ROCCurve

Planned · @experimental

Planned Component

This component is part of the faststrap[plots] / faststrap[ml] optional extra and has not been implemented yet. The documentation is a preview of the planned API.

Renders a Receiver Operating Characteristic (ROC) curve with AUC annotation. Supports binary and multi-class classification with overlaid curves.

Quick Start

from faststrap.components.advance import ROCCurve
from sklearn.metrics import roc_curve, auc

fpr, tpr, _ = roc_curve(y_true, y_scores)

ROCCurve(
    fpr=fpr,
    tpr=tpr,
    auc_score=auc(fpr, tpr),
    title="ROC Curve — Binary Classifier",
)

Features

Single or multi-class ROC curves on the same axes
AUC score annotation
Diagonal random-classifier reference line
Theme-aware colors (light/dark)
Accepts pre-computed fpr/tpr arrays or sklearn RocCurveDisplay data

Parameters

Parameter	Type	Default	Description
`fpr`	`list \\| np.ndarray`	required (or `curves`)	False positive rate array
`tpr`	`list \\| np.ndarray`	required (or `curves`)	True positive rate array
`curves`	`list[dict] \\| None`	`None`	Multi-class: list of `{fpr, tpr, label, auc}` dicts
`auc_score`	`float \\| None`	`None`	AUC score to annotate
`title`	`str \\| None`	`None`	Optional chart title
`show_random_line`	`bool`	`True`	Show diagonal reference line
`responsive`	`bool`	`True`	Full-width responsive wrapper
`height`	`int \\| str`	`350`	Chart height
`**kwargs`	`Any`		Extra wrapper attributes

Notes

Pure server-rendered — no JS required for static curves.
Marked @experimental.