Documentation Index
Fetch the complete documentation index at: https://docs.cyberwave.com/llms.txt
Use this file to discover all available pages before exploring further.
stub — This page will be curated before publishing. Content below reflects the current implementation.
Wire frame layout
Every sample on the Cyberwave data bus (Zenoh or filesystem) uses one framing:| Field | Type | Size |
|---|---|---|
header_len | u32, LE | 4 bytes |
ts | f64, LE | 8 bytes |
seq | i64, LE | 8 bytes |
| header JSON | UTF-8 | variable |
| payload | bytes | variable |
header_len— 4-byte LE uint32 coveringts+seq+ JSON.ts— 8-byte LE float64, acquisition timestamp (Unix epoch seconds). Binary-packed for O(1) read.seq— 8-byte LE int64, per-channel sequence number for drop detection.- Header JSON — compact UTF-8 JSON with at least
content_typeand channel-specific metadata. - Payload — raw bytes (numpy buffer) or UTF-8 JSON.
HeaderTemplate class pre-encodes the JSON once and splices only ts/seq per sample (<500 ns overhead).
Key expression pattern
| Segment | Description |
|---|---|
cw | Prefix (configurable via BackendConfig.key_prefix). |
twin_uuid | UUID-4 of the digital twin. |
data | Fixed literal. |
channel | Channel name (frames, joint_states, etc.). |
sensor_name | Optional qualifier (default, wrist, left_arm). |
Usage
Encoding / decoding with HeaderTemplate (hot path)
One-shot encode / decode
Building and parsing key expressions
Well-known channels
| Channel | Pattern | Encoding |
|---|---|---|
frames | Stream | binary |
depth | Stream | binary |
audio | Stream | binary |
pointcloud | Stream | binary |
imu | Stream | JSON |
force_torque | Stream | JSON |
joint_states | Latest value | JSON |
position | Latest value | JSON |
attitude | Latest value | JSON |
gps | Latest value | JSON |
end_effector_pose | Latest value | JSON |
gripper_state | Latest value | JSON |
map | Latest value | binary |
battery | Latest value | JSON |
temperature | Latest value | JSON |
telemetry | Latest value | JSON |