feat: the document-first surface closes — show read-back, typed stops, bare plateau

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
This commit is contained in:
2026-07-21 12:50:10 +02:00
parent 4ed6455b64
commit 757e3ac1bd
5 changed files with 517 additions and 100 deletions
+42 -91
View File
@@ -487,12 +487,13 @@ fn resolve_campaign_name(name: &str, env: &aura_runner::project::Env) -> NameRes
// `aura_runner::family` (#295 Task 8) — imported below by plain name so
// call sites are unchanged.
/// Parse a `--select` token: `argmax` | `plateau:mean` | `plateau:worst`. Unknown
/// tokens are a usage error (the caller maps `Err(())` to exit 2).
/// Parse a `--select` token: `argmax` | `plateau` (alias for `plateau:mean`) |
/// `plateau:mean` | `plateau:worst`. Unknown tokens are a usage error (the
/// caller maps `Err(())` to exit 2).
fn parse_select(s: &str) -> Result<Selection, ()> {
match s {
"argmax" => Ok(Selection::Argmax),
"plateau:mean" => Ok(Selection::Plateau(PlateauMode::Mean)),
"plateau" | "plateau:mean" => Ok(Selection::Plateau(PlateauMode::Mean)),
"plateau:worst" => Ok(Selection::Plateau(PlateauMode::Worst)),
_ => Err(()),
}
@@ -508,22 +509,6 @@ fn select_rule_of(sel: Selection) -> aura_research::SelectRule {
}
}
/// Parse a comma-separated list of `T` (each item parsed via `FromStr`), rejecting
/// any item that fails to parse — the shared validator for the r-sma grid flags
/// (#137). The caller folds the `Err` into the subcommand `usage()` so a malformed
/// `--fast 2,x` is the strict usage path, not a downstream panic. `str::split(',')`
/// always yields at least one item, and an empty item (`""`, or a trailing comma)
/// fails its per-item `parse`, so a non-empty result needs no separate empty-list
/// check — the empty/blank inputs are rejected by the per-item parse itself.
fn parse_csv_list<T: std::str::FromStr>(s: &str) -> Result<Vec<T>, ()> {
let items: Vec<&str> = s.split(',').collect();
let mut out = Vec::with_capacity(items.len());
for item in items {
out.push(item.parse::<T>().map_err(|_| ())?);
}
Ok(out)
}
/// Render a family-member stdout line: the assigned `family_id` plus the embedded
/// `RunReport`. The report is emitted in its own declaration key order (manifest
/// leads with `commit`, C18) so the line is byte-identical to the stored
@@ -1001,7 +986,7 @@ fn parse_scalar_csv(csv: &str) -> Option<Vec<Scalar>> {
// `--version`, `--flag=value`, `--`, and long-option abbreviation; the `dispatch_*`
// handlers below convert each `*Cmd` into the argument shapes the existing execution
// fns accept, reusing the value helpers (`Selection`,
// `DataSource::from_choice`, `parse_scalar_csv`, `parse_csv_list`, `parse_select`,
// `DataSource::from_choice`, `parse_scalar_csv`, `parse_select`,
// `parse_param_cells`). The four subcommands with an optional `[blueprint]` positional
// dispatch on `is_blueprint_file`: a first-positional naming an existing `.json` file
// selects the loaded-blueprint branch. There is no second built-in grammar anymore
@@ -1358,13 +1343,14 @@ struct WalkforwardCmd {
/// mutually exclusive with --name).
#[arg(long)]
trace: Option<String>,
/// Campaign-path stop length (single value; --real mode).
/// Campaign-path stop length (--real mode).
#[arg(long)]
stop_length: Option<String>,
/// Campaign-path stop-k multiple (single value; --real mode).
stop_length: Option<i64>,
/// Campaign-path stop-k multiple (--real mode).
#[arg(long)]
stop_k: Option<String>,
/// In-sample winner selection: argmax | plateau:mean | plateau:worst (default argmax).
stop_k: Option<f64>,
/// In-sample winner selection: argmax | plateau (alias for plateau:mean) |
/// plateau:mean | plateau:worst (default argmax).
#[arg(long)]
select: Option<String>,
/// Blueprint IS-refit axis `<name>=<csv>` (repeatable, >=1 required; .json mode).
@@ -1395,12 +1381,12 @@ struct McCmd {
/// Blueprint IS-refit axis `<name>=<csv>` (repeatable, >=1 required; --real mode).
#[arg(long)]
axis: Vec<String>,
/// Campaign-path stop length (single value; --real mode).
/// Campaign-path stop length (--real mode).
#[arg(long)]
stop_length: Option<String>,
/// Campaign-path stop-k multiple (single value; --real mode).
stop_length: Option<i64>,
/// Campaign-path stop-k multiple (--real mode).
#[arg(long)]
stop_k: Option<String>,
stop_k: Option<f64>,
/// Moving-block bootstrap block length (R-bootstrap; default 1).
#[arg(long)]
block_len: Option<usize>,
@@ -1442,41 +1428,16 @@ fn run_data_from(real: Option<&str>, from: Option<i64>, to: Option<i64>) -> RunD
}
}
/// Resolve a single optional stop knob (`--stop-length`/`--stop-k`): an absent flag
/// defaults to `default`; a present one parses via [`parse_csv_list`] and refuses
/// unless it holds exactly one value (the stop is a risk regime, not a swept axis).
/// Single-sourced across `walkforward_args_from`/`mc_args_from` (#217 follow-up) —
/// `generalize_args_from`'s typed `Option<i64>`/`Option<f64>` fields need no parse
/// step, so it keeps its own plain `unwrap_or`.
fn stop_knob_or<T: Copy + std::str::FromStr>(
raw: Option<&str>,
default: T,
regime: &impl Fn() -> String,
) -> Result<T, String> {
match raw {
None => Ok(default),
Some(s) => {
let v: Vec<T> = parse_csv_list(s).map_err(|()| regime())?;
if v.len() != 1 {
return Err(regime());
}
Ok(v[0])
}
}
}
/// Convert `WalkforwardCmd` into the resolved argument shape the walkforward sugar
/// consumes (the dissolved `.json --real` branch). Single instrument, single-value
/// stop (Fork A: the stop is a risk regime, not a swept axis); `parse_csv_list` is
/// generic over `T: FromStr`, so the same helper parses the i64 length and the f64
/// k. `--stop-length`/`--stop-k` are each OPTIONAL (#217): a missing flag defaults
/// independently to the single-sourced [`R_SMA_STOP_LENGTH`]/[`R_SMA_STOP_K`]
/// regime; a present-but-multi-value stop still refuses (the local `regime`
/// closure, captureless → `Copy`, is reused across both arms). `--name`/`--trace`
/// are mutually exclusive, matching the flag's own documented contract and the
/// inline path's `name_persist` refusal; an omitted flag defaults the family name
/// to "walkforward". The IS-refit `--axis` grid is parsed separately at the
/// dispatch site, not here.
/// stop (Fork A: the stop is a risk regime, not a swept axis); `--stop-length`/
/// `--stop-k` are typed `Option<i64>`/`Option<f64>` (clap rejects a multi-value
/// spelling itself, "invalid value") and each OPTIONAL (#217): a missing flag
/// defaults independently to the single-sourced [`R_SMA_STOP_LENGTH`]/
/// [`R_SMA_STOP_K`] regime. `--name`/`--trace` are mutually exclusive, matching
/// the flag's own documented contract and the inline path's `name_persist`
/// refusal; an omitted flag defaults the family name to "walkforward". The
/// IS-refit `--axis` grid is parsed separately at the dispatch site, not here.
#[allow(clippy::type_complexity)]
fn walkforward_args_from(
a: &WalkforwardCmd,
@@ -1485,9 +1446,8 @@ fn walkforward_args_from(
None | Some("") => return Err("walkforward dissolves only over --real <SYMBOL>".to_string()),
Some(s) => s.to_string(),
};
let regime = || "walkforward: the stop is a single risk regime; --stop-length and --stop-k take one value each".to_string();
let stop_length = stop_knob_or(a.stop_length.as_deref(), R_SMA_STOP_LENGTH, &regime)?;
let stop_k = stop_knob_or(a.stop_k.as_deref(), R_SMA_STOP_K, &regime)?;
let stop_length = a.stop_length.unwrap_or(R_SMA_STOP_LENGTH);
let stop_k = a.stop_k.unwrap_or(R_SMA_STOP_K);
let (name, trace) = match (a.name.as_deref(), a.trace.as_deref()) {
(Some(_), Some(_)) => {
return Err("walkforward: --name and --trace are mutually exclusive".to_string());
@@ -1501,16 +1461,16 @@ fn walkforward_args_from(
/// Convert `McCmd` into the resolved argument shape the mc sugar consumes (the
/// `--real` campaign branch). Single instrument, single-value stop (Fork A: the
/// stop is a risk regime, not a swept axis); `--stop-length`/`--stop-k` are each
/// OPTIONAL (#217), independently defaulting to the single-sourced
/// [`R_SMA_STOP_LENGTH`]/[`R_SMA_STOP_K`] regime when omitted — a present-but-
/// multi-value stop still refuses. `--block-len`/`--resamples`/`--seed`
/// default to `1`/`1000`/`1` — the same defaults the retired real-R dispatch
/// used, so an omitted flag produces the same document either way. The
/// usize->u32 conversion lands HERE, at the argv boundary where the CLI's
/// `usize` meets the document's `u32` vocabulary (`StageBlock::MonteCarlo`).
/// `--name`/`--trace` are rejected: the R-bootstrap records without a family
/// name (the campaign name is a constant "mc").
/// stop is a risk regime, not a swept axis); `--stop-length`/`--stop-k` are typed
/// `Option<i64>`/`Option<f64>` (clap rejects a multi-value spelling itself,
/// "invalid value") and each OPTIONAL (#217), independently defaulting to the
/// single-sourced [`R_SMA_STOP_LENGTH`]/[`R_SMA_STOP_K`] regime when omitted.
/// `--block-len`/`--resamples`/`--seed` default to `1`/`1000`/`1` — the same
/// defaults the retired real-R dispatch used, so an omitted flag produces the
/// same document either way. The usize->u32 conversion lands HERE, at the argv
/// boundary where the CLI's `usize` meets the document's `u32` vocabulary
/// (`StageBlock::MonteCarlo`). `--name`/`--trace` are rejected: the R-bootstrap
/// records without a family name (the campaign name is a constant "mc").
#[allow(clippy::type_complexity)]
fn mc_args_from(
a: &McCmd,
@@ -1522,9 +1482,8 @@ fn mc_args_from(
if a.name.is_some() || a.trace.is_some() {
return Err("mc --real: --name/--trace are not accepted (the R-bootstrap records without a family name)".to_string());
}
let regime = || "mc: the stop is a single risk regime; --stop-length and --stop-k take one value each".to_string();
let stop_length = stop_knob_or(a.stop_length.as_deref(), R_SMA_STOP_LENGTH, &regime)?;
let stop_k = stop_knob_or(a.stop_k.as_deref(), R_SMA_STOP_K, &regime)?;
let stop_length = a.stop_length.unwrap_or(R_SMA_STOP_LENGTH);
let stop_k = a.stop_k.unwrap_or(R_SMA_STOP_K);
let block_len = a.block_len.map(|v| v as u32).unwrap_or(1);
let resamples = a.resamples.map(|v| v as u32).unwrap_or(1000);
let seed = a.seed.unwrap_or(1);
@@ -2319,7 +2278,7 @@ fn dispatch_sweep(a: SweepCmd, env: &aura_runner::project::Env) {
/// #220: arbitrary user blueprint + axes, formerly the welded r-sma branch).
fn dispatch_walkforward(a: WalkforwardCmd, env: &aura_runner::project::Env) {
// Single-sourced: both arms below read this one closure (house style, #179).
let usage = || format!("Usage: aura walkforward <blueprint.json> --axis <name>=<csv> [--axis …] [--select <argmax|plateau:mean|plateau:worst>] [--name <n>] | aura walkforward <blueprint.json> --real <SYMBOL> --axis <name>=<csv> [--axis …] [--stop-length <n> (default {R_SMA_STOP_LENGTH})] [--stop-k <x> (default {R_SMA_STOP_K:.1})] [--from <ms>] [--to <ms>] [--name <n> | --trace <n>]");
let usage = || format!("Usage: aura walkforward <blueprint.json> --axis <name>=<csv> [--axis …] [--select <argmax|plateau|plateau:mean|plateau:worst>] [--name <n>] | aura walkforward <blueprint.json> --real <SYMBOL> --axis <name>=<csv> [--axis …] [--stop-length <n> (default {R_SMA_STOP_LENGTH})] [--stop-k <x> (default {R_SMA_STOP_K:.1})] [--from <ms>] [--to <ms>] [--name <n> | --trace <n>]");
match is_blueprint_file(&a.blueprint) {
Some(path) => {
let doc = std::fs::read_to_string(path).unwrap_or_else(|e| {
@@ -2763,6 +2722,10 @@ mod tests {
assert!(matches!(parse_select("argmax").unwrap(), Selection::Argmax), "default is argmax");
assert!(matches!(parse_select("plateau:mean").unwrap(), Selection::Plateau(PlateauMode::Mean)));
assert!(matches!(parse_select("plateau:worst").unwrap(), Selection::Plateau(PlateauMode::Worst)));
assert!(matches!(
parse_select("plateau").unwrap(),
Selection::Plateau(PlateauMode::Mean)
));
assert!(parse_select("bogus").is_err(), "unknown --select token is a usage error");
}
@@ -3820,18 +3783,6 @@ mod tests {
assert_eq!(stop_k, R_SMA_STOP_K, "omitted --stop-k defaults to the regime");
}
#[test]
fn mc_args_from_refuses_a_multi_value_stop() {
let a = McCmd {
real: Some("GER40".to_string()),
stop_length: Some("14,20".to_string()),
stop_k: Some("2.0".to_string()),
..bare_mc_cmd()
};
let err = mc_args_from(&a).unwrap_err();
assert!(err.contains("single risk regime"), "stop-regime refusal: {err}");
}
/// Property (#217): a missing `--stop-length`/`--stop-k` no longer refuses —
/// each independently defaults to the single-sourced [`R_SMA_STOP_LENGTH`]/
/// [`R_SMA_STOP_K`] regime (length 3, k 2.0).
+34
View File
@@ -29,6 +29,8 @@ pub enum ProcessSub {
Introspect(DocIntrospectCmd),
/// Register a valid process document into the store under the runs root.
Register { file: PathBuf },
/// Print a registered process document's canonical bytes to stdout.
Show { id: String },
}
#[derive(clap::Args)]
@@ -208,6 +210,7 @@ pub fn process_cmd(cmd: ProcessCmd, env: &Env) {
introspect_process(i)
}
ProcessSub::Register { file } => register_process(file, env),
ProcessSub::Show { id } => show_process(id, env),
};
if let Err(m) = result {
eprintln!("aura: {m}");
@@ -259,6 +262,21 @@ fn register_process(file: &PathBuf, env: &Env) -> Result<(), String> {
Ok(())
}
fn show_process(id: &str, env: &Env) -> Result<(), String> {
let registry = env.registry();
match registry.get_process(id).map_err(|e| e.to_string())? {
Some(bytes) => {
// `print!`, not `println!`: `bytes` is the content-addressed canonical
// form the store keyed `id` on — the CLI is a transport and must not
// frame the canonical bytes with a trailing newline (#164, mirroring
// `graph_construct::build_cmd`).
print!("{bytes}");
Ok(())
}
None => Err(format!("process {id} not found in the project store{}", prefix_hint(id))),
}
}
fn introspect_process(cmd: &DocIntrospectCmd) -> Result<(), String> {
if cmd.metrics {
print_metric_roster();
@@ -341,6 +359,8 @@ pub enum CampaignSub {
/// List stored campaign realizations, or dump one campaign's records
/// (the bare store lines, not the run-emit wrapper).
Runs { campaign: Option<String>, run: Option<usize> },
/// Print a registered campaign document's canonical bytes to stdout.
Show { id: String },
}
/// The one authoring trap the referential tier can name outright: a ref that
@@ -460,6 +480,7 @@ pub fn campaign_cmd(cmd: CampaignCmd, env: &Env) {
}
CampaignSub::Register { file } => register_campaign(file, env),
CampaignSub::Runs { campaign, run } => campaign_runs(campaign.as_deref(), *run, env),
CampaignSub::Show { id } => show_campaign(id, env),
CampaignSub::Run { .. } => unreachable!("handled above"),
};
if let Err(m) = result {
@@ -562,6 +583,19 @@ fn register_campaign(file: &PathBuf, env: &Env) -> Result<(), String> {
Ok(())
}
fn show_campaign(id: &str, env: &Env) -> Result<(), String> {
let registry = env.registry();
match registry.get_campaign(id).map_err(|e| e.to_string())? {
Some(bytes) => {
// `print!`, not `println!`: see `show_process` — the CLI must not
// frame the canonical bytes with a trailing newline (#164).
print!("{bytes}");
Ok(())
}
None => Err(format!("campaign {id} not found in the project store{}", prefix_hint(id))),
}
}
fn introspect_campaign(cmd: &DocIntrospectCmd) -> Result<(), String> {
if cmd.metrics {
print_metric_roster();
+423 -2
View File
@@ -3220,7 +3220,7 @@ fn walkforward_dissolved_refuses_a_multi_value_stop() {
.expect("spawn aura");
assert_eq!(out.status.code(), Some(2), "multi-value stop is a usage refusal");
let stderr = String::from_utf8_lossy(&out.stderr);
assert!(stderr.contains("single risk regime"), "stop-regime refusal: {stderr}");
assert!(stderr.contains("invalid value"), "stop-regime refusal: {stderr}");
}
/// #215: --select plateau is no longer refused on the dissolved path — it now
@@ -3647,6 +3647,96 @@ fn walkforward_stopless_defaults_the_regime_byte_identically_to_explicit_3_2() {
);
}
/// #227: bare `--select plateau` is the documented default aggregation
/// (`plateau:mean`), not a distinct spelling. Same dedup-pin idiom as
/// `walkforward_stopless_defaults_the_regime_byte_identically_to_explicit_3_2`:
/// running the same grid-sweep blueprint in the same store first with
/// `--select plateau`, then with `--select plateau:mean`, must produce
/// byte-identical campaign documents — content-addressed storage collapses
/// them onto EXACTLY ONE stored campaign — while each invocation still
/// records its own campaign-run line. Gated on the shared GER40 archive;
/// skips cleanly on a data refusal.
#[test]
fn walkforward_bare_plateau_generates_the_same_campaign_as_plateau_mean() {
const FROM_MS: &str = "1735689600000";
const TO_MS: &str = "1747353600000";
let (cwd, _g) = fresh_project();
let fixture = format!("{}/examples/r_sma.json", env!("CARGO_MANIFEST_DIR"));
// Bare `plateau`: today this is an unknown-token usage refusal (exit 2);
// the feature makes it an alias for `plateau:mean`.
let bare = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args([
"walkforward", &fixture, "--real", "GER40",
"--axis", "sma_signal.fast.length=3,5", "--axis", "sma_signal.slow.length=12,20",
"--stop-length", "14", "--stop-k", "2.0",
"--select", "plateau",
"--from", FROM_MS, "--to", TO_MS,
])
.current_dir(&cwd)
.output()
.expect("spawn aura");
if bare.status.code() == Some(1) {
let stderr = String::from_utf8_lossy(&bare.stderr);
assert!(
stderr.contains("no local data") || stderr.contains("no recorded geometry"),
"exit 1 must be a data refusal, got: {stderr}"
);
eprintln!("skip: no local GER40 data");
return;
}
assert_eq!(
bare.status.code(),
Some(0),
"bare --select plateau must no longer be refused (it aliases plateau:mean): {}",
String::from_utf8_lossy(&bare.stderr)
);
// Explicit `plateau:mean`: names the very aggregation the bare spelling
// resolves to. Same blueprint, same axes, same window, same stop.
let explicit = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args([
"walkforward", &fixture, "--real", "GER40",
"--axis", "sma_signal.fast.length=3,5", "--axis", "sma_signal.slow.length=12,20",
"--stop-length", "14", "--stop-k", "2.0",
"--select", "plateau:mean",
"--from", FROM_MS, "--to", TO_MS,
])
.current_dir(&cwd)
.output()
.expect("spawn aura");
assert_eq!(
explicit.status.code(),
Some(0),
"the explicit --select plateau:mean walkforward must succeed: {}",
String::from_utf8_lossy(&explicit.stderr)
);
// Byte-identity: bare `plateau` IS `plateau:mean`, so both invocations
// generate the same campaign document and content-addressed storage
// collapses them onto one content id — exactly one campaign document in
// the store, but two independent campaign-run lines.
let count = |sub: &str| {
std::fs::read_dir(cwd.join("runs").join(sub))
.map(|d| d.count())
.unwrap_or(0)
};
assert_eq!(
count("campaigns"),
1,
"bare plateau binds the SAME aggregation as plateau:mean, so the two campaign \
documents are byte-identical and dedup to one content id"
);
let runs_log = std::fs::read_to_string(cwd.join("runs").join("campaign_runs.jsonl"))
.expect("campaign_runs.jsonl exists after two sugar runs");
assert_eq!(
runs_log.lines().count(),
2,
"each invocation still records its own campaign run (a run is an event, a document \
is content)"
);
}
/// Property: `aura generalize` is now thin sugar over the one campaign path — a
/// successful run durably auto-registers exactly one generated process document,
/// one generated campaign document (carrying the `--name` handle and the stop as
@@ -5665,7 +5755,7 @@ fn mc_dissolved_refuses_a_multi_value_stop() {
.expect("spawn aura");
assert_eq!(out.status.code(), Some(2), "multi-value stop is a usage refusal");
let stderr = String::from_utf8_lossy(&out.stderr);
assert!(stderr.contains("single risk regime"), "stop-regime refusal: {stderr}");
assert!(stderr.contains("invalid value"), "stop-regime refusal: {stderr}");
}
/// Property: an `--axis` name that is not among the loaded blueprint's sweepable
@@ -6286,3 +6376,334 @@ fn data_list_enumerates_archive_symbols_one_per_line_sorted() {
// scoping an instrument matrix.
assert_eq!(stdout, "SYMA\nSYMB\n", "sorted symbol enumeration: {stdout:?}");
}
/// Property (#300): `aura campaign show <CID>` prints the registered
/// document's exact canonical bytes to stdout — round-tripping them back
/// through `aura campaign register` re-derives the SAME content id (canonical
/// bytes are content-id-identical), and the store stays deduplicated at one
/// document. Gated on the shared GER40 archive; skips cleanly on a data
/// refusal (same guard shape as
/// `walkforward_stopless_defaults_the_regime_byte_identically_to_explicit_3_2`).
#[test]
fn campaign_show_round_trips_the_registered_document() {
const FROM_MS: &str = "1735689600000";
const TO_MS: &str = "1767225599000";
let (cwd, _g) = fresh_project();
let fixture = format!("{}/examples/r_sma.json", env!("CARGO_MANIFEST_DIR"));
let run = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args([
"walkforward", &fixture, "--real", "GER40",
"--axis", "sma_signal.fast.length=3", "--axis", "sma_signal.slow.length=12",
"--select", "plateau:mean",
"--from", FROM_MS, "--to", TO_MS,
])
.current_dir(&cwd)
.output()
.expect("spawn aura");
if run.status.code() == Some(1) {
let stderr = String::from_utf8_lossy(&run.stderr);
assert!(
stderr.contains("no local data") || stderr.contains("no recorded geometry"),
"exit 1 must be a data refusal, got: {stderr}"
);
eprintln!("skip: no local GER40 data");
return;
}
assert_eq!(
run.status.code(),
Some(0),
"walkforward must succeed: {}",
String::from_utf8_lossy(&run.stderr)
);
let campaigns_dir = cwd.join("runs").join("campaigns");
let cid = std::fs::read_dir(&campaigns_dir)
.expect("campaigns dir exists")
.next()
.expect("exactly one campaign document")
.expect("readable dir entry")
.path()
.file_stem()
.expect("campaign filename has a stem")
.to_string_lossy()
.into_owned();
let show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["campaign", "show", &cid])
.current_dir(&cwd)
.output()
.expect("spawn aura campaign show");
assert_eq!(
show.status.code(),
Some(0),
"campaign show must succeed: {}",
String::from_utf8_lossy(&show.stderr)
);
// Directly pin the #164 no-framing invariant: `show` must print the
// stored canonical bytes byte-for-byte, with no added trailing newline
// (a `println!` regression in `show_campaign` would still pass the
// re-register check below via `campaign_to_json` -> `content_id_of`
// stripping it back out before hashing, so this comparison is the only
// thing that would catch it).
let stored_bytes = std::fs::read(campaigns_dir.join(format!("{cid}.json"))).expect("read stored campaign document");
assert_eq!(
show.stdout, stored_bytes,
"campaign show must print the stored canonical bytes byte-for-byte, no framing"
);
let roundtrip_path = cwd.join("roundtrip.campaign.json");
std::fs::write(&roundtrip_path, &show.stdout).expect("write roundtrip campaign doc");
let register = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["campaign", "register", roundtrip_path.to_str().expect("utf-8 path")])
.current_dir(&cwd)
.output()
.expect("spawn aura campaign register");
assert_eq!(
register.status.code(),
Some(0),
"re-registering the shown document must succeed: {}",
String::from_utf8_lossy(&register.stderr)
);
let register_out = String::from_utf8_lossy(&register.stdout).into_owned();
assert!(
register_out.contains(&format!("registered campaign {cid}")),
"canonical bytes must re-register content-id-identically to {cid}: {register_out}"
);
let count = std::fs::read_dir(&campaigns_dir).expect("campaigns dir exists").count();
assert_eq!(count, 1, "the shown bytes dedup onto the same one stored document");
}
/// Property (#300): identical to `campaign_show_round_trips_the_registered_document`
/// but over the generated process document — `aura process show <PID>` prints
/// canonical bytes that re-register content-id-identically through `aura
/// process register`, with the store staying deduplicated at one document.
#[test]
fn process_show_round_trips_the_registered_document() {
const FROM_MS: &str = "1735689600000";
const TO_MS: &str = "1767225599000";
let (cwd, _g) = fresh_project();
let fixture = format!("{}/examples/r_sma.json", env!("CARGO_MANIFEST_DIR"));
let run = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args([
"walkforward", &fixture, "--real", "GER40",
"--axis", "sma_signal.fast.length=3", "--axis", "sma_signal.slow.length=12",
"--select", "plateau:mean",
"--from", FROM_MS, "--to", TO_MS,
])
.current_dir(&cwd)
.output()
.expect("spawn aura");
if run.status.code() == Some(1) {
let stderr = String::from_utf8_lossy(&run.stderr);
assert!(
stderr.contains("no local data") || stderr.contains("no recorded geometry"),
"exit 1 must be a data refusal, got: {stderr}"
);
eprintln!("skip: no local GER40 data");
return;
}
assert_eq!(
run.status.code(),
Some(0),
"walkforward must succeed: {}",
String::from_utf8_lossy(&run.stderr)
);
let processes_dir = cwd.join("runs").join("processes");
let pid = std::fs::read_dir(&processes_dir)
.expect("processes dir exists")
.next()
.expect("exactly one process document")
.expect("readable dir entry")
.path()
.file_stem()
.expect("process filename has a stem")
.to_string_lossy()
.into_owned();
let show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["process", "show", &pid])
.current_dir(&cwd)
.output()
.expect("spawn aura process show");
assert_eq!(
show.status.code(),
Some(0),
"process show must succeed: {}",
String::from_utf8_lossy(&show.stderr)
);
// Directly pin the #164 no-framing invariant (see the sibling campaign
// test's comment): `show` must print the stored canonical bytes
// byte-for-byte, with no added trailing newline.
let stored_bytes = std::fs::read(processes_dir.join(format!("{pid}.json"))).expect("read stored process document");
assert_eq!(
show.stdout, stored_bytes,
"process show must print the stored canonical bytes byte-for-byte, no framing"
);
let roundtrip_path = cwd.join("roundtrip.process.json");
std::fs::write(&roundtrip_path, &show.stdout).expect("write roundtrip process doc");
let register = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["process", "register", roundtrip_path.to_str().expect("utf-8 path")])
.current_dir(&cwd)
.output()
.expect("spawn aura process register");
assert_eq!(
register.status.code(),
Some(0),
"re-registering the shown document must succeed: {}",
String::from_utf8_lossy(&register.stderr)
);
let register_out = String::from_utf8_lossy(&register.stdout).into_owned();
assert!(
register_out.contains(&format!("registered process {pid}")),
"canonical bytes must re-register content-id-identically to {pid}: {register_out}"
);
let count = std::fs::read_dir(&processes_dir).expect("processes dir exists").count();
assert_eq!(count, 1, "the shown bytes dedup onto the same one stored document");
}
/// Property (#300): `show` over an unknown id refuses with the same
/// project-store-not-found prose the referential tier already uses for a
/// missing `process.ref`/`strategy.ref` (`ref_fault_prose`'s
/// `ProcessNotFound`/campaign-equivalent hint shape) — no project run needed,
/// so this test needs no archive gate.
#[test]
fn show_refuses_an_unknown_id_with_the_store_hint() {
let (cwd, _g) = fresh_project();
let campaign_show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["campaign", "show", "deadbeef"])
.current_dir(&cwd)
.output()
.expect("spawn aura campaign show");
assert_eq!(campaign_show.status.code(), Some(1), "unknown campaign id is a runtime refusal");
let campaign_stderr = String::from_utf8_lossy(&campaign_show.stderr);
assert!(
campaign_stderr.contains("campaign deadbeef not found in the project store"),
"campaign show stderr: {campaign_stderr}"
);
let process_show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["process", "show", "deadbeef"])
.current_dir(&cwd)
.output()
.expect("spawn aura process show");
assert_eq!(process_show.status.code(), Some(1), "unknown process id is a runtime refusal");
let process_stderr = String::from_utf8_lossy(&process_show.stderr);
assert!(
process_stderr.contains("process deadbeef not found in the project store"),
"process show stderr: {process_stderr}"
);
}
/// Property (#300): `show`'s not-found prose reuses the existing `prefix_hint`
/// authoring-trap detector (#194) — an id pasted WITH the display-only
/// `content:` prefix (the form `ref_fault_prose` already prints elsewhere for
/// an unresolved `process.ref`/`strategy.ref`) still gets the "drop the
/// 'content:' prefix" hint through the new `show` code path, not a bare
/// not-found. This is the one authoring mistake the store can name outright,
/// and it must survive `show` reusing `ref_fault_prose`'s helper rather than
/// re-deriving its own prose.
#[test]
fn show_hints_the_content_prefix_authoring_mistake() {
let (cwd, _g) = fresh_project();
let campaign_show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["campaign", "show", "content:deadbeef"])
.current_dir(&cwd)
.output()
.expect("spawn aura campaign show");
assert_eq!(campaign_show.status.code(), Some(1), "unresolvable id is a runtime refusal");
let campaign_stderr = String::from_utf8_lossy(&campaign_show.stderr);
assert!(
campaign_stderr.contains("drop the 'content:' prefix"),
"campaign show must hint the content: authoring mistake: {campaign_stderr}"
);
let process_show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["process", "show", "content:deadbeef"])
.current_dir(&cwd)
.output()
.expect("spawn aura process show");
assert_eq!(process_show.status.code(), Some(1), "unresolvable id is a runtime refusal");
let process_stderr = String::from_utf8_lossy(&process_show.stderr);
assert!(
process_stderr.contains("drop the 'content:' prefix"),
"process show must hint the content: authoring mistake: {process_stderr}"
);
}
/// Property (#300): `show` is scoped to its own document-kind store — a
/// content id that IS a real, registered document in one store (here: a
/// minimal sweep-only process) does not resolve through the OTHER verb
/// (`campaign show`). `show_process`/`show_campaign` are two near-identical
/// functions written in the same change (both `registry.get_*(id) -> Option
/// -> print! the bytes or format! a not-found`); this pins that they read
/// from their own store method and neither accidentally shares the other's
/// namespace or falls back to a generic content-blob lookup.
#[test]
fn show_is_scoped_to_its_own_document_store() {
let (cwd, _g) = fresh_project();
const SWEEP_ONLY_PROCESS: &str = r#"{"format_version":1,"kind":"process","name":"knob-sweep-only","pipeline":[{"block":"std::sweep","metric":"sqn_normalized","select":"argmax"}]}"#;
std::fs::write(cwd.join("knob.process.json"), SWEEP_ONLY_PROCESS).expect("write process doc");
let register = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["process", "register", "knob.process.json"])
.current_dir(&cwd)
.output()
.expect("spawn aura process register");
assert_eq!(
register.status.code(),
Some(0),
"process register must succeed: {}",
String::from_utf8_lossy(&register.stderr)
);
let register_out = String::from_utf8_lossy(&register.stdout).into_owned();
let pid = register_out
.lines()
.find(|l| l.starts_with("registered process "))
.expect("register line")
.trim_start_matches("registered process ")
.split(' ')
.next()
.expect("process id")
.to_string();
let process_show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["process", "show", &pid])
.current_dir(&cwd)
.output()
.expect("spawn aura process show");
assert_eq!(
process_show.status.code(),
Some(0),
"process show must find the just-registered process: {}",
String::from_utf8_lossy(&process_show.stderr)
);
let campaign_show = std::process::Command::new(env!("CARGO_BIN_EXE_aura"))
.args(["campaign", "show", &pid])
.current_dir(&cwd)
.output()
.expect("spawn aura campaign show");
assert_eq!(
campaign_show.status.code(),
Some(1),
"a process id must NOT resolve through campaign show: {}",
String::from_utf8_lossy(&campaign_show.stdout)
);
let campaign_stderr = String::from_utf8_lossy(&campaign_show.stderr);
assert!(
campaign_stderr.contains(&format!("campaign {pid} not found in the project store")),
"campaign show stderr: {campaign_stderr}"
);
}
+7 -4
View File
@@ -40,6 +40,9 @@ use crate::member::{
use crate::project::Env;
use crate::translate::{R_SMA_STOP_K, R_SMA_STOP_LENGTH};
/// The demo default stop regime, shared by every family builder in this module.
const DEFAULT_STOP: StopRule = StopRule::Vol { length: R_SMA_STOP_LENGTH, k: R_SMA_STOP_K };
/// The in-sample winner-selection objective for walk-forward's per-window IS
/// refit (cycle 0077). `Argmax` is the bare-best pick deflated for trials
/// (#144, the default); `Plateau` argmaxes the neighbourhood-smoothed surface
@@ -427,7 +430,7 @@ pub fn blueprint_sweep_family(
.sweep(|point| {
// fresh per-member graph (Composite is !Clone, reload per member) run through
// the shared reduce-mode member path — the same fn reproduction re-runs.
run_blueprint_member(reload(doc), point, &space, data.run_sources(env, &binding.columns()), window, 0, pip, &topo, env, StopRule::Vol { length: R_SMA_STOP_LENGTH, k: R_SMA_STOP_K }, &binding, &[], NO_INSTRUMENT_CONTEXT)
run_blueprint_member(reload(doc), point, &space, data.run_sources(env, &binding.columns()), window, 0, pip, &topo, env, DEFAULT_STOP, &binding, &[], NO_INSTRUMENT_CONTEXT)
})
// render the sweep terminal's BindError to prose (#247), the fn's String error
// contract — never the raw Debug struct.
@@ -476,7 +479,7 @@ pub fn blueprint_sweep_over(
binder.sweep_with_lattice(|point| {
let sources = data.windowed_sources(from, to, env, &binding.columns());
let window = window_of(&sources).expect("non-empty in-sample window");
run_blueprint_member(reopen_all(reload(doc), &overrides), point, &space, sources, window, 0, pip, &topo, env, StopRule::Vol { length: R_SMA_STOP_LENGTH, k: R_SMA_STOP_K }, binding, &[], NO_INSTRUMENT_CONTEXT)
run_blueprint_member(reopen_all(reload(doc), &overrides), point, &space, sources, window, 0, pip, &topo, env, DEFAULT_STOP, binding, &[], NO_INSTRUMENT_CONTEXT)
})
}
@@ -502,7 +505,7 @@ pub fn run_oos_blueprint(
let pip = data.pip_size();
let sources = data.windowed_sources(from, to, env, &binding.columns());
let window = window_of(&sources).expect("non-empty out-of-sample window");
let report = run_blueprint_member(reload, params, space, sources, window, 0, pip, topo, env, StopRule::Vol { length: R_SMA_STOP_LENGTH, k: R_SMA_STOP_K }, binding, &[], NO_INSTRUMENT_CONTEXT);
let report = run_blueprint_member(reload, params, space, sources, window, 0, pip, topo, env, DEFAULT_STOP, binding, &[], NO_INSTRUMENT_CONTEXT);
(Vec::new(), report)
}
@@ -642,7 +645,7 @@ pub fn blueprint_mc_family(
let family = monte_carlo(&base_point, &seeds, |seed, _base| {
let sources = synthetic_walk_sources(seed);
let window = window_of(&sources).expect("non-empty synthetic walk");
run_blueprint_member(reload(doc), &[], &space, sources, window, seed, pip, &topo, env, StopRule::Vol { length: R_SMA_STOP_LENGTH, k: R_SMA_STOP_K }, &binding, &[], NO_INSTRUMENT_CONTEXT)
run_blueprint_member(reload(doc), &[], &space, sources, window, seed, pip, &topo, env, DEFAULT_STOP, &binding, &[], NO_INSTRUMENT_CONTEXT)
});
// Silent-vacuous MC guard (refuse-don't-guess, C10): with >= 2 seeds, if every draw's
// metrics are bit-identical to the first, no seed reached a distinguishable realization —
+11 -3
View File
@@ -1598,6 +1598,10 @@ by design (the 0106 fieldtest corpus stays as the historical record). Known debt
metric-roster triplication (still hand-maintained, but drift from the shipped
`aura-analysis` types is now test-caught by a cross-crate guard, #190;
single-source removal waits on #147), deflation-constant duplication (#199).
#300 (2026-07-21) closed the store's read loop: `aura process|campaign show
<content-id>` prints a registered document's canonical bytes — the
generate → retrieve → hand-extend → re-register cycle needs no direct store
filesystem access.
**Realization (cycle 0108, #200 — the annotator stages execute).** The v2 executable
shape is `std::sweep [std::gate]* [std::walk_forward]? [std::monte_carlo]?
[std::generalize]?` — an ordered optional annotator suffix, each at most once,
@@ -2526,9 +2530,13 @@ vocabulary (blueprints, process/campaign documents, registry records) is the
canonical layer, and every control surface — the one-shot CLI's executor
verbs, any future long-running host, any MCP face, a World program — is a
projection/executor over those artifacts, never a second home for intent.
Which projection is built next (document-first flag-vocabulary completion, a
typed-protocol host, an MCP face, or none until demand) is deliberately open
— recorded as an open direction fork on #295.
The owner resolved the direction 2026-07-21 (minuted on #295): **document-first
completion first** — delivered by #300: the executor verb set is settled
(`run`, the four thin per-verb generators, and the document verbs
validate/introspect/register/show/run/runs, `show` being #300's read-back
addition), the verbs' per-verb identity re-ratified (#300 F8 — reduction, not
grammar collapse), and a typed-protocol host or MCP face remains demand-driven,
unbuilt.
### C26 — Harness input binding: role names bind archive columns
**Guarantee.** A strategy blueprint declares WHICH data it consumes as the NAMES