MicPy 0.4 Changelog
Performance Improvements
New decompression backend for MICRESS binary field data
- Introduced a redesigned backend enabling parallel data reading.
- Added persistent on-disk indexing for field data, allowing random-access indices to be reused across sessions.
- Significantly improved load times and access performance for large datasets, especially in repeated or interactive workflows.
Improved robustness and maintainability
- Index creation and management are now handled by a dedicated backend library1.
- Reduced internal code complexity, improving overall stability and long-term maintainability.
Visualization and Data Exchange
Native VTK support
- Added direct conversion of MICRESS field data (NumPy arrays) to VTK image data.
- Supports both CellData and PointData representations.
- Optional interpolation from cell-centered data to point-based values.
- Enables straightforward export to VTI files and direct use with VTK-based Python tools such as PyVista.
Enhanced support for interactive Python workflows
- Improved integration with Jupyter notebooks for 2D and 3D visualization and analysis.
- Facilitates seamless data exchange between MicPy and modern visualization libraries.
ParaView Integration
New ParaView plugin based on MicPy
- Allows direct loading of MICRESS binary field data into ParaView without intermediate conversion steps.
- Leverages the new decompression and indexing backend for smooth, responsive interaction with large datasets.
- Provides full access to ParaView’s filtering and data processing pipelines, with strong support for 3D visualization workflows.
-
Maximilian Knespel and Holger Brunst, “Rapidgzip: Parallel Decompression and Seeking in Gzip Files Using Cache Prefetching,” in Proceedings of the 32nd International Symposium on High-Performance Parallel and Distributed Computing (HPDC ’23) (New York: ACM, 2023), 295–307, https://doi.org/10.1145/3588195.3592992. ↩↩
