A suite of libraries for data science & high performance computation in Swift
.package(url: "https://github.com/saeta/penguin.git", from: "sr-12940")
Explore the ideas of data frames, accelerated compute, tables, data structures, and
distributed data in Swift.
Below is the aspirational roadmap (at an extremely high level) for the high-level goals:
- Parse from CSV.
- Finish up base API & document it. (Incomplete list of needs: appending rows, support in-place mutations everywhere, more powerful sorting, remove unnecessary APIs)
- Refactor internals around a parallelism model from
- Optimize backing store for
PTypedColumn, including adding support for larger-than-RAM & file formats like Arrow.
- Add querying / group-by support (including support for a
PTableGroup --- alternate names wanted).
- Parallelize the implementation of the operators.
- Investigate hardware acceleration & JIT code-gen.
- Distributed orchestration of computation.
Extensions (help wanted):
- Connect to databases.
- Parse additional file formats.
This is not an officially supported Google product.