Files
ruvnet--RuView/python/Cargo.toml
T
ruv 189ac9dfb0 feat(adr-185): P2 MERIDIAN bindings (wifi_densepose.meridian) + parity harness
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.
2026-07-21 16:38:17 -07:00

105 lines
5.5 KiB
TOML
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
[package]
name = "wifi-densepose-py"
version = "2.0.0-alpha.1"
# The `python/` crate is intentionally OUTSIDE the `v2/` Cargo
# workspace (ADR-117 §5.2) so maturin's `python-source` + `module-name`
# config stays self-contained and `cargo test --workspace` in v2/
# doesn't have to compile pyo3. Hence no `*.workspace = true`
# inheritance here — every field is local.
edition = "2021"
license = "MIT"
authors = ["rUv <ruv@ruv.net>", "WiFi-DensePose Contributors"]
description = "PyO3 bindings for the WiFi-DensePose Rust core — ships as the `wifi-densepose` PyPI wheel (ADR-117)"
repository = "https://github.com/ruvnet/RuView"
# ADR-117 §5.2: the Python wheel's compiled module name is
# `wifi_densepose._native` (the leading underscore marks it as an internal
# implementation detail re-exported by the pure-Python facade in
# `wifi_densepose/__init__.py`). Keeping the name distinct from the crate
# avoids the maturin gotcha where `wifi_densepose-py` would collide with
# the user-facing `wifi_densepose` package on import.
[lib]
name = "wifi_densepose_native"
crate-type = ["cdylib", "rlib"]
path = "src/lib.rs"
# ADR-185 §3.1 — optional pip extras map to Cargo features so the
# default wheel links none of the SOTA subsystems. P1 wires `aether`.
[features]
default = []
# ADR-185 P1 — AETHER contrastive CSI embeddings. Pulls the backing
# `wifi-densepose-sensing-server` crate (see the honest wheel-size note
# on that dep below).
aether = ["dep:wifi-densepose-sensing-server"]
# ADR-185 P2 — MERIDIAN domain generalization. Binds the tch-free
# inference/adaptation path only (see the wheel-size note on the deps
# below). `wifi-densepose-train` is depended on WITHOUT `tch-backend`,
# so no libtorch is linked.
meridian = ["dep:wifi-densepose-train", "dep:wifi-densepose-signal"]
[dependencies]
# PyO3 with abi3-py310 — one compiled binary covers Python 3.10, 3.11,
# 3.12, 3.13, and any future 3.x that keeps the stable ABI (ADR-117 §5.4).
# Without abi3 we'd need a separate wheel per Python minor version × OS
# × arch, blowing up the cibuildwheel matrix.
pyo3 = { version = "0.22", features = ["extension-module", "abi3-py310"] }
# Re-export the Rust core types through PyO3 #[pyclass] wrappers in P2.
# Default-features-off keeps the wheel size below the 5 MB ADR-117 §5.4
# budget by avoiding optional BLAS/openssl chains.
wifi-densepose-core = { version = "0.3.0", path = "../v2/crates/wifi-densepose-core" }
# P3 — vitals extraction (HR/BR via the 4-stage pipeline). Pure-sync;
# no tokio (Q5 audited 2026-05-24); safe to wrap in py.allow_threads.
wifi-densepose-vitals = { version = "0.3.0", path = "../v2/crates/wifi-densepose-vitals" }
# ADR-118 BFLD core — PrivacyClass enum + identity_risk scoring +
# privacy gate. Exposed to Python via bindings/privacy_gate.rs so the
# c6-presence-watcher.py runtime (currently using a Python port of the
# same semantics) can switch to the canonical Rust implementation when
# the wheel ships. ADR-125 §2.1.d invariant enforcement lives here.
wifi-densepose-bfld = { version = "0.3.0", path = "../v2/crates/wifi-densepose-bfld" }
# numpy bridge — needed for P3.5 BfldFrame (Complex64 ndarray) and for
# the future P3 CsiFrame numpy round-trip.
numpy = "0.22"
# ADR-185 P1 — AETHER backing crate (contrastive `embedding` module,
# ADR-024). Optional + gated behind the `aether` feature.
#
# HONEST WHEEL-SIZE NOTE (ADR-185 §9 / Open Q §11.1): this crate is an
# Axum/tokio server crate. `default-features = false` does NOT drop
# tokio/axum here — they are non-optional deps (only `mqtt`/`matter`
# are features). So an `[aether]` wheel currently links the full
# sensing-server dependency tree, breaking the ADR-117 §5.4 ≤5 MB
# budget. The ADR-sanctioned fix is to hoist `embedding.rs` (+ its
# `graph_transformer`/`sona` siblings) into a leaf crate so the wheel
# links only pure compute — a Rust-side refactor inside
# `wifi-densepose-sensing-server`, owned by another agent this session.
# P1 binds against the real code and proves parity; the leaf-crate
# hoist is the required pre-release follow-up.
wifi-densepose-sensing-server = { version = "0.3.0", path = "../v2/crates/wifi-densepose-sensing-server", optional = true, default-features = false }
# ADR-185 P2 — MERIDIAN backing crates (optional, `meridian`-gated).
#
# HONEST WHEEL-SIZE NOTE (ADR-185 §9 / §1.2): unlike AETHER, the libtorch
# risk is AVOIDED here — `wifi-densepose-train`'s `tch` dep is properly
# optional (feature `tch-backend`, OFF by default), so no libtorch links.
# BUT `wifi-densepose-train` still carries NON-optional deps: `tokio` (rt
# subset), the five `ruvector-*` crates, `wifi-densepose-nn`, petgraph,
# memmap2, indicatif, ndarray-npy, csv, toml, clap. So a `[meridian]`
# wheel is heavier than the ≤5 MB ADR-117 §5.4 budget (though far 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-free leaf crate.
# `wifi-densepose-signal` is depended on `default-features = false` to
# drop the optional ndarray-linalg/BLAS chain (Windows-friendly).
wifi-densepose-train = { version = "0.3.0", path = "../v2/crates/wifi-densepose-train", optional = true, default-features = false }
wifi-densepose-signal = { version = "0.3.0", path = "../v2/crates/wifi-densepose-signal", optional = true, default-features = false }
[dev-dependencies]
# ADR-185 §4.1 parity harness — SHA-256 the native-Rust reference
# embedding and read the committed golden fixture.
sha2 = "0.10"
serde_json = "1"