The aura-research leaf crate: ProcessDoc/CampaignDoc with strict parsing
(hand-rolled StageBlock deserializer — deny_unknown_fields does not compose
with internally-tagged enums; Axis declares its ScalarKind once with bare
values, per the #189 user veto), canonical form + content_id_of (the
SHA-256 primitive moved from aura-cli, which now delegates — the id pin set
incl. the independent sweep-store recompute stayed green untouched),
intrinsic validation for both document types (P1 constraints structural),
and the introspection contract (schema tables single-source parse
strictness AND vocabulary/describe/open-slots — the Blockly litmus made
checkable). aura-registry: content-addressed processes/ + campaigns/
stores on the blueprint-store pattern (Ok(None) treat-as-empty) and the
referential tier (validate_campaign_refs: process/strategy refs incl.
identity-scan, axis-name + declared-kind checks against param_space).
Plan deviations, verified by hand and folded back into the plan file:
- Task 9's fixture could not use aura-composites as planned: every shipped
composite with an open param routes through LinComb, which the zero-arg
std_vocabulary roster deliberately excludes, so none round-trips through
blueprint_from_json with the by-type-name resolver. The test hand-builds
a minimal Bias composite instead (generic over param_space, as planned);
the unused pro-forma aura-composites dev-dep from the repair pass is
dropped. This is why the loop reported task 9 BLOCKED
(review-loop-exhausted on the literal code block); the tree itself is
complete and green.
- A plan-verbatim test assertion false-matched ("deflate" as substring of
"deflated-positive"); tightened to the JSON key.
- OpenSlot doc comment backticked (rustdoc read strategies[0] as a link;
doc gate back to 0 warnings).
Verification: cargo test --workspace 898/0; clippy -D warnings clean;
cargo doc --no-deps 0 warnings. Tasks 10-12 (CLI verb families, final
gates) follow.
refs #189
User veto (2026-07-03) of the per-cell tagged Scalar draft: the kind is a
property of the parameter, not of each value. An axis is now
{"kind": "I64", "values": [8, 12, 16]} — mixed-kind axes are
unrepresentable, canonical form normalizes value spelling, and the
referential kind check is one comparison per axis. Op-script binds keep
their tagged form (single-value slots). Decision log: #189.
refs #189
The campaign example wrote bare JSON numbers in the axes while the spec's
own typed-cell clause (op-script bind parity) requires the tagged Scalar
wire form. Example now matches the clause; no assumption changed.
refs #189
Two doc-tightenings from the green milestone fieldtest (F3, F6 in
fieldtests/milestone-project-environment/FINDINGS.md): the glossary
manifest entry now names the topology hash and the project provenance
block {namespace, dylib_sha256, commit} a project-hosted run carries;
project-layout.md's Aura.toml annotation now matches what the scaffold
actually emits today ([paths] runs only — the data archive root joins
once real data enters).
refs #180
Resolves the #181 design pass (option c: ledger declaration, no lockstep
mechanism). The aura-new templates (scaffold.rs) are the canonical
authoring shape and evolve with the engine; the cycle-0102 demo-project
fixture is an intentionally frozen known-good twin for the load-boundary
tests. Deliberately NOT lockstep-guarded: no consumer requires the copies
to match (skeptic-verified — nothing treats the fixture as what aura new
emits), each is e2e-guarded on its fitness for purpose (build/load/
charter/deterministic run; blueprint wiring is pinned in neither, stated
honestly after an adversarial skeptic pass refuted the stronger claim),
and an equality guard would convert deliberate template improvements into
forced churn of the frozen fixture — the same cross-purpose coupling that
rules out regenerating the fixture from the scaffolder.
closes#181, refs #180
Architect drift review (range 7b429f9..57f401f): drift_found (low/medium
only), resolved inline as close-fixes:
- crates/aura-std/src/vocabulary.rs — the roster-site doc now names BOTH
count pins a node addition trips (the in-crate shape test and aura-cli's
cross-boundary --vocabulary count e2e); the singular phrasing understated
the cross-crate lockstep the extra e2e introduced.
- docs/design/INDEX.md — the C24 enforcement-shift passage gains the 0105
delivery note: resolver-vs-list drift closed by construction via the
roster macro; the un-rostered-node residual stays fail-safe.
Noted, no action: spec/plan predicted 884 (no new test) but delivery added
one e2e (885) — consciously superseded, disclosed in the iter commit body;
the artifacts are ephemeral and removed below.
What holds (architect): byte-preserving refactor confirmed against the
diff (22 keys/order exact, signatures and module doc untouched, every
consumer unchanged); invariant 9 / C24 preserved (compile-time expansion,
no registry); the new e2e count pin is genuinely roster-tied.
Regression gates (all green): cargo build clean; cargo test --workspace
885 passed / 0 failed (884 baseline + 1 new e2e); clippy --all-targets
-D warnings clean; cargo doc --no-deps 0 warnings.
Ephemeral cycle artifacts removed per project convention (git rm):
docs/specs/0105-std-vocabulary-roster-macro.md,
docs/plans/0105-std-vocabulary-roster-macro.md.
refs #180
Two tasks: (1) the private std_vocabulary_roster! macro + single 22-key
invocation replacing the two hand-written fns in aura-std/vocabulary.rs
(byte-identical signatures/keys/order; docs generated with the fns), plus
the two unit tests reshaped to iterate the generated list instead of
carrying an inline roster copy (oracle: PrimitiveBuilder::label(); count
pin kept as deliberate friction); (2) full regression — suite green
UNCHANGED at 884 (no test added or removed).
Recon-verified isolation: lib.rs re-export untouched (private in-module
macro, same pub fns), no other roster copy anywhere in the workspace, no
doc/ledger reference stales, first macro_rules in the crate, no
missing_docs gate.
refs #180, refs #160
Boss-signed via grounding-check PASS (fresh-context agent): all seven
load-bearing current-behaviour assumptions ratified by named green tests —
the two vocabulary unit tests (label round-trip + negatives; count==22 +
non-members), the PrimitiveBuilder::label() oracle, and the four consumer
pins (graph introspect --vocabulary e2e, engine loader UnknownNodeType,
the 0102 merged-vocabulary charter suite, the canonical golden).
Design basis: the mechanism recorded on #160 (declarative macro as single
source for resolver match + enumerable list) plus the cycle-0105
reconciliation comment there (single pair-generating invocation; the
test's inline roster copy replaced by iterating the generated list with
label() as independent oracle; the un-rostered-node residual accepted as
fail-safe and documented; public surface byte-identical).
refs #180, refs #160
Four tasks: (1) engine identity projection — factor build_doc/serialize_doc
out of blueprint_to_json (byte-preserving, guarded by the canonical golden),
add blueprint_identity_json + strip_debug_symbols + 4 property tests
(renamed twins, openness, bound values, edge swap); (2) CLI --identity-id
sibling flag — composite_from_str factored out of build_from_str, the
exactly-one introspect dispatch deliberately relaxed so the two id flags
combine (one build, both ids, content id first), plus the in-crate
cross-path twin test (Rust sma_signal vs op-script twin: distinct content
ids, one identity id — placed beside the existing cross-surface pins since
sma_signal is private to the bin); (3) four e2e tests incl. the previously
uncovered count!=1 usage path; (4) full regression (873 -> 882 expected).
Plan-recon corrections folded in: the spec sketch's BlueprintData does not
exist — the real DTO is CompositeData; engine re-export + CLI import are a
lockstep pair.
refs #180, refs #171
Boss-signed via grounding-check PASS (fresh-context agent): all five
load-bearing assumptions about current behaviour are ratified by named
currently-green tests — the byte-canonical golden
(signal_serializes_to_canonical_golden), the --content-id CLI contract
(graph_introspect_content_id_is_deterministic_and_distinguishes and the
bad-document refusal), the three topology_hash roles (introspect
content id, blueprint-store round-trip key, reproduce fetch anchor —
each with its own green test), the I64:4-vs-5 distinguishing property,
and the bound-array openness encoding (bound slot textually present,
open slot absent).
Design carried from the #171 reconciliation (adversarially reviewed):
the identity hash is an additive SIBLING of topology_hash —
introspection-only this cycle, no manifest field, no store key; the
byte-canonical form and every consumer of it stay byte-for-byte
untouched. Reviewer note carried to planning: allowing --content-id
and --identity-id together relaxes the current exactly-one flag
dispatch in graph_construct.rs — the usage tests must be adjusted
deliberately, not incidentally.
refs #180, refs #171
Architect review of b672a37..HEAD. What holds: emitted templates match
the settled contracts (Aura.toml paths-only per C17, cdylib + empty
[workspace] per C16, ::-namespaced ids per the charter, the C14 exit
partition in dispatch_new); the load-bypass guard is narrow and sound
(only the new-project verb exempted, the load arm byte-unchanged for
every other verb); e2e coverage is real (scaffold -> build -> run twice
byte-identical, namespace stamped, introspection, four refusals, the
unbuilt-tree bypass, plus the additive namespace-override test).
Regression: cargo build --workspace rc=0; cargo test --workspace 873
passed / 0 failed (project_new 6/6); clippy --workspace --all-targets
-D warnings rc=0; cargo doc --workspace --no-deps 0 warnings. No
baselines exist to move, so nothing to ratify.
Drift resolved in this commit: three doc mirrors still framing the
scaffolder as future — project-layout.md's 'the future aura new
scaffolder will emit this line; until then, add it by hand', the C24
status paragraph's 'what remains open is the aura new scaffolder (the
milestone's next cycle)', and the #109 open-thread bullet's 'the aura
new scaffolder and the experiment-builder API remain open' — all
rewritten to the landed reading (only the experiment-builder API stays
open in that layer).
Carried debt, filed forward: the scaffold templates and the
demo-project fixture encode one project shape twice with no lockstep
guard and deliberate cosmetic divergences (#181, idea — needs a design
pass, not a mechanical edit).
Process note: one mid-cycle plan-fix commit (derive(Debug) on a
plan-prescribed struct whose own test code required it) — the
implement loop's spec gate correctly blocked the necessary deviation;
resolution followed the fix-plan/commit-subset/re-run path.
Ephemeral cycle artifacts removed per convention: spec 0103, plan 0103.
refs #180
The Task-1 unit tests call unwrap_err() on Result<ScaffoldSpec, String>,
which requires ScaffoldSpec: Debug; the prescribed struct lacked the
derive, so the plan's own bytes could not compile. The implement loop
correctly blocked on the necessary deviation (spec-compliance holds
verbatim plan code); this aligns the plan with the only compiling form.
refs #180
Grounding-check PASS as the autonomous signature: 8 load-bearing
current-behaviour assumptions each ratified by a named green test
(the templates are the cycle-0102 fixture verbatim, itself e2e-green);
greenfield scaffolder surfaces excluded per contract. Fork decisions:
the cycle-0103 reconciliation comment on the milestone reference issue.
refs #180
Architect review of d7c935d..HEAD. What holds: C13/C16 separation
verified by diff (descriptor in aura-core::project, loader entirely in
aura-cli::project, aura-engine gains only the additive provenance
field); the C-tier-before-Rust-tier discipline is enforced in load()
(magic -> version -> stamps, fn pointers touched only after Ok);
C17/C24/invariant-9 hold (paths-only Aura.toml, injected merged
resolver, charter at the one seam); outside-a-project byte-identity
confirmed by the full green suite incl. every golden-bearing test.
Regression: cargo build --workspace rc=0; cargo test --workspace 862
passed / 0 failed (cli_run 133, project_load 7 e2e); clippy
--workspace --all-targets -D warnings rc=0; cargo doc --workspace
--no-deps 0 warnings. No baselines exist to move, so nothing to
ratify.
Drift resolved in this commit: two stale project-as-crate ledger
mirrors — the C24 status paragraph still calling the layer
sequencing-coupled/unbuilt, and the #109 open-thread bullet naming it
an unbuilt half — rewritten to record the landed load boundary (the
aura new scaffolder and the experiment-builder API stay open).
Carried holds (plan-prescribed bytes, cosmetic): the unreachable
rsplit fallback in node_name, the unreachable! build closure in one
aura-core test.
Ephemeral cycle artifacts removed per convention: spec 0102, plan 0102.
refs #180
A research project is now a loadable external cdylib crate. Inside a
directory whose ancestry holds an Aura.toml, aura discovers the project
root cargo-style, locates the compiled dylib via cargo metadata (debug
default, --release opt-in), loads it load-and-hold, and refuses
mismatches before trusting anything: the AURA_PROJECT descriptor
(aura-core::project, #[repr(C)]) carries a C-ABI stamp prefix (rustc +
aura-core version, baked per consuming build by the new aura-core
build.rs) validated before any Rust-ABI field is read. The vocabulary
charter gates the merged resolution: project type ids are ::-namespaced
(std stays bare), duplicates refuse, and the enumerable type-id list
must agree with the resolver, so introspection can never silently omit
a project type.
All blueprint verbs resolve through the merged project + std vocabulary
via a per-invocation Env threaded through the dispatch chains;
registry, trace-store, and data paths anchor at the project runs root
(Aura.toml [paths], paths-only by design — instrument geometry stays
the recorded sidecar, C15). RunManifest gains the Tier-1 project
provenance field (namespace + dylib sha256 + best-effort commit),
stamped beside topology_hash on the blueprint-run paths; pre-0102
registry lines load unchanged. Default node names strip the namespace,
so :: never reaches the param-path address space.
Proven by the demo-project fixture (built by the e2e via cargo,
path-dep on this workspace): run twice bit-identical, provenance
recorded, introspection lists demo::* beside std, registry anchors at
the discovered root from a subdirectory; the badcharter fixture proves
the charter refusal through the real libloading path; a never-built
project refuses with a cargo-build hint. Outside a project every path
collapses to the previous literals — goldens and manifest pins
byte-identical.
Verification: cargo build --workspace clean; cargo test --workspace 862
passed / 0 failed (incl. 7 project_load e2e); clippy -D warnings clean
(one precedent-matching allow(too_many_arguments) on run_oos_blueprint,
whose arity the Env threading raised to 8); doc build unchanged.
Docs/ledger aligned: Aura.toml field lists are paths-only in
project-layout.md, glossary, C16/C17; new C13 realization note records
the per-invocation-reload reading and the load-and-hold one-shot scope
boundary.
New deps, per-case review (aura-cli leaf binary only, never the frozen
artifact): libloading, toml.
refs #180
Grounding-check PASS as the autonomous signature: 7 load-bearing
current-behaviour assumptions each ratified by a named green test;
greenfield surfaces (descriptor, loader, Aura.toml) excluded per
contract. Design decisions and provenance: the decision-log on the
milestone reference issue.
refs #180
Architect review of 8814e295..HEAD. What holds: the cosmetic-only claim is
verified (every -/+ pair in the code diff is a string literal; no exit(),
channel, or control-flow byte moved; the C14 usage=2/runtime=1 partition
preserved at every touched site); grep gates exact (zero lowercase usage:
literals, 11 'Usage: aura' source lines = 10 edited + the pre-conforming
run_args_from); fieldtests corpus and refusal diagnostics byte-identical.
The recorded spec deviation (3 additive house-style pins) strengthens
coverage over a real pin gap and is decision-logged on the issue.
Regression: cargo build rc=0, cargo test --workspace rc=0 (cli_run 133
passed / 0 failed), clippy -D warnings rc=0, doc build rc=0. No baselines
exist to move, so nothing to ratify.
Drift resolved in this commit: the C14 ledger note still described the
casing normalization as deferred after #179 landed - rewritten to record
the landed house style (usage lines 'Usage: aura <verb> ...', refusal
diagnostics deliberately unprefixed).
Ephemeral cycle artifacts removed per convention: spec 0101, plan 0101.
refs #179
One lockstep task: 10 byte-exact literal edits (6 lowercase usage: flips,
4 bare closures gain the 'Usage: aura ' prefix, mc built-in gets the program
name in both alternatives) + the single casing-sensitive pin flip in
cli_run.rs, then build / targeted test / full suite / clippy / grep gates /
diff-shape gate / sweep smoke.
refs #179
Single house style for every hand-rolled usage line: 'Usage: aura <verb> ...'
(clap's capital prefix + program name), 10 literal edits + the one
casing-sensitive test pin in lockstep. Refusal diagnostics, the conforming
run_args_from site, clap output, and the fieldtests corpus stay byte-identical.
Grounding-check PASS is the autonomous signature; scope-fork decisions are
recorded on the issue.
refs #179
Architect review of 88e9c9a..HEAD. What holds: token purity of the rename
(removed + mapping == added, 33 sanctioned prose rewordings, numeric golden
bytes untouched); CLAUDE.md invariant 7 now matches the ledger C10 section
clause-by-clause; no silent alias for the old selectors (unknown-selector
exit-2 path is tested).
Regression: cargo build / clippy -D warnings / doc build clean; cargo test
--workspace rc=0, 51 targets, 0 failures. No baselines exist to move, so
nothing to ratify.
Drift resolved in this commit: the C10 terminology note (INDEX.md) still
called Stage-1 a surviving identifier name after the glossary recorded the
rename - aligned to name the r-family rename (#174). Remaining low items:
the spec's own survivor-list contradiction was ratified and recorded on the
issue; the RGrid const-seeding debt is recorded on the harness-retirement
issue (refs #159).
Ephemeral cycle artifacts removed per convention: spec 0100, plan 0100, and
the spent prior-milestone fieldtest spec.
refs #174
The stage1 ordinal has been dead vocabulary since the C10 reframe dropped
the two-stage research model: a 1 structurally implies a 2 that no longer
exists. The family is renamed by its live discriminator - the R yardstick -
with members named by their signal, uniform with their signal-named
siblings (sma, macd, momentum):
- selectors: stage1-r -> r-sma, stage1-breakout -> r-breakout,
stage1-meanrev -> r-meanrev (old tokens are usage errors, exit 2 - no
silent alias)
- identifiers: Strategy::RSma/RBreakout/RMeanRev, HarnessKind::RSma,
r_sma_*/r_breakout_*/r_meanrev_*, wrap_r, run_signal_r, RGrid, R_SMA_*
- persisted identity: the sma_signal composite (param prefix
sma_signal.fast.length / .slow.length; fixtures regenerated; content-ids
shift - no test pins a literal hash, the registry parses no record names)
- e2e test files git-mv'd to r_sma_e2e / r_breakout_e2e / r_meanrev_e2e
- dead Stage-1/Stage-2 prose reworded to post-C10 vocabulary across
rustdoc, Cargo.tomls, ledger live lines, and the glossary (historical
entries stay; fieldtests corpus untouched)
- CLAUDE.md invariant 7 rewritten to record the ratified C10 reframe
faithfully (the token-swap alone would have laundered the retired
gated-currency/realistic-broker design into unmarked live prose); the
unbacked account-mode clause dropped
Verification: cargo build/test --workspace green (51 targets), clippy
-D warnings clean, doc build clean, acceptance grep gate leaves exactly
the one resampling-stage false positive (harness.rs), smoke: --harness
r-sma runs, --harness stage1-r exits 2 with the new usage line.
Decision log: forks and rationale recorded on the issue (reconciliation
+ implementation-phase comments).
closes#174
Five ordered tasks: aura-cli source + include_str-coupled fixtures, the
cli_run suite, engine e2e git-mvs + doc-refs, dead-ordinal prose across
rustdoc/CLAUDE.md/glossary/ledger live lines, and a workspace grep gate.
Substitutions are ordered longest-token-first; every task carries its own
verification gate.
refs #174
The stage1 ordinal is dead vocabulary since the C10 reframe; the spec
renames the selector/identifier family to r-sma / r-breakout / r-meanrev
(signal-named members under the R-yardstick marker), reworders dead
Stage-1/Stage-2 doc prose, and pins the survivors (history, fieldtests).
Signed autonomously under /boss on the grounding-check PASS; fork
decisions recorded on the reference issue.
refs #174
Three polish items surfaced by the World/C21 milestone fieldtest:
- A sweep of a fully-bound (closed) blueprint is refused up front with a clear
"this blueprint is fully bound; nothing to sweep" message — the symmetric
inverse of blueprint_mc_family's closed-blueprint guard — instead of leaking a
misleading UnknownKnob for a knob that is not unknown but bound out.
blueprint_sweep_family's error contract widens to String; the sweep terminal's
BindError (a genuine unknown/kind-mismatch axis) still surfaces verbatim.
- mc reproduce lines now carry each member's seed=<N> label. Monte-Carlo members
hold no tuning params, so the member label was blank; the seed is each draw's
realization identity. Sweep / walk-forward labels (which echo params) unchanged.
- docs (design ledger + glossary): state that every open knob --list-axes prints
is mandatory on sweep / walkforward — a subset is refused with the missing knob
named (BindError::MissingKnob), there is no default; pin a knob you do not want
to vary with a single-value axis.
RED-first for the two behaviour items. One existing unknown-axis E2E is re-pointed
from the closed fixture (which now hits the new "nothing to sweep" pre-check) to
the open fixture; it still exercises UnknownKnob("nope") specifically — resolve
raises UnknownKnob in phase 1 before MissingKnob in phase 2 — so the property is
preserved, not weakened.
closes#178
Architect drift review (scope 780d823..HEAD, the sole gate — no regression
script) returned one medium ledger-gap, now resolved; otherwise drift-clean with
a positive What-holds:
- invariant 8 / C16 engine-domain separation preserved: clap is confined to
crates/aura-cli/Cargo.toml; the whole diff touches only aura-cli/ + Cargo.lock
+ ephemeral docs; no engine/std/composites/analysis/registry/core/ingest crate
mentions clap, so the frozen deploy artifact structurally cannot pick it up
(aura-cli is a leaf binary). The 16 new lock entries are exactly clap's tree.
- C16 per-case admission honest: clap admitted with a documented rationale at the
Cargo.toml site (replaces the hand-rolled parser + ten help surfaces; dev-loop
tax, not frozen-artifact tax), the same admission form as sha2/libc.
- C14 automation-face served, execution layer intact: the usage=2/runtime=1
partition, scoped --help, --version, and GNU flags match the spec; the domain
fns keep their signatures (arg-plumbing via thin *_from adapters); new
behaviour is pinned (partition property test, version, scoped-help, dual-grammar
stray-positional guard).
Ledger refreshed: the C14 realization block now records the clap adoption + the
exit-code partition (usage=2/runtime=1) as a durable automation contract, before
the ephemeral spec that first stated it is removed (the drift item the architect
raised).
Deferred/forward: error-message casing normalization (#179, idea); the
machine-first help surface stays on the #157/C21 track (Fork B settled this cycle
as human/GNU convention compliance).
Verified green (orchestrator, this session): cargo build --workspace clean; cargo
clippy --workspace --all-targets -- -D warnings clean; cargo test --workspace
green. Cycle drift-clean.
closes#175
Iteration-2 plan for #175: apply the clean exit-code partition (deviation #8) on top of the clap migration — runtime failures exit 1, usage errors exit 2, no same-class inconsistency. Uses the attribution principle (command-line fault = 2, missing-environment/bad-stdin = 1) to resolve the 6 boundary cases the spec under-specified. refs #175
Iteration-1 plan for #175: replace the hand-rolled aura-cli argv parser with a
clap derive parser (scoped `aura <sub> --help`, `--version`/`-V`, per-flag
Options section, `--flag=value`/`--`/long-opt abbreviation), behaviour-preserving
on exit codes. Task 1 adds clap + RED tests for the new behaviour; Task 2 is the
atomic migration + renegotiation of the help/usage pins the migration inverts
(the four dual-grammar subcommands mapped as optional [blueprint] positional +
post-parse is_file() dispatch; execution layer untouched). The exit-code split
is iteration 2.
refs #175
Replace the hand-rolled aura-cli argv parser with a clap derive parser:
scoped `aura <sub> --help`, `--version`/`-V`, a per-flag Options section, and
`--flag=value`/`--`/long-opt abbreviation for free. Two iterations — (1) clap
migration, behaviour-preserving on exit codes; (2) exit-code split (usage=2,
runtime=1). clap is admitted under the C16 per-case review (research-side CLI,
invariant 8 untouched).
Forks ratified by the user (clap over hand-rolled; GNU-compliance over the
machine-first reframe; exit-code Option A) and the three spec-time derived
decisions (dual-grammar clap mapping, parse-vs-runtime exit classification,
two-iteration split) are recorded on #175.
Boss-signed on grounding-check PASS (independent, fresh-context; 13
current-behaviour assumptions all ratified by named green tests).
refs #175
Milestone-close validation of the World/C21 blueprint-data family surface, run
from the PUBLIC interface only (design ledger + op-script grammar; blueprints
authored as a consumer via `aura graph build`, no crates/ source read). Binary
built from 4710fcc.
The milestone's FUNCTIONAL PROMISE HOLDS (3 working findings): a consumer-authored
blueprint round-trips the whole verb family (graph build → list-axes → sweep →
rank → reproduce; walkforward → reproduce; mc → reproduce); all three FamilyKinds
(Sweep 4/4, MonteCarlo 5/5, WalkForward 3/3) reproduce bit-identically through the
one shared content-addressed store (the C18 core promise); list-axes→sweep→rank
composes and self-corrects.
Findings surfaced for follow-on (NOT fixed here — filed to the tracker): 2 bugs
(aura run on an open blueprint PANICS exit 101 instead of a clean exit-2 like mc;
closed→walkforward --axis prints the rejection twice), 2 friction (sweep-side
closed error terser than mc's exemplary message; mc reproduce lines omit the
seed), 2 spec-gaps (every open knob is mandatory on sweep but the docs frame
list-axes as a menu; the whole blueprint-data verb family is invisible from
`--help` / no README — empirically confirms #159). Detail per finding in the spec.
fieldtests/world-c21-milestone/ carries the op-scripts, authored blueprints,
captured outputs, and run_all.sh (regenerates the gitignored runs/ telemetry).
refs #170
cycle 0097 tidy (clean). Architect drift review over 485f3c8..HEAD: no code
drift, no debt. Regression: no dedicated script (architect is the sole gate);
its "What holds" is non-empty (design commitments confirmed against the diff):
- Inv. 8 / 9: the cycle touches only crates/aura-cli (main.rs + cli_run.rs +
the ephemeral spec/plan); walk_forward/WindowRoller/FamilyKind::WalkForward/
append_family/put_blueprint are read-only reuse, no engine/registry change.
- Inv. 1 / 2: blueprint_walkforward_family reuses the generic walk_forward roller
(windows disjoint-parallel, across-sim); each IS-sweep/OOS-run sees only its
window's windowed_sources; IS precedes OOS per window — no look-ahead.
- Reproduce lockstep: the reproduce_family_in WalkForward arm rebuilds each OOS
slice from stored.manifest.window (before the _ => run_sources fallthrough that
would rebuild the wrong full-window source); winner params via the shared
manifest->cells recovery; pinned by the reproduces_bit_identically E2E.
- Structural mirror + panic-safety: one-substitution-deep twin of the hard-wired
Stage1R arm (select_winner + walk_forward reused, metric sqn_normalized);
run_blueprint_walkforward follows the content-addressed sibling pattern (no
ensure_name_free); blueprint_sweep_over returns Result<_,BindError>, a bad axis
is in-closure exit 2 (never a panic); run_oos_blueprint derives pip internally
(7 args, no #[allow]).
Ledger: refreshed the C24 realization block (docs/design/INDEX.md) — replaced the
now-false "Walk-forward is deferred (#173, dep #169)" sentence with the "IS-refit
walk-forward shipped (cycle 0097)" note; and extended the C18 reproduction note —
reproduce now spans the third family kind (WalkForward), so all three FamilyKind
variants persist AND reproduce through the one shared topology_hash+put_blueprint
hook.
Milestone: the World/C21 blueprint-data family surface (run/sweep/mc/walkforward/
list-axes/reproduce) is coherent and complete per the architect review — no
structural gap. This audit proves drift-clean; a green end-to-end MILESTONE
FIELDTEST remains the deliberate gate to formally close the World/C21 milestone.
Ephemeral spec/plan git-rm'd per the lifecycle convention.
refs #173
Four tasks, one iteration: (1) the windowed IS-sweep + OOS-run helpers +
blueprint_walkforward_family (mirrors the hard-wired Stage1R walkforward arm one
substitution deep — loaded blueprint via blueprint_axis_probe); (2) the
--axis/--select parse grammar + IO wrapper (persist via put_blueprint +
append_family, content-addressed like aura mc) + the .json walkforward dispatch
arm; (3) the reproduce_family_in WalkForward branch (persist+reproduce land
together); (4) the E2Es (persist shape, reproduce bit-identical, no-axis reject,
malformed-safety) + the full clippy/suite gate.
Panic-safety: blueprint_sweep_over returns Result<_, BindError> (user --axis
fallible); a bad axis → clean in-closure exit 2 (mirrors the hard-wired arm's
select_winner handling), never a panic. reduce-mode members carry R-metrics (no
raw pip curve) → empty stitched equity, oos_r is the meaningful summary (C10
R-first). Closes the World/C21 milestone's last piece.
refs #173
Add `aura walkforward <blueprint.json> --axis <name>=<csv> [--axis …]`: a true
in-sample-refit walk-forward over a loaded blueprint (Arm A, #173). Per IS
window it re-optimizes the blueprint's params over the user's --axis grid,
selects the winner by sqn_normalized, runs it out-of-sample, and aggregates the
per-window OOS reports into a FamilyKind::WalkForward family that is
content-addressed (put_blueprint) and aura-reproduce-aware. Closes the World/C21
milestone's last piece (the walk-forward half of #170).
Design (Arm A settled by the user; sub-forks derived and recorded on #173): the
loaded-blueprint WF mirrors the hard-wired stage1-r walk-forward arm one
substitution deep — the generic engine walk_forward driver + select_winner +
the 24/12/12 synthetic schedule are reused verbatim; only the per-window IS-sweep
and OOS-run source the loaded blueprint (wrap_stage1r via cycle-0096's
blueprint_axis_probe) over the user --axis grid. Persist + reproduce ship
together (coupled — a persisted family with no store hook is unreproducible, the
#170 hazard); the reproduce WalkForward branch reconstructs each OOS window's
windowed source + winner params from the member manifest, mirroring the 0095
MonteCarlo branch. Loaded members are reduce-mode (R-metrics, no raw pip curve),
so the WF is measured in R (oos_r), stitched pip-equity empty — consistent with
the C10 R-first yardstick.
Boss-signed on grounding-check PASS (independent, fresh-context; all 8
load-bearing current-behaviour assumptions ratified by named green tests —
incl. an empirically-confirmed one: a loaded-blueprint member carries and ranks
by sqn_normalized. One spec-internal tension flagged + fixed before sign).
Engine/registry untouched (invariants 1/2/8/9).
refs #173
cycle 0096 tidy (clean). Architect drift review over ebebc24..HEAD: no drift,
no debt. Regression: no dedicated script (architect is the sole gate); its
"What holds" is non-empty (design commitments confirmed against the diff + a
green test run, not the commit body):
- Inv. 8 / 9: the cycle touches only crates/aura-cli (main.rs + cli_run.rs +
the ephemeral spec/plan) — no engine/registry/blueprint.rs change; those are
read-only callees.
- Inv. 1: --list-axes is a pure read-only probe (short-circuits before
run_blueprint_sweep; builds the wrapped composite, reads param_space, prints
— no run, no seed, no engine execution).
- Behaviour-preserving single-source: blueprint_axis_probe is token-identical to
the two former inline probes; three copies (sweep/mc/list) collapse to one; the
reload closures survive (topology_hash + per-member run); clippy -D warnings
clean, no dead code; the subsumed tests
(blueprint_sweep_member_equals_single_run_and_shares_topology_hash,
blueprint_mc_family_rejects_an_open_blueprint) stayed green.
- The blueprint_axis_probe .expect is not a latent panic: the ["sweep",..] arm
validates blueprint_from_json first (exit 2), pinned by the shipped
aura_sweep_list_axes_rejects_malformed_blueprint E2E (no panic).
Ledger: refreshed the C24 realization block (docs/design/INDEX.md) — reclassified
the "axis-name discovery surface" from tracked debt to shipped (cycle 0096),
recording the single-sourced blueprint_axis_probe and the listed==swept-by-
construction property. Per-member trace-writing (#168) remains the one tracked
debt in that block.
Milestone World/C21 stays OPEN — the walk-forward half (#173, IS-refit, now
unblocked by this #169) remains. Ephemeral spec/plan git-rm'd per the lifecycle
convention.
refs #169
One iteration, three tasks: (1) extract blueprint_axis_probe and subsume the
two existing inline copies of the wrapped-namespace probe
(blueprint_sweep_family + blueprint_mc_family) into it — the single source the
spec's rationale requires; (2) the --list-axes surface (SweepBlueprintArgs
field, valueless-flag parse + standalone guard, list_blueprint_axes, dispatch
short-circuit); (3) four E2Es (open lists prefixed names, closed empty,
combined-flags reject, malformed-blueprint safety — no panic).
refs #169
Add `aura sweep <blueprint.json> --list-axes`: a read-only query sub-mode
that lists a loaded blueprint's open sweepable knobs (one `<name>:<kind>`
per line, param_space() order), then exits. Closes the discovery gap #169
surfaced — today the only way to learn the axis names is to guess one and
read a MissingKnob/UnknownKnob error, and that error names only the
rejected knob, never the valid set. Prerequisite for #173 (IS-refit
walk-forward must know which axes to re-fit per in-sample window).
Derived surface decision (recorded on #169): the query lands on the SWEEP
verb, not `graph introspect`. The axis names `--axis` accepts are produced
by the sweep's own harness-wrapping — the loaded signal (name
"stage1_signal") is nested via wrap_stage1r as a BlueprintNode::Composite,
so collect_params prefixes the names (stage1_signal.fast.length), NOT the
raw param_space (fast.length; empirically: the bare name → UnknownKnob).
Only the sweep verb, which owns the wrapping, can print the names the user
will actually pass; a shared blueprint_axis_probe single-sources the probe
for both the sweep terminal and --list-axes, so listed == swept by
construction (and stays correct across #159's harness retirement).
Boss-signed on grounding-check PASS (independent, fresh-context; one BLOCK
raised + fixed: the malformed-sweep .expect panic-path is now pinned by a
this-cycle E2E over unknown_node.json, not an untested current-behaviour
claim). Engine/registry untouched (invariants 1/8/9).
refs #169
Architect drift review (scope cec7811..HEAD, the sole gate — no regression
script) returned drift-clean with positive What-holds:
- invariants 8/9: the whole cycle is aura-cli (main.rs + cli_run.rs + ephemeral
docs); nothing entered aura-engine/aura-registry; FamilyKind::MonteCarlo,
monte_carlo, append_family, mc_member_reports are pre-existing seams reused.
- invariant 1 / C1: draws run disjoint-parallel via the engine monte_carlo seam
(seed-input order, no within-sim concurrency); bit-identity proven by the
reproduce round-trip (unit + E2E).
- the C1 lockstep is a single shared fn: synthetic_walk_sources (one fixed
SyntheticSpec) is called by BOTH the persist path and the reproduce MonteCarlo
branch; pip is lockstepped too (both pass DataSource::Synthetic).
- the sweep-sibling pattern matches: Result<McFamily,String> + boundary render.
- scope honesty: WF deferral (#173, dep #169) and synthetic-walk-DGP (#172) are
named in the spec + feat body, not silent.
Carry-on debt (does NOT block close, filed forward, NOT new-this-cycle drift):
synthetic_walk_sources' len:60 is hand-calibrated to the shallow stage1 warm-up;
a deep-lookback closed blueprint warms poorly -> silent-vacuous draws with no
error, and the anti-degenerate test only exercises the shallow fixture. Folded
into #172 (real-data DGP retires the length calibration) per the architect's
recommendation, rather than a new thread.
Ledger refreshed: the C21 realization block now records the MC verb shipped
(cycle 0095) with WF deferred; the C18 reproduction block notes aura reproduce
now spans FamilyKind::MonteCarlo.
Verified green (orchestrator, this session): cargo build --workspace clean;
cargo clippy --workspace --all-targets -- -D warnings clean; cargo test
--workspace green. Cycle is drift-clean (not a milestone close — #170's WF half
remains, milestone stays open).
refs #170
4 tasks: parse_mc_blueprint_args, blueprint_mc_family (closed-blueprint
guard + seed-driven synthetic walks), the reproduce MonteCarlo realization
branch, and run_blueprint_mc + the .json dispatch arm + E2E. All in
aura-cli/main.rs (engine/registry untouched).
The MC half of #170 (World/C21): aura mc <closed blueprint.json> --seeds N
builds a FamilyKind::MonteCarlo family from a loaded closed blueprint, each
seed drawing a distinct synthetic price walk (the mc_family pattern), sharing
one topology_hash, written once to the content-addressed store (the 0094 hook)
so aura reproduce re-derives it bit-identically.
Scope narrowed from #170's two verbs by two read-only review gates:
- spec-skeptic: a loaded-blueprint 'walkforward' with no in-sample refit under
the verb name that hard-wired walkforward uses for real IS-optimization is a
guarantee collision -> walk-forward deferred to #173 (deps #169 axis-discovery).
- grounding-check: there is no per-knob declared-default mechanism and the raw
bind path panics on arity mismatch -> MC runs a CLOSED blueprint over the empty
point (as aura run does), rejecting an open blueprint with a named error + exit 2.
Grounding-check PASS on the revised spec; decisions logged on #170.
refs #170
Architect drift review (4751d4b..HEAD): drift_found — ledger-currency only; the
code is clean.
What holds (architect, confirmed by diff + green tests):
- Inv 8 / C9: aura-engine byte-untouched; content_id + sha2 in aura-cli, the
store in aura-registry.
- Inv 9: Registry::{put,get}_blueprint is a dumb bytes-by-key store (opaque
String keyed by String hash; no sha2, no parse, no sha256==hash verify — the
caller owns the hash). No domain knowledge crept into the engine or registry.
- C1: the run_blueprint_member extraction is behaviour-preserving (keystone
blueprint_sweep_member_equals_single_run_and_shares_topology_hash unchanged +
green); serde_json/float_roundtrip serves determinism (disk==recomputed), not a
mask (1-ULP canary green).
- C11/C12: one blueprint stored per family, keyed by the shared topology_hash.
Resolutions:
- [ledger] Refreshed C18 + C24: added the cycle-0094 Realization (the
content-addressed store runs/blueprints/<hash>.json + aura reproduce), and
retired the now-false "content-addressed identity / full reproduction deferred"
claims (INDEX.md:1261, C24 deferred block). Only the structural-axis /
whole-harness content-id (and the debug-name-in-id question) genuinely remains,
filed forward (#171, #170).
- [debt->closed] Added the two untested refuse-don't-guess E2E arms
(aura_reproduce_rejects_a_non_generated_family — a hard-wired sweep member has
no topology_hash; aura_reproduce_rejects_a_missing_stored_blueprint — the store
was deleted), and a cross-surface unit pin
(topology_hash_is_the_content_id_of_the_canonical_form) so the two acc-1 paths
cannot silently drift.
- [scaling] MC/walk-forward-from-blueprint (#170) will need the same put_blueprint
hook or reproduce won't cover them — noted on #170.
No regression scripts in this project (Test is the gate): full workspace suite
green (51 suites), clippy -D warnings clean, binary verified end-to-end.
Ephemeral spec/plan git-rm'd per the lifecycle convention; durable rationale is in
the ledger. The World/C21 milestone stays open (#170 remains).
closes#158
Iteration 1 (acceptance 2, the heart): the content-addressed blueprint store
(Registry::put_blueprint/get_blueprint under runs/blueprints/), the store write
at the sweep persist seam, the extracted shared run_blueprint_member, and
`aura reproduce <family-id>` re-deriving every persisted member bit-identically
from the stored blueprint. acceptance 1 (graph introspect --content-id) and 3
(Tier-1 id stability) deferred to plan 0094b.
refs #158
Restore C18 reproducibility for generated (blueprint-sweep) runs. Since cycle
0093 a sweep member's topology is a generated data value, so commit alone no
longer identifies the graph and the manifest is not a complete re-derivation
recipe (#158).
Design: content-address the topology. The canonical blueprint JSON is stored
once, keyed by the topology_hash the manifest already carries (cycle 0092),
under runs/blueprints/<hash>.json (a dumb bytes-by-key store in aura-registry,
no sha2 dep — the reproduce path's bit-identical compare is the integrity
check). `aura reproduce <family-id>` loads each member's blueprint by hash,
re-bootstraps with the member's params/window/seed, re-runs, and asserts the
metrics are bit-identical (C1). acc-1 (two paths -> same content id) via a
minimal `graph introspect --content-id`; acc-3 (Tier-1-stable id) composes the
#156 omit-defaults pin.
Derived forks (logged on #158): content-address (not carry-inline) — keeps the
manifest tiny (C18) and dedups one topology across a family (C11/C12); hash
domain = signal-only this cycle (the scaffolding is not yet blueprint-data,
C24), boundary documented, whole-harness id deferred (#156/#159).
Engine untouched (C9); store + hash off the frozen engine (invariant 8).
boss-signed on a grounding-check PASS (9/9 current-behaviour assumptions
ratified by green tests).
refs #158
Cycle-close drift review (architect) for cycle 0093 / #166 (aura sweep
<blueprint.json>). Drift-found on debt only; no contract violation.
What holds (architect, evidence of review): invariant 9 fail-clean (resolve_axes
runs before the sweep closure → named BindError UnknownKnob/MissingKnob/KindMismatch,
exit 2; UnknownNodeType on load; no registry); C1/C12/C18 shape verbatim from
stage1_r_sweep_family (members disjoint + enumeration-ordered; shared topology_hash
correct, consistent with the cycle-1 #158 signal-only scope; append_family/
sweep_member_reports reused); reload-per-member is sound (Composite !Clone, #164);
stdout uses family_member_line with family_id, run_sweep-consistent.
Resolved: ledger refreshed (C24 Status — the cycle-0093 realization: the World now
orchestrates FAMILIES from blueprint-data; the open-vs-bound fixture fact + the
reload-per-member rationale recorded). Drift filed forward: #168 (--trace writes no
per-member traces — silently like --name, debt-med), #169 (no axis-name discovery
for loaded sweeps, debt-low), #170 (MC/walk-forward over a loaded blueprint — the
next family-verb cycle, assigned to the milestone).
Ephemeral 0093 spec + plan git-rm'd per the lifecycle convention.
closes#166
Three-task decomposition of #166 (sweep a loaded blueprint): the parser
(parse_sweep_blueprint_args + SweepBlueprintArgs + --axis i64-then-f64 lexing),
the member-build (run_blueprint_sweep — a structural mirror of stage1_r_sweep_family
with three deviations: by-name axes, reload-per-member since Composite is !Clone,
per-member topology_hash) + the .json dispatch arm, and the tests (grammar +
equivalence-vs-run_signal_stage1r + the binary sweep-family E2E + unknown-axis
fail-clean). Purely additive; reuses the engine sweep machinery + cycle-1's
wrap_stage1r unchanged. Grounding decisions (Q1-Q5) recorded on #166.
refs #166
Cycle 2 of the World/C21 milestone (#166): `aura sweep <blueprint.json> --axis
<name>=<csv>` sweeps a loaded signal over its named param-space axes, reusing the
existing stage1-r sweep machinery with the signal source swapped to cycle-1's
wrap_stage1r seam, aggregating to a FamilyKind::Sweep family; each member manifest
carries the shared topology_hash.
Derivable extension, not a fork (recon + grounding confirm): the sweep is by-name
param-space-driven (SweepBinder.axis / resolve_axes / bootstrap_with_cells), so a
loaded blueprint's param_space slots in; wrap_stage1r nests an arbitrary signal;
append_family is blueprint-agnostic. Scoped to sweep (the keystone family); MC /
walk-forward over a loaded blueprint are the same machinery per verb and follow as
a fast follow-on. Decisions recorded on #166.
Boss-signed on a grounding-check PASS: all six load-bearing current-behaviour
assumptions ratified by named green tests (wrap_stage1r arbitrary-signal nesting,
the by-name param-space sweep, stage1_r_sweep_family, append_family/FamilyKind::Sweep,
topology_hash, the .json-discriminator dispatch precedent).
refs #166
Cycle-close drift review (architect) for cycle 0092 / #165 (aura run
<blueprint.json>). Drift-found on ledger currency only; the code is clean and
correct.
What holds (architect, evidence of review): the stage1_signal/wrap_stage1r split
is behaviour-preserving (golden metrics byte-unchanged; loaded_signal bit-identical
green; the STAGE1_R_SMA_* constants keep the hashed signal == the executed signal);
sha2 + the topology_hash helper live in aura-cli only, aura-engine stays sha2-free
(invariant 8 / C16); the resolver resolves the closed std_vocabulary set, unknown
type -> UnknownNodeType, no registry (invariant 9); topology_hash is a correct
Tier-1 optional and the RunManifest <-> RunManifestRead read-mirror both carry it.
Resolved: ledger refreshed (C24 Status — the cycle-0092 run-from-blueprint
realization + the topology_hash #158 anchor landing; #158 reframed from "premature"
to "field landed, full reproduction remains"). Drift filed forward: #167
(single-source the stage1-r suffix-remapping lockstep — debt-med), and a scope note
on #158 (the hash currently covers the signal only; a structural-variant content-id
is the #158/#166 concern). The spec's illustrative instrument:"GER40" example
(note-low) is moot — the ephemeral spec is removed here.
Ephemeral 0092 spec + 0092/0092b plans git-rm'd per the lifecycle convention.
closes#165
Brummel