All file operations are handled by a pool of managed processes created by spawning, which make them safe for fetching and processing remote data files asynchronously.
0.0.10
Key Updates: 1. Amazon S3 file system is now supported. A RamData object stored in Amazon S3 can be analyzed as if it is in the local file system.
2. Synchronization of simultaneous modification attempts through file-locking - a class ZarrSpinLockServer was implemented to support file-locking.
0.0.7
features: - "combined" mode combined analysis of multiple RamData objects as components of a larger, combined RamData.
- Embedding to reference data: Deep-learning based embedding/classifier KNN-index based embedding/classifier
0.0.5
performance was optimized to match that of scanpy for 1M ~ 10M cells
0.0.1
this version was used for the comparison with different single-cell data analysis platforms (Scanpy, Seurat, and Monocle3) for the memory consumption benchmark of the platforms.