module

Moses

Constants

EPSILON_SCORE = Float64::EPSILON
VERSION = "0.1.0"
VERY_BEST_SCORE = Float64::MAX

Special score values

VERY_WORST_SCORE = Float64::MIN

Class methods

compare_candidates(a : Candidate, b : Candidate) : Int32

Helper method to safely compare candidates by score. Relies on score_or_worst to normalize NaN; the || 0 guards the nilable return type of Float64#<=> (which is Int32?).

initialize

Initialize the MOSES subsystem

main(args = ARGV)

Command-line interface for MOSES

run_boolean_demo

Run boolean classification example

run_demo

Run a comprehensive MOSES demonstration

run_moses(params : MosesParams) : MosesResult

Main MOSES evolutionary search function This is the primary entry point for running MOSES optimization

run_regression_demo

Run regression example

run_test_suite

Run MOSES test suite

score_or_worst(candidate : Candidate) : Score

Helper method to get score or worst score. NaN scores (e.g. from Inf - Inf in penalized_score) are normalized to VERY_WORST_SCORE so that downstream sort/max operations remain transitive and treat incomparable candidates consistently as worst.

Nested types