format_multiview_data_for_pca

lightning_pose.utils.pca.format_multiview_data_for_pca(data_arr: Tensor, {'__torchtyping__': True, 'details': ('batch', 'num_keypoints', '2'), 'cls_name': 'TensorType'}], mirrored_column_matches: ListConfig | list) → Tensor, {'__torchtyping__': True, 'details': ('batch_times_num_selected_keypoints', 'two_times_num_views',), 'cls_name': 'TensorType'}]

Reformat multiview data so each observation is a single body part across views.

Parameters:

• data_arr – keypoints from training data

• mirrored_column_matches – one element for each camera view; each element is itself a list that contains indices into the overall ordering of the keypoints

Returns:

formatted data to run pca