raise warning if labels and target_names have different length in classification_report #7751
Description
Summarizing #3959. If target_names doesn't match the length of (inferred) labels, probably something is wrong. The following code should raise a warning and point the user to the labels parameter.
y_true = [0, 0, 2, 0, 0]
y_pred = [0, 2, 2, 0, 0]
target_names = ['class 0', 'class 1', 'class 2']
print(metrics.classification_report(y_true, y_pred, target_names=target_names))
== console ==
precision recall f1-score support
class 0 1.00 0.75 0.86 4
class 1 0.50 1.00 0.67 1
avg / total 0.90 0.80 0.82 5