Is it possible to have sim agent metrics function in py_metrics_ops ?

[pengzhenghao]

We have this function for computing motion prediction metrics:

py_metrics_ops.motion_metrics(config=eval_config.SerializeToString(), **input_dict)

wherein the entries of input_dict are all tf tensor.

Do we also have the function like py_metrics_ops.sim_agents_metrics so that we can avoid using the protobuf to compute the metrics?