The UR force() method reduce the linear forces with F = norm([Fx, Fy, Fz]), so you could do the same with sensor linear forces.
If the part is non-uniform, you probably dont know about its center of gravity so it would be difficult to use the moments.
If I had more precision on the task (insertion, masse validation, ..) I could suggest how to use the moments to be more precise.
Has anyone developed a script or formula for calculating the resultant force that the sensor is seeing resolved into a single vector? Just looking for the resultant force magnitude similar to what UR does with the force() method where they are reducing the 3 linear forces and 3 moments that are reported from get_actual_tcp_force() into a single magnitude. This is very handy when you are dealing with non-uniform parts that might have a Z only component one time but then a Z component and a moment another time