The verifier called `Validation::set_issuer` and listed `iss` in
`set_required_spec_claims`. Cognitum access tokens have **no `iss` claim** — the
real claim set is typ, sub, account_id, org_id, workspace_id, client_id, scope,
family_id, jti, iat, exp, setup, workload. So the verifier rejected every
genuine token with a flat 401.
The whole 41-test suite was green throughout, because the test fixtures included
an `iss` the real thing does not have. The tests validated my assumption instead
of the platform. Found only by pointing a real token at a real server; no amount
of additional unit testing against the same wrong fixture would have caught it.
`valid_claims()` now mirrors production exactly, with a comment saying why, so
the next person cannot reintroduce the drift by "fixing" an incomplete-looking
fixture.
What binds a token to its issuer, then: the JWKS. We accept only signatures made
by a key served from the configured jwks_uri, so a valid signature IS proof of
issuer. meta-llm's verifier — the org's only other resource-side verifier —
checks neither `iss` nor `aud`, for the same reason. `VerifierConfig.issuer`
stays, now documented as JWKS-derivation and logging rather than a claim
assertion.
Three tests replace the two that asserted issuer behaviour that cannot exist:
a real-shaped (iss-less) token verifies; an unrelated `iss` changes nothing (so
identity adding one later cannot silently start failing tokens); and a token
signed by a different key is still refused — the complement that proves removing
the issuer check did not remove the boundary.
Also: report the granted scope from the token's `scope` claim when the
/oauth/token envelope omits it, which identity's does. `login` and `whoami` said
"(not reported by the server)" while the authoritative answer sat inside the
token they had just stored. The claim-peek helper is display-only and documented
at length as NOT verification — a client reading its own freshly issued token is
a different situation from a server reading a stranger's.
VERIFIED END TO END against production (gate G-3, previously open):
no credential GET /api/v1/models -> 401
garbage bearer GET /api/v1/models -> 401
real sensing:read token GET /api/v1/models -> 200
real sensing:read token GET /api/v1/recording/list-> 200
real sensing:read token POST /api/v1/train/start -> 401
real sensing:read token POST /api/v1/train/stop -> 401
Token obtained by a human through `wifi-densepose login` against
auth.cognitum.one; server ran with RUVIEW_OAUTH_ISSUER set, JWKS fetched live
at boot (key_count=1).
Tests: 82 with --features login (58 unit + 22 matrix + 2 doctests).
Co-Authored-By: Ruflo & AQE
The ADR said the login flow was "not in this crate". It is now, gated behind a
non-default `login` feature. The original line existed to keep the sensing
server lean; a feature gate achieves that without forcing the Tauri desktop app
to grow a second copy of a PKCE + rotating-refresh implementation — the kind of
duplication that drifts and then disagrees about something subtle.
Recording the change rather than letting the ADR quietly go stale, which is the
exact failure this session hit twice in other repos' ADRs.
Co-Authored-By: Ruflo & AQE
Phase 1 could verify a token and phase 3 could gate on one, but there was no
way for a user to OBTAIN one. This closes that: sign in with a Cognitum account
and get a token a RuView sensing server accepts, instead of everyone sharing
one static RUVIEW_API_TOKEN string.
Lives in `ruview-auth` behind a non-default `login` feature rather than in the
CLI, so the Tauri desktop app can reuse it instead of growing a second copy.
A server built with default features still gets the verifier and nothing else —
no reqwest, no tokio net, no browser launcher. (This amends ADR-271's "no login
flow in this crate" note; the reason for that line was to keep the server lean,
and a feature gate achieves it without duplication.)
Ported from meta-proxy's src/oauth/, cross-checked against musica's
cognitum_provider.rs — two independent implementations against this same AS.
Where they agree, this follows both: redirect path EXACTLY /oauth/callback,
60-second refresh skew, OOB fallback on SSH/CONTAINER//.dockerenv.
Refresh is the part with teeth. Identity rotates refresh tokens with reuse
detection, so presenting a spent one revokes the whole session family. Both
obvious implementations are wrong: refreshing concurrently looks like replay,
and retrying a failed refresh with the same token IS the replay. So
`Session::ensure_fresh` holds an async mutex across the await, re-checks expiry
after acquiring it (the waiter usually finds the work already done), persists
the rotated token BEFORE returning it, and never retries. A missing expires_at
counts as expired rather than being given a guessed default.
Least scope by default: `login` requests `sensing:read`. `--admin` adds
`sensing:admin` explicitly, and requests both because there is no scope
hierarchy server-side. A session that streams poses should not casually hold
the capability to delete the model it streams through.
Credentials are written atomically and 0600 (temp file, chmod BEFORE rename) —
the same discipline the seed applies to its cloud key. `logout` is local-only
and says so: it makes this machine unable to act as you, but revoking the
session everywhere is an account-level action.
Also `whoami`, which reports whether the stored token is live — an
expired-looking session is the most common reason a command starts 401ing, and
it should be visible directly rather than inferred from a failure elsewhere.
Verified against PRODUCTION, not just locally: authorize URLs built by this
exact code path return HTTP 200 from auth.cognitum.one for both
`sensing:read` and `sensing:read sensing:admin`, which exercises the real
client_id, scope encoding, PKCE parameters and redirect_uri shape.
Tests: 74 with --features login (51 unit + 21 verifier matrix + 2 doctests),
including the RFC 7636 Appendix B vector, multi-scope URL encoding (a space
that is hand-formatted rather than encoded silently truncates the request), a
real TCP callback round-trip, callback timeout, 0600 permissions asserted on
disk, atomic-save leaving no temp file, and refresh-window boundaries.
Unchanged: 43 with default features, 501 in the sensing server.
Co-Authored-By: Ruflo & AQE
SensingClient now accepts a `token=` argument, defaulting to the
RUVIEW_API_TOKEN env var, and sets `Authorization: Bearer <token>`
directly on the WebSocket upgrade — no browser-style ticket workaround,
since Python can set the header on the handshake. Empty/unset token
means no auth (auth-disabled path unchanged).
websockets version-compat: the header keyword was renamed in the
`>=12.0` range this package pins (`extra_headers` <= 13,
`additional_headers` >= 14). Resolved by runtime detection —
`_select_header_kwarg()` inspects `websockets.connect`'s signature and
passes the correct keyword — rather than raising the floor to `>=14`.
Chosen deliberately so existing users on older websockets are not forced
to upgrade; keeps the `websockets>=12.0` pin valid.
Scope: only ws.py connects to the auth-enabled sensing-server. mqtt.py
already supports username/password (its own auth convention); ha.py and
primitives.py do no network I/O. OAuth (ADR-271) is out of scope.
Tests: constructor token, env-var token, constructor-overrides-env,
no-token/empty-token (no header), a parametrized test across both
websockets kwarg conventions, and an end-to-end test asserting the
bearer reaches the in-process server's upgrade request.
Wires `ruview-auth` into `bearer_auth.rs`. `RUVIEW_OAUTH_ISSUER` enables it;
unset, nothing changes.
Layering, in order:
1. `RUVIEW_API_TOKEN` set and the bearer matches exactly -> allow. Byte-for-
byte today's behaviour.
2. Otherwise, if OAuth is configured, verify the bearer as a Cognitum access
token and require the scope the route needs.
3. Otherwise 401.
The static compare goes first for compatibility, not security: a matching
static token is not a JWT and a JWT never matches the static token. It means an
existing deployment behaves identically even with OAuth switched on.
Scope gate (`required_scope_for`), split by blast radius per ADR-060 —
"can this destroy something", not how many routes it covers:
sensing:admin /api/v1/train/* (hours of Pi CPU, writes models)
DELETE /api/v1/models/{id} (irreversible)
DELETE /api/v1/recording/{id} (irreversible)
sensing:read everything else
Deliberately NOT admin: model load/unload and recording start. They mutate
server state but destroy nothing, and gating them would push routine dashboard
use into requesting delete capability — the opposite of least privilege.
The legacy static token stays un-scope-gated. It predates scopes and carries no
claims, so narrowing it would be a silent breaking change to deployments using
it; migrating to OAuth is how an operator opts into the finer split.
FAIL CLOSED at boot. If OAuth is requested but cannot work — empty issuer, or a
JWKS we cannot fetch — the server logs why and exits rather than serving.
Starting anyway would silently downgrade an operator who asked for OAuth to
either an open API or a shared-secret one, with no signal it happened. The JWKS
is warmed eagerly for the same reason: a bad `jwks_uri` should die at boot with
a legible message, not surface as a puzzling 401 an hour later.
The verified `Principal` is attached to request extensions, so handlers and
audit logs can attribute a request (`sub`, `account_id`, `org_id`,
`workspace_id`, `jti`) instead of knowing only "someone had the secret". That
is the point of moving off a shared bearer.
Verification failures are logged with the reason and returned as a flat 401 —
the reason is useful to an operator and equally useful to an attacker probing
for which claim to forge next.
Also aligns `ruview_auth::extract_bearer` to match the scheme
case-insensitively (RFC 7235 §2.1). The sensing server has always done this
deliberately, with a comment saying why; the two layers disagreeing about what
a valid header looks like would be a latent bug.
Tests: 16 new in `bearer_auth::oauth_tests`, driving a real Router end to end
(request -> middleware -> verifier -> handler) with ES256 tokens signed by a
runtime-generated key. Covers the scope policy as a pure function, read-scoped
tokens refused on delete and train, admin-scoped tokens allowed, an
`inference`-only token from another Cognitum product refused on every route,
garbage and absent bearers, both legacy-token layering directions, the
principal reaching a handler, and the unset case remaining a no-op.
`cargo test -p wifi-densepose-sensing-server --lib --no-default-features`:
501 passed, 0 failed. `ruview-auth`: 43 passed across both feature configs.
Co-Authored-By: Ruflo & AQE
The previous commit referenced ADR-271 in five places without the ADR existing.
This writes it.
Records the decision and, more importantly, the direction — RuView verifies
tokens users present, it does not obtain tokens to call Cognitum. Every other
Cognitum OAuth integration in the org is the client side of a plane RuView does
not have; the sole relevant precedent is meta-llm's oauthBearer.ts.
Covers: why offline verification is a requirement rather than an optimisation
(Pi-class hosts lose WAN, and no introspection endpoint exists); why the
accept-rule is a port and not a design; why long-lived setup/workload
credentials are refused (no database to check revocation); why scope — not
`aud`, not `client_id` — is the capability boundary; and the alternatives,
including the `cog_`-minting approach that was tried org-wide and found broken
against production.
Co-Authored-By: Ruflo & AQE
RuView's `/api/v1/*` is gated today by `RUVIEW_API_TOKEN`: one shared secret,
no expiry, no per-user attribution. This adds the verifier half of replacing
that with Cognitum identity — RuView as an OAuth **resource server**, so a user
signs in to their own sensing server with their Cognitum account.
Note the direction: RuView does not call Cognitum. It never obtains a token for
its own use; it verifies tokens users present. Every other Cognitum OAuth
integration in the org (meta-proxy, musica, metaharness, the dashboard CLI) is
the client side of a plane RuView doesn't have. The one existing resource-server
verifier is `meta-llm/src/auth/oauthBearer.ts`, and this crate is a port of its
accept-rule — divergence would be a bug, not a preference: a token meta-llm
rejects must not be one RuView accepts.
Verification is OFFLINE, and that is a requirement rather than an optimisation.
RuView runs on Pi-class hardware that loses WAN, and identity publishes no
introspection endpoint even when the network is up. So: ES256 over identity's
published JWKS, cached by `kid`.
What it accepts, mirroring oauthBearer.ts:
typ == "access" AND NOT setup AND NOT workload AND account_id non-empty AND
exp in the future AND iss matches verbatim AND the required scope is held.
Long-lived setup (365-day) and workload credentials are refused outright for
identity's own stated reason: their revocation lives in `oauth_setup_tokens`,
and RuView — like meta-llm — has no database to check it. A 15-minute access
token needs no revocation round-trip because it expires faster than revocation
propagates; a 365-day one does.
Scope is the capability boundary, and it has to be. Cognitum access tokens carry
no `aud`, and `client_id` cannot substitute because clients borrow each other's
registrations (musica ships DEFAULT_CLIENT_ID = "meta-proxy"). `sensing:read`
covers streams and inference; `sensing:admin` covers training, model delete and
recording delete. No hierarchy — admin does not imply read; consent means what
it said. Registered in identity migration 0016 (cognitum-one/dashboard#116).
Design notes:
- `ureq`, not `reqwest`: the sensing server deliberately chose ureq as "the
smallest" HTTP client, and pulling reqwest in here would reverse that for the
whole graph. Transport sits behind a `JwksFetcher` trait, so a host can supply
its own and take no HTTP dependency at all (feature `ureq-transport`, default).
- A JWKS refetch failure is survivable while a key set is already cached: a key
that verified a minute ago has not stopped being valid because the network
blipped, and failing closed there logs every user out of their own sensing
server whenever their internet wobbles. We fail closed in exactly one case —
no key set has ever been fetched.
- Unknown `kid` forces one refetch so rotation is picked up without waiting out
the TTL, rate-limited so junk-kid tokens can't become a request amplifier
aimed at identity.
- Expiry is reported distinctly from a bad signature: on an RTC-less Pi that is
usually a clock-sync fault, and an operator must be able to tell them apart.
- Test keypairs are generated at runtime, never committed. This repo tracks zero
`.pem` files and committed key material shouldn't start here — it also means
no fixture can drift out of sync with the JWKS it's served by.
Tests: 41 green (19 unit + 21 matrix + 1 doctest) under BOTH
`cargo test --no-default-features` (the repo's canonical gate) and default
features. The matrix signs real ES256 tokens rather than asserting on strings,
and covers alg:none, forged signature, spliced payload, unknown kid, expired,
leeway boundaries both sides, wrong issuer, issuer differing only by a trailing
slash, missing/!=access typ, setup and workload smuggled onto typ=access,
missing and empty account_id, and scope escalation.
The highest-value case is `g2_a_genuinely_valid_token_from_another_cognitum_
product_cannot_reach_the_sensing_surface`: a correctly signed, unexpired,
right-issuer, right-typ token with client_id=meta-proxy and scope=inference is
rejected. Nothing about its signature or identity claims distinguishes it — only
scope does. A naive verifier accepts it, and an `inference` token becomes a key
to someone's home sensor.
No wiring into the sensing server yet; this crate is verification only, with no
login flow and no outbound Cognitum calls.
Co-Authored-By: Ruflo & AQE
The python/ PyO3+maturin wheel (wifi-densepose, ADR-117/185) had ZERO
per-PR CI coverage: pip-release.yml only fires on release triggers, so a
PR could break the wheel build, a native-Rust parity test, or the wheel
budget and nothing in the gating CI would catch it until release day.
New .github/workflows/python-ci.yml (path-scoped to python/** + the
backing crates, following the repo convention cf. firmware-ci.yml --
GitHub has no per-job paths filter and no workflow here uses a
change-detection action, so a dedicated path-scoped workflow is the
idiomatic, no-new-dependency choice):
* parity-tests: cargo build of the aether/mat/train backing crates
(fast-fail attribution), then `maturin develop --features sota` +
`pytest python/tests/ -q`. sota builds the aether/meridian/mat
compiled submodules so their SHA-256 parity tests against native
Rust actually run (those test modules import the submodules at
collection time -- without the features they error, not skip).
* wheel-size-budget: builds the DEFAULT (no-features) wheel and fails
if it exceeds the ADR-117 5.4 5 MiB budget -- the real numeric
enforcement a string fix-marker cannot provide.
Two fix-marker guards (scripts/fix-markers.json):
* RuView#1387-rvf-filename-collision: training_api.rs next_model_id()
must keep the microsecond timestamp (%6f) + AtomicU64 counter; the
forbid uses a negative lookahead so it fires only on a bare
second-resolution format lacking the disambiguator. Guards the
CI-caught collision fixed on this branch in 8409f434c.
* RuView#1387-default-wheel-budget-config: default Cargo features
empty, SOTA deps optional, maturin strip=true -- the config that
keeps the default wheel small. Bytes are enforced by the CI job
above, not a string marker (per the marker system's limits).
Verification:
- python scripts/check_fix_markers.py -> All 27 fix markers present (0 regressed)
- all 26 .github/workflows/*.yml parse (yaml.safe_load, utf-8)
- cargo build -p wifi-densepose-aether -> Finished (exit 0)
- forbid regexes confirmed: fire on a simulated regression, clean on the current tree
Co-Authored-By: claude-flow <ruv@ruv.net>
Reverts the OIDC-only publish path from cc153e8b5. Restores password:
${{ secrets.PYPI_API_TOKEN }} on all four publish steps and drops the
id-token: write / environment: pypi blocks from publish-v2 and
publish-tombstone. OIDC-only is a net regression until the manual
pypi.org Trusted Publisher registration exists — publishing would fail
'no trusted publisher configured'. Token auth is the active working path.
OIDC design retained as documented TODO(ADR-184 P1b) in the header; the
switch becomes a dedicated gated follow-up once the owner confirms
registration. Restores compliance with fix-marker RuView#786-pypi-token-auth.
Refs ADR-184 P1.
Closes the ADR-185 §13.a follow-up (the tractable, data-independent one of
the three §6.7 gaps): the bound EmbeddingExtractor was random-Xavier-init
only, with NO way to load real weights — so it was structurally untrained.
This adds the *capability* to load weights whenever a trained checkpoint
exists. It does NOT itself produce trained/SOTA embeddings and does NOT
close §6.7 (still no trained checkpoint, no labeled data, no eval harness).
Serialization: not greenfield — EmbeddingExtractor already had
flatten_weights()/unflatten_weights() (flat Vec<f32>). wifi-densepose-aether
is a deliberately dependency-free std-only leaf crate (ADR-185 §13), so
rather than add safetensors/serde/bincode (which would undo the zero-dep
property), the on-disk format is raw little-endian f32 with a 12-byte header
(magic "AETHERW1" + u32 param count) — zero new deps.
Native (v2/crates/wifi-densepose-aether/src/embedding.rs):
- EmbeddingExtractor::save_weights(path) / load_weights(path). load_weights
never panics: errors on unreadable file, short/oversized payload, bad
magic, or a param-count mismatch (delegated to unflatten_weights).
- Default (no weights) construction is UNCHANGED — still random init,
clearly labeled untrained. Purely additive.
Python binding (python/src/bindings/aether.rs, .pyi):
- EmbeddingExtractor.load_weights(path) / save_weights(path) / param_count,
GIL-released, ValueError on bad input.
Tests (both Rust and Python, per the task):
- Rust unit (embedding.rs): load_weights_actually_replaces_weights_and_
round_trips proves the loaded weights MOVE the embedding away from the
random-init baseline (not a silent no-op), match the source extractor
(round-trip), and are bit-identical after the file round-trip; plus a
bad-magic/wrong-count rejection test.
- Cross-language golden (aether_weights_parity.rs + test_aether.py): a
shared deterministic weight formula (w[i]=k/65536-0.5, exact in f32+f64)
written to the AETHER format; both native Rust and the Python binding load
it and must produce the byte-identical embedding SHA-256
(tests/golden/aether_loaded_embedding.sha256) — proving the binding's
load path is bit-identical to native, and (vs the random baseline) that
the loaded weights are actually used.
Verified:
cargo test -p wifi-densepose-aether 98 passed, 0 failed
cargo test --features aether --test aether_parity
--test aether_weights_parity 3 passed, 0 failed
maturin develop --features aether + pytest test_aether.py 13/13 pass
default cargo build (no aether feature) clean
Root cause: the trained-model filename was `trained-{type}-{%Y%m%d_%H%M%S}` — a
**second-resolution** timestamp. Two training runs of the same type that finish
in the same wall-clock second compute the SAME path and silently overwrite each
other's artifact. This is a real robustness gap (a second rapid run clobbers the
first model), and it surfaced as a flaky test on the Linux CI runner:
`http_train_start_produces_model_and_streams` asserts (via a before/after
directory diff) that a new `.rvf` appeared, but the concurrent
`training_job_streams_real_progress_and_writes_model` test — same `supervised`
type, same second — wrote to the identical path and then deleted it in its own
cleanup, so the diff saw no new file. Windows scheduling happened to avoid the
same-second overlap; Linux CI hit it (the sibling `TRAIN_ENV_LOCK` holders then
failed via std::sync::Mutex poison-cascade after the first panic).
Reproduced locally by isolating the three model-writing tests (which forces the
overlap): ~10/20 runs flaked before; 0/30 after.
Fix: `next_model_id()` appends microseconds (`%6f`) plus a monotonic per-process
`AtomicU64` counter, so every run gets a distinct path even for same-microsecond
concurrent completions. Pinned by `model_ids_are_unique_per_call` (1000 ids, all
distinct) so the scheme can't regress to non-unique. No test-level band-aid
(no added sleeps / serialization) — the collision is removed at the source.
Verified: cargo test -p wifi-densepose-sensing-server -p wifi-densepose-train
--no-default-features — bin 218 passed / 0 failed, all suites 0 failed; the
isolated-tests stress loop is 0/15 flakes post-fix.
ADR-185 §13 flagged the `[aether]` wheel as linking the whole
`wifi-densepose-sensing-server` Axum/tokio/worldgraph/ruvector tree via the
non-optional deps of that crate. Hoist the pure-compute AETHER stack out so the
binding depends only on pure Rust.
New crate `wifi-densepose-aether` (std-only, ZERO external deps) holds the four
self-contained modules the AETHER surface needs — `embedding`, `graph_transformer`,
`sona`, `sparse_inference` (verified: no serde/tokio/axum/crate:: refs outside the
set; the four form a closed dependency graph). `wifi-densepose-sensing-server`
now `pub use`s them from the leaf crate, so its own code (`crate::embedding`, …)
and public API (`wifi_densepose_sensing_server::embedding`, …) are unchanged —
`main.rs`/`trainer.rs` untouched. The Python binding + parity test + `[aether]`
feature repoint at the leaf crate; the sensing-server dep is dropped from python.
MEASURED wheel size (maturin build --release --features aether --strip; each
wheel install-verified to actually contain `AetherConfig`):
BEFORE (aether -> sensing-server): 369,782 B (~361 KB)
AFTER (aether -> leaf crate): 319,719 B (~312 KB)
HONEST CORRECTION to the §9/§13 premise: the pre-hoist wheel was NOT over the
ADR-117 §5.4 5 MB budget — it was ~361 KB, ~14x under. Linker dead-code
elimination (--gc-sections on the pyo3 cdylib) already strips the unreached
server tree because the binding only reaches pure-compute symbols. So this is
"blows the budget" CLAIMED, not MEASURED. The hoist is still worthwhile and
lands the real, measured wins:
- `[aether]` build time 71s -> 12s (no server tree to compile),
- python/Cargo.lock -1238 lines (server tree no longer resolved),
- ~50 KB smaller wheel,
- honest, minimal declared dependency surface + removes the latent risk that a
future change makes some server code reachable and *then* bloats the wheel.
Verified this session (real counts):
- cargo test --features aether --test aether_parity: 2 passed / 0 failed
- pytest tests/test_aether.py (maturin develop --features aether): 9 passed
- cargo test -p wifi-densepose-sensing-server --no-default-features: 217 bin +
388 lib, 0 failed (no regression)
- cargo test -p wifi-densepose-aether: 96 passed (the relocated unit tests)
The §9/§13 wheel-size narrative in ADR-185 (owned by the ADR author) should be
corrected to reflect the measured reality — flagged to the team lead.
ADR-184 P2 (version promotion + changelog) — version-metadata prep only;
triggers NO publish (P3's real PyPI upload is still gated on the manual
Trusted Publisher registration on pypi.org).
- python/pyproject.toml: version 2.0.0a1 -> 2.0.0; classifier
"Development Status :: 3 - Alpha" -> "5 - Production/Stable".
- python/ruview-meta/pyproject.toml: version 2.0.0a1 -> 2.0.0; same
classifier bump; repointed the wifi-densepose dependency pins
(base + [client]) from ==2.0.0a1 to ==2.0.0 so the stable ruview
meta-package depends on the stable wifi-densepose, not the alpha.
- python/wifi_densepose/__init__.py: runtime __version__ 2.0.0a1 -> 2.0.0
(caught during the sanity check — it would otherwise contradict the
stable wheel version reported by `wifi_densepose.__version__`).
- CHANGELOG.md: [Unreleased] entry recording the promotion.
Sanity-checked that the core package genuinely deserves "Production/Stable"
(not just a version-string change):
maturin build --release --strip -> default wheel 279 KB
pytest python/tests/ (excluding [aether]/[meridian]/[mat] extra modules)
-> 185 passed, 0 failed (smoke/keypoint/pose/vitals/bfld/security/
WS+MQTT client)
post-bump: wheel installs as 2.0.0, __version__ = 2.0.0, smoke 6/6 pass.
No publish/tag/release commands were run.
Root cause (investigated, same rigor as the MAT fix): `wifi-densepose-train`
declared `wifi-densepose-nn` as a dependency but NEVER imported it anywhere
in src/ or bin/ — a dead dependency. (The tch-backend model path uses the
`tch` crate directly, not wifi-densepose-nn; `grep -r wifi_densepose_nn
src/` returns nothing.) That dead dep pulled `ort` (ONNX Runtime) +
reqwest/hyper into every downstream consumer, including the ADR-185
`[meridian]` wheel, which binds only train's pure-compute geometry /
rapid_adapt / eval modules + signal::hardware_norm.
`cargo tree -i wifi-densepose-nn` confirmed train was the SOLE path pulling
nn into the wheel; after removal, the package is absent from the graph.
Fix: remove the dead dep line from train's Cargo.toml (no code change, no
feature-gating needed — nothing referenced it). Not a hoist: the coupling
turned out to be nonexistent, not deep.
Measured (maturin build --release --features meridian --strip):
BEFORE: 1.8 MB AFTER: 1.7 MB (both already within the ADR-117 §5.4
<=5 MB budget). Honest note: unlike MAT — which instantiates OnnxSession
and thus bundles the ort native runtime (8.4 MB) — train never references
ort, so the native lib was dead-stripped and the wheel was already under
budget. The real win is removing ort/reqwest/hyper from train's dependency
graph and compile for ALL consumers, not a dramatic wheel shrink.
Verified:
cargo build -p wifi-densepose-train (default=[]) OK
cargo tree -i wifi-densepose-nn -> no longer in the graph
cargo test --features meridian --test meridian_parity 2/2 pass
maturin develop --features meridian + pytest test_meridian 13/13 pass
Root cause (investigated, not assumed): the heavy deps in a [mat] wheel
were NOT on the survivor-detection/triage path the Python binding uses.
`wifi-densepose-mat` pulled `wifi-densepose-nn` as a NON-optional dep, and
nn's default `onnx` feature drags in `ort` (ONNX Runtime) + its
download/reqwest/hyper stack. But nn is used ONLY by mat's `src/ml/`
module (debris + vital-signs ONNX classifiers), which is an OPTIONAL
runtime enhancement: `DetectionPipeline.ml_pipeline` is `Option<_>`,
created only when `DetectionConfig::enable_ml` is set, and
`from_disaster_config` (the DisasterResponse path) never enables it. So
the ONNX stack was compiled-and-linked but never executed by the binding.
Fix = feature-gating (not a hoist — a hoist was unnecessary):
- `wifi-densepose-nn` made optional; new `ml` Cargo feature gates it.
- `default = [..., "ml"]` so existing consumers are unchanged; `api`
implies `ml` (the REST surface exposes `ml_ready`).
- `#[cfg(feature = "ml")]` on the `ml` module, its crate-root + prelude
re-exports, the `MatError::Ml` variant, and every ml touchpoint in
detection/pipeline.rs (ml_config/ml_pipeline fields, constructor,
process_zone enhancement block, enhance_with_ml/get_ml_results/
initialize_ml/ml_ready/ml_pipeline accessors, update_config).
- Fixed a latent feature-unification bug surfaced by dropping nn:
integration/hardware_adapter.rs uses `tokio::select!`, which needs
`tokio/macros`. That was only satisfied transitively via nn; now
declared explicitly on mat's own tokio dep so --no-default-features
builds compile.
The ADR-185 [mat] wheel already depended on mat with
`default-features = false, features = ["std"]`, so no python/ change was
needed — dropping `ml` from the default now actually removes ort.
Measured (maturin build --release --features mat --strip):
BEFORE (ml/ort in wheel): 8.4 MB (over the ADR-117 §5.4 <=5 MB budget)
AFTER (ml gated off): 2.0 MB (under budget) — 76% smaller
Verified:
cargo build -p wifi-densepose-mat (default+ml) OK
cargo build -p wifi-densepose-mat --no-default-features --features std OK
cargo test --features mat --test mat_parity 2/2 pass
maturin develop --features mat + pytest tests/test_mat.py 7/7 pass
(same detection result: 1 survivor, triage Delayed — behavior-transparent)
Completes ADR-186 phases P5 (dashboard honesty / fallback guarantee) and P6
(tests + witness) on top of the P1–P4 reconnection, and flips the ADR to Accepted.
P5 — dashboard honesty:
- Backend (training_api): a runtime enablement gate (`RUVIEW_DISABLE_SERVER_TRAINING`)
makes the three start endpoints return a structured `{enabled:false, reason,
cli:"wifi-densepose train-room"}` HTTP 409 (never a silent success) when server
training is disabled; `/api/v1/train/status` now carries an `enabled` flag. A runtime
flag (not a Cargo feature) so `--no-default-features` builds keep training ON (resolves
§9.4). Handlers return `Response` to carry the 409.
- Frontend (TrainingPanel.js): reads `enabled` off the status payload and disables the
Start/Pretrain/LoRA buttons with a CLI tooltip when server training is off; a rejected
start tears down the optimistically-opened WS and refreshes. The enabled path was
already fully wired (opens the WS before POST, renders live epoch/loss/ETA + terminal
state) — verified.
P6 — tests & witness:
- New HTTP-level tests in a `#[cfg(test)] mod adr186_http_tests` built on a minimal
`AppStateInner::minimal()` test-state helper: a live-socket test that completes a
genuine 101 WebSocket handshake (tokio-tungstenite) and receives a real progress frame
after a POST start; a 426-not-404 route-wired check; a full POST→poll-status→`.rvf`-exists
round-trip; and the disabled-409 structured-response test. Added `tokio-tungstenite`
dev-dep (version already in the workspace lock).
- CHANGELOG updated; ADR §6 ledger + §7 acceptance criteria checked off with evidence;
Status → Accepted. README/CLAUDE have no training route table (no edit needed).
Fixed a test-only parallelism race found under `cargo test --workspace`: two
model-writing tests deleted `.rvf`s by directory-diff and could remove a file a third
test asserted existed. Each test now cleans only its own artifact (data/models is
gitignored).
Verified this session: `cargo test -p wifi-densepose-sensing-server -p wifi-densepose-train
--no-default-features` — sensing-server bin 217 passed / 0 failed, all train suites 0
failed. Full `--workspace` re-run in progress to reconfirm untouched crates.
ADR-185 §4.2 pytest-benchmark micro-benchmarks + runnable examples +
README extras table for the aether/meridian/mat bindings.
- python/bench/test_bench_{aether,meridian,mat}.py — follow the existing
test_bench_vitals.py pattern (skipped by default; --benchmark-only).
- python/examples/{reid_from_csi,cross_room_calibrate,mat_triage}.py —
typed, runnable, mypy --strict clean.
- python/README.md — SOTA extras table + example links.
Measured on a RELEASE wheel (maturin develop --release --features sota),
reference machine per ADR-117 §10:
AETHER embed() mean ~150 us/window (target <2 ms) PASS
batch scaling 1/8/64: 140 / 1091 / 8509 us (linear, no O(n^2)) PASS
MERIDIAN normalize() mean ~2.2 us/frame (target <200 us) PASS
MERIDIAN encode() mean ~6.9 us (target <200 us) PASS
MAT ingest+scan_once() mean ~40 ms/256-frame (< 500 ms interval) PASS
Acceptance self-verification (ADR-185 §6), all run just now:
§6.1 default wheel 279 KB (<=5 MB); build_features has no p6-* feature PASS
§6.2 pytest tests/test_aether.py 9/9 PASS
§6.3 pytest tests/test_meridian.py 13/13 PASS
§6.4 pytest tests/test_mat.py 7/7 PASS
§6.5 benchmarks meet all targets (above) PASS
§6.6 parity harness: 3/3 SHA golden gates green (cargo test --features
sota, 6/6); CI *wiring* as a release gate is out of python/ scope PARTIAL
§6.7 SOTA accuracy bars on labeled fixtures: NOT met (no labeled
fixtures / trained models available; parity proves path-equality,
not accuracy) OPEN
§6.8 .pyi stubs present for all three; mypy --strict on the 3 examples PASS
§6.9 base wheel `import wifi_densepose.{aether,meridian,mat}` raises a
clear ImportError naming the extra PASS
No regression: 76 pre-existing tests pass on the default wheel.
Status NOT flipped to Accepted: §6.7 (accuracy bars) is unmet, §6.6 CI
wiring is pending, and the per-extra wheel-size hoists (sensing-server /
train / mat leaf crates) remain follow-ups. docs/adr/ is owned by another
agent this session, so the ADR ledger edit is deferred to that owner.
Bind the ADR-024 MAT (Mass Casualty Assessment Tool) disaster-survivor
detection + START triage surface into the wheel behind a gated [mat]
extra / Cargo `mat` feature, mirroring the upstream disaster/ML gating.
Also adds the [sota] superset extra (aether+meridian+mat).
Surface (bound against the REAL code at HEAD, not the ADR wishlist):
- DisasterType (9 variants) / TriageStatus (5, START) enums
- DisasterConfig (builder-backed, continuous_monitoring forced off)
- DisasterResponse: initialize_event / add_zone / push_csi_data /
scan_once / survivors / survivors_by_triage
- Survivor (id, triage_status, confidence, location, latest_vitals)
- VitalSignsReading (breathing/heartbeat rate, movement, confidence)
- ScanZone.rectangle / ScanZone.circle
push_csi_data + scan_once are GIL-released.
Honest deviations from ADR section 3.4 (documented in module header):
- ADR proposed adding a Rust-side sync scan_once() (section 11.3). That was
UNNECESSARY: the public async start_scanning() runs exactly one
scan_cycle and returns when continuous_monitoring == false. The binding
forces that flag off and drives one cycle on a private current-thread
tokio runtime -- NO change to wifi-densepose-mat.
- scan_cycle requires an active event + Active zone, which the ADR surface
omitted; initialize_event + add_zone are bound as required additions.
- Survivor.vital_signs is a *history* in the real code; bound as
Survivor.latest_vitals -> Optional[VitalSignsReading].
- DisasterType has 9 variants at HEAD (adds Landslide/MineCollapse/
Industrial/TunnelCollapse); all bound.
Parity (section 4.1, release-blocking): committed fixture mat_input.json
(synthetic breathing-modulated CSI stream) -> native Rust reference
(tests/mat_parity.rs, drives DisasterResponse directly) locks
tests/golden/mat_result.sha256 over a canonical
`count=<K>;triage_priorities=<sorted>` string (survivor UUIDs/timestamps
excluded as non-deterministic); pytest (tests/test_mat.py) runs the same
stream through the binding and asserts the identical hash. Both detect
exactly 1 survivor, triage Delayed. Honest: synthetic fixture proves
binding==native path equality, NOT live detection accuracy.
Verified:
cargo test --features mat --test mat_parity -> 2/2 pass
maturin develop --features mat + pytest tests/test_mat.py -> 7/7 pass
default cargo build clean, 0 mat/tokio refs in the default dep graph.
WHEEL-SIZE FINDING (ADR-185 section 9): default-features=false drops MAT's
`api` (axum) and `ruvector` features, but MAT still carries NON-optional
tokio (rt/sync/time), wifi-densepose-nn (ort/ONNX + reqwest/hyper),
rustfft, geo, ndarray. So a [mat] wheel exceeds the ADR-117 section 5.4
<=5 MB budget -- same leaf-crate-hoist follow-up as AETHER/MERIDIAN. The
default wheel is untouched (feature-gated).
Bind the ADR-027 MERIDIAN cross-environment domain-generalization surface
into the wheel behind a gated [meridian] extra / Cargo `meridian` feature.
Inference/adaptation path only (tch-free), per ADR-185 section 3.3.
Surface (bound against the REAL code at HEAD, not the ADR wishlist):
- HardwareType / HardwareNormalizer / CanonicalCsiFrame (from
wifi-densepose-signal::hardware_norm)
- MeridianGeometryConfig / GeometryEncoder (64-dim, permutation-invariant)
- RapidAdaptation / AdaptationResult (push_frame + adapt, LoRA deltas)
- CrossDomainEvaluator + mpjpe (from wifi-densepose-train, NO tch-backend)
All compute paths GIL-released (py.allow_threads).
Honest deviations from ADR section 3.3 (documented in the module header):
- ADR's RapidAdaptation.calibrate(windows) and AdaptationResult.converged
DO NOT EXIST. Real API is push_frame + adapt(); result carries
{lora_weights, final_loss, frames_used, adaptation_epochs}. Bound as-is.
- ADR's HardwareType.detect exposed as a staticmethod delegating to the
real HardwareNormalizer::detect_hardware.
- ADR's normalize(frame: CsiFrame, hw) is really normalize(amplitude,
phase, hw) over f64 vectors returning Result; bound faithfully.
- CanonicalCsiFrame fields are singular amplitude/phase (ADR said plural).
Training-time types (DomainFactorizer, GradientReversalLayer,
VirtualDomainAugmentor) are out of P6 scope (need the libtorch tier).
Parity (section 4.1, release-blocking): committed fixture
meridian_input.json -> native Rust reference (tests/meridian_parity.rs,
calls hardware_norm + geometry + rapid_adapt directly) locks
tests/golden/meridian_output.sha256 over the concatenated f32 outputs
(esp32+intel canonical frames, 64-dim geometry vector, rapid-adapt LoRA
weights); pytest (tests/test_meridian.py) runs the same fixture through
the binding and asserts the identical SHA-256. Both pass.
Verified:
cargo test --features meridian --test meridian_parity -> 2/2 pass
maturin develop --features meridian + pytest tests/test_meridian.py
-> 13/13 pass
default cargo build clean, 0 train/signal/sensing-server refs in the
default dep graph (gate keeps the base wheel lean).
WHEEL-SIZE FINDING (ADR-185 section 9 / section 1.2): the libtorch risk
the ADR feared is AVOIDED -- wifi-densepose-train's `tch` dep is properly
optional (feature tch-backend, OFF), so no libtorch links. BUT train
still carries NON-optional deps: tokio (rt subset), the five ruvector-*
crates, and wifi-densepose-nn (which itself pulls `ort` / ONNX Runtime +
reqwest/hyper). So a [meridian] wheel exceeds the ADR-117 section 5.4
<=5 MB budget (though lighter than AETHER's axum/tokio server tree). The
clean fix is the same leaf-crate hoist: move the pure inference modules
(geometry, rapid_adapt, eval, hardware_norm) into a tch/tokio/ort-free
leaf crate. A required pre-release follow-up, not a functional blocker;
P2 binds real code and proves parity today.
The training pipeline in `training_api.rs` was written and committed but never
declared as a module (no `mod training_api;`), so it was dead, uncompiled code.
The live `/api/v1/train/start` was a stub that flipped a status string and
logged one line without starting any job, and `/ws/train/progress` 404'd
(issue #1233). This wires the real trainer into the live server.
- main.rs: declare `mod training_api;` (+ `mod path_safety;` for its load
guard), replace the `training_status`/`training_config` stub fields on
AppStateInner with `training_state: training_api::TrainingState` +
`training_progress_tx`, delete the three stub handlers, and
`.merge(training_api::routes())` before `.with_state` so `/api/v1/train/*`
(bearer-gated) and `/ws/train/progress` resolve against the shared state.
- training_api.rs: decouple the training core (`run_training_job`) from the
~60-field AppStateInner via a shared status handle (`Arc<Mutex<TrainingStatus>>`)
+ cooperative cancel flag (`Arc<AtomicBool>`) + an owned frame_history
snapshot, making it unit-testable. Self-contained input schema (local
`RecordedFrame`/`RECORDINGS_DIR`) instead of coupling to the still-orphaned
`recording.rs`. Consolidate the single-job guard + spawn into
`spawn_training_job`.
- Tests: real end-to-end test (synthetic CSI dataset -> real progress frames
stream + an actual `.rvf` artifact is written), plus path-traversal-rejection
and cancellation tests.
Verified: cargo test -p wifi-densepose-sensing-server -p wifi-densepose-train
--no-default-features -> 0 failed.
P5 (feature-flagged disabled-build honesty) deferred as follow-up; the
default/enabled build no longer silently no-ops.
Bind the ADR-024 contrastive CSI embedding surface into the wheel behind
a gated [aether] extra / Cargo `aether` feature, per ADR-185 section 3.2.
Surface (bound against the REAL code at HEAD, not the ADR wishlist):
- AetherConfig -> EmbeddingConfig {d_model,d_proj,temperature,normalize}
- CsiAugmenter -> augment_pair(window, seed)
- EmbeddingExtractor -> embed(csi): 128-dim L2-normed, GIL-released
- info_nce_loss, cosine_similarity (module fns)
ADR-185 section 3.2 also names aether_loss/VICReg components,
alignment_metric, uniformity_metric, forward_dual, and vicreg_* config
fields. None exist in embedding.rs at HEAD, so they are intentionally NOT
bound (a Rust-side gap, not a binding gap) rather than fabricated.
Parity (section 4.1, release-blocking): committed fixture
aether_input.json -> native Rust reference (tests/aether_parity.rs, calls
sensing-server EmbeddingExtractor directly) locks
tests/golden/aether_embedding.sha256; pytest (tests/test_aether.py) runs
the same fixture through the binding and asserts the identical SHA-256 of
the LE f32 bytes. Both pass.
Verified: cargo test --features aether --test aether_parity -> 2/2 pass;
maturin develop --features aether + pytest tests/test_aether.py -> 9/9
pass; default cargo build clean with 0 sensing-server refs in the dep
graph (gate keeps the base wheel lean).
HONEST WHEEL-SIZE FINDING (ADR-185 section 9 / Open Q 11.1, unresolved):
wifi-densepose-sensing-server is an Axum/tokio crate whose tokio/axum/
worldgraph/ruvector deps are NON-optional (only mqtt/matter are
features), so default-features=false does NOT drop them. An [aether]
wheel therefore links the full server tree and BREAKS the ADR-117 section
5.4 <=5 MB budget. The fix is the ADR-sanctioned hoist of embedding.rs
(+ its graph_transformer/sona siblings) into a leaf crate so the wheel
links pure compute only -- a change INSIDE wifi-densepose-sensing-server,
owned by another agent this session, so deferred as a required
pre-release follow-up. P1 binds real code and proves parity today; the
hoist is a wheel-size optimization, not a functional blocker.
Note: building required initializing the worldgraph/rufield/ruvector
submodules (absent in the fresh worktree).
Verified all 6 acceptance criteria against current file contents and checked them.
Set Status to Accepted; marked P1/P2/P3 DONE (1fb5397dd); left P4 as ACCEPTED-FUTURE (#645).
To make criterion 4 genuinely true (live ESP32 17-keypoint feature nowhere advertised
without the first-cut/below-target/runtime-stub caveat), added the caveat to the three
remaining live-pose advertisements in README.md: the recommended-hardware capability
table, the hero image caption, and the Live-ESP32-pipeline note.
Refs #509, #1125
Drop all four PYPI_API_TOKEN password inputs from the publish-v2 and
publish-tombstone jobs, grant id-token: write, and bind both jobs to the
pypi GitHub environment so pypa/gh-action-pypi-publish uses secretless
Trusted Publishing. Rewrite the header + add step comments documenting
the blocking manual pypi.org trusted-publisher registration (ADR-184
§3.1) and the token fallback (§3.2). Build matrix untouched.
Refs ADR-184 P1.
Follow-up to the calibration deadlock fix. With the status gate unstuck,
maybe_feed_calibration reached feed_calibration, but a real ESP32 HT40 node
streams 128-wide amplitude frames while the single-link FieldModel is the
canonical 56-tone grid. LinkStats::update returned DimensionMismatch,
feed_calibration bubbled it, and maybe_feed_calibration swallowed it at debug
level — so frame_count stayed pinned at 0 on live hardware (presence/vitals,
which read the global history, were unaffected).
Resample each frame onto the model's canonical 56-tone grid via
HardwareNormalizer::resample_to_canonical before feeding — the same length-only
canonicalization the multistatic fusion path uses (#1170).
Pinned by maybe_feed_calibration_resamples_wide_frames_and_accumulates
(128-wide -> Collecting + count 1). sensing-server bin: 173 passed, 0 failed.
Co-Authored-By: claude-flow <ruv@ruv.net>
POST /api/v1/calibration/start created the FieldModel in Uncalibrated, but
field_bridge::maybe_feed_calibration only fed frames while already Collecting
— and the only thing that sets Collecting is feed_calibration on its first
fed frame. The two gates deadlocked: no first frame was ever fed, so
calibration_frame_count stayed 0 and status never left Uncalibrated. Observed
live on a streaming ESP32 node as {"status":"Uncalibrated","frame_count":0}
that never advanced.
- field_bridge::maybe_feed_calibration: feed while Uncalibrated | Collecting
so the first frame flips the model to Collecting and the count advances.
- calibration_stop: return structured {success:false, frame_count,
frames_needed} instead of an opaque 500 when finalized with too few frames.
- FieldModel::min_calibration_frames() accessor for the guard above.
- Regression test: maybe_feed_calibration_advances_uncalibrated_to_collecting.
Presence/motion/vitals were unaffected (separate auto rolling baseline).
Co-Authored-By: claude-flow <ruv@ruv.net>
Lower the STA auth threshold from WPA2_PSK to WPA_PSK so routers running
WPA/WPA2-mixed compatibility mode aren't rejected with
WIFI_REASON_NO_AP_FOUND_IN_AUTHMODE_THRESHOLD (reason=211), and log the
disconnect reason/rssi instead of retrying blind. Also document the thermal
risk of running this firmware's continuous-radio, tier-2 DSP pipeline on
coin-sized clone boards (ESP32-S3-Zero, SuperMini) with minimal PCB copper
and budget regulators, after a field report of three such boards failing to
power on following a normal session.
Co-Authored-By: claude-flow <ruv@ruv.net>
Dragan Spiridonov