Precision@k is removed after v0.11.0

[donglihe-hub]

Hi TorchMetrics team,

I am thankful for this powerful tool and really appreciate your hard work.

I'm working on evaluating multilabel classification results with precision@k, which is logged by Lightning loggers. I noticed that since v0.11.0, Precison module cannot generalize to Precision@K anymore.

An pseudo example of preds and target in my case is:

preds:  [0.1, 0.9, 0.3, 0.4, 0.5, 0.1]
target: [0, 0, 1, 0, 1, 0]

To calculate precision@k, I obtain the indices of the top k elements in preds, apply thresholding to them, and calculate precision using the tp and fp obtained from the top_k preds-target pair.

I can achieve my purpose with Precision before 0.11.0. Currently the version of TorchMetrics in my project is stuck to 0.10.3. And I'd like to know if there is an workaround for my problem. Thank you very much.