Crate u_numflow 

u-numflow

Foundational mathematical primitives.

This crate provides foundational mathematical, statistical, and probabilistic building blocks that are domain-agnostic. It knows nothing about scheduling, nesting, geometry, or any consumer domain.

Modules

• stats — Descriptive statistics with numerical stability guarantees
• distributions — Probability distributions (Uniform, Triangular, PERT, Normal, LogNormal, Weibull)
• special — Special mathematical functions (normal CDF, inverse normal CDF)
• matrix — Dense matrix operations (multiply, determinant, inverse, Cholesky)
• random — Random number generation, shuffling, and weighted sampling
• collections — Specialized data structures (Union-Find)
• transforms — Box-Cox power transformation
• fourier — Discrete Fourier transform of any length (radix-2 / Bluestein)

Design Philosophy

• Numerical stability first: Welford’s algorithm for variance, Neumaier summation for accumulation
• Reproducibility: Seeded RNG support for deterministic experiments
• Property-based testing: Mathematical invariants verified via proptest

Modules

collections

Specialized data structures for optimization algorithms.

distributions

Probability distributions.

fourier

Discrete Fourier transform.

matrix

Dense matrix operations.

random

Random number generation, shuffling, and weighted sampling.

special

Special mathematical functions.

stats

Descriptive statistics with numerical stability guarantees.

transforms

Data transformations for statistical analysis.

units

Units of measure: dimensional analysis and conversion between units of the same dimension.