1 rri_metrics.plotting

rri_metrics.plotting

Build candidate-aligned Plotly figures for RRI and distance diagnostics.

This module visualizes RriResult labels, squared point–mesh components, and selected candidate geometry. It owns presentation only: candidate order, validity, actor-visible inputs, and oracle-label construction remain with their source payloads and callers. Distance chart values are in square metres when the underlying geometry is metric world-frame data.

1.1 Classes

Name	Description
RriOrdinalPreview	Finite RRI population, fitted quantile edges, and resulting class counts.

1.1.1 RriOrdinalPreview

rri_metrics.plotting.RriOrdinalPreview(values, edges, class_counts, num_classes)

Finite RRI population, fitted quantile edges, and resulting class counts.

1.2 Functions

Name	Description
build_rri_ordinal_preview	Fit the production ordinal binner and return immutable preview evidence.
plot_pm_accuracy	Plot pre/post mean-squared point-to-mesh accuracy errors in square metres.
plot_pm_completeness	Plot pre/post mean-squared mesh-to-point completeness errors in square metres.
plot_pm_distances	Plot broadcast pre-view and per-candidate post-view bidirectional errors in square metres.
plot_rri_scores	Build one dimensionless oracle-RRI bar per candidate label.
rri_color_map	Assign colors deterministically in input-label order, cycling the palette when needed.
rri_ordinal_preview_figures	Build raw-value/edge and class-count figures from one fitted preview.

1.2.1 build_rri_ordinal_preview

rri_metrics.plotting.build_rri_ordinal_preview(values, *, num_classes)

Fit the production ordinal binner and return immutable preview evidence.

1.2.2 plot_pm_accuracy

rri_metrics.plotting.plot_pm_accuracy(
    rri,
    labels,
    color_map,
    *,
    baseline_label='-1',
    title,
)

Plot pre/post mean-squared point-to-mesh accuracy errors in square metres.

1.2.3 plot_pm_completeness

rri_metrics.plotting.plot_pm_completeness(
    rri,
    labels,
    color_map,
    *,
    baseline_label='-1',
    title,
)

Plot pre/post mean-squared mesh-to-point completeness errors in square metres.

1.2.4 plot_pm_distances

rri_metrics.plotting.plot_pm_distances(
    rri,
    labels,
    color_map,
    *,
    baseline_label='-1',
    title,
)

Plot broadcast pre-view and per-candidate post-view bidirectional errors in square metres.

1.2.5 plot_rri_scores

rri_metrics.plotting.plot_rri_scores(rri, labels, color_map, *, title)

Build one dimensionless oracle-RRI bar per candidate label.

1.2.5.1 Parameters

Name	Type	Description	Default
rri	RriResult	Candidate-aligned oracle result.	required
labels	Sequence[str]	Display labels in the same order as rri.rri.	required
color_map	Mapping[str, str]	Color lookup containing every display label.	required
title	str	Figure title.	required

1.2.5.2 Returns

Name	Type	Description
	go.Figure	Plotly bar figure whose x-axis is candidate identity and whose y-axis
	go.Figure	is dimensionless RRI. Negative bars are valid low-utility labels, not
	go.Figure	invalid-candidate markers.

1.2.6 rri_color_map

rri_metrics.plotting.rri_color_map(labels, *, palette=None)

Assign colors deterministically in input-label order, cycling the palette when needed.

1.2.7 rri_ordinal_preview_figures

rri_metrics.plotting.rri_ordinal_preview_figures(
    preview,
    *,
    histogram_bins,
    rri_label,
)

Build raw-value/edge and class-count figures from one fitted preview.