Implemented Features
For now only a small subset of the features implemented in scanpy are available in SingleRust. We're working on expanding the feature set as quickly as possible. While improving on performance and memory usage.
These are the features currently implemented in the SingleRust library:
| Feature | Description | In-Memory Status | Backed Status |
|---|---|---|---|
| Statistics Calculations | Basic statistics calculations of the matrix. Sum/Var/Non-Zero/Min-Max | ✅ Fully implemented (whole/chunked) | ✅ Fully implemented (whole/chunked) |
| Normalization | Basic normalization of the data via target-sum | ✅ Fully implemented, no multithreading yet | ❌ Not implemented, because of streaming issues |
| Log-Transformation | Logarithmic transformation of the data | ✅ Fully implemented, no multithreading yet | ❌ Not implemented, because of streaming issues |
| Filtering | Filtering of cells and genes based on various criteria | ✅ Fully implemented, no multithreading yet | ❌ Not implemented yet, but in the works |
| Highly Variable Genes | Calculation of highly variable genes | ✅ Fully implemented, no multithreading yet | ❌ Not implemented yet, but in the works |
| PCA | Principal Component Analysis of the data | ✅ Fully implemented, partial multithreading | ❌ Not implemented yet, but in the works |
More features will be added in the future. For most features we've already implemented the internals in single-algebra, but we're still working on the integration into SingleRust. Take a look at the Roadmap for further information on the state of the individual features.