Changes in version 0.1.2 (2026-09-29)                  

This is the first public release of rbiogeme.

The package is distributed under the MIT License.

This version is the CRAN resubmission with the example tree installed
under inst/examples and a package word list for legitimate proper names.

This version also avoids initializing an unavailable native runtime
during package checks.

CRAN review fixes

  - Removed the redundant “An R” wording from the package title and
    description.
  - Removed default current-working-directory destinations from writing
    APIs. Native reports, checkpoints, diagnostics, Pareto files, and
    sampled-data files now require explicit user-selected paths or
    directories.
  - Routed native generated files through the requested output directory
    and returned stable absolute paths without changing the R working
    directory.
  - Updated the examples under inst/examples, vignettes, and tests to
    use explicit output paths or session temporary directories;
    distributed examples no longer call setwd().
  - Ensured temporary working-directory changes retained by integration
    tests are restored with on.exit().

User-facing interface

  - Added a complete R-facing interface to the native Biogeme engine.
  - Added symbolic data, parameter, probability, likelihood, panel,
    draw, integration, and post-estimation interfaces.
  - Added biogeme_setup() for automatic native-runtime provisioning and
    biogeme_check() for actionable environment diagnostics.
  - Added native model validation with validate_model().
  - Added native probability prediction with predict() and scenario
    data.
  - Added self-contained R examples mirroring the supported Biogeme
    example groups.

Runtime and reproducibility

  - The default native requirement is biogeme==3.3.5.
  - Numerical estimation, differentiation, integration, optimization,
    simulation, and reporting remain delegated to native Biogeme.
  - Ordinary estimate() calls do not silently recycle previous YAML or
    iteration files.
  - Random-draw and sampling workflows document their reproducibility
    controls and remaining simulation noise.

Attribution

The whole package was implemented by ChatGPT 5.6 Luna under the
supervision of Michel Bierlaire.