Currently the split file requires each line to be of the format <pdb_id>_final, and the dataset pipeline constructs paths to the file with <BASE_DIR>/<pdb_id>/<pdb_id>_final.pdb. A cleaner way to do this would likely be have the user specify <BASE_DIR> and each line on the split file should be a path to the pdb/cif file relative to the base dir.
Currently the split file requires each line to be of the format
<pdb_id>_final, and the dataset pipeline constructs paths to the file with<BASE_DIR>/<pdb_id>/<pdb_id>_final.pdb. A cleaner way to do this would likely be have the user specify<BASE_DIR>and each line on the split file should be a path to the pdb/cif file relative to the base dir.