Crate wifi_densepose_worldmodel 

wifi-densepose-worldmodel — OccWorld thin-client bridge (ADR-147).

Bridges wifi_densepose_worldgraph PersonTrack history to the OccWorld Python inference subprocess and returns TrajectoryPriors that can be injected into the Kalman pose tracker.

Quick start

use wifi_densepose_worldmodel::{
    OccWorldBridge, OccupancyWorldModelRequest, OccupancyGrid3D,
    SceneBoundsJson, worldgraph_to_occupancy,
};
use wifi_densepose_worldmodel::occupancy::{PersonPosition, SceneBounds};

let bridge = OccWorldBridge::new("/tmp/occworld.sock");

let bounds = SceneBounds { min_e: -10.0, min_n: -10.0, max_e: 10.0, max_n: 10.0 };
let persons = vec![
    PersonPosition { track_id: 1, east_m: 2.0, north_m: 3.0, up_m: 1.0 },
];
let frame = worldgraph_to_occupancy(&persons, &bounds, 0.1);

let request = OccupancyWorldModelRequest {
    past_frames: vec![frame],
    voxel_resolution_m: 0.1,
    scene_bounds: SceneBoundsJson {
        min_e: bounds.min_e, min_n: bounds.min_n,
        max_e: bounds.max_e, max_n: bounds.max_n,
    },
    prediction_steps: 15,
};

let response = bridge.predict(request).await?;
println!("confidence={:.2}", response.confidence);
for prior in &response.trajectory_priors {
    println!("track {} has {} waypoints", prior.track_id, prior.waypoints.len());
}

Re-exports

pub use bridge::default_socket_path;

pub use bridge::OccWorldBridge;

pub use error::WorldModelError;

pub use occupancy::worldgraph_to_occupancy;

Modules

bridge

Async Unix-socket client that sends an OccupancyWorldModelRequest to the OccWorld Python inference server and receives an OccupancyWorldModelResponse (ADR-147).

error

Error types for the OccWorld world-model bridge (ADR-147).

occupancy

Converts WorldGraph PersonTrack ENU positions into an OccupancyGrid3D tensor suitable for submission to the OccWorld inference server (ADR-147).

Structs

OccupancyGrid3D

A 3-D occupancy grid whose voxel values are class indices (u8).

OccupancyWorldModelRequest

JSON request sent from the Rust bridge to the OccWorld Python server.

OccupancyWorldModelResponse

JSON response returned by the OccWorld Python server.

SceneBoundsJson

Axis-aligned scene footprint sent to the OccWorld server in the IPC request. Mirrors occupancy::SceneBounds but derives Serialize / Deserialize for direct inclusion in the JSON payload.

TrajectoryPrior

Predicted future trajectory for one tracked person.

TrajectoryWaypoint

A single point on a predicted trajectory, with a relative timestamp.

Functions

persons_from_worldgraph

Extracts all PersonPositions from a WorldGraph by serialising the graph to its canonical JSON form (via WorldGraph::to_json) and scanning the nodes array for PersonTrack entries.