bagx 0.1.3

Post-processing analysis engine for ROS2 rosbag data

bagx

One command to check if your rosbag data is ready — for SLAM, Nav2, Autoware, or MoveIt.

pip install bagx
bagx eval your_bag.db3

bagx analyzes your rosbag and gives you:

• Exact IMU noise values to paste into your SLAM config
• Sync delay warnings so you know when to enable deskew
• "Use LiDAR-only" when your IMU is too noisy for LIO

The problem

You recorded a rosbag. You run SLAM. It diverges. Now you spend hours asking:

• "Is my IMU noise too high for LIO?" → bagx eval tells you
• "What noise value should I set in the SLAM config?" → bagx gives you the exact number
• "Are my LiDAR and IMU actually synchronized?" → bagx sync shows the delay
• "Where did the data quality drop?" → bagx anomaly pinpoints the exact timestamp
• "Which of my 10 bags is the best for benchmarking?" → bagx batch eval ranks them all
• "Is my odometry fast enough for Nav2?" → bagx checks rate and warns if too slow
• "Are all Autoware sensing topics alive?" → bagx verifies camera, LiDAR, GNSS rates

Example: catching a real problem

$ bagx eval ouster_os0-32.db3

Overall Score: 76.6/100

Recommendations:
  ✔ IMU accel noise 0.087 m/s² — good for LIO, set imu_acc_noise_density to 0.087
  ⚠ IMU gyro noise 0.022 rad/s — consider lowering IMU integration weight
  ⚠ IMU rate 50Hz is low — 200Hz+ recommended for tightly-coupled LIO
  ⚠ LiDAR↔IMU sync delay 23ms — enable per-point deskew in SLAM

Without bagx, you'd discover these issues after hours of failed SLAM runs.

Install

pip install bagx

Works without ROS2 — reads .db3 files directly via SQLite.

Commands

Command	One-liner
bagx eval bag.db3	Is this bag ready? Auto-detects SLAM / Nav2 / Autoware / MoveIt
bagx compare A.db3 B.db3	Which sensor config is better?
bagx sync bag.db3 /imu /lidar	Are my sensors synchronized?
bagx anomaly bag.db3	Where did sensor quality drop?
bagx scenario bag.db3	Find GNSS-lost and high-dynamics segments
bagx export bag.db3 --ai	Export to Parquet/JSON for ML
bagx batch eval *.db3 --csv	Rank an entire dataset
bagx ask bag.db3 "question"	Ask questions via LLM

Tested on public datasets

Dataset	IMU	Sync	Score	Insight
Newer College	84	100	92	Best data, ideal for SLAM benchmarking
Livox MID-360	97	70	84	Great IMU, but 25ms sync — needs deskew
NTU VIRAL	60	100	80	IMU too noisy for LIO, use LiDAR-only
Ouster OS0-32	80	74	77	50Hz IMU too slow, add external IMU

Auto-detects your framework

bagx recognizes topic patterns and gives framework-specific advice:

$ bagx eval nav2_robot.db3
Nav2 topics detected
  ✔ Odometry (/odom) at 50Hz — good for Nav2
  ⚠ LaserScan (/scan) at 5Hz — 10Hz+ recommended for costmap updates
  ⚠ cmd_vel at 8Hz — control loop may be too slow

$ bagx eval autoware_vehicle.db3
Autoware topics detected
  ✔ LiDAR (/sensing/lidar/points) at 10Hz
  ✔ Camera (/sensing/camera/image) at 30Hz
  ✔ GNSS (/sensing/gnss/fix)

$ bagx eval moveit_arm.db3
MoveIt topics detected
  ✔ JointState (/joint_states) at 500Hz — good for motion planning

Links

• Documentation
• PyPI
• Changelog