gd_dl package

Subpackages

Submodules

gd_dl.lib_pdb_mol2 module

class gd_dl.lib_pdb_mol2.Atom(line)

Bases: Residue

R()
atmName()
i_atm()
class gd_dl.lib_pdb_mol2.MOL2(mol2_fn)

Bases: object

read(read_end=None, read_hydrogen=False)
write(model_index_start=0, model_index_end=None)
class gd_dl.lib_pdb_mol2.MOL2_UNIT(model_no)

Bases: object

add_hydrogen_index(index)
append_atom_index(index)
append_atom_mol2_type(mol2_type)
append_coordinates(tmp_crd_list)
get_atom_index_list()
get_atom_mol2_type_list()
get_bond_dict()
get_coordinates_np_array()
get_hydrogen_set()
read_line(line)
update_bond(start, end, bond_type)
write()
class gd_dl.lib_pdb_mol2.Model(model_no=0)

Bases: object

append(X)
get_residue_lines(res_range=[])
get_residues(res_range=[])
index(key)
write(exclude_remark=False, exclude_symm=False, exclude_missing_bb=False, exclude_nucl=False, exclude_SSbond=False, remark_s=[], chain_id=None)
class gd_dl.lib_pdb_mol2.PDB(pdb_fn, read=True, read_het=True)

Bases: object

read(read_het=True)
write(exclude_remark=True, exclude_symm=False, exclude_missing_bb=True, model_index=[], remark_s=[])
class gd_dl.lib_pdb_mol2.PDBline(line)

Bases: object

isAtom()
isHetatm()
isResidue()
startswith(key)
class gd_dl.lib_pdb_mol2.Residue(line)

Bases: object

R(atmName=None, atmIndex=None)
append(line)
atmIndex(atmName)
atmName()
chainID()
check_bb()
exists(atmName)
get_CB()
get_backbone()
get_heavy()
get_sc()
i_atm(atmName=None, atmIndex=None)
isAtom()
isHetatm()
isResidue()
resName()
resNo()
resNo_char()
write()

gd_dl.models module

class gd_dl.models.Cgc_block(node_dim=32, edge_dim=16)

Bases: PatchedModule

forward(args)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class gd_dl.models.Rerank_model(node_dim_in=29, node_dim_hidden=64, edge_dim_in=17, edge_dim_hidden=32, ligand_only=True, readout='mean')

Bases: PatchedModule

forward(G, n_atom)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class gd_dl.models.Weight_CGConv(channels: int | Tuple[int, int], dim: int = 0, aggr: str = 'add', batch_norm: bool = False, bias: bool = True, **kwargs)

Bases: MessagePassing

The crystal graph convolutional operator from the “Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties” paper We modifed the crystal graph convolutional operator to incorporate manual edge weights

forward(x: Tensor | Tuple[Tensor, Tensor], edge_index: Tensor | SparseTensor, edge_attr: Tensor | None = None, edge_weight: Tensor | None = None) → Tensor

Runs the forward pass of the module.

message(x_i, x_j, edge_attr: Tensor | None, edge_weight: Tensor | None) → Tensor

Constructs messages from node \(j\) to node \(i\) in analogy to \(\phi_{\mathbf{\Theta}}\) for each edge in edge_index. This function can take any argument as input which was initially passed to propagate(). Furthermore, tensors passed to propagate() can be mapped to the respective nodes \(i\) and \(j\) by appending _i or _j to the variable name, .e.g. x_i and x_j.

reset_parameters()

Resets all learnable parameters of the module.

gd_dl.path_setting module

gd_dl.rerank_model module

class gd_dl.rerank_model.Cgc_block(node_dim=32, edge_dim=16)

Bases: PatchedModule

forward(args)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class gd_dl.rerank_model.Rerank_model(node_dim_in=29, node_dim_hidden=32, edge_dim_in=17, edge_dim_hidden=16, ligand_only=True, readout='mean')

Bases: PatchedModule

forward(G, n_atom)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class gd_dl.rerank_model.Weight_CGConv(channels: int | Tuple[int, int], dim: int = 0, aggr: str = 'add', batch_norm: bool = False, bias: bool = True, **kwargs)

Bases: MessagePassing

The crystal graph convolutional operator from the “Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties” paper

forward(x: Tensor | Tuple[Tensor, Tensor], edge_index: Tensor | SparseTensor, edge_attr: Tensor | None = None, edge_weight: Tensor | None = None) → Tensor
message(x_i, x_j, edge_attr: Tensor | None) → Tensor

Constructs messages from node \(j\) to node \(i\) in analogy to \(\phi_{\mathbf{\Theta}}\) for each edge in edge_index. This function can take any argument as input which was initially passed to propagate(). Furthermore, tensors passed to propagate() can be mapped to the respective nodes \(i\) and \(j\) by appending _i or _j to the variable name, .e.g. x_i and x_j.

reset_parameters()

Resets all learnable parameters of the module.

gd_dl.utils module

gd_dl.utils.set_prep(mode, valid_idx, train_val_ratio)
gd_dl.utils.str2bool(v)

Module contents