Public API

EvoLib provides a flat import surface through evolib/api.py. Instead of navigating deep into subpackages, you can import the most relevant classes, functions, and utilities directly:

from evolib import Pop, Indiv, Vector, EvoNet, NetVector
from evolib import plot_fitness, rastrigin, mse_loss
from evolib import save_checkpoint, resume_from_checkpoint

Core Classes

  • Pop (evolib.core.population.Population)

  • Indiv (evolib.core.individual.Indiv)

  • Vector (evolib.representation.vector.Vector)

  • EvoNet (evolib.representation.evonet.EvoNet)

  • NetVector (evolib.representation.netvector.NetVector)

  • HistoryLogger (evolib.utils.history_logger.HistoryLogger)

Fitness Functions

  • Type alias: evolib.interfaces.types.FitnessFunction

Benchmark Functions: - sphere (sphere_2d, sphere_3d) - rastrigin (rastrigin_2d, rastrigin_3d) - ackley (ackley_2d, ackley_3d) - rosenbrock (rosenbrock_2d, rosenbrock_3d) - griewank (griewank_2d, griewank_3d) - schwefel (schwefel_2d, schwefel_3d) - simple_quadratic

Loss Functions

  • mse_loss • mae_loss • huber_loss • bce_loss • cce_loss

Plotting Utilities

  • plot_fitness • plot_approximation • plot_history • plot_diversity

  • plot_mutation_trends • plot_fitness_comparison • save_combined_net_plot

Checkpointing

  • save_checkpoint • resume_from_checkpoint • resume_or_create

  • save_best_indiv • load_best_indiv

Notes

  • This page mirrors what is re-exported in evolib/api.py (__all__).

  • For detailed docs, see the dedicated pages (e.g., api_core_individual).

  • Importing via the public API is the recommended way for end users.