In the evaluation of binary metrics, precision_score, recall_score and f1_score get label and score in wrong order:
**

def compute_binary_metrics(anomaly_pred, anomaly_label, adjustment=False):
if not adjustment:
eval_anomaly_pred = anomaly_pred
metrics = {
"f1": f1_score(eval_anomaly_pred, anomaly_label),
"pc": precision_score(eval_anomaly_pred, anomaly_label),
"rc": recall_score(eval_anomaly_pred, anomaly_label),
}

**

It should be:
**

def compute_binary_metrics(anomaly_pred, anomaly_label, adjustment=False):
if not adjustment:
eval_anomaly_pred = anomaly_pred
metrics = {
"f1": f1_score(anomaly_label,eval_anomaly_pred),
"pc": precision_score(anomaly_label, eval_anomaly_pred),
"rc": recall_score(anomaly_label, eval_anomaly_pred),
}

**