-run green after the tidy edits.
refs #300
The #300 keystone cycle: the residual run-quintet flags were already
document vocabulary on the --real arm (#210/#220); this closes the loop
and ratifies the surface.
- `aura process|campaign show <content-id>` prints a registered
document's canonical bytes (print!, no framing — #164), so
generate -> retrieve -> hand-extend -> re-register needs no direct
store filesystem access. Refusals reuse the store-hint prose, exit 1.
- walkforward/mc --stop-length/--stop-k become clap-typed i64/f64
(generalize's form); stop_knob_or and its parse_csv_list helper are
deleted (multi-value/invalid input is now clap's exit-2 rejection —
the two dissolved-verb refusal tests retarget, the type-dead unit
test is deleted).
- --select accepts bare `plateau` as the documented default
plateau:mean (#227 carry); document schema untouched — the two
spellings generate content-id-identical campaigns (test-pinned).
- family.rs's quadruplicated demo stop literal collapses onto one
DEFAULT_STOP const (maintainer-lens drift finding).
- Ledger: C25's "which projection next" line resolved (document-first
delivered, per-verb identity re-ratified — F8; host/MCP stay
demand-driven); C18 records the read-back.
Design triage, fork decisions F1-F8, and the auto-signed spec live on
the reference issue. Verification: full workspace suite 1482 passed /
0 failed with the real GER40 archive exercised (the byte-identity
round-trips ran, not skipped); clippy -D warnings clean.
closes#300
GREEN for 78e68e6's RED. aura_runner::reproduce::load_family resolved
only the derived C18 handle '{family}-{run}' while the registry's
enumeration exposes the bare name factor — the natural
list-then-reproduce library workflow dead-ended (found by the #295
fieldtest; predates the extraction byte-identically). Resolution now:
exact-handle match keeps first precedence; a bare name naming exactly
one stored run resolves as fallback; an ambiguous name (several runs,
or a name/handle collision) refuses without guessing, listing the
candidate handles. Two additional tests pin the ambiguity refusal and
the handle-precedence-on-collision semantics.
Ledger record-reality rider: the C18 #158 paragraph claimed reproduce
refuses an unknown id with exit 2; the code has always exited 1 —
correct per C14's partition (the id names missing recorded state, a
runtime failure, not a usage error). The paragraph now records the
actual behaviour and the new name-fallback semantics.
Verification: cargo test --workspace green (1477 passed, 0 failed);
clippy --workspace --all-targets -D warnings clean.
closes#298
Registry::load_family_members() exposes the bare name factor while
aura_runner::reproduce::load_family exact-matches only the derived C18
handle '{family}-{run}' — the natural list-then-reproduce workflow
dead-ends on its first attempt (found by the #295 fieldtest; predates
the extraction byte-identically). The test pins: the enumerated
identity of an unambiguous single-run family resolves through
load_family to the same family the canonical handle names.
refs #298
Downstream-consumer corpus for the #295 library surface: a standalone
workspace (path-deps, no aura-cli in the link graph — cargo tree
verified) exercising the four cycle axes as a World program would,
public interface only: a DefaultMemberRunner member run over real
GER40, a hand-authored campaign document through
aura_campaign::execute, the IC + registry deflation dispatch as
library calls, and load_family/reproduce round-trips. Run transcripts
committed beside the sources.
Findings routed separately: the family-id mismatch between the
registry enumeration and the reproduce keying (list-then-reproduce
dead-ends) to debug; the reproduce-API friction (no report on the
simple path; DataSource/pip re-specification the manifest already
holds, wrong pip silently DIVERGED) to a follow-up. Four working
confirmations, chief among them the cycle's acceptance claim: all
four axes compiled first-try from the public docs alone.
refs #295
Cycle-close audit (architect: drift_found; design core clean). What
holds, architect-confirmed with the guard run and the diff read in
full: the shell boundary is real and enforced (full-workspace
c28_layering incl. completeness + shell-content checks, acceptance
grep clean — no member-run symbol left under crates/aura-cli/src);
aura-campaign keeps zero production dependency on the runner; the IC
move is verbatim against the anchor; the new library-only E2E pins C1
(two independent runners, byte-identical RunReport).
Drift items, all resolved as fixes in this commit:
- design ledger: the C28 status sentence "No crate exists yet for
measurement" contradicted this cycle's own phase-3 done line — now
records aura-measurement as seeded (#295), execution still unbuilt.
- the #147 registry-dispatch acceptance tests (IC deflation null,
seeded determinism, cross-vocabulary refusal) were pure library
properties stranded in the shell's test module — relocated to
aura-measurement/tests/registry_dispatch.rs (engine/registry as
dev-only, layering-exempt edges), per the cycle's own tests-move-
with-their-module rule.
- aura-runner's direct production edge on the external data-server
tree was structurally invisible (the direction table checks aura-*
keys only) and unrecorded — the C28 assembly prose now names it, and
a new c28_layering test pins the external tree's entry points to
exactly {aura-ingest, aura-runner, aura-cli}.
- the C28 shell prose named three buckets that did not cover two real
shell residents — it now names the op-script construction front-end
(translation) and the `aura new` scaffolder (authoring-tooling,
shell-resident like rendering until a second consumer wants it).
Deliberately deferred, tracker homes exist (not drift): #297
(process::exit sites in aura-runner's single-run verb paths), #294
(IC duplicate-timestamp semantics), #288-era rustdoc unresolved-link
warnings in aura-std/aura-backtest (pre-existing byte-identically at
the anchor).
No regression scripts are configured (the bench is report-only); the
architect review is the gate. Verification: cargo test --workspace
green (1474 passed, 0 failed) incl. the relocated registry-dispatch
tests and the new data-server guard; clippy --workspace --all-targets
-D warnings clean.
refs #295
Completes the shell-boundary cycle (tasks 10-13):
- c28_layering now enumerates the FULL workspace: every crate row is
reconciled against its real production [dependencies], a completeness
assertion pins the table to the on-disk crates/ set (a new crate can no
longer escape the guard silently), the shell/assembly imported-by-
nothing rule is asserted, and a shell-content check pins aura-cli to
no-[lib] plus a closed allow-list of argv/translation/presentation
modules. A new domain module in the shell now fails the suite.
- Ledger amendments: C28 gains the assembly position (aura-runner) and
the corrected import-rule prose (the ratified aura-campaign ->
aura-backtest production edge, #291/#292); phase 3 is marked done with
the #297 process::exit residual named; a provenance note records this
as structural-debt closure, not a demonstrated downstream blocker.
C25 records the control-surface decision: the text artifact vocabulary
is canonical, every control surface (CLI executor verbs, a future
host, an MCP face, a World program) is a projection/executor over it —
the which-projection-next ranking deliberately open on #295. C14 gets
the executor-face amendment, C26 the binding-module relocation.
- aura_campaign::MemberRunner's doc names the shipped default
implementor (aura-runner::DefaultMemberRunner) while the column keeps
zero dependency on it.
- New library-only E2E fixture (aura-runner/tests/world_member_run_e2e):
runs the canonical member recipe over a tiny synthetic archive with no
aura-cli in the link graph, and pins C1 determinism (two independently
constructed runners produce a byte-identical RunReport).
- campaign_run.rs module doc: intra-doc MemberRunner link re-anchored to
aura_campaign::MemberRunner (the impl moved out with part 1).
Verification: cargo test --workspace green (1473 passed, 0 failed);
clippy --workspace --all-targets -D warnings clean; cargo doc clean of
NEW warnings (the five unresolved-link warnings in aura-std/aura-backtest
predate this cycle byte-identically at the anchor — #288-era doc drift,
left for the cycle audit).
refs #295
The shell no longer defines what an aura backtest is. Tasks 1-9 of the
shell-boundary cycle — structural extraction, behaviour byte-identical:
- aura-runner (new; the C28 assembly position): input binding (the C26
module, moved whole), the C1-load-bearing param<->config translators,
harness assembly (wrap_r / run_signal_r / run_blueprint_member and the
probe/reopen cluster), axis + risk-regime conventions (bind_axes et
al.), the campaign family builders + MC guards, reproduce
(process::exit -> RunnerError, shell remaps to identical bytes), the
measurement-run orchestration, project loading (Env / cdylib load /
charter / staleness, moved whole), the coverage gap-walk (deduplicated
onto interior_gap_months), and DefaultMemberRunner — the public
implementation of aura_campaign::MemberRunner (ex CliMemberRunner).
The MemberRunner trait stays in aura-campaign; the column still
imports no harness/data-binding machinery.
- aura-measurement (new; seeds C28 rung 3): the IcMetrics/IcKey
vocabulary + information_coefficient, verbatim incl. serde derives
(#294 duplicate-timestamp semantics move as-is, unresolved).
- aura-backtest: the pure per-run scaffold (point_from_params,
wf_ms_sizes / fit_wf_ms_sizes, intersect_shared_window).
- shell residue: main.rs / campaign_run.rs keep argv/dispatch,
argv->document translation, and presentation only;
walkforward_summary_json_from_reports now calls the public
aura_engine::param_stability instead of its inline twin.
- worked example (crates/aura-runner/examples/world_member_run.rs): a
World program runs a member backtest through the library alone.
Held quality findings, adjudicated: the plan's literal pub-visibility
list for binding.rs kept (cosmetic); ~20 refusal sites inside
aura-runner (family/member/measure/translate) still process::exit —
behaviour-identical today, conversion to RunnerError is filed forward
(refs #297); rustfmt line-width drift on re-pathed call sites left for
a repo-wide fmt decision.
Verification: cargo test --workspace green (1471 passed, 0 failed);
cargo clippy --workspace --all-targets -D warnings clean. Byte-identity
is pinned by the untouched shell E2E suites (cli_run, measure_ic,
run_measurement, research_docs, run_refuses_unrunnable_blueprint,
tap_recording — zero assertion edits).
Remaining in this cycle: full-workspace c28_layering + shell-content
check, ledger amendments (C28 assembly position, C25/C14 control-surface
consequence, C26 realization note).
refs #295
Cycle-close audit (architect: drift_found, design core clean). What holds,
architect-confirmed: C18 bit-identity for RunMetrics (prose derived
byte-identically, rng order preserved, registry suite unchanged); the C10
wall (monomorphic R-gates, r_based in the R vocabulary, cross-vocabulary
refusal tested both ways — no leak); C28 direction (trait in aura-analysis,
vocabulary supplied from the outer rungs, zero Cargo edge changes); C1
determinism pinned through the generic path.
Drift items, all resolved as fixes in this commit:
- design ledger: the C28 #147 disposition now records item 2 SHIPPED (the
A1 cut) with A2 still deliberately deferred; the #136 one-implementor
clause carries a supersession note (the IC is the second implementor).
- seven stale doc comments describing the pre-#147 or mid-cycle state
(analysis trait + estimator docs, registry check_r_metric C9 claim,
engine re-export note, campaign PER_MEMBER_METRICS roster note,
member-seam guard comment, research vocabulary note) updated to the
shipped state.
- R_BASED_METRICS is now oracle-pinned against RunMetricKey::r_based()
in the vocabulary test (it feeds the NonRMetric refusal prose; a
divergence would have misreported the R-gate silently).
Noted, not amended (history stays): commit 6744f67's body says 'three
syntax-only edits' where the test module actually took five.
No regression scripts are configured (the bench is report-only); the
architect review is the gate. Verification: extended vocabulary test,
campaign suites, analysis/research suites green; clippy --workspace
-D warnings clean.
refs #147
The A1 cut of #147 item 2, part 2 of 2 (the consumers), completing what
part 1 (the MetricVocabulary substrate) set up:
- aura-campaign's RANKABLE_METRICS hand-copy becomes a re-export of the
backtest vocabulary's roster; the #190 guard test extends to pin the
roster nesting (rankable == vocabulary roster, rankable within
per-member, per-member within the research vocabulary) and the legacy
exposure_sign_flips alias, so any drift fails a test loudly. The
research and per-member rosters themselves stay hand-copies by ratified
isolation (#190 accepted residual) — now oracle-pinned instead of trusted.
- aura-cli gains IcMetrics + IcKey implementing MetricVocabulary: the
second PRODUCTION implementor. The standalone `aura measure ic`
reduction now computes its permutation p via the shared null_draw +
one_sided_p_laplace building blocks (independent draws instead of
chained shuffles — statistically equivalent, deterministic per seed;
emitted p values may differ in the last digits from #290's).
- Acceptance proven at the registry boundary: an engineered
SweepFamily<IcMetrics> deflates through optimize_deflated to a LOW
overfit probability and a noise family to a HIGH one (the permutation
null, dispatched by the vocabulary — the deflation machinery is no
longer baked-in R-only); same seed reproduces the identical
FamilySelection; cross-vocabulary names are refused with the family's
own roster in the prose (the C10 guard). A new cli_run E2E exercises
the legacy alias through the real `aura runs family … rank` boundary.
Deviation from the plan, forced by the compiler: IcMetrics additionally
derives Debug (unwrap_err on the deflation Result needs it; mirrors
RunMetrics).
Verification: full workspace suite green (no FAILED, no compile errors),
clippy -D warnings clean, C18 goldens (cli_run, measure_ic, c28_layering)
green unchanged. A2 — measurement runs as sweep-family citizens — stays
deliberately deferred until a concrete campaign demand exists.
closes#147
The A1 cut of #147 item 2, part 1 of 2 (the substrate): the registry's
ranking/deflation machinery no longer bakes in the R metric vocabulary.
- aura-analysis gains the deliberately narrow MetricVocabulary trait (name
resolution, roster, direction, value read, one null draw) and the
single-sourced one_sided_p_laplace estimator; re-exported through
aura-engine's existing seam (no Cargo.toml edge changes anywhere).
- aura-backtest now OWNS the R vocabulary: RunMetricKey (was the registry's
private Metric enum, exposure_sign_flips alias preserved), r_based (the
C10 classification), R_BASED_METRICS/RANKABLE_METRICS rosters, the
centred moving-block null_draw, and member_metric_from_rs.
- aura-registry: metric_cmp/rank_by/null_best_of_k/member_sd generic over
M: MetricVocabulary; optimize/optimize_plateau/optimize_deflated over
M: MetricVocabulary + Clone (SweepPoint<M>: Clone bounds on M: Clone);
the private vocabulary deleted; check_r_metric/generalization stay
monomorphic on the RunMetrics vocabulary (the C10 wall by type);
RegistryError::UnknownMetric carries the vocabulary's roster so the
refusal prose derives instead of being hand-strung (byte-identical for R).
Bit-identity (C18): rng consumption order, centring, Laplace arithmetic,
floors, tie rules, and refusal prose are all unchanged for RunMetrics
inputs — the full registry suite (75 tests) passes unchanged except three
syntax-only edits for the UnknownMetric variant reshape; clippy -D warnings
clean on the three touched crates. Alternatives considered: a vocabulary
table in aura-core (rejected: inverts C28's supplied-from-outer-rungs
direction and cannot host the null model); a parallel measurement-side
deflation path (rejected: duplicates the machinery and adds a fifth
roster). Part 2 lands the consumers (campaign roster single-sourcing, the
IC implementor, acceptance tests).
refs #147
Give a measurement run a standalone post-run quality score: the Information
Coefficient (IC), corr(signal_t, forward_return_{t+h}), with a permutation
null model. Before this a measurement run persisted only tap names and series
(MeasurementReport) — inspectable but not rankable or deflatable. IC is the
first metric measurement supplies: the tap-to-scalar reduction the C28
process-column / metric-interface seam anticipates ("a named metric vocabulary
supplied by measurement and backtest instead of baked in R-only").
Surface: `aura measure ic <run> --signal <tap> --price <tap> [--horizon]
[--permutations] [--seed]`. It reads the run's two recorded tap traces, builds
the forward-return series offline over the recorded price spine, aligns
signal -> return by exact timestamp, computes the IC, and prints it with a
one-sided permutation-null overfit probability (Laplace-smoothed, mirroring the
R deflation arm's formula in optimize_deflated).
Placement follows the C28 ladder: the generic pieces — pearson_corr and a
Fisher-Yates permute over SplitMix64 — land in aura-analysis (the domain-free
statistics foundation, beside resample_block); the IC semantics (forward-return
horizon, the offline join) and the CLI verb land in aura-cli (the shell). The
run path (run_measurement), MeasurementReport, the trace store, and the registry
metric vocabulary are untouched — existing `aura run` output stays byte-identical
(the C18 golden run_prints_json_and_exits_zero stays green).
Causality (C2): the forward-return read close_{t+h} is over a completed run's
recorded trace, not an in-graph node. A look-ahead signal is structurally
impossible in a causal run, so the engineered-vs-noise acceptance property is
unit-tested over hand-built offline series (perfect signal -> IC ~= 1, p < 0.05;
a varying but exactly-uncorrelated signal -> IC = 0, not significant;
constant/degenerate -> the (0.0, 1.0) floor), while the E2E validates the CLI
wiring, report well-formedness, determinism, and the error paths over a real run.
This lands the second, structurally distinct null-model computation (permutation,
beside the moving-block bootstrap behind R). That makes the deferred registry
deflation-vocabulary abstraction (#147 item 2) demand-driven rather than
speculative — the next cycle.
Verified: cargo test --workspace green (incl. the new aura-analysis unit tests,
the aura-cli ic_* unit tests, and the measure_ic E2E); cargo clippy --workspace
--all-targets -D warnings clean; the C18 byte-identity golden unchanged.
closes#290
refs #147
C28 phase 2 (Stratification); realizes item 1 of the deferred #147. The
engine's production surface no longer names a backtest-metric type:
- RunReport becomes generic over its metric payload M; sweep/mc/walkforward/
blueprint thread the parameter (SweepPoint<M>, SweepFamily<M>, WindowRun<M>,
WalkForwardResult<M>). RunManifest stays concrete and engine-owned (its
selection: Option<FamilySelection> embeds the foundation-grade analysis type).
- summarize and the MC assembly (McDraw/McFamily/McAggregate/RBootstrap/
r_bootstrap/monte_carlo) move to aura-backtest - McAggregate::from_draws reads
RunMetrics fields by name, so generifying it is the phase-6 metric-vocabulary
abstraction (#147 item 2), still deferred; wholesale relocation is the honest
cut. The concrete instantiation lives in aura-backtest as
`type RunReport = aura_engine::RunReport<RunMetrics>` + sibling aliases.
- the statistics kernel (MetricStats/quantile/resample_block/SplitMix64) moves
to the aura-analysis foundation; the engine re-imports it (inner->foundation,
legal) and re-exports it so existing consumers stay source-compatible.
Dependency inversion in one commit: aura-engine drops aura-backtest from
[dependencies] (back to dev-deps for its SimBroker/RunMetrics test fixtures);
aura-backtest gains aura-engine. Cycle-free for lib targets - the cycle closes
only through the engine's dev-dep edge, the pattern aura-vocabulary already
uses. aura-backtest reaches the kernel transitively through the engine
re-export, so no aura-backtest -> aura-analysis edge exists (the C28 ladder
permits backtest -> {core, engine} only). run_indexed / SplitMix64::next_f64
widened pub(crate) -> pub for cross-crate use.
Consumers (registry/campaign/cli/composites/ingest/bench) rewired by import
path only, no call-site logic changed. The c28_layering structural test extends
to the full ladder: aura-analysis (no aura-* deps), aura-engine ⊆ {core,
analysis}, aura-backtest ⊆ {core, engine}.
Behaviour-preserving: 1448/0 tests, clippy -D warnings clean, serde shapes
byte-identical (C18 - RunReport<M> keeps field order manifest,metrics; the
CLI pre-serialized-splice contract unchanged), moved code traceable via git
rename detection. Cycle-introduced broken intra-doc links fixed.
closes#292
C28 phase 5 (Stratification): the backtest reductions - RunMetrics, RMetrics,
summarize_r, r_metrics_from_rs, and the position-event table - move verbatim
into aura-backtest::metrics, beside the position_management producer whose
record layout they read (the r_col/cost_col lockstep contract is now
intra-crate; its stale "aura-std" comment corrected). aura-analysis is reduced
to the domain-free half: the multiple-comparison hurdle math (inv_norm_cdf,
expected_max_of_normals) and the selection-provenance types
(FamilySelection/SelectionMode - kept beside the statistics so
RunManifest.selection embeds a foundation-grade type), [dependencies] = serde
only.
The engine re-export surface is name-unchanged (report.rs re-imports from both
crates), so no indirect consumer needed a source edit; aura-backtest moves from
aura-engine's dev-dependencies into [dependencies] as a commented TRANSIENT
widening of the C28 violation, removed by the phase-2 edge cut on this branch.
The campaign drift guard imports its pinned types from their new source
crates. Behaviour-preserving: 1448/0 tests, clippy -D warnings clean, serde
shapes byte-identical (C18), moved code traceable via git copy detection.
closes#291
Phase 4 of the Stratification milestone. aura-std held four C28 ladder
layers in one roster; this cuts them into layer-aligned, aura-core-only
node crates so the import direction is enforced by the crate graph:
- aura-std — engine nodes only (arithmetic/logic/rolling + sinks)
- aura-market — session, resample
- aura-strategy — bias, stops, sizer, cost-model machinery
- aura-backtest — sim_broker, position_management
- aura-vocabulary — the relocated closed std_vocabulary roster
Node modules move verbatim (byte-identical renames); consumers are
rewired by import path only. A new structural test
(aura-vocabulary/tests/c28_layering.rs) asserts each node crate's
[dependencies] stay within its C28-permitted inner set, catching the
acyclic-but-outward violation the compiler misses.
Behaviour byte-identical: full workspace suite green (1448 tests), no
golden edited, clippy -D warnings clean. C28 Status block updated.
closes#288
Phase 3 of the Stratification milestone (contract C28). `aura run` now
shape-dispatches on the loaded blueprint: a `bias` output takes the
existing strategy path (wrap_r + R evaluation), byte-identical; a
blueprint that declares taps but exposes no `bias` runs bare via the new
`run_measurement` (bootstrap, run, drain the declared taps) with no
SimBroker, no risk executor, and no per-cycle equity/exposure/r
recorders; a blueprint exposing neither is refused. This gives the
measurement layer its own run verb and removes the O(cycles) recording
retention (the measured ~2 GB peak RSS over 12y UK100 M1) the mandatory
R scaffold imposed on a measurement-shaped run.
A new `MeasurementReport { manifest, taps }` (aura-engine::report, beside
RunReport) is the stdout/record shape — no R metrics, since a bare run
has no broker. The manifest's `broker` carries an interim "measurement"
sentinel; the honest `Option<String>` model is deferred to the #147
metric-genericity block, keeping this phase #147-free and the
strategy-path C18 goldens byte-identical.
`run_signal_r` is untouched (the tap machinery is duplicated into
`run_measurement`, not extracted). The measurement arm's closed-blueprint
guard uses `signal.param_space()` rather than `blueprint_axis_probe`,
which welds wrap_r and would panic on a no-`bias` signal.
refs #286
claude