Changes in version 0.3.1                        

Release maintenance

  - Registers the September 2026 ecosystem-methods article in the
    pkgdown article index so the complete documentation site builds
    successfully.
  - Adds exact-tag R CMD check triggering for immutable v* release tags
    and extends the matrix with Ubuntu oldrel-1.
  - Preserves the complete 0.3.0 scientific API, defaults, estimands,
    analysis contracts, modelling semantics, and interpretation
    boundaries.
  - The existing annotated v0.3.0 tag remains immutable and unchanged.

                 Changes in version 0.3.0 (2026-08-23)                  

  - Added CRAN status, download, DOI, GitHub Actions, pkgdown, and
    license badges; added canonical CRAN, website, source, issue, and
    CRAN-check links; and documented CRAN and GitHub installation paths.

  - Synchronized README, package description, citation metadata, and
    pkgdown-facing release status for 0.3.0; README counts are now
    derived from the public NAMESPACE and vignette sources.

  - Added a common analysis-contract and provenance foundation for
    systematic 0.3.0 hardening.

  - Added sequence_capabilities(), audit_sequence_analysis(), and
    compare_sequence_analysis_results() for machine-readable capability,
    contract, and reproducibility audits.

  - Added deterministic contract, metamorphic, and adversarial test
    foundations plus a machine-readable public-API validation ledger.

  - Added auditable longitudinal and panel-sequence preparation,
    summaries, change metrics, and plots.

  - Added bounded non-contiguous subsequence enumeration, prevalence
    summaries, group comparisons, and plots.

  - Added optional mgcv-based time-varying state and transition models
    with participant random-effect support.

  - Added dependency-light multichannel and covariate-dependent
    categorical hidden Markov models.

  - Added design-aware permutation and bootstrap comparisons with
    explicit associational and causal guardrails.

  - Added six focused visualisation helpers, deterministic tests, an
    installed-package smoke test, and six synthetic articles.

                        Changes in version 0.2.0                        

  - Released the advanced standalone sequence-analysis API with 46
    exported functions and 47 manual pages.

  - Added four integrated workflow articles, bringing the package
    website to nine synthetic, reproducible articles.

  - Prepared the complete reference and article collection for
    publication at the main package website.

  - Configured pkgdown so future validated master updates publish
    directly to the main website.

  - Added four integrated, user-oriented articles covering getting
    started, validation and preparation, method selection, and a
    reproducible synthetic case study.

  - Added the complete advanced sequence-analysis roadmap.

  - Added deterministic aligned-position consensus sequences and
    agreement summaries.

  - Added descriptive between-group state, transition, and length
    comparisons.

  - Added Levenshtein, LCS, optimal-matching, and transition-profile
    distances.

  - Added hierarchical, PAM, and CLARA clustering interfaces, validation
    metrics, representatives, ensembles, and stability resampling.

  - Added transition networks, centrality, communities, higher-order
    transition models, prediction, and sequence-level bootstrap
    intervals.

  - Added dependency-light categorical HMM and mixture-HMM estimation,
    decoding, summaries, and model-comparison tables.

  - Added optional adapters for TraMineR, arulesSequences, GrpString,
    seqHMM, igraph, and common gp3tools-style data frames.

  - Added four synthetic workflow articles, deterministic tests, and an
    end-to-end smoke test.

  - Initial public release.

  - Created the independent gp3sequences package scaffold.

  - Added MIT licensing and package-level documentation.

  - Added the initial testthat edition 3 infrastructure.

  - Defined the initial neutral scope and interpretation boundary.

  - Added the initial README and pkgdown website configuration.

  - Added a multi-platform GitHub Actions R CMD check workflow for
    macOS, Windows, Ubuntu release, and Ubuntu devel.

  - Added a GitHub Actions pkgdown workflow for automated website
    building and deployment.

  - Added audit_sequence_data() and validate_sequence_data() for
    structured input diagnostics.

  - Added prepare_sequence_data() with explicit missing-state,
    duplicate-position, repeated-state, duration, and unknown-state
    policies.

  - Added encode_sequence_data() for deterministic state dictionaries
    and transparent state codes.

  - Added summarise_sequence_states(), summarise_sequence_transitions(),
    and format_sequence_paths() for structural sequence summaries.

  - Added extract_sequence_ngrams() for deterministic contiguous motif
    enumeration with explicit overlap handling.

  - Added summarise_sequence_motifs(), filter_sequence_motifs(), and
    format_sequence_motifs() for transparent motif prevalence,
    filtering, and reporting.

  - Added summarise_sequence_motif_positions() and
    format_sequence_motif_positions() for deterministic structural
    location summaries.

  - Added dependency-free plot_sequence_motifs() and
    plot_sequence_motif_positions() visualisations plus a synthetic
    contiguous-motif workflow article.