Loading deepchem/feat/__init__.py +2 −1 Original line number Diff line number Diff line Loading @@ -3,7 +3,8 @@ Making it easy to import in classes. """ from deepchem.feat.base_classes import Featurizer from deepchem.feat.base_classes import MolecularFeaturizer from deepchem.feat.base_classes import CrystalFeaturizer from deepchem.feat.base_classes import StructureFeaturizer from deepchem.feat.base_classes import CompositionFeaturizer from deepchem.feat.base_classes import ComplexFeaturizer from deepchem.feat.base_classes import UserDefinedFeaturizer from deepchem.feat.graph_features import ConvMolFeaturizer Loading deepchem/feat/base_classes.py +130 −29 Original line number Diff line number Diff line Loading @@ -208,17 +208,17 @@ class MolecularFeaturizer(Featurizer): return self.featurize(molecules) class CrystalFeaturizer(Featurizer): class StructureFeaturizer(Featurizer): """ Abstract class for calculating a set of features for a crystal structure. Abstract class for calculating a set of features for an inorganic crystal structure. The defining feature of a `CrystalFeaturizer` is that it operates on 3D crystals with periodic boundary conditions. Inorganic crystal structures are represented by Pymatgen composition and structure The defining feature of a `StructureFeaturizer` is that it operates on 3D crystal structures with periodic boundary conditions. Inorganic crystal structures are represented by Pymatgen structure objects. Featurizers for inorganic crystal structures that are subclasses of this class should plan to process input which comes as composition strings or pymatgen structure dictionaries. this class should plan to process input which comes as pymatgen structure objects. Child classes need to implement the _featurize method for calculating features for a single crystal. Loading @@ -230,14 +230,16 @@ class CrystalFeaturizer(Featurizer): """ def featurize(self, crystals: Iterable, log_every_n: int = 1000) -> np.ndarray: """Calculate features for crystals. def featurize(self, structures: Iterable[dict], log_every_n: int = 1000) -> np.ndarray: """Calculate features for crystal structures. Parameters ---------- crystals: Iterable Iterable sequence of composition strings, pymatgen structure dictionaries, or another crystal representation. structures: Iterable[dict] Iterable sequence of pymatgen structure dictionaries. Json-serializable dictionary representation of pymatgen.core.structure https://pymatgen.org/pymatgen.core.structure.html log_every_n: int, default 1000 Logging messages reported every `log_every_n` samples. Loading @@ -245,23 +247,122 @@ class CrystalFeaturizer(Featurizer): ------- features: np.ndarray A numpy array containing a featurized representation of `crystals`. `structures`. """ # Special case handling of single crystal if not isinstance(crystals, Iterable): crystals = [crystals] # Special case handling of single crystal structure if not isinstance(structures, Iterable): structures = [structures] else: # Convert iterables to list crystals = list(crystals) structures = list(structures) try: from pymatgen import Structure except ModuleNotFoundError: raise ValueError("This class requires pymatgen to be installed.") features = [] for idx, structure in enumerate(structures): if idx % log_every_n == 0: logger.info("Featurizing datapoint %i" % idx) try: s = Structure.from_dict(structure) features.append(self._featurize(s)) except: logger.warning( "Failed to featurize datapoint %i. Appending empty array" % idx) features.append(np.array([])) features = np.asarray(features) return features def _featurize(self, structure: "pymatgen.Structure"): """Calculate features for a single crystal structure. Parameters ---------- structure: pymatgen.Structure object Structure object with 3D coordinates and periodic lattice. """ raise NotImplementedError('Featurizer is not defined.') def __call__(self, structures: Iterable[dict]): """Calculate features for crystal structures. Parameters ---------- structures: Iterable[dict] An iterable of crystal structure dictionaries. """ return self.featurize(structures) class CompositionFeaturizer(Featurizer): """ Abstract class for calculating a set of features for an inorganic crystal composition. The defining feature of a `CompositionFeaturizer` is that it operates on 3D crystal chemical compositions. Inorganic crystal compositions are represented by Pymatgen composition objects. Featurizers for inorganic crystal compositions that are subclasses of this class should plan to process input which comes as Pymatgen composition objects. Child classes need to implement the _featurize method for calculating features for a single composition. Notes ----- Some subclasses of this class will require pymatgen and matminer to be installed. """ def featurize(self, compositions: Iterable[str], log_every_n: int = 1000) -> np.ndarray: """Calculate features for crystal compositions. Parameters ---------- compositions: Iterable[str] Iterable sequence of composition strings, e.g. "MoS2". log_every_n: int, default 1000 Logging messages reported every `log_every_n` samples. Returns ------- features: np.ndarray A numpy array containing a featurized representation of `compositions`. """ # Special case handling of single crystal composition if not isinstance(compositions, Iterable): compositions = [compositions] else: # Convert iterables to list compositions = list(compositions) try: from pymatgen import Composition except ModuleNotFoundError: raise ValueError("This class requires pymatgen to be installed.") features = [] for idx, crystal in enumerate(crystals): for idx, composition in enumerate(compositions): if idx % log_every_n == 0: logger.info("Featurizing datapoint %i" % idx) try: features.append(self._featurize(crystal)) c = Composition(composition) features.append(self._featurize(c)) except: logger.warning( "Failed to featurize datapoint %i. Appending empty array" % idx) Loading @@ -270,29 +371,29 @@ class CrystalFeaturizer(Featurizer): features = np.asarray(features) return features def _featurize(self, crystal): """Calculate features for a single crystal. def _featurize(self, composition: "pymatgen.Composition"): """Calculate features for a single crystal composition. Parameters ---------- crystal: crystal representation Crystal. composition: pymatgen.Composition object Composition object for 3D inorganic crystal. """ raise NotImplementedError('Featurizer is not defined.') def __call__(self, crystals: Iterable): """Calculate features for crystals. def __call__(self, compositions: Iterable[str]): """Calculate features for crystal compositions. Parameters ---------- crystals: Iterable An iterable of crystal representations. compositions: Iterable[str] An iterable of crystal compositions. """ return self.featurize(crystals) return self.featurize(compositions) class UserDefinedFeaturizer(Featurizer): Loading deepchem/feat/materials_featurizers.py +21 −37 Original line number Diff line number Diff line Loading @@ -4,11 +4,11 @@ Featurizers for inorganic crystals. import numpy as np from deepchem.feat import CrystalFeaturizer from deepchem.feat import StructureFeaturizer, CompositionFeaturizer from deepchem.utils import pad_array class ElementPropertyFingerprint(CrystalFeaturizer): class ElementPropertyFingerprint(CompositionFeaturizer): """ Fingerprint of elemental properties from composition. Loading Loading @@ -50,14 +50,14 @@ class ElementPropertyFingerprint(CrystalFeaturizer): self.data_source = data_source def _featurize(self, comp): def _featurize(self, composition: "pymatgen.Composition"): """ Calculate chemical fingerprint from crystal composition. Parameters ---------- comp : str Reduced formula of crystal. composition: pymatgen.Composition object Composition object. Returns ------- Loading @@ -66,27 +66,22 @@ class ElementPropertyFingerprint(CrystalFeaturizer): stoichiometry. Some values may be NaN. """ try: from pymatgen import Composition from matminer.featurizers.composition import ElementProperty except ModuleNotFoundError: raise ValueError("This class requires pymatgen and matminer to be installed.") # Get pymatgen Composition object c = Composition(comp) raise ValueError("This class requires matminer to be installed.") ep = ElementProperty.from_preset(self.data_source) try: feats = ep.featurize(c) feats = ep.featurize(composition) except: feats = [] return np.array(feats) class SineCoulombMatrix(CrystalFeaturizer): class SineCoulombMatrix(StructureFeaturizer): """ Calculate sine Coulomb matrix for crystals. Loading Loading @@ -129,15 +124,15 @@ class SineCoulombMatrix(CrystalFeaturizer): self.max_atoms = int(max_atoms) self.flatten = flatten def _featurize(self, struct): def _featurize(self, struct: "pymatgen.Structure"): """ Calculate sine Coulomb matrix from pymatgen structure. Parameters ---------- struct : dict Json-serializable dictionary representation of pymatgen.core.structure https://pymatgen.org/pymatgen.core.structure.html struct : pymatgen.Structure A periodic crystal composed of a lattice and a sequence of atomic sites with 3D coordinates and elements. Returns ------- Loading @@ -148,16 +143,13 @@ class SineCoulombMatrix(CrystalFeaturizer): """ try: from pymatgen import Structure from matminer.featurizers.structure import SineCoulombMatrix as SCM except ModuleNotFoundError: raise ValueError("This class requires pymatgen and matminer to be installed.") s = Structure.from_dict(struct) raise ValueError("This class requires matminer to be installed.") # Get full N x N SCM scm = SCM(flatten=False) sine_mat = scm.featurize(s) sine_mat = scm.featurize(struct) if self.flatten: eigs, _ = np.linalg.eig(sine_mat) Loading @@ -172,7 +164,7 @@ class SineCoulombMatrix(CrystalFeaturizer): return features class StructureGraphFeaturizer(CrystalFeaturizer): class StructureGraphFeaturizer(StructureFeaturizer): """ Calculate structure graph features for crystals. Loading Loading @@ -218,9 +210,9 @@ class StructureGraphFeaturizer(CrystalFeaturizer): Parameters ---------- struct : dict Json-serializable dictionary representation of pymatgen.core.structure https://pymatgen.org/pymatgen.core.structure.html struct : pymatgen.Structure A periodic crystal composed of a lattice and a sequence of atomic sites with 3D coordinates and elements. Returns ------- Loading @@ -230,15 +222,7 @@ class StructureGraphFeaturizer(CrystalFeaturizer): """ try: from pymatgen import Structure except ModuleNotFoundError: raise ValueError("This class requires pymatgen to be installed.") # Get pymatgen structure object s = Structure.from_dict(struct) features = self._get_structure_graph_features(s) features = self._get_structure_graph_features(struct) features = np.array(features) return features Loading @@ -249,7 +233,7 @@ class StructureGraphFeaturizer(CrystalFeaturizer): Parameters ---------- struct : pymatgen.core.structure struct : pymatgen.Structure A periodic crystal composed of a lattice and a sequence of atomic sites with 3D coordinates and elements. Loading docs/featurizers.rst +25 −12 Original line number Diff line number Diff line Loading @@ -161,23 +161,17 @@ AtomConvFeaturizer .. autoclass:: deepchem.feat.NeighborListComplexAtomicCoordinates :members: CrystalFeaturizer ----------------- StructureFeaturizer ------------------- Crystal Featurizers are those that work with datasets of crystals with Structure Featurizers are those that work with datasets of crystals with periodic boundary conditions. For inorganic crystal structures, these featurizers operate on chemical compositions (e.g. "MoS2"), or on a featurizers operate on pymatgen.Structure objects, which include a lattice and 3D coordinates that specify a periodic crystal structure. They should be applied on systems that have periodic boundary conditions. Crystal featurizers are not designed to work with molecules. .. autoclass:: deepchem.feat.CrystalFeaturizer :members: ElementPropertyFingerprint ^^^^^^^^^^^^^^^^^^^^^^^^^^ Structure featurizers are not designed to work with molecules. .. autoclass:: deepchem.feat.ElementPropertyFingerprint .. autoclass:: deepchem.feat.StructureFeaturizer :members: SineCoulombMatrix Loading @@ -192,6 +186,25 @@ StructureGraphFeaturizer .. autoclass:: deepchem.feat.StructureGraphFeaturizer :members: CompositionFeaturizer --------------------- Composition Featurizers are those that work with datasets of crystal compositions with periodic boundary conditions. For inorganic crystal structures, these featurizers operate on chemical compositions (e.g. "MoS2"). They should be applied on systems that have periodic boundary conditions. Composition featurizers are not designed to work with molecules. .. autoclass:: deepchem.feat.CompositionFeaturizer :members: ElementPropertyFingerprint ^^^^^^^^^^^^^^^^^^^^^^^^^^ .. autoclass:: deepchem.feat.ElementPropertyFingerprint :members: BindingPocketFeaturizer ----------------------- Loading Loading
deepchem/feat/__init__.py +2 −1 Original line number Diff line number Diff line Loading @@ -3,7 +3,8 @@ Making it easy to import in classes. """ from deepchem.feat.base_classes import Featurizer from deepchem.feat.base_classes import MolecularFeaturizer from deepchem.feat.base_classes import CrystalFeaturizer from deepchem.feat.base_classes import StructureFeaturizer from deepchem.feat.base_classes import CompositionFeaturizer from deepchem.feat.base_classes import ComplexFeaturizer from deepchem.feat.base_classes import UserDefinedFeaturizer from deepchem.feat.graph_features import ConvMolFeaturizer Loading
deepchem/feat/base_classes.py +130 −29 Original line number Diff line number Diff line Loading @@ -208,17 +208,17 @@ class MolecularFeaturizer(Featurizer): return self.featurize(molecules) class CrystalFeaturizer(Featurizer): class StructureFeaturizer(Featurizer): """ Abstract class for calculating a set of features for a crystal structure. Abstract class for calculating a set of features for an inorganic crystal structure. The defining feature of a `CrystalFeaturizer` is that it operates on 3D crystals with periodic boundary conditions. Inorganic crystal structures are represented by Pymatgen composition and structure The defining feature of a `StructureFeaturizer` is that it operates on 3D crystal structures with periodic boundary conditions. Inorganic crystal structures are represented by Pymatgen structure objects. Featurizers for inorganic crystal structures that are subclasses of this class should plan to process input which comes as composition strings or pymatgen structure dictionaries. this class should plan to process input which comes as pymatgen structure objects. Child classes need to implement the _featurize method for calculating features for a single crystal. Loading @@ -230,14 +230,16 @@ class CrystalFeaturizer(Featurizer): """ def featurize(self, crystals: Iterable, log_every_n: int = 1000) -> np.ndarray: """Calculate features for crystals. def featurize(self, structures: Iterable[dict], log_every_n: int = 1000) -> np.ndarray: """Calculate features for crystal structures. Parameters ---------- crystals: Iterable Iterable sequence of composition strings, pymatgen structure dictionaries, or another crystal representation. structures: Iterable[dict] Iterable sequence of pymatgen structure dictionaries. Json-serializable dictionary representation of pymatgen.core.structure https://pymatgen.org/pymatgen.core.structure.html log_every_n: int, default 1000 Logging messages reported every `log_every_n` samples. Loading @@ -245,23 +247,122 @@ class CrystalFeaturizer(Featurizer): ------- features: np.ndarray A numpy array containing a featurized representation of `crystals`. `structures`. """ # Special case handling of single crystal if not isinstance(crystals, Iterable): crystals = [crystals] # Special case handling of single crystal structure if not isinstance(structures, Iterable): structures = [structures] else: # Convert iterables to list crystals = list(crystals) structures = list(structures) try: from pymatgen import Structure except ModuleNotFoundError: raise ValueError("This class requires pymatgen to be installed.") features = [] for idx, structure in enumerate(structures): if idx % log_every_n == 0: logger.info("Featurizing datapoint %i" % idx) try: s = Structure.from_dict(structure) features.append(self._featurize(s)) except: logger.warning( "Failed to featurize datapoint %i. Appending empty array" % idx) features.append(np.array([])) features = np.asarray(features) return features def _featurize(self, structure: "pymatgen.Structure"): """Calculate features for a single crystal structure. Parameters ---------- structure: pymatgen.Structure object Structure object with 3D coordinates and periodic lattice. """ raise NotImplementedError('Featurizer is not defined.') def __call__(self, structures: Iterable[dict]): """Calculate features for crystal structures. Parameters ---------- structures: Iterable[dict] An iterable of crystal structure dictionaries. """ return self.featurize(structures) class CompositionFeaturizer(Featurizer): """ Abstract class for calculating a set of features for an inorganic crystal composition. The defining feature of a `CompositionFeaturizer` is that it operates on 3D crystal chemical compositions. Inorganic crystal compositions are represented by Pymatgen composition objects. Featurizers for inorganic crystal compositions that are subclasses of this class should plan to process input which comes as Pymatgen composition objects. Child classes need to implement the _featurize method for calculating features for a single composition. Notes ----- Some subclasses of this class will require pymatgen and matminer to be installed. """ def featurize(self, compositions: Iterable[str], log_every_n: int = 1000) -> np.ndarray: """Calculate features for crystal compositions. Parameters ---------- compositions: Iterable[str] Iterable sequence of composition strings, e.g. "MoS2". log_every_n: int, default 1000 Logging messages reported every `log_every_n` samples. Returns ------- features: np.ndarray A numpy array containing a featurized representation of `compositions`. """ # Special case handling of single crystal composition if not isinstance(compositions, Iterable): compositions = [compositions] else: # Convert iterables to list compositions = list(compositions) try: from pymatgen import Composition except ModuleNotFoundError: raise ValueError("This class requires pymatgen to be installed.") features = [] for idx, crystal in enumerate(crystals): for idx, composition in enumerate(compositions): if idx % log_every_n == 0: logger.info("Featurizing datapoint %i" % idx) try: features.append(self._featurize(crystal)) c = Composition(composition) features.append(self._featurize(c)) except: logger.warning( "Failed to featurize datapoint %i. Appending empty array" % idx) Loading @@ -270,29 +371,29 @@ class CrystalFeaturizer(Featurizer): features = np.asarray(features) return features def _featurize(self, crystal): """Calculate features for a single crystal. def _featurize(self, composition: "pymatgen.Composition"): """Calculate features for a single crystal composition. Parameters ---------- crystal: crystal representation Crystal. composition: pymatgen.Composition object Composition object for 3D inorganic crystal. """ raise NotImplementedError('Featurizer is not defined.') def __call__(self, crystals: Iterable): """Calculate features for crystals. def __call__(self, compositions: Iterable[str]): """Calculate features for crystal compositions. Parameters ---------- crystals: Iterable An iterable of crystal representations. compositions: Iterable[str] An iterable of crystal compositions. """ return self.featurize(crystals) return self.featurize(compositions) class UserDefinedFeaturizer(Featurizer): Loading
deepchem/feat/materials_featurizers.py +21 −37 Original line number Diff line number Diff line Loading @@ -4,11 +4,11 @@ Featurizers for inorganic crystals. import numpy as np from deepchem.feat import CrystalFeaturizer from deepchem.feat import StructureFeaturizer, CompositionFeaturizer from deepchem.utils import pad_array class ElementPropertyFingerprint(CrystalFeaturizer): class ElementPropertyFingerprint(CompositionFeaturizer): """ Fingerprint of elemental properties from composition. Loading Loading @@ -50,14 +50,14 @@ class ElementPropertyFingerprint(CrystalFeaturizer): self.data_source = data_source def _featurize(self, comp): def _featurize(self, composition: "pymatgen.Composition"): """ Calculate chemical fingerprint from crystal composition. Parameters ---------- comp : str Reduced formula of crystal. composition: pymatgen.Composition object Composition object. Returns ------- Loading @@ -66,27 +66,22 @@ class ElementPropertyFingerprint(CrystalFeaturizer): stoichiometry. Some values may be NaN. """ try: from pymatgen import Composition from matminer.featurizers.composition import ElementProperty except ModuleNotFoundError: raise ValueError("This class requires pymatgen and matminer to be installed.") # Get pymatgen Composition object c = Composition(comp) raise ValueError("This class requires matminer to be installed.") ep = ElementProperty.from_preset(self.data_source) try: feats = ep.featurize(c) feats = ep.featurize(composition) except: feats = [] return np.array(feats) class SineCoulombMatrix(CrystalFeaturizer): class SineCoulombMatrix(StructureFeaturizer): """ Calculate sine Coulomb matrix for crystals. Loading Loading @@ -129,15 +124,15 @@ class SineCoulombMatrix(CrystalFeaturizer): self.max_atoms = int(max_atoms) self.flatten = flatten def _featurize(self, struct): def _featurize(self, struct: "pymatgen.Structure"): """ Calculate sine Coulomb matrix from pymatgen structure. Parameters ---------- struct : dict Json-serializable dictionary representation of pymatgen.core.structure https://pymatgen.org/pymatgen.core.structure.html struct : pymatgen.Structure A periodic crystal composed of a lattice and a sequence of atomic sites with 3D coordinates and elements. Returns ------- Loading @@ -148,16 +143,13 @@ class SineCoulombMatrix(CrystalFeaturizer): """ try: from pymatgen import Structure from matminer.featurizers.structure import SineCoulombMatrix as SCM except ModuleNotFoundError: raise ValueError("This class requires pymatgen and matminer to be installed.") s = Structure.from_dict(struct) raise ValueError("This class requires matminer to be installed.") # Get full N x N SCM scm = SCM(flatten=False) sine_mat = scm.featurize(s) sine_mat = scm.featurize(struct) if self.flatten: eigs, _ = np.linalg.eig(sine_mat) Loading @@ -172,7 +164,7 @@ class SineCoulombMatrix(CrystalFeaturizer): return features class StructureGraphFeaturizer(CrystalFeaturizer): class StructureGraphFeaturizer(StructureFeaturizer): """ Calculate structure graph features for crystals. Loading Loading @@ -218,9 +210,9 @@ class StructureGraphFeaturizer(CrystalFeaturizer): Parameters ---------- struct : dict Json-serializable dictionary representation of pymatgen.core.structure https://pymatgen.org/pymatgen.core.structure.html struct : pymatgen.Structure A periodic crystal composed of a lattice and a sequence of atomic sites with 3D coordinates and elements. Returns ------- Loading @@ -230,15 +222,7 @@ class StructureGraphFeaturizer(CrystalFeaturizer): """ try: from pymatgen import Structure except ModuleNotFoundError: raise ValueError("This class requires pymatgen to be installed.") # Get pymatgen structure object s = Structure.from_dict(struct) features = self._get_structure_graph_features(s) features = self._get_structure_graph_features(struct) features = np.array(features) return features Loading @@ -249,7 +233,7 @@ class StructureGraphFeaturizer(CrystalFeaturizer): Parameters ---------- struct : pymatgen.core.structure struct : pymatgen.Structure A periodic crystal composed of a lattice and a sequence of atomic sites with 3D coordinates and elements. Loading
docs/featurizers.rst +25 −12 Original line number Diff line number Diff line Loading @@ -161,23 +161,17 @@ AtomConvFeaturizer .. autoclass:: deepchem.feat.NeighborListComplexAtomicCoordinates :members: CrystalFeaturizer ----------------- StructureFeaturizer ------------------- Crystal Featurizers are those that work with datasets of crystals with Structure Featurizers are those that work with datasets of crystals with periodic boundary conditions. For inorganic crystal structures, these featurizers operate on chemical compositions (e.g. "MoS2"), or on a featurizers operate on pymatgen.Structure objects, which include a lattice and 3D coordinates that specify a periodic crystal structure. They should be applied on systems that have periodic boundary conditions. Crystal featurizers are not designed to work with molecules. .. autoclass:: deepchem.feat.CrystalFeaturizer :members: ElementPropertyFingerprint ^^^^^^^^^^^^^^^^^^^^^^^^^^ Structure featurizers are not designed to work with molecules. .. autoclass:: deepchem.feat.ElementPropertyFingerprint .. autoclass:: deepchem.feat.StructureFeaturizer :members: SineCoulombMatrix Loading @@ -192,6 +186,25 @@ StructureGraphFeaturizer .. autoclass:: deepchem.feat.StructureGraphFeaturizer :members: CompositionFeaturizer --------------------- Composition Featurizers are those that work with datasets of crystal compositions with periodic boundary conditions. For inorganic crystal structures, these featurizers operate on chemical compositions (e.g. "MoS2"). They should be applied on systems that have periodic boundary conditions. Composition featurizers are not designed to work with molecules. .. autoclass:: deepchem.feat.CompositionFeaturizer :members: ElementPropertyFingerprint ^^^^^^^^^^^^^^^^^^^^^^^^^^ .. autoclass:: deepchem.feat.ElementPropertyFingerprint :members: BindingPocketFeaturizer ----------------------- Loading
