o ߥi, @s dZddlZddlZddlmZmZmZddlZddl m Z ddl m Z ee ejfZee efZdZeeZejddGd d d Zdd e d ee d efddZd e fddZded e fddZded ejfddZ ddededeejd efddZ ddedeejd efddZdS)zProtein data type.N)AnyMappingOptional) PDBParser)residue_constants>ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789T)frozenc@sVeZdZUdZejed<ejed<ejed<ejed<ejed<ejed<dd Zd S) Proteinz!Protein structure representation.atom_positionsaatype atom_mask residue_index chain_index b_factorscCs(tt|jtkrtdtddS)Nz(Cannot build an instance with more than z6 chains because these cannot be written to PDB format.)lennpuniquerPDB_MAX_CHAINS ValueError)selfrb/home/ubuntu/.local/lib/python3.10/site-packages/modelscope/models/science/unifold/data/protein.py __post_init__<s  zProtein.__post_init__N)__name__ __module__ __qualname____doc__rndarray__annotations__rrrrrr s        r pdb_strchain_idreturnc st|}tdd}|d|}t|}t|dkr&tdt|d|d}g}g}g} g} g} g} |D]} |durD| j|krDq8| D]}|jd d kr^td | jd |jdd t j |j d}t j |t j}tt jdf}tt jf}tt jf}|D]#}|jt jvrq|j|t j|j<d|t j|j<|j|t j|j<qt|dkrqF||||| || |jd| | j| |qFq8t| }ddt|Dtfdd| D}tt|t| t|t| |t| dS)aTakes a PDB string and constructs a Protein object. WARNING: All non-standard residue types will be converted into UNK. All non-standard atoms will be ignored. Args: pdb_str: The contents of the pdb file chain_id: If chain_id is specified (e.g. A), then only that chain is parsed. Otherwise all chains are parsed. Returns: A new `Protein` parsed from the pdb contents. T)QUIETnonez,Only single model PDBs are supported. Found z models.rN z(PDB contains an insertion code at chain z and residue index z. These are not supported.X??cSsi|]\}}||qSrr).0ncidrrr z#from_pdb_string..csg|]}|qSrr)r+r-chain_id_mappingrr sz#from_pdb_string..)r r r r rr)ioStringIOr get_structurelist get_modelsrridr restype_3to1getresname restype_order restype_numrzeros atom_type_numname atom_typescoord atom_orderbfactorsumappendr enumeratearrayr )rr pdb_fhparser structuremodelsmodelr r r r chain_idsrchainres res_shortname restype_idxposmask res_b_factorsatomunique_chain_idsrrr0rfrom_pdb_stringCs~             rXcCs*d}|d|dd|dd|d|dS) NTER<6>5z >3r&>1>4r) atom_index end_resname chain_namer chain_endrrr _chain_ends rcprotcs|tjdgfdd}tj}g}|j}|j}|j}|jtj }|j tj }|j } t |tj kr7tdi} t|D]} | tkrLtdtdt| | | <q>|dd} |d } t|jd D]} | || kr|t| ||| d| || d|| d|| } | d7} ||| }t||| || | | D]n\}}}}|d krqd }t|d kr|nd |}d}d}d}|d }d}|d| dd |d|d|dd | || d|| d|dd|d d|dd|dd|d|dd|d|d}||| d7} qqe|t| ||d| |d|d|d|ddd |D}d!|d!S)"zConverts a `Protein` instance to a PDB string. Args: prot: The protein to convert to PDB. Returns: PDB string. r'cstj|dS)NUNK)r restype_1to3r:)rrestypesrrres_1to3szto_pdb..res_1to3zInvalid aatypes.z The PDB format supports at most z chains.z MODEL 1r$rr*ATOMr&r)rZr[z<4r]r\r^z z>8.3fr%z>6.2fz z>2ENDMDLENDcSsg|]}|dqS)P)ljust)r+linerrrr2r/zto_pdb.. )rrirAr r r r astyperint32rranyr=rrr PDB_CHAIN_IDSrFrangeshapercziprjoin)rdrjrA pdb_linesr r r r rrrNir_last_chain_index res_name_3 atom_namerSrTb_factor record_typer@alt_locinsertion_code occupancyelementcharge atom_linerrhrto_pdbs              rcCs tj|jS)asComputes an ideal atom mask. `Protein.atom_mask` typically is defined according to the atoms that are reported in the PDB. This function computes a mask according to heavy atoms that should be present in the given sequence of amino acids. Args: prot: `Protein` whose fields are `numpy.ndarray` objects. Returns: An ideal atom mask. )rSTANDARD_ATOM_MASKr )rdrrrideal_atom_masks rfeaturesresultrcCsbd|vr |dd}nt|d}|durt|d}t|d|d|d|dd||dS) aAssembles a protein from a prediction. Args: features: Dictionary holding model inputs. fold_output: Dictionary holding model outputs. b_factors: (Optional) B-factors to use for the protein. Returns: A protein instance. asym_idr$r Nfinal_atom_maskfinal_atom_positionsr r r r r rrr zeros_liker )rrrrrrrfrom_predictions rcCsbd|vr |dd}nt|d}|durt|d}t|d|d|d|dd||dS) zAssembles a standard pdb from input atom positions & mask. Args: features: Dictionary holding model inputs. b_factors: (Optional) B-factors to use for the protein. Returns: A protein instance. rr$r N all_atom_maskall_atom_positionsr rr)rrrrrr from_feature's  r)N)r dataclassesr3typingrrrnumpyrBio.PDBr&modelscope.models.science.unifold.datarstrr FeatureDict ModelOutputrxrr dataclassr rXrcrrrrrrrrs>     "N` !