Changelog

All notable changes to dimtensor.

[0.3.1] - 2026-01-08

Added

• Documentation site: Full mkdocs documentation with Material theme
• Display configuration: New config module for controlling output
• config.set_display(precision=N) - set display precision
• config.precision(N) - context manager for temporary precision
• config.options(...) - context manager for multiple options
• config.reset_display() - reset to defaults

[0.3.0] - 2026-01-08

Added

• Module-level array functions (NumPy-style API):
• concatenate(arrays, axis) - join arrays along existing axis
• stack(arrays, axis) - join arrays along new axis
• split(array, indices_or_sections, axis) - split array into sub-arrays
• Linear algebra functions:
• dot(a, b) - dot product with dimension multiplication
• matmul(a, b) - matrix multiplication with dimension multiplication
• norm(array, ord, axis, keepdims) - vector/matrix norm, preserves units
• Reshaping methods on DimArray:
• reshape(shape) - reshape preserving units
• transpose(axes) - permute dimensions preserving units
• flatten() - flatten to 1D preserving units
• Statistics method:
• var(axis, keepdims) - variance with squared units (m -> m^2)
• Searching methods:
• argmin(axis) - return indices of minimum values
• argmax(axis) - return indices of maximum values

Changed

• Array functions enforce same dimension for all input arrays
• Linear algebra functions properly multiply dimensions (L * L = L^2)

[0.2.0] - 2025-01-08

Added

• Unit simplification: Compound units now display as their SI derived equivalents
• kg*m/s^2 -> N (newton)
• kg*m^2/s^2 -> J (joule)
• kg*m^2/s^3 -> W (watt)
• m/s*s -> m (cancellation)
• Format string support: Use f-strings with DimArray
• f"{distance:.2f}" -> "1234.57 m"
• f"{energy:.2e}" -> "1.23e+03 J"
• NumPy ufunc integration: Use numpy functions directly
• np.sin(angle), np.cos(angle) - require dimensionless input
• np.exp(x), np.log(x) - require dimensionless input
• np.sqrt(area) - halves dimension exponents
• np.abs(velocity) - preserves units
• np.add(a, b), np.multiply(a, b) - dimension-aware arithmetic

Changed

• Unit display now uses simplified symbols by default

[0.1.2] - 2025-01-08

Fixed

• Corrected GitHub repository URLs in PyPI metadata

[0.1.1] - 2025-01-08

Added

• sqrt() method on DimArray for square root with proper dimension handling

[0.1.0] - 2025-01-08

Added

• Initial release
• DimArray class wrapping numpy arrays with unit metadata
• Dimension class with full algebra (multiply, divide, power)
• SI base units: meter, kilogram, second, ampere, kelvin, mole, candela
• SI derived units: newton, joule, watt, pascal, volt, hertz, etc.
• Common non-SI units: km, mile, hour, eV, atm, etc.
• Unit conversion with .to() method
• Dimensional error catching at operation time
• Arithmetic operations with automatic dimension checking
• Comparison operations between compatible dimensions
• Array indexing, slicing, and iteration
• Reduction operations (sum, mean, std, min, max)