Skip to content

CHANGELOG.md

Latest commit

 

History

History
76 lines (60 loc) · 3.16 KB

CHANGELOG.md

File metadata and controls

76 lines (60 loc) · 3.16 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

0.1.0 - 2025-09-XX

Added

  • Initial release of Fugue, a monadic probabilistic programming library for Rust.
  • Core probabilistic programming framework:
    • Model<A> abstraction for composable probabilistic programs.
    • Monadic operations: bind, map, and_then, pure for program composition.
    • sample, observe, factor, guard primitives for probabilistic modeling.
  • Type-safe distribution system with natural return types:
    • Bernoulli distribution returning bool (eliminates == 1.0 comparisons).
    • Poisson and Binomial distributions returning u64 (natural counting).
    • Categorical distribution returning usize (safe array indexing).
    • Continuous distributions (Normal, Beta, Gamma, etc.) returning f64.
    • 10 built-in distributions with parameter validation and numerical stability.
  • Ergonomic macros for probabilistic programming:
    • prob! macro for Haskell-style do-notation.
    • plate! macro for vectorized operations over collections.
    • addr! and scoped_addr! macros for hierarchical addressing.
  • Multiple inference algorithms:
    • MCMC: Adaptive Metropolis-Hastings with convergence diagnostics.
    • SMC: Sequential Monte Carlo with multiple resampling methods.
    • VI: Mean-field variational inference with ELBO optimization.
    • ABC: Approximate Bayesian Computation with distance functions.
  • Effect handler system:
    • Handler trait for pluggable model interpreters.
    • 5 built-in handlers: PriorHandler, ReplayHandler, ScoreGivenTrace, SafeReplayHandler, SafeScoreGivenTrace.
    • Type-safe execution preserving distribution return types.
  • Trace system for execution history:
    • Complete recording of random choices and log-weights.
    • Type-safe value access with get_f64(), get_bool(), get_u64(), get_usize().
    • Three-component log-weight decomposition (prior, likelihood, factors).
  • Production features:
    • Comprehensive error handling with FugueError and error codes.
    • Numerically stable algorithms with overflow protection.
    • Convergence diagnostics: R-hat, effective sample size, Geweke tests.
    • Statistical validation against analytical solutions.
  • Documentation and examples:
    • Comprehensive user guide with 20+ tutorial and how-to pages.
    • Complete API documentation with rustdoc.
    • 14 examples covering foundation concepts, statistical modeling, and advanced patterns.
    • 158+ doctests ensuring example correctness.
  • Testing infrastructure:
    • 82+ unit tests across all modules.
    • 9+ integration tests for end-to-end workflows.
    • Property-based testing with proptest.
    • Continuous integration with format, lint, and test enforcement.

Changed

  • N/A (initial release)

Deprecated

  • N/A (initial release)

Removed

  • N/A (initial release)

Fixed

  • N/A (initial release)

Security

  • N/A (initial release)