Bagzel is an efficient and reproducible data extraction pipeline for ROS 1 bag files, built on Bazel. It supports generating structured datasets—including image sequences, GPS logs, metadata, and map visualizations—from recorded robotic data.
Make sure the following are installed on your system:
- Bazel (installation guide)
- git
- Git Large File Storage (LFS) (Git LFS installation guide)
To install Git LFS:
git lfs installTo obtain and prepare the source code:
git clone https://github.com/UniBwTAS/bagzel.git
cd bagzel
git lfs pull
⚠️ Ensure you rungit lfs pullto fetch large files such as bag files tracked via Git LFS.
To process a minimal dataset and generate all related build artifacts, run:
bazel build //data:rosbags_sub1_allThe resulting artifacts will be available under:
bazel-bin/data
To explore available build targets, use:
bazel query //data:all-
docs/declare-rosbag-artifacts.md – Overview of the generated Bazel targets and data structure.
-
docs/cluster-deployment.md – Instructions for deploying the pipeline on a computing cluster.
-
docs/bazel-profiling.md – Performance profiling with Bazel.
-
docs/update-python-deps.md – Managing and updating Python dependencies using Bazel.
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.