SciArx is a scientific computing library for ArxLang.
This repository contains the initial SciArx implementation for the current Arx toolchain. The implementation follows the roadmap in phased steps, beginning with focused math, stats, and linalg modules.
Early bootstrap stage. sciarx.math includes numeric scalar helpers,
sciarx.stats includes scalar statistical kernels, and sciarx.linalg includes
initial dense scalar kernels plus runtime-shaped tensor entry points. Arx
collection types such as tensor, dataframe, and series are covered by
compatibility tests, but SciArx does not expose placeholder wrapper modules for
them yet.
Full tensor-returning APIs such as zeros, sum, matmul, and solve are
specified in docs/phase-0-2-proposal.md, but remain gated on Arx source
support for unsized tensor values outside parameter positions.
sciarx.mathsciarx.statssciarx.linalgsciarx.signal
See ROADMAP.md for the full plan.
The test suite uses the Arx test runner:
arx test testsCurrent tests cover Phase 0 collection smoke paths, Phase 1 math/stats helpers,
Phase 2 linalg kernels, and the sciarx.stats smoke helpers.
This repository includes a minimal .arxproject.toml manifest for arxpm.
Current intent:
- package/library-style repository
- sources live under
src/ - build artifacts go to
build/
Useful future commands may include:
arxpm doctorarxpm installarxpm buildarxpm pack
If later SciArx needs an executable smoke target, add a dedicated src/main.x
and update [build].entry accordingly.