Issue with script - 'plDDT is not a file in the archive'

[GarlicPaasta]

Hi,

I am trying to use ipsae.py to calculate scores for designs generated using the BoltzGen pipeline (runs Boltz-2 under the hood). I understood that providing the .npz and .cif output from the refolding step would be compatible with the Boltz-1 usage mode of this script.

However, when I run the script on the files in intermediate_designs_inverse_folded/fold_out_npz/, I encounter a KeyError because the script expects full plddt and pae arrays, which seem to be missing from the BoltzGen output archives.

Command:
python3 ipsae.py \ .../intermediate_designs_inverse_folded/fold_out_npz/O14936_PF00595_domain_0001.npz \ .../intermediate_designs_inverse_folded/des_refold_pdbs/O14936_PF00595_domain_0001.cif \ 10 10

Error:
KeyError: 'plddt is not a file in the archive'

Content of the .npz file: I inspected the keys in the .npz file, and it only contains scalar summary metrics, not the full matrices:

['token_index', 'mol_type', 'res_type', 'atom_resolved_mask', 'coords', 'min_interaction_pae', 'iptm', 'design_iptm', 'design_ptm', ...]

Question: Is there a way to run ipsae.py with these summary files (from boltzgen)?

Thanks!