Rosbag Resurrector v0.8 brings pandas-like ML export tools to ROS 2 bags

Rosbag Resurrector, released by robotics developer vikramnagashoka, is a new analysis workbench for ROS 2 MCAP bag files that sidesteps the usual 3D-viewer approach. Instead, it treats every bag as a DataFrame, giving users pandas-like access to topics, signals, and metadata. The tool builds a DuckDB index across many bags, enabling cross-bag queries and health/QC validation, plus semantic frame search to locate moments based on meaning rather than raw timestamps. A WebSocket bridge links it to PlotJuggler for visualization, while a one-command export pipeline outputs to LeRobot, RLDS, or Parquet. Crucially, it installs via pip with no ROS installation required, lowering the barrier for data scientists and ML engineers working with robotics logs.

The latest v0.8 release introduces a notebook-style web UI that unifies the workflow: users can open a bag, plot any signal, brush over a weird spike, and receive a grounded explanation of what happened at that moment. They can overlay two runs to compare behavior, then export the relevant slices directly into a training dataset—all without leaving the browser. The tool is actively developed and positioned as a complement to existing bag workflows rather than a replacement: it doesn't fight PlotJuggler for plotting duties, and it deliberately skips heavy 3D rendering. For robotics teams drowning in raw MCAP files, rosbag-resurrector offers a practical bridge between data exploration and machine learning, making bag data as easily queryable as a SQL table and as export-ready as a pandas DataFrame.