Rerun 0.30 Enables Plotting of Arbitrary Scalars and Custom GPU Visualizers Without Code Forking
Key Takeaways
- ▸Rerun 0.30 now supports plotting any scalar value, including data from MCAP files and custom components, without requiring Rerun-specific semantics
- ▸Multiple scalars can be logged to the same entity, significantly reducing .rrd file sizes and improving data organization
- ▸Supports wide range of data types including Float32/64, Int8-64, UInt8-64, Boolean, and nested Arrow structs
Summary
Rerun has released version 0.30.0 of its multimodal and temporal data visualization toolbox, introducing significant improvements to data plotting capabilities. The headline feature allows users to plot any scalar value, even without Rerun-specific semantics, enabling visualization of data from sources like MCAP files. The update supports a wide range of data types including Float32/64, various integer formats (Int8-64, UInt8-64), Boolean values, and nested Arrow structs.
The release addresses a major workflow bottleneck by allowing multiple scalars to be logged to the same entity, which can dramatically reduce .rrd file sizes. Users can now visualize arbitrary scalar components logged through AnyValues or DynamicArchetype, with time series views accessible through context menus, blueprint panels, or SDK specifications. The selection panel features a completely redesigned visualizer section with a dropdown menu for component selection and quick navigation.
Rerun, which positions itself as an "easy-to-use database and visualization toolbox for multimodal and temporal data," continues to be available across multiple platforms including Python (via pip), Rust (via cargo), and C++ (via FetchContent). The update maintains backward compatibility while expanding the tool's flexibility for handling diverse data sources and custom visualization requirements.
- Completely redesigned visualizer section in selection panel with improved component selection and navigation features
- Custom scalar time series views can be spawned via context menus, blueprint panels, or SDK specifications
Editorial Opinion
This release represents a significant maturation of Rerun's visualization capabilities, addressing a common pain point in data analysis workflows where custom or third-party data formats often require extensive preprocessing. By supporting arbitrary scalar plotting and reducing file sizes through multi-scalar entities, Rerun is positioning itself as a more flexible alternative to traditional visualization tools. The focus on MCAP file support is particularly strategic, given the format's growing adoption in robotics and autonomous systems development.
