Crate cont_arith_code

source ·

Modules§

Structs§

  • Turns a model and binary code into a value
  • Byte interface for Decoder
  • Turns a model and a value into binary code
  • Byte interface for Encoder
  • A single sample from a distribution
  • A vector of values from a distribution. MLE for best results. The model is static and therefore not optimal (values don’t get removed from the distribution, so like a bag “with replacement”)

Traits§

  • Models produce distributions and consume their resolutions (as i64s) until a value T pops out. Used by both Encoder and Decoder symmetrically.
  • A univariate distribution with a truncated probability mass between a lower- and upper-bound.
  • A univariate distribution that can be truncated.

Type Aliases§

  • Bins of a univariate distribution are indexed by i64s