Substractive registration refinement

[DamienGilliard]

This PR concerns the last step of the process, where we identify the individual joints in a beam and propose an analysis of each one, through local registrations

1) assume two inputs: i_cloud and i_joints. i_cloud is the point cloud of the whole object, i_joints is a list of list the meshes of the joints. [[face_1_joint_A, face_2_joint_A, ...], [face_1_joint_B, face_2_joint_B, ...], ...]
2) segment the joints from the i_cloud and store them in a list that follows the same structure as i_joints: [joint_A, joint_B, ...]
3) for each joint point cloud, perform a registration by populating the corresponding mesh faces of each joint
4) apply the registration to the segmented joint point cloud and re-segment the joints
5) the result is a list of list of the segmented joints, where each segment is locally registered to the corresponding mesh faces: [[face_1_joint_A_arfter_reg_a, face_2_joint_A_after_reg_a, ...], [face_1_joint_B_after_reg_b, face_2_joint_B_after_reg_b, ...], ...]
--> that way we can analyse the joints individually (face 1 of joint A is too much to the left, ...)

[DamienGilliard]

UPDATE: in the current situation, the component isolates and merges the clusters of each joint, creates a reference point cloud of each joint, computes the transformation between both, and applies it. Therefore, the transformation applied to each joint is slightly different.

[DamienGilliard]

@9and3 , The branch is not yet fully ready for review, but close

[DamienGilliard]

We have a component !! This branch is ready for review @9and3 !! since it is almost only python stuff, I also include @eleniv3d

[9and3]

Thanks for this 🦖 ! Left few comments, the main one is about the synthax for the component, the rest seem good (also some questions) have a look!

[9and3]

noice!