NEWS.md

In niarules: Numerical Association Rule Mining using Population-Based Nature-Inspired Algorithms

niarules 0.3.1

New (experimental)

• render_coral_rgl_experimental(): a feature-rich 3D renderer for coral plots. Theming, legends, and flexible mappings for edge width/color/alpha. Status: experimental — arguments and defaults may change before 0.4.0.
• metric_domains(): compute [min, max] ranges for rule metrics (support, confidence, lift) from a parsed object or a data frame. Status: experimental.
• coral_list_themes() and coral_get_theme(): list and retrieve theme objects for the experimental renderer. Status: experimental.

Enhancements

• build_coral_plots()
• New optional arguments: bin_breaks, bin_digits (defaults preserve prior behavior).
• Returns additional node fields: node_id, is_root, coral_id, interval_brackets, bin_index.
• Adds a new return component bin_legend (or NULL) to describe bins.
• The original nodes, edges, and grid_size components are unchanged.
• parse_rules()
• More permissive input: accepts common column aliases (lift/Fitness, lhs/rhs, case-insensitive).
• Output now carries class "parsed" to enable S3 methods like metric_domains(parsed).

Bug fixes

• Interval parsing (C++): mixed bracket intervals like [lo, hi) and (lo, hi] are parsed correctly; bracketed intervals now consistently use rel_op = "in".
• LHS parsing (C++): single outer braces are stripped before splitting, so {A,B} is treated as A, B.

Internal

• Rcpp entry points (build_layout_cpp(), parse_rules_cpp()) unchanged.

niarules documentation built on Sept. 15, 2025, 5:08 p.m.