gt4sd.algorithms.conditional_generation.molgx.implementation module

Implementation of MolGX conditional generators.

Summary

Classes:

MolGXGenerator

Interface for MolGX generator.

Reference

class MolGXGenerator(resources_path, tag_name, homo_energy_value=-0.25, lumo_energy_value=0.08, use_linear_model=True, number_of_candidates=2, maximum_number_of_candidates=3, maximum_number_of_solutions=3, maximum_number_of_nodes=50000, beam_size=2000, without_estimate=True, use_specific_rings=True, use_fragment_const=False)[source]

Bases: object

Interface for MolGX generator.

__init__(resources_path, tag_name, homo_energy_value=-0.25, lumo_energy_value=0.08, use_linear_model=True, number_of_candidates=2, maximum_number_of_candidates=3, maximum_number_of_solutions=3, maximum_number_of_nodes=50000, beam_size=2000, without_estimate=True, use_specific_rings=True, use_fragment_const=False)[source]

Instantiate a MolGX generator.

Parameters

  • resources_path (str) – path to the resources for model loading.

  • tag_name (str) – tag for the pretrained model.

  • homo_energy_value (float) – target HOMO energy value. Defaults to -0.25.

  • lumo_energy_value (float) – target LUMO energy value. Defaults to 0.08.

  • use_linear_model (bool) – linear model usage. Defaults to True.

  • number_of_candidates (int) – number of candidates to consider. Defaults to 2.

  • maximum_number_of_candidates (int) – maximum number of candidates to consider. Defaults to 3.

  • maximum_number_of_solutions (int) – maximum number of solutions. Defaults to 3.

  • maximum_number_of_nodes (int) – maximum number of nodes in the graph exploration. Defaults to 50000.

  • beam_size (int) – size of the beam during search. Defaults to 2000.

  • without_estimate (bool) – disable estimates. Defaults to True.

  • use_specific_rings (bool) – flag to indicate whether specific rings are used. Defaults to True.

  • use_fragment_const (bool) – using constant fragments. Defaults to False.

Raises

RuntimeError – in the case extras are disabled.

static load_molgx(resource_path, tag_name)[source]

Load MolGX model.

Parameters

  • resource_path (str) – path to the resources for model loading.

  • tag_name (str) – tag for the pretrained model.

Return type

MolgxSdk

Returns

MolGX model SDK.

_create_parameters_dictionary()[source]

Create parameters dictionary.

Return type

Dict[str, Any]

Returns

the parameters to run MolGX.

property homo_energy_value: float
Return type

float

property lumo_energy_value: float
Return type

float

property use_linear_model: bool
Return type

bool

property number_of_candidates: int
Return type

int

property maximum_number_of_candidates: int
Return type

int

property maximum_number_of_solutions: int
Return type

int

__dict__ = mappingproxy({'__module__': 'gt4sd.algorithms.conditional_generation.molgx.implementation', '__doc__': 'Interface for MolGX generator.', '__init__': <function MolGXGenerator.__init__>, 'load_molgx': <staticmethod(<function MolGXGenerator.load_molgx>)>, '_create_parameters_dictionary': <function MolGXGenerator._create_parameters_dictionary>, 'homo_energy_value': <property object>, 'lumo_energy_value': <property object>, 'use_linear_model': <property object>, 'number_of_candidates': <property object>, 'maximum_number_of_candidates': <property object>, 'maximum_number_of_solutions': <property object>, 'maximum_number_of_nodes': <property object>, 'beam_size': <property object>, 'without_estimate': <property object>, 'use_specific_rings': <property object>, 'use_fragment_const': <property object>, 'parameters': <property object>, 'generate': <function MolGXGenerator.generate>, '__dict__': <attribute '__dict__' of 'MolGXGenerator' objects>, '__weakref__': <attribute '__weakref__' of 'MolGXGenerator' objects>, '__annotations__': {}})
__doc__ = 'Interface for MolGX generator.'
__module__ = 'gt4sd.algorithms.conditional_generation.molgx.implementation'
__weakref__

list of weak references to the object (if defined)

property maximum_number_of_nodes: int
Return type

int

property beam_size: int
Return type

int

property without_estimate: bool
Return type

bool

property use_specific_rings: bool
Return type

bool

property use_fragment_const: bool
Return type

bool

property parameters: Dict[str, Any]
Return type

Dict[str, Any]

generate()[source]

Sample random molecules.

Return type

List[str]

Returns

sampled molecule (SMILES).