Fix the metrics for a rare disease classifier

An AI evaluated a rare disease classifier, with about 1.5% positive cases, using accuracy alone. It printed 98.6% and called the model "highly performant", but a model that says "negative" to everyone scores 98.5%. The fix reports precision, recall and the majority baseline next to accuracy.

An assistant wrote evaluate_rare_class(y_true, y_pred). It is in the editor. It has a bug. Find it and fix it.

Input and output:

• y_true: the real labels. y_pred: the model's predictions, at the same index.
• Return a list of four numbers: accuracy, precision, recall and baseline.

Rules the tests check:

• If the lists differ in length, use only the first n pairs, where n is the shorter length. If n is 0, return [0, 0, 0, 0].
• A label is positive only when it is exactly 1. Every other value is negative.
• Accuracy: the share of pairs where the prediction and the real label are both positive or both negative.
• Precision: of the pairs predicted positive, the share that are really positive. 0 if nothing is predicted positive.
• Recall: of the pairs that are really positive, the share predicted positive. 0 if there are no real positives.
• Baseline: the accuracy of always predicting the larger class, which is the larger of the positive and negative counts, divided by n.
• Answers within 0.000001 of the expected values pass.