Tags

Tags give the ability to mark specific points in history as being important

2.0.0

ad510ec8 · Merge branch 'dev' into 'main' · May 27, 2026

Release: 2.0.0

What's Changed

This release modernizes the project's packaging and runtime requirements.
The classification API (`Classification.predict` / `train`) is unchanged.

Breaking changes
  - Python 3.14+ is now required (previously 3.10+).
  - numpy and pandas upgraded to the 2.x series (previously pinned <2.0);
    matplotlib raised to 3.9+.

Packaging
  - Migrated from `setup.py` to a PEP 621 `pyproject.toml` with a pinned
    `uv.lock` for reproducible installs. `setup.py` has been removed.
  - The project is now managed with [uv](https://docs.astral.sh/uv/).

Documentation
  - README documents uv-based installation, dev/docs dependency groups, and how
    to add vehi-cl-e as a library in another uv-managed project.

Tooling
  - `.gitignore` updated for the uv environment (`.venv/`) and editor history.

v1.0.0

b1868e2e · Merge branch 'dev' into 'main' · Feb 10, 2026

First stable release of the VEHI-CL-E vehicle classification library.

Classifies light duty vehicles into European segments (A-I standard, MPV, SUV, N-I/N-II/N-III vans) using a Bayesian statistical method based on physical characteristics.

Highlights:
  - Pydantic v2 validation for input data and configuration
  - Multiple input formats: DataFrames, NumPy arrays, dictionaries
  - Train custom classification boundaries from labelled datasets
  - 34 pytest tests with reference dataset validation
  - Tutorial notebook with practical use-case examples

Reference: Laveneziana et al. (2025), "A science-based approach to classifying light vehicles in Europe", Scientific Reports, 15, 9099. DOI: 10.1038/s41598-025-90625-9