[0.2.6] - 2026-06-07

Added

• plot_marginal_probability(axis='recency', ...) method for visualizing marginal empirical revisit
  probability along the recency or frequency axis as a line chart.
  Returns a matplotlib.figure.Figure for inline display in Jupyter Lab / Colab.
  Use this method to check monotonicity before calling optimize().
• Marginal probability attributes populated by fit() / fit_period():
  • recency_probability_ / frequency_probability_: DataFrames with aggregated N, cv, probability
  • R2N, R2CV, R2Prob: dicts mapping recency rank to sample count, conversion count, probability
  • F2N, F2CV, F2Prob: dicts mapping frequency to sample count, conversion count, probability
• title, figsize, fontsize parameters to plot_probability_surface() and plot_marginal_probability()
  for publication-ready output.
• recency_label, frequency_label, probability_label parameters to plot_probability_surface()
  for customizing axis labels.
• xlabel, probability_label parameters to plot_marginal_probability()
  for customizing axis labels.
• Optional [ja] extra dependency: pip install rfscorer[ja] installs japanize-matplotlib
  to enable Japanese axis labels and titles in both plot methods.

Changed

• img/ reference PNGs updated to reflect the new plot style (black wireframe, transparent panes).