Italbytz.AI.Learning.Abstractions 1.0.0-preview.5

nuget-ai

nuget-ai is the target repository for the refactored Italbytz.AI.* package family.

Which package should I use?

• Use Italbytz.AI.Abstractions for general solver and metrics contracts.
• Use Italbytz.AI for shared runtime helpers such as metrics and thread-safe randomness.
• Use Italbytz.AI.Agent.Abstractions when you only need agent-facing contracts.
• Use Italbytz.AI.Agent for basic agent and environment helper implementations.
• Use Italbytz.AI.Search.Abstractions for search/problem contracts.
• Use Italbytz.AI.Search for concrete search helpers such as SearchAgent, GeneralProblem, breadth-first search, and uniform-cost search.
• Use Italbytz.AI.CSP.Abstractions for constraint-satisfaction contracts such as ICSP, IAssignment, IDomain, and IConstraint.
• Use Italbytz.AI.CSP for concrete CSP building blocks and solvers such as CSP, MapCSP, FlexibleBacktrackingSolver, TreeCspSolver, and MinConflictsSolver.
• Use Italbytz.AI.Planning.Abstractions for planning contracts such as IActionSchema, IState, IPlanningProblem, and the minimal FOL term/literal abstractions they rely on.
• Use Italbytz.AI.Planning for planning parsers and implementations such as ActionSchema, State, PlanningProblem, HierarchicalSearchAlgorithm, and PlanningProblemFactory.
• Use Italbytz.AI.Learning.Abstractions for dataset, attribute, and learner contracts such as IDataSet, IExample, ILearner, IParameterizedLearner, and ICrossValidation.
• Use Italbytz.AI.Learning for concrete learning helpers and algorithms such as DataSetFactory, DataSetSpecification, MajorityLearner, DecisionTreeLearner, DecisionTree, and CrossValidation.
• Use Italbytz.AI.Evolutionary.Abstractions for evolutionary-algorithm contracts such as IFitnessFunction, IFitnessValue, IIndividual, IGenotype, ISearchSpace, ILiteral, IMonomial, and IPolynomial.
• Use Italbytz.AI.Evolutionary for core evolutionary implementations such as SingleFitnessValue, OneMax, BitString, BitStringGenotype, SetLiteral, WeightedMonomial, WeightedPolynomial, and WeightedPolynomialGenotype.
• Use Italbytz.AI.ML.Core for shared ML.NET helpers and configuration DTOs such as ThreadSafeMLContext, DataExcerpt, CategoricalFeature, NumericalFeature, TrainingConfiguration, TabularFileDataSourceV3, metrics, default column names, and lookup-map utilities.
• Use Italbytz.AI.ML for Learning-backed ML.NET integrations, trainers, and model-inspection helpers such as DecisionTreeBinaryTrainer, DecisionTreeMulticlassTrainer, LeastSquaresTrainer, Explainer, and Interpreter.
• Use Italbytz.AI.ML.UciDatasets for curated UCI-style dataset loaders and preprocessing helpers such as IrisDataset, HeartDiseaseDataset, WineQualityDataset, BreastCancerWisconsinDiagnosticDataset, and the shared Data registry.

Documentation

API documentation is generated with docfx and can be published via GitHub Pages:

• https://italbytz.github.io/nuget-ai/

The doc site now also carries forward the most useful orientation material from the older AI/AIMA repositories:

• an architecture guide for the historic Ports-and-Adapters split
• an AIMA-oriented algorithm index
• an ML quickstart for LeastSquaresTrainer, Explainer, and Interpreter

If the URL still returns 404, wait until the CI workflow on main has completed the first Pages publish run.

Quality checks

This repository includes:

• a GitHub Actions workflow in .github/workflows/ci.yml
• automated restore, build, test, pack, and docs generation
• a docfx setup under docfx/

Release model

• the current nuget-ai line stays on 1.0.0-preview.* while the broader AI migration is still in progress
• a pushed tag such as v1.0.0-preview.2 triggers the release-ready pipeline in GitHub Actions
• if the repository secret NUGET_API_KEY is configured, the workflow also publishes .nupkg and .snupkg files to NuGet

Local validation

dotnet restore nuget-ai.sln
dotnet test nuget-ai.sln -v minimal
dotnet pack nuget-ai.sln -c Release -v minimal
dotnet tool restore
dotnet tool run docfx docfx/docfx.json