Point-LIO native module + Virtual Livox

[jeff-hykin]

Adds virtual mid360, and pointlio as alternative to fastlio

Usage / Testing

Pcap to DB

# snippet that normally diverges
PCAP_PATH="$(python -c "from dimos.utils.data import get_data; print(get_data('ruwik2_part3/ruwik2_part3.pcap'))")"

# gen .db from pcap
python -m dimos.hardware.sensors.lidar.pointlio.tools.pcap_to_db --pcap "$PCAP_PATH"
# add to existing .db
python -m dimos.hardware.sensors.lidar.pointlio.tools.pcap_to_db --db mem2.db  --pcap "$PCAP_PATH"

Record a Mid-360 pcap

# Allow using tcpdump without sudo (only need to do once)
sudo setcap cap_net_raw,cap_net_admin=eip "$(which tcpdump)"
# set these (env var or python module config)
export DIMOS_MID360_PCAP_IFACE=enp2s0
export DIMOS_MID360_LIDAR_IP
export DIMOS_MID360_HOST_IP
export DIMOS_MID360_NETNS=lidar

from dimos.core.coordination.blueprints import autoconnect
from dimos.core.coordination.module_coordinator import ModuleCoordinator
from dimos.hardware.sensors.lidar.livox.virtual_mid360.recorder import Mid360PcapRecorder

record = autoconnect(
    Mid360PcapRecorder.blueprint(
        pcap_path="recordings/run1.pcap",
        # lidar_ip="192.168.1.155",
        # iface="enp2s0",
    ),
)
ModuleCoordinator.build(record).loop()

Replay a pcap to a module

from dimos.core.coordination.blueprints import autoconnect
from dimos.core.coordination.module_coordinator import ModuleCoordinator
from dimos.hardware.sensors.lidar.livox.virtual_mid360.module import VirtualMid360
from dimos.hardware.sensors.lidar.pointlio.module import PointLio
from dimos.visualization.vis_module import vis_module

replay = autoconnect(
    VirtualMid360.blueprint(
        pcap="recordings/run1.pcap",
        # lidar_ip="192.168.1.155",
    ),
    PointLio.blueprint(
        # lidar_ip="192.168.1.155"
    ),
    vis_module("rerun"),
).global_config(n_workers=3)
ModuleCoordinator.build(replay).loop()

[codecov]

Codecov Report

❌ Patch coverage is 70.45455% with 39 lines in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
dimos/hardware/sensors/lidar/pointlio/module.py	61.76%	39 Missing ⚠️

@@            Coverage Diff             @@
##             main    #2486      +/-   ##
==========================================
- Coverage   70.09%   69.97%   -0.13%     
==========================================
  Files         838      842       +4     
  Lines       74337    74469     +132     
  Branches     6667     6672       +5     
==========================================
- Hits        52110    52109       -1     
- Misses      20552    20658     +106     
- Partials     1675     1702      +27     

Flag	Coverage Δ
OS-ubuntu-24.04-arm	63.84% <70.45%> (+0.01%)	⬆️
OS-ubuntu-latest	64.68% <70.45%> (+0.01%)	⬆️
Py-3.10	64.67% <70.45%> (+0.01%)	⬆️
Py-3.11	64.67% <70.45%> (+<0.01%)	⬆️
Py-3.12	64.67% <70.45%> (+0.01%)	⬆️
Py-3.13	64.67% <70.45%> (+0.01%)	⬆️
Py-3.14	64.68% <70.45%> (+0.01%)	⬆️
Py-3.14t	64.67% <70.45%> (+<0.01%)	⬆️
SelfHosted-Large	30.29% <ø> (-0.06%)	⬇️
SelfHosted-Linux	38.21% <ø> (ø)
SelfHosted-macOS	36.98% <ø> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...e/sensors/lidar/livox/virtual_mid360/blueprints.py	100.00% <100.00%> (ø)
...dware/sensors/lidar/livox/virtual_mid360/module.py	100.00% <100.00%> (ø)
...ware/sensors/lidar/pointlio/pointlio_blueprints.py	100.00% <100.00%> (ø)
dimos/robot/all_blueprints.py	100.00% <ø> (ø)
dimos/hardware/sensors/lidar/pointlio/module.py	61.76% <61.76%> (ø)

... and 4 files with indirect coverage changes

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[greptile-apps]

Greptile Summary

This PR adds two new modules: VirtualMid360, a Rust binary that replays a recorded Livox Mid-360 pcap file over a virtual network interface with a synthesized Livox SDK2 control handshake, and PointLio, a Python NativeModule wrapping a C++ binary that runs Point-LIO IESKF SLAM against Livox SDK2 input. A pcap_to_db.py tool orchestrates both to write odometry and point-cloud output into a SQLite database. All previously flagged issues (velocity silently discarded, shutdown race, cp.y comment-bug, falsy-zero ts fallbacks, wrong TF frame IDs, process-scoped cleanup, DB ownership on setup failure) appear to have been addressed in this revision.

• VirtualMid360 (Rust): synthesizes Livox SDK2 discovery/control handshake and replays timestamped point/IMU/status UDP packets with inter-packet timing, suitable for use with the live PointLio SDK path.
• PointLio (Python + C++): full IESKF SLAM pipeline feeding from Livox SDK2 callbacks, with per-point deskew and velocity forwarding; startup/shutdown ordering correctly guards against SDK-thread races.
• pcap_to_db.py: orchestrates netns/alias isolation, inner-process coordinator lifecycle, startup and stagnation timeouts, and db ownership restoration in a single tool.

Confidence Score: 5/5

Safe to merge; all previously identified blocking issues have been addressed in this revision, and only two minor logic nits remain in the Rust virtual sensor binary.

Every issue flagged in prior review rounds — velocity forwarding, cp.y coordinate assignment, TF frame IDs, ts falsy-zero fallbacks, shutdown ordering, cleanup scope, and DB ownership — is correctly resolved in the current diff. The two remaining findings (a double-sleep when the handshake times out with a non-zero delay, and unicast vs. broadcast in the discovery ACK) are both in the Rust virtual-sensor replay path, affect only non-default configurations or unusual network topologies, and do not corrupt data or block the primary use case.

dimos/hardware/sensors/lidar/livox/virtual_mid360/src/main.rs — the two minor fixes in spawn_stream and spawn_discovery.

Important Files Changed

Filename	Overview
dimos/hardware/sensors/lidar/livox/virtual_mid360/src/main.rs	Rust virtual lidar replayer with synthesized Livox SDK2 handshake; two minor logic issues in spawn_stream (double-sleep on timeout) and spawn_discovery (ACK broadcast vs unicast).
dimos/hardware/sensors/lidar/pointlio/cpp/main.cpp	C++ IESKF SLAM wrapper: velocity forwarding, point coordinate assignments, and shutdown ordering all correctly implemented vs. previous comments.
dimos/hardware/sensors/lidar/pointlio/module.py	Python NativeModule wrapper; TF uses config frame IDs, ts uses msg.ts directly (no falsy fallback), network validation is thorough.
dimos/hardware/sensors/lidar/pointlio/tools/pcap_to_db.py	Orchestration tool with startup timeout, stagnant-detection, DB ownership restoration inside try/finally, and namespace-scoped cleanup; all previously flagged issues addressed.
dimos/hardware/sensors/lidar/livox/virtual_mid360/recorder.py	tcpdump-based Mid-360 recorder with AppArmor-aware signal escalation; lidar_ip field still requires explicit passing (no env-var fallback unlike VirtualMid360Config).
dimos/hardware/sensors/lidar/livox/virtual_mid360/module.py	Thin NativeModule wrapper around the Rust binary with correct env-var defaults for all config fields.
dimos/hardware/sensors/lidar/pointlio/config/default.yaml	Point-LIO YAML config tuned for Livox Mid-360 with correct lidar_type, extrinsics, and acc_norm for the g-unit IMU path.

Sequence Diagram

%%{init: {'theme': 'neutral'}}%%
sequenceDiagram
    participant User
    participant Outer as pcap_to_db outer
    participant Inner as inner Python coord
    participant PLIO as PointLio binary
    participant VM as VirtualMid360 binary
    participant DB as SQLite DB

    User->>Outer: python -m pcap_to_db --pcap file.pcap
    Outer->>Outer: setup netns/alias iface via sudo
    Outer->>Inner: spawn in drv netns
    Inner->>PLIO: spawn via NativeModule
    PLIO->>PLIO: LivoxLidarSdkStart, wait for sensor
    Outer->>Outer: sleep warmup_sec
    Outer->>VM: spawn, write JSON config to stdin
    VM->>VM: listen on discovery port 56000 and control port 56100
    PLIO->>VM: discovery broadcast
    VM->>PLIO: discovery ACK cmd 0x0000
    PLIO->>VM: work-mode cmd 0x0100
    VM->>PLIO: work-mode ACK arm stream
    VM->>PLIO: replay point and IMU UDP at configured rate
    PLIO->>PLIO: IESKF SLAM feed_lidar feed_imu process
    PLIO->>Inner: LCM odometry and point cloud
    Inner->>DB: append ts-aligned rows
    Inner->>Inner: poll stagnant and startup timeout
    Inner->>Outer: exit when pcap drained or timeout
    Outer->>Outer: teardown netns alias restore db ownership
    Outer->>User: print stats and exit

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
sequenceDiagram
    participant User
    participant Outer as pcap_to_db outer
    participant Inner as inner Python coord
    participant PLIO as PointLio binary
    participant VM as VirtualMid360 binary
    participant DB as SQLite DB

    User->>Outer: python -m pcap_to_db --pcap file.pcap
    Outer->>Outer: setup netns/alias iface via sudo
    Outer->>Inner: spawn in drv netns
    Inner->>PLIO: spawn via NativeModule
    PLIO->>PLIO: LivoxLidarSdkStart, wait for sensor
    Outer->>Outer: sleep warmup_sec
    Outer->>VM: spawn, write JSON config to stdin
    VM->>VM: listen on discovery port 56000 and control port 56100
    PLIO->>VM: discovery broadcast
    VM->>PLIO: discovery ACK cmd 0x0000
    PLIO->>VM: work-mode cmd 0x0100
    VM->>PLIO: work-mode ACK arm stream
    VM->>PLIO: replay point and IMU UDP at configured rate
    PLIO->>PLIO: IESKF SLAM feed_lidar feed_imu process
    PLIO->>Inner: LCM odometry and point cloud
    Inner->>DB: append ts-aligned rows
    Inner->>Inner: poll stagnant and startup timeout
    Inner->>Outer: exit when pcap drained or timeout
    Outer->>Outer: teardown netns alias restore db ownership
    Outer->>User: print stats and exit

Loading

_{Reviews (30): Last reviewed commit: "Merge remote-tracking branch 'origin/mai..." | Re-trigger Greptile}

[greptile-apps]

lidar_ip required but env-var fallback missing

Mid360PcapRecorderConfig.lidar_ip is a bare str with no default, so Pydantic requires it at construction time. The PR description shows it commented out (# lidar_ip="192.168.1.155") and the setup section advertises export DIMOS_MID360_LIDAR_IP, but unlike VirtualMid360Config (which reads DIMOS_MID360_LIDAR_IP via a Field(default_factory=...)) the recorder config has no env-var hook. Any user who follows the documented example without explicitly passing lidar_ip will get a ValidationError: Field required before the module even starts.

[aclauer]

Why eprintln! instead of just using tracing?