url-kaist/multi-mapcher ? reverse-engineered prompt

Build me a ROS based LiDAR SLAM project that can take recorded driving data and make clean maps across multiple sessions, even when the LiDAR sensors are different.

I want it to support both a single session run and a multi session run, with the same core logic working in ROS 1 Noetic and ROS 2 Humble. It should read a dataset YAML file, run from a launch file, and produce the mapping and pose graph outputs in a way that is easy to inspect and reuse for the next session. Please make it practical for real autonomous vehicle data, with robust registration so it does not depend on loop closure detection. If useful, include a Docker setup and clear README instructions for getting started with a sample dataset and the expected folder layout. Look up current ROS and dependency docs online if you need to.