v0.0.1

[0.0.1] - 2026-03-31

Added

• Runtime shape and dtype validation for NumPy, JAX, PyTorch, and CuPy arrays,
  powered by typing.Annotated and beartype validators.
• Symbolic dimension syntax including named dimensions, anonymous dimensions,
  Scalar, arithmetic dimension expressions, and constrained Value(...)
  checks.
• Tree[...] validation helpers, explicit @shapix.check support, and
  check_context() for shared manual bearability checks.
• Multi-checker typing coverage across pyright, mypy, and ty, with CI, tox, and
  pre-commit validation for runtime and typing behavior.
• A documentation website covering installation, API boundaries, supported
  backends, and practical usage patterns.

Notes

• The PyPI distribution for this release is shapix-rt, where rt means
  runtime; the import path remains shapix.
• The root shapix module intentionally stays lightweight and
  optional-dependency-safe; backend-specific aliases and factories live in
  shapix.numpy, shapix.jax, shapix.torch, and shapix.cupy.
• CuPy support remains optional at install time and requires a compatible CuPy
  environment when used at runtime.