The current implementation of proximity search is generic, allowing it to work with any pseudo-metric regardless of the nature of the data. However, the most common use case for Mapper involves tabular data or numpy arrays.
Proximity search, as implemented in proximity-net, could be enhanced by developing a specialized version of vptree optimized for numpy arrays. Potential improvements could include using numba or a Rust backend for vptrees.
The current implementation of proximity search is generic, allowing it to work with any pseudo-metric regardless of the nature of the data. However, the most common use case for Mapper involves tabular data or numpy arrays.
Proximity search, as implemented in proximity-net, could be enhanced by developing a specialized version of vptree optimized for numpy arrays. Potential improvements could include using numba or a Rust backend for vptrees.