Bagel 2.0 Simplifies ROS Data Querying with LLMs
TL;DR. Bagel 2.0 introduces an open-source tool enabling plain English queries of robotics data, bypassing traditional scripting for analysis. - The system uses LLMs to interpret natural language requests and generate SQL queries for robotic 'rosbag' data. - A key feature is data reduction, allowing users to extract relevant snippets from large datasets based on natural language conditions. - Bagel 2.0 prioritizes data accuracy by writing queries to Arrow and running DuckDB SQL, showing the audit trail.
- Bagel 2.0 allows natural language queries on robotics data (rosbags).
- It uses LLMs (like Claude, Ollama) to generate SQL queries for data analysis.
- The tool includes a data reduction pipeline to extract relevant data segments.
- Bagel 2.0 focuses on data accuracy and auditability by displaying generated SQL.
- The system can identify specific events and conditions within robotics sensor data.
Sources
- Bagel 2.0: ask your rosbags questions in plain English, reduce them to the seconds that matter (open source) — discourse.openrobotics.org