feat(research): 0106 tasks 1-9 — document layer, content-id move, stores + referential tier

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

Cargo.lock

@@ -114,6 +114,7 @@ dependencies = [
  "aura-engine",
  "aura-ingest",
  "aura-registry",
+ "aura-research",
  "aura-std",
  "clap",
  "data-server",
@@ -174,10 +175,22 @@ version = "0.1.0"
 dependencies = [
  "aura-core",
  "aura-engine",
+ "aura-research",
+ "aura-std",
  "serde",
  "serde_json",
 ]
 
+[[package]]
+name = "aura-research"
+version = "0.1.0"
+dependencies = [
+ "aura-core",
+ "serde",
+ "serde_json",
+ "sha2",
+]
+
 [[package]]
 name = "aura-std"
 version = "0.1.0"

Cargo.toml

@@ -17,6 +17,7 @@ members = [
     "crates/aura-cli",
     "crates/aura-ingest",
     "crates/aura-registry",
+    "crates/aura-research",
 ]
 
 [workspace.package]

crates/aura-cli/Cargo.toml

@@ -16,6 +16,7 @@ aura-engine = { path = "../aura-engine" }
 # r-sma harness wires (kept out of aura-engine so the engine stays domain-free).
 aura-composites = { path = "../aura-composites" }
 aura-registry = { path = "../aura-registry" }
+aura-research = { path = "../aura-research" }
 aura-std = { path = "../aura-std" }
 aura-ingest = { path = "../aura-ingest" }
 # data-server: the local M1 archive `aura run --real` streams from. Mirrors the

crates/aura-cli/src/main.rs

@@ -2585,9 +2585,7 @@ fn sma_signal(fast_len: Option<i64>, slow_len: Option<i64>) -> Composite {
 /// two surfaces agree by construction over the same canonical bytes. Research-side
 /// (aura-cli), off the frozen engine (invariant 8).
 fn content_id(canonical_json: &str) -> String {
-    use sha2::{Digest, Sha256};
+    aura_research::content_id_of(canonical_json)
-    let digest = Sha256::digest(canonical_json.as_bytes());
-    digest.iter().map(|b| format!("{b:02x}")).collect()
 }
 
 /// SHA256 (hex) of the canonical (#164, no-trailing-newline) serialization of a

crates/aura-registry/Cargo.toml

@@ -10,11 +10,17 @@ publish.workspace = true
 # back (admitted under the amended C16 per-case policy, INDEX.md). RunReport
 # derives serde in aura-engine.
 aura-engine = { path = "../aura-engine" }
+# the document stores' content-id primitive (put_process/put_campaign key on
+# aura_research::content_id_of, the same hash the doc types canonicalize to);
+# also PrimitiveBuilder, the referential tier's resolver signature.
+aura-core = { path = "../aura-core" }
+aura-research = { path = "../aura-research" }
 # the lineage record types (FamilyKind / FamilyRunRecord) derive serde; admitted
 # under the same per-case policy as serde_json (INDEX.md).
 serde = { workspace = true, features = ["derive"] }
 serde_json = { workspace = true }
 
 [dev-dependencies]
-# fixtures construct RunReport/RunManifest literals, whose window needs Timestamp
+# the referential-tier test's generic param-space fixture (a real vocabulary
-aura-core = { path = "../aura-core" }
+# resolver + a zero-arg node, not hand-rolled).
+aura-std = { path = "../aura-std" }

crates/aura-registry/src/lib.rs

@@ -17,10 +17,13 @@ use std::fs;
 use std::io::Write;
 use std::path::{Path, PathBuf};
 
+use aura_core::PrimitiveBuilder;
 use aura_engine::{
-    expected_max_of_normals, r_metrics_from_rs, resample_block, FamilySelection, MetricStats,
+    blueprint_from_json, blueprint_identity_json, expected_max_of_normals, r_metrics_from_rs,
-    RunReport, SelectionMode, SplitMix64, SweepFamily, SweepPoint,
+    resample_block, FamilySelection, MetricStats, RunReport, SelectionMode, SplitMix64,
+    SweepFamily, SweepPoint,
 };
+use aura_research::{CampaignDoc, DocRef};
 
 mod compat;
 
@@ -131,6 +134,161 @@ impl Registry {
             Err(e) => Err(RegistryError::Io(e)),
         }
     }
+
+    /// The content-addressed document-store dir for one document kind —
+    /// a sibling of the runs store, like `blueprints/`.
+    fn doc_dir(&self, kind: &str) -> PathBuf {
+        self.path.with_file_name(kind)
+    }
+
+    /// The single id→path mapping every document put/get routes through
+    /// (the `blueprint_path` write-one/read-one lockstep discipline).
+    fn doc_path(&self, kind: &str, content_id: &str) -> PathBuf {
+        self.doc_dir(kind).join(format!("{content_id}.json"))
+    }
+
+    /// Content-addressed put: computes the content id from the canonical
+    /// bytes (unlike `put_blueprint`, whose caller owns the hash — document
+    /// callers always hold canonical bytes, so self-keying is safe here)
+    /// and writes `<kind>/<id>.json`. Idempotent.
+    fn put_doc(&self, kind: &str, canonical_json: &str) -> Result<String, RegistryError> {
+        let id = aura_research::content_id_of(canonical_json);
+        fs::create_dir_all(self.doc_dir(kind))?;
+        fs::write(self.doc_path(kind, &id), canonical_json)?;
+        Ok(id)
+    }
+
+    /// Read a stored document by content id; `Ok(None)` if absent (the
+    /// `get_blueprint` treat-as-empty discipline).
+    fn get_doc(&self, kind: &str, content_id: &str) -> Result<Option<String>, RegistryError> {
+        match fs::read_to_string(self.doc_path(kind, content_id)) {
+            Ok(s) => Ok(Some(s)),
+            Err(e) if e.kind() == std::io::ErrorKind::NotFound => Ok(None),
+            Err(e) => Err(RegistryError::Io(e)),
+        }
+    }
+
+    /// Store a canonical process document; returns its content id.
+    pub fn put_process(&self, canonical_json: &str) -> Result<String, RegistryError> {
+        self.put_doc("processes", canonical_json)
+    }
+
+    /// Load a stored process document by content id (`Ok(None)` if absent).
+    pub fn get_process(&self, content_id: &str) -> Result<Option<String>, RegistryError> {
+        self.get_doc("processes", content_id)
+    }
+
+    /// Store a canonical campaign document; returns its content id.
+    pub fn put_campaign(&self, canonical_json: &str) -> Result<String, RegistryError> {
+        self.put_doc("campaigns", canonical_json)
+    }
+
+    /// Load a stored campaign document by content id (`Ok(None)` if absent).
+    pub fn get_campaign(&self, content_id: &str) -> Result<Option<String>, RegistryError> {
+        self.get_doc("campaigns", content_id)
+    }
+}
+
+/// Referential-validation findings for a campaign document. By-identifier
+/// and Display-free (the CLI phrases them).
+#[derive(Clone, Debug, PartialEq)]
+pub enum RefFault {
+    ProcessNotFound(String),
+    StrategyNotFound(String),
+    IdentityUnmatched(String),
+    StrategyUnloadable { id: String, error: String },
+    AxisNotInParamSpace { strategy: String, axis: String },
+    AxisKindMismatch { strategy: String, axis: String },
+}
+
+impl Registry {
+    /// Referential tier: resolve the campaign's references against the
+    /// project's stores and check each axis (name AND declared kind) against
+    /// the referenced strategy's param space. IO faults are Errors; semantic
+    /// findings are RefFaults.
+    pub fn validate_campaign_refs(
+        &self,
+        doc: &CampaignDoc,
+        resolve: &dyn Fn(&str) -> Option<PrimitiveBuilder>,
+    ) -> Result<Vec<RefFault>, RegistryError> {
+        let mut faults = Vec::new();
+
+        if let DocRef::ContentId(id) = &doc.process.r#ref
+            && self.get_process(id)?.is_none()
+        {
+            faults.push(RefFault::ProcessNotFound(id.clone()));
+        }
+
+        for entry in &doc.strategies {
+            let (label, blueprint_json) = match &entry.r#ref {
+                DocRef::ContentId(id) => match self.get_blueprint(id)? {
+                    Some(json) => (id.clone(), Some(json)),
+                    None => {
+                        faults.push(RefFault::StrategyNotFound(id.clone()));
+                        (id.clone(), None)
+                    }
+                },
+                DocRef::IdentityId(id) => match self.find_blueprint_by_identity(id, resolve)? {
+                    Some(json) => (id.clone(), Some(json)),
+                    None => {
+                        faults.push(RefFault::IdentityUnmatched(id.clone()));
+                        (id.clone(), None)
+                    }
+                },
+            };
+            let Some(json) = blueprint_json else { continue };
+            let composite = match blueprint_from_json(&json, resolve) {
+                Ok(c) => c,
+                Err(e) => {
+                    faults.push(RefFault::StrategyUnloadable {
+                        id: label.clone(),
+                        error: format!("{e:?}"),
+                    });
+                    continue;
+                }
+            };
+            let space = composite.param_space();
+            for (axis, ax) in &entry.axes {
+                match space.iter().find(|p| &p.name == axis) {
+                    None => faults.push(RefFault::AxisNotInParamSpace {
+                        strategy: label.clone(),
+                        axis: axis.clone(),
+                    }),
+                    Some(spec) => {
+                        if ax.kind != spec.kind {
+                            faults.push(RefFault::AxisKindMismatch {
+                                strategy: label.clone(),
+                                axis: axis.clone(),
+                            });
+                        }
+                    }
+                }
+            }
+        }
+        Ok(faults)
+    }
+
+    /// Scan the blueprint store for a blueprint whose identity id matches.
+    fn find_blueprint_by_identity(
+        &self,
+        identity_id: &str,
+        resolve: &dyn Fn(&str) -> Option<PrimitiveBuilder>,
+    ) -> Result<Option<String>, RegistryError> {
+        let entries = match fs::read_dir(self.blueprints_dir()) {
+            Ok(e) => e,
+            Err(_) => return Ok(None), // no store yet -> nothing matches
+        };
+        for entry in entries {
+            let entry = entry?;
+            let json = fs::read_to_string(entry.path())?;
+            let Ok(composite) = blueprint_from_json(&json, resolve) else { continue };
+            let Ok(identity_json) = blueprint_identity_json(&composite) else { continue };
+            if aura_research::content_id_of(&identity_json) == identity_id {
+                return Ok(Some(json));
+            }
+        }
+        Ok(None)
+    }
 }
 
 /// Which metric a best-first comparison keys on. Resolves a metric *name* (a
@@ -711,6 +869,21 @@ mod tests {
         assert_eq!(reg.get_blueprint("never-written").expect("get"), None);
     }
 
+    #[test]
+    fn document_stores_round_trip_by_content_id() {
+        let reg = Registry::open(temp_family_dir("document_store_round_trip"));
+        let process = r#"{"format_version":1,"kind":"process","name":"p","pipeline":[{"block":"std::generalize","metric":"sqn"}]}"#;
+        let campaign = r#"{"format_version":1,"kind":"campaign","name":"c"}"#;
+        let pid = reg.put_process(process).expect("put process");
+        assert_eq!(pid, aura_research::content_id_of(process));
+        assert_eq!(reg.get_process(&pid).expect("get process"), Some(process.to_string()));
+        let cid = reg.put_campaign(campaign).expect("put campaign");
+        assert_eq!(reg.get_campaign(&cid).expect("get campaign"), Some(campaign.to_string()));
+        assert_ne!(pid, cid);
+        // absent id is Ok(None), not an error (treat-as-empty discipline)
+        assert_eq!(reg.get_process("never-written").expect("get"), None);
+    }
+
     #[test]
     fn corrupt_line_is_a_parse_error_with_line_number() {
         let path = temp_path("corrupt");
@@ -1291,6 +1464,121 @@ mod tests {
         assert_eq!(g1, g2, "a pure fold must be deterministic (C1)");
     }
 
+    /// Property: `validate_campaign_refs` resolves campaign document
+    /// references (content-id and identity-id) against the registry's
+    /// stores and cross-checks each declared tuning axis (name AND kind)
+    /// against the resolved strategy's real, generic param space — not a
+    /// hardcoded param name. Faults are reported per finding kind
+    /// (ProcessNotFound / StrategyNotFound / AxisNotInParamSpace /
+    /// AxisKindMismatch), and a fully-resolved, kind-correct document
+    /// yields no faults at all.
+    #[test]
+    fn referential_tier_resolves_refs_and_checks_axes() {
+        use aura_engine::{blueprint_to_json, Composite};
+        use aura_research::{Axis, DocRef};
+
+        let reg = Registry::open(temp_family_dir("referential_tier"));
+        let resolve = |t: &str| aura_std::std_vocabulary(t);
+
+        // A minimal fixture composite with one OPEN param, built from a
+        // zero-arg `aura-std` node (`Bias`) so `std_vocabulary` can actually
+        // resolve it on load. `aura-composites`' shipped composites (vol_stop,
+        // risk_executor*) all route through `LinComb`, whose builder needs a
+        // structural arity argument and is therefore deliberately absent from
+        // the zero-arg `std_vocabulary` roster (see aura-std/src/vocabulary.rs)
+        // — they cannot round-trip through `blueprint_from_json` with this
+        // resolver, so this fixture stands in as the "real, open-param,
+        // vocabulary-loadable composite" the test needs.
+        let composite = Composite::new(
+            "fixture",
+            vec![aura_std::Bias::builder().named("b").into()],
+            vec![],
+            vec![],
+            vec![],
+        );
+        let space = composite.param_space();
+        let real = space.first().expect("fixture composite has an open param");
+        let real_name = real.name.clone();
+        let real_kind = real.kind;
+
+        let blueprint_json = blueprint_to_json(&composite).expect("serializes");
+        let bp_id = aura_research::content_id_of(&blueprint_json);
+        reg.put_blueprint(&bp_id, &blueprint_json).expect("seed blueprint");
+        let process = r#"{"format_version":1,"kind":"process","name":"p","pipeline":[{"block":"std::generalize","metric":"sqn"}]}"#;
+        let proc_id = reg.put_process(process).expect("seed process");
+
+        // one kind-correct bare value for the real param
+        let bare = match real_kind {
+            aura_core::ScalarKind::I64 => "4",
+            aura_core::ScalarKind::F64 => "1.5",
+            aura_core::ScalarKind::Bool => "true",
+            aura_core::ScalarKind::Timestamp => "0",
+        };
+        let kind_tag = format!("{real_kind:?}"); // unit-variant Debug == serde string
+        let campaign_text = format!(
+            concat!(
+                r#"{{"format_version":1,"kind":"campaign","name":"c","#,
+                r#""data":{{"instruments":["GER40"],"windows":[{{"from_ms":1,"to_ms":2}}]}},"#,
+                r#""strategies":[{{"ref":{{"content_id":"{bp}"}},"#,
+                r#""axes":{{"{axis}":{{"kind":"{kind}","values":[{val}]}}}}}}],"#,
+                r#""process":{{"ref":{{"content_id":"{proc}"}}}},"#,
+                r#""seed":1,"presentation":{{"persist_taps":[],"emit":["family_table"]}}}}"#
+            ),
+            bp = bp_id,
+            axis = real_name,
+            kind = kind_tag,
+            val = bare,
+            proc = proc_id,
+        );
+        let doc = aura_research::parse_campaign(&campaign_text).expect("campaign parses");
+        assert_eq!(reg.validate_campaign_refs(&doc, &resolve).expect("io ok"), Vec::new());
+
+        // unknown process + unknown strategy ids
+        let mut missing = doc.clone();
+        missing.process.r#ref = DocRef::ContentId("00".into());
+        missing.strategies[0].r#ref = DocRef::ContentId("11".into());
+        let faults = reg.validate_campaign_refs(&missing, &resolve).expect("io ok");
+        assert!(faults.contains(&RefFault::ProcessNotFound("00".into())));
+        assert!(faults.contains(&RefFault::StrategyNotFound("11".into())));
+
+        // unknown axis name
+        let mut bad_axis = doc.clone();
+        let ax = bad_axis.strategies[0].axes.remove(&real_name).unwrap();
+        bad_axis.strategies[0].axes.insert("nope".into(), ax);
+        assert!(reg
+            .validate_campaign_refs(&bad_axis, &resolve)
+            .expect("io ok")
+            .iter()
+            .any(|f| matches!(f, RefFault::AxisNotInParamSpace { axis, .. } if axis == "nope")));
+
+        // declared kind != the param's kind
+        let wrong_kind = if matches!(real_kind, aura_core::ScalarKind::I64) {
+            aura_core::ScalarKind::F64
+        } else {
+            aura_core::ScalarKind::I64
+        };
+        let mut mismatched = doc.clone();
+        // values were parsed under the right kind; re-kind the axis only —
+        // the check compares DECLARED kind vs the param's kind
+        let kept_values = mismatched.strategies[0].axes[&real_name].values.clone();
+        mismatched.strategies[0].axes.insert(
+            real_name.clone(),
+            Axis { kind: wrong_kind, values: kept_values },
+        );
+        assert!(reg
+            .validate_campaign_refs(&mismatched, &resolve)
+            .expect("io ok")
+            .iter()
+            .any(|f| matches!(f, RefFault::AxisKindMismatch { axis, .. } if axis == &real_name)));
+
+        // identity ref resolves via the store scan
+        let identity_json = blueprint_identity_json(&composite).expect("identity form");
+        let identity_id = aura_research::content_id_of(&identity_json);
+        let mut by_identity = doc.clone();
+        by_identity.strategies[0].r#ref = DocRef::IdentityId(identity_id);
+        assert_eq!(reg.validate_campaign_refs(&by_identity, &resolve).expect("io ok"), Vec::new());
+    }
+
     #[test]
     fn check_r_metric_accepts_r_and_refuses_pip() {
         assert!(check_r_metric("expectancy_r").is_ok());

crates/aura-research/Cargo.toml

@@ -0,0 +1,14 @@
+[package]
+name = "aura-research"
+edition.workspace = true
+version.workspace = true
+license.workspace = true
+publish.workspace = true
+
+[dependencies]
+aura-core = { path = "../aura-core" }
+serde = { workspace = true, features = ["derive"] }
+serde_json = { workspace = true }
+# sha2: content-addressed ids for research documents (mirrors aura-cli's
+# topology_hash use, per-case C16 review, SHA256 user-settled).
+sha2 = "0.10"

docs/plans/0106-research-artifact-vocabularies.md

@@ -1724,27 +1724,42 @@ import.)
 
 - [ ] **Step 2: Dev-deps for the fixture, if absent**
 
-In `crates/aura-registry/Cargo.toml` `[dev-dependencies]` add whichever of
+In `crates/aura-registry/Cargo.toml` `[dev-dependencies]` add if missing:
-these is missing:
 
 ```toml
 aura-std = { path = "../aura-std" }
-aura-composites = { path = "../aura-composites" }
 ```
 
+(NOT aura-composites: its shipped composites — vol_stop, risk_executor* —
+all route through `LinComb`, whose builder needs a structural arity
+argument and is therefore deliberately absent from the zero-arg
+`std_vocabulary` roster, so none of them can round-trip through
+`blueprint_from_json` with the by-type-name resolver this test uses. The
+fixture below hand-builds a minimal loadable composite instead.)
+
 - [ ] **Step 3: Append the referential test (tests module). The fixture is
   GENERIC over the composite's real param space — no hardcoded param names**
 
 ```rust
     #[test]
     fn referential_tier_resolves_refs_and_checks_axes() {
+        use aura_engine::{blueprint_to_json, Composite};
         use aura_research::{Axis, DocRef};
 
         let reg = Registry::open(temp_family_dir("referential_tier"));
         let resolve = |t: &str| aura_std::std_vocabulary(t);
 
-        // a real public composite with an open param space
+        // A minimal fixture composite with one OPEN param, built from a
-        let composite = aura_composites::vol_stop(4, 1.5);
+        // zero-arg `aura-std` node (`Bias`) so `std_vocabulary` can actually
+        // resolve it on load (see Step 2's note on why no aura-composites
+        // composite qualifies).
+        let composite = Composite::new(
+            "fixture",
+            vec![aura_std::Bias::builder().named("b").into()],
+            vec![],
+            vec![],
+            vec![],
+        );
         let space = composite.param_space();
         let real = space.first().expect("fixture composite has an open param");
         let real_name = real.name.clone();
@@ -1833,10 +1848,8 @@ Anchor notes: `blueprint_to_json`/`blueprint_identity_json` — import from
 the engine re-export the way the crate already imports engine items; if
 `aura_core::ScalarKind`/`Scalar` are not yet dev-visible, they come through
 the existing `aura-engine` dependency's re-exports or a direct
-`aura-core` path dep (mirror whichever the crate already uses). If
+`aura-core` path dep (mirror whichever the crate already uses). The test
-`vol_stop(4, 1.5).param_space()` turns out EMPTY, use
+is generic and needs only ONE open param from the fixture composite.
-`aura_composites::risk_executor_vol_open(1.0)` instead — the test is
-generic and needs only ONE open param.
 
 - [ ] **Step 4: Run the test**