AUC score legend not plotted when score=True

[yztxwd]

🐛 Bug

When calling the .plot(score=True) on curve metrics like BinaryPrecisionRecall, BinaryROC.... The AUC score is not shown.

To Reproduce

Steps to reproduce the behavior...

from torch import randn, randint
from torchmetrics.classification import BinaryPrecisionRecallCurve
import torch.nn.functional as F
preds = F.softmax(randn(20, 2), dim=1)
target = randint(2, (20,))
metric = BinaryPrecisionRecallCurve()
metric.update(preds[:, 1], target)
fig_, ax_ = metric.plot(score=True)

Code sample

Expected behavior

Expect a legend showing the AUC score.

Environment

• TorchMetrics version (and how you installed TM, e.g. conda, pip, build from source): 1.01, from pip
• Python & PyTorch Version (e.g., 1.0): torch 1.13, python 3.10.11
• Any other relevant information such as OS (e.g., Linux): Linux

Additional context

The bug seems be due to the score is computed when curve is None and score is True, however curve is computed before this expression, so curve is always not None and score will never be computed, here is an example of BinaryROC:

# line 156-164 in src/torchmetrics/classification/roc.py
        curve = curve or self.compute()
        score = _auc_compute_without_check(curve[0], curve[1], 1.0) if not curve and score is True else None
        return plot_curve(
            curve,
            score=score,
            ax=ax,
            label_names=("False positive rate", "True positive rate"),
            name=self.__class__.__name__,
        )

it works as expected after changing it to:

        curve_computed = curve or self.compute()
        score = _auc_compute_without_check(curve_computed[0], curve_computed[1], 1.0) if not curve and score is True else None
        return plot_curve(
            curve_computed,
            score=score,
            ax=ax,
            label_names=("False positive rate", "True positive rate"),
            name=self.__class__.__name__,
        )

[github-actions]

Hi! thanks for your contribution!, great first issue!