The built-in sample that `aura graph` renders and `aura sweep` runs is now a
"trend + momentum, weighted blend" strategy that exercises four authoring/viewer
capabilities the single-level sample left dark:
- multiply-nested composites: root -> signals -> {trend = sma_cross, momentum = macd};
- a multi-param node inside a composite: blend = LinComb(3) (weights[0]/[1] shown);
- a multi-output node: momentum (the macd composite) re-exports macd/signal/histogram,
two of which the blend consumes;
- bind(): blend.weights[2] is fixed as a structural constant, so it drops out of the
sweepable param surface (and renders absent).
The new `signals` composite reuses the existing sma_cross/macd builders unchanged; the
sweep re-paths its axes to the eight free params (still a 4-point grid) and runs an
18-tick warm-up stream (showcase_prices) so the MACD EMA-of-EMA chain and the all-Any
LinComb join warm up. The render fixture (sample-model.json) is re-captured, and a new
headless guard (viewer_nested_depth) pins that the viewer renders the nesting to depth 2
with valid Graphviz ids. Pure authoring over already-shipped primitives — no engine or
graph-viewer.js change.
The flat run_sample, run_macd/macd(), the inline model_golden test, and the viewer are
untouched. A third sweep-param pin in tests/cli_run.rs (the aura sweep E2E) was re-pathed
in lockstep with its in-crate twin.
closes#62
A leaf primitive built with `.named("fast")` now renders its `aura graph` viewer
box head as `fast: SMA[length]` — the instance name as a `:` declaration prefix.
An unnamed leaf keeps the bare `SMA[length]`.
Engine half: a new raw `PrimitiveBuilder::instance_name() -> Option<&str>` (the
explicit name, default not resolved) feeds a conditional leading `"name"` field
in prim_record, present only for an explicitly-named node; both golden twins
(inline model_golden + sample-model.json) re-captured. Viewer half: adaptNodes
carries `name`, the genDot leaf-emit forwards it, and cellLabel prepends the
`name: ` prefix when present (composites stay bare). `named()`'s non-empty
debug_assert is unchanged, its doc re-grounded to the now load-bearing Some/None
invariant (knob-address segment + prefix switch).
The name is a render/model-only debug symbol — dropped at lowering (C23), and the
model stays deterministic (C14). Parameter ganging (one knob, several nodes) was
considered and spun off as an explicit composite-shared-param idea (#61), not via
name collision (param_space is injective, C12/C19).
closes#58
Observed bug in the graph viewer shipped in 66dff88: clicking `[+]` to
inline-expand a composite in `aura graph`'s HTML output renders a broken graph —
the interior nodes' edges collapse onto a single phantom node "0" with uncoloured
arrows, and the real pin docking is lost. The collapsed view and the body-drill
view both render correctly; only inline-expand breaks.
Root cause: in graph-viewer.js, genDot's `build` forms a node id by path-joining
the model keys (`cid = cpath.join("__")`). The aura model keys interior nodes by
numeric index (C23), so an inline-expanded composite yields ids like `0__0`. A
DOT identifier that begins with a digit but is not a pure numeral is invalid —
Graphviz parses `0__0` as the numeral `0` plus a leftover token, collapsing
`0__0`/`0__1`/`0__2` onto one phantom node `0`. The collapsed view survives only
because its ids are single-digit numerals (accidentally valid); the prototype
survived because it used name keys, never digit-prefixed ones. The bug surfaced
only against the real index-keyed model.
Fix: letter-prefix the id (`cid = "n" + cpath.join("__")`) so every id is a valid
identifier by construction. The fix lives in the viewer (DOT-syntax is its
responsibility), not in model_to_json — the model's index keys are correct, and
leaking the Graphviz identifier constraint into the model would be a leaky
abstraction.
RED-first guard (this establishes the project's first JS test infra, since the
viewer now carries real logic — normalizeModel + genDot — that the model golden
does not cover):
- tests/viewer_dot_ids.mjs: a headless node guard that loads the real viewer
module, discovers the expandable composites from a collapsed pass, inline-
expands them, and asserts every node id in the produced DOT is a valid Graphviz
identifier (quoted | pure numeral | bare identifier). Version-independent — it
checks the identifier grammar, not a render. The expand keys are DISCOVERED via
genDot's `expandable`, never hard-coded, so the guard cannot go vacuously green
when the fix reshapes the id (a hard-coded "0" key would stop matching "n0" and
silently test only the collapsed view — a false-green this guard refuses).
- tests/viewer_dot.rs: a Rust integration test that shells out to `node`,
wiring the guard into `cargo test --workspace`. `node` is required: if it is
absent the test fails (a skipped guard is not a guard).
- tests/fixtures/sample-model.json: the triggering fixture (numeric index keys,
a nested composite).
- graph-viewer.js: made node-loadable (typeof window/document guards + a
module.exports) so the guard can exercise the pure generator headless. This is
a test-enabler, behaviour-neutral in the browser — verified by the existing
render_html / graph_emits self-containment tests staying green and by the
expand re-rendering correctly through the vendored WASM-Graphviz.
Verified RED against the unfixed viewer (the guard trips on `0__0`/`0__1`/`0__2`
and the Rust test fails), GREEN after the prefix; full `cargo test --workspace`
green; clippy clean. Rendered the expanded state through the vendored
WASM-Graphviz to confirm the interior wires up correctly (price -> both SMA
series, SMA value -> Sub lhs/rhs, Sub -> signal, all f64-blue).
Brummel