Nice work! It takes decent dedication and effort to maintain an underappreciated low-level mathematics library.

Unfortunately, Google seemed to have developed its own multi-dimensional array library as well, featuring virtual views, handling of large arrays that cannot fit into main memory, compression and more: https://github.com/google/tensorstore https://ai.googleblog.com/2022/09/tensorstore-for-high-perfo...

What are `librapid`'s strengths and unique features compared to Google's `Tensor Store`? Being lightweight might be one.