Loading deepchem/models/atomic_conv.py +3 −113 Original line number Original line Diff line number Diff line Loading @@ -17,102 +17,6 @@ from deepchem.utils.typing import KerasActivationFn, LossFn, OneOrMany logger = logging.getLogger(__name__) logger = logging.getLogger(__name__) # class AtomicConvScore(Layer): # """The scoring function used by the atomic convolution models.""" # def __init__(self, atom_types, layer_sizes, **kwargs): # super(AtomicConvScore, self).__init__(**kwargs) # self.atom_types = atom_types # self.layer_sizes = layer_sizes # def build(self, input_shape): # self.type_weights = [] # self.type_biases = [] # self.output_weights = [] # self.output_biases = [] # n_features = int(input_shape[0][-1]) # layer_sizes = self.layer_sizes # num_layers = len(layer_sizes) # weight_init_stddevs = [1 / np.sqrt(x) for x in layer_sizes] # bias_init_consts = [0.0] * num_layers # for ind, atomtype in enumerate(self.atom_types): # prev_layer_size = n_features # self.type_weights.append([]) # self.type_biases.append([]) # self.output_weights.append([]) # self.output_biases.append([]) # for i in range(num_layers): # weight, bias = initializeWeightsBiases( # prev_layer_size=prev_layer_size, # size=layer_sizes[i], # weights=tf.random.truncated_normal( # shape=[prev_layer_size, layer_sizes[i]], # stddev=weight_init_stddevs[i]), # biases=tf.constant( # value=bias_init_consts[i], shape=[layer_sizes[i]])) # self.type_weights[ind].append(weight) # self.type_biases[ind].append(bias) # prev_layer_size = layer_sizes[i] # weight, bias = initializeWeightsBiases(prev_layer_size, 1) # self.output_weights[ind].append(weight) # self.output_biases[ind].append(bias) # def call(self, inputs): # frag1_layer, frag2_layer, complex_layer, frag1_z, frag2_z, complex_z = inputs # atom_types = self.atom_types # num_layers = len(self.layer_sizes) # def atomnet(current_input, atomtype): # prev_layer = current_input # for i in range(num_layers): # #layer = tf.nn.bias_add( # # tf.matmul(prev_layer, self.type_weights[atomtype][i]), # # self.type_biases[atomtype][i]) # #layer = tf.nn.relu(layer) # layer = Dense(100)(prev_layer) # prev_layer = layer # #output_layer = tf.squeeze( # # tf.nn.bias_add( # # tf.matmul(prev_layer, self.output_weights[atomtype][0]), # # self.output_biases[atomtype][0])) # print("self.output_weights[atomtype][0].shape") # print(self.output_weights[atomtype][0].shape) # output_layer = Dense( # self.output_weights[atomtype][0].shape[0])(prev_layer) # return output_layer # frag1_zeros = tf.zeros_like(frag1_z, dtype=tf.float32) # frag2_zeros = tf.zeros_like(frag2_z, dtype=tf.float32) # complex_zeros = tf.zeros_like(complex_z, dtype=tf.float32) # frag1_atomtype_energy = [] # frag2_atomtype_energy = [] # complex_atomtype_energy = [] # for ind, atomtype in enumerate(atom_types): # frag1_outputs = tf.map_fn(lambda x: atomnet(x, ind), frag1_layer) # frag2_outputs = tf.map_fn(lambda x: atomnet(x, ind), frag2_layer) # complex_outputs = tf.map_fn(lambda x: atomnet(x, ind), complex_layer) # cond = tf.equal(frag1_z, atomtype) # frag1_atomtype_energy.append(tf.where(cond, frag1_outputs, frag1_zeros)) # cond = tf.equal(frag2_z, atomtype) # frag2_atomtype_energy.append(tf.where(cond, frag2_outputs, frag2_zeros)) # cond = tf.equal(complex_z, atomtype) # complex_atomtype_energy.append( # tf.where(cond, complex_outputs, complex_zeros)) # frag1_outputs = tf.add_n(frag1_atomtype_energy) # frag2_outputs = tf.add_n(frag2_atomtype_energy) # complex_outputs = tf.add_n(complex_atomtype_energy) # frag1_energy = tf.reduce_sum(frag1_outputs, 1) # frag2_energy = tf.reduce_sum(frag2_outputs, 1) # complex_energy = tf.reduce_sum(complex_outputs, 1) # binding_energy = complex_energy - (frag1_energy + frag2_energy) # return tf.expand_dims(binding_energy, axis=1) class AtomicConvModel(KerasModel): class AtomicConvModel(KerasModel): """Implements an Atomic Convolution Model. """Implements an Atomic Convolution Model. Loading Loading @@ -203,7 +107,7 @@ class AtomicConvModel(KerasModel): learning_rate: float learning_rate: float Learning rate for the model. Learning rate for the model. """ """ # TODO: Turning off queue for now. Safe to re-activate? self.complex_num_atoms = complex_num_atoms self.complex_num_atoms = complex_num_atoms self.frag1_num_atoms = frag1_num_atoms self.frag1_num_atoms = frag1_num_atoms self.frag2_num_atoms = frag2_num_atoms self.frag2_num_atoms = frag2_num_atoms Loading Loading @@ -260,16 +164,11 @@ class AtomicConvModel(KerasModel): regularizer = tf.keras.regularizers.l2(weight_decay_penalty) regularizer = tf.keras.regularizers.l2(weight_decay_penalty) else: else: regularizer = None regularizer = None # score = AtomicConvScore(self.atom_types, layer_sizes)([ # self._frag1_conv, self._frag2_conv, self._complex_conv, frag1_z, # frag2_z, complex_z # ]) # print("score") # print(score) prev_layer = concat prev_layer = concat # dropout_switch = Input(shape=tuple()) prev_size = concat.shape[0] prev_size = concat.shape[0] next_activation = None next_activation = None # Add the dense layers # Add the dense layers for size, weight_stddev, bias_const, dropout, activation_fn in zip( for size, weight_stddev, bias_const, dropout, activation_fn in zip( Loading @@ -278,9 +177,6 @@ class AtomicConvModel(KerasModel): layer = prev_layer layer = prev_layer if next_activation is not None: if next_activation is not None: layer = Activation(next_activation)(layer) layer = Activation(next_activation)(layer) print("size") print(size) # layer = Dense(100)(layer) layer = Dense( layer = Dense( size, size, kernel_initializer=tf.keras.initializers.TruncatedNormal( kernel_initializer=tf.keras.initializers.TruncatedNormal( Loading @@ -305,11 +201,6 @@ class AtomicConvModel(KerasModel): bias_initializer=tf.constant_initializer( bias_initializer=tf.constant_initializer( value=bias_init_consts[-1]))(prev_layer)) value=bias_init_consts[-1]))(prev_layer)) loss: Union[dc.models.losses.Loss, LossFn] loss: Union[dc.models.losses.Loss, LossFn] # prev_layer = Dense(100)(prev_layer) # output = Dense(1)(prev_layer) # print("output") # print(output) # loss = dc.models.losses.L2Loss() model = tf.keras.Model( model = tf.keras.Model( inputs=[ inputs=[ Loading @@ -317,7 +208,6 @@ class AtomicConvModel(KerasModel): frag2_nbrs_z, frag2_z, complex_X, complex_nbrs, complex_nbrs_z, frag2_nbrs_z, frag2_z, complex_X, complex_nbrs, complex_nbrs_z, complex_z complex_z ], ], # outputs=score) outputs=output) outputs=output) super(AtomicConvModel, self).__init__( super(AtomicConvModel, self).__init__( model, L2Loss(), batch_size=batch_size, **kwargs) model, L2Loss(), batch_size=batch_size, **kwargs) Loading deepchem/models/tests/test_atomic_conv.py +36 −21 Original line number Original line Diff line number Diff line Loading @@ -6,14 +6,26 @@ import os import pytest import pytest import deepchem import numpy as np import tensorflow as tf import unittest import numpy as np import numpy as np from deepchem.models import atomic_conv from deepchem.models import atomic_conv from deepchem.data import NumpyDataset from deepchem.data import NumpyDataset from deepchem.feat import ComplexNeighborListFragmentAtomicCoordinates from deepchem.feat import AtomicConvFeaturizer def test_atomic_conv_initialize(): """Quick test of AtomicConv.""" acm = atomic_conv.AtomicConvModel( n_tasks=1, batch_size=1, layer_sizes=[ 1, ], frag1_num_atoms=5, frag2_num_atoms=5, complex_num_atoms=10) assert acm.complex_num_atoms == 10 assert len(acm.atom_types) == 15 @pytest.mark.slow @pytest.mark.slow Loading @@ -26,6 +38,7 @@ def test_atomic_conv(): atomic_convnet = atomic_conv.AtomicConvModel( atomic_convnet = atomic_conv.AtomicConvModel( n_tasks=1, n_tasks=1, batch_size=batch_size, batch_size=batch_size, layer_sizes=[10], frag1_num_atoms=5, frag1_num_atoms=5, frag2_num_atoms=5, frag2_num_atoms=5, complex_num_atoms=10, complex_num_atoms=10, Loading @@ -40,7 +53,6 @@ def test_atomic_conv(): frag1_z = np.random.randint(10, size=(N_atoms)) frag1_z = np.random.randint(10, size=(N_atoms)) frag2_coords = np.random.rand(N_atoms, 3) frag2_coords = np.random.rand(N_atoms, 3) frag2_nbr_list = {0: [], 1: [], 2: [], 3: [], 4: []} frag2_nbr_list = {0: [], 1: [], 2: [], 3: [], 4: []} #frag2_z = np.random.rand(N_atoms, 3) frag2_z = np.random.randint(10, size=(N_atoms)) frag2_z = np.random.randint(10, size=(N_atoms)) system_coords = np.random.rand(2 * N_atoms, 3) system_coords = np.random.rand(2 * N_atoms, 3) system_nbr_list = { system_nbr_list = { Loading @@ -63,20 +75,13 @@ def test_atomic_conv(): features = np.asarray(features) features = np.asarray(features) labels = np.random.rand(batch_size) labels = np.random.rand(batch_size) train = NumpyDataset(features, labels) train = NumpyDataset(features, labels) #atomic_convnet.fit(train, nb_epoch=300) atomic_convnet.fit(train, nb_epoch=150) atomic_convnet.fit(train, nb_epoch=150) print("labels") print(labels) print("atomic_convnet.predict(train)") print(atomic_convnet.predict(train)) assert np.allclose(labels, atomic_convnet.predict(train), atol=0.01) assert np.allclose(labels, atomic_convnet.predict(train), atol=0.01) @pytest.mark.slow @pytest.mark.slow def test_atomic_conv_variable(): def test_atomic_conv_variable(): """A simple test that initializes and fits an AtomicConvModel on variable input size.""" """A simple test that initializes and fits an AtomicConvModel on variable input size.""" # For simplicity, let's assume both molecules have same number of # atoms. frag1_num_atoms = 1000 frag1_num_atoms = 1000 frag2_num_atoms = 1200 frag2_num_atoms = 1200 complex_num_atoms = frag1_num_atoms + frag2_num_atoms complex_num_atoms = frag1_num_atoms + frag2_num_atoms Loading @@ -84,6 +89,9 @@ def test_atomic_conv_variable(): atomic_convnet = atomic_conv.AtomicConvModel( atomic_convnet = atomic_conv.AtomicConvModel( n_tasks=1, n_tasks=1, batch_size=batch_size, batch_size=batch_size, layer_sizes=[ 10, ], frag1_num_atoms=frag1_num_atoms, frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms) complex_num_atoms=complex_num_atoms) Loading @@ -108,6 +116,9 @@ def test_atomic_conv_variable(): labels = np.zeros(batch_size) labels = np.zeros(batch_size) train = NumpyDataset(features, labels) train = NumpyDataset(features, labels) atomic_convnet.fit(train, nb_epoch=1) atomic_convnet.fit(train, nb_epoch=1) preds = atomic_convnet.predict(train) assert preds.shape == (1, 1, 1) assert np.count_nonzero(preds) > 0 @pytest.mark.slow @pytest.mark.slow Loading @@ -117,28 +128,29 @@ def test_atomic_conv_with_feat(): ligand_file = os.path.join(dir_path, ligand_file = os.path.join(dir_path, "../../feat/tests/data/3zso_ligand_hyd.pdb") "../../feat/tests/data/3zso_ligand_hyd.pdb") protein_file = os.path.join(dir_path, protein_file = os.path.join(dir_path, "../../feat/tests/data/3zso_protein.pdb") "../../feat/tests/data/3zso_protein_noH.pdb") # Pulled from PDB files. For larger datasets with more PDBs, would use # Pulled from PDB files. For larger datasets with more PDBs, would use # max num atoms instead of exact. # max num atoms instead of exact. frag1_num_atoms = 44 # for ligand atoms frag1_num_atoms = 44 # for ligand atoms frag2_num_atoms = 2336 # for protein atoms frag2_num_atoms = 2334 # for protein atoms complex_num_atoms = 2380 # in total complex_num_atoms = 2378 # in total max_num_neighbors = 4 max_num_neighbors = 4 # Cutoff in angstroms # Cutoff in angstroms neighbor_cutoff = 4 neighbor_cutoff = 4 complex_featurizer = ComplexNeighborListFragmentAtomicCoordinates( complex_featurizer = AtomicConvFeaturizer(frag1_num_atoms, frag2_num_atoms, frag1_num_atoms, frag2_num_atoms, complex_num_atoms, max_num_neighbors, complex_num_atoms, neighbor_cutoff) max_num_neighbors, neighbor_cutoff) # arbitrary label # arbitrary label labels = np.array([0]) labels = np.array([0]) features, _ = complex_featurizer.featurize([ligand_file], [protein_file]) features = complex_featurizer.featurize([(ligand_file, protein_file)]) dataset = deepchem.data.DiskDataset.from_numpy(features, labels) dataset = NumpyDataset(features, labels) batch_size = 1 batch_size = 1 print("Constructing Atomic Conv model") print("Constructing Atomic Conv model") atomic_convnet = atomic_conv.AtomicConvModel( atomic_convnet = atomic_conv.AtomicConvModel( n_tasks=1, n_tasks=1, batch_size=batch_size, batch_size=batch_size, layer_sizes=[10], frag1_num_atoms=frag1_num_atoms, frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms) complex_num_atoms=complex_num_atoms) Loading @@ -146,3 +158,6 @@ def test_atomic_conv_with_feat(): print("About to call fit") print("About to call fit") # Run a fitting operation # Run a fitting operation atomic_convnet.fit(dataset) atomic_convnet.fit(dataset) preds = atomic_convnet.predict(dataset) assert preds.shape == (1, 1, 1) assert np.count_nonzero(preds) > 0 docs/source/api_reference/featurizers.rst +3 −3 Original line number Original line Diff line number Diff line Loading @@ -229,10 +229,10 @@ RdkitGridFeaturizer :members: :members: :inherited-members: :inherited-members: AtomConvFeaturizer AtomicConvFeaturizer ^^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^ .. autoclass:: deepchem.feat.NeighborListComplexAtomicCoordinates .. autoclass:: deepchem.feat.AtomicConvFeaturizer :members: :members: :inherited-members: :inherited-members: Loading Loading
deepchem/models/atomic_conv.py +3 −113 Original line number Original line Diff line number Diff line Loading @@ -17,102 +17,6 @@ from deepchem.utils.typing import KerasActivationFn, LossFn, OneOrMany logger = logging.getLogger(__name__) logger = logging.getLogger(__name__) # class AtomicConvScore(Layer): # """The scoring function used by the atomic convolution models.""" # def __init__(self, atom_types, layer_sizes, **kwargs): # super(AtomicConvScore, self).__init__(**kwargs) # self.atom_types = atom_types # self.layer_sizes = layer_sizes # def build(self, input_shape): # self.type_weights = [] # self.type_biases = [] # self.output_weights = [] # self.output_biases = [] # n_features = int(input_shape[0][-1]) # layer_sizes = self.layer_sizes # num_layers = len(layer_sizes) # weight_init_stddevs = [1 / np.sqrt(x) for x in layer_sizes] # bias_init_consts = [0.0] * num_layers # for ind, atomtype in enumerate(self.atom_types): # prev_layer_size = n_features # self.type_weights.append([]) # self.type_biases.append([]) # self.output_weights.append([]) # self.output_biases.append([]) # for i in range(num_layers): # weight, bias = initializeWeightsBiases( # prev_layer_size=prev_layer_size, # size=layer_sizes[i], # weights=tf.random.truncated_normal( # shape=[prev_layer_size, layer_sizes[i]], # stddev=weight_init_stddevs[i]), # biases=tf.constant( # value=bias_init_consts[i], shape=[layer_sizes[i]])) # self.type_weights[ind].append(weight) # self.type_biases[ind].append(bias) # prev_layer_size = layer_sizes[i] # weight, bias = initializeWeightsBiases(prev_layer_size, 1) # self.output_weights[ind].append(weight) # self.output_biases[ind].append(bias) # def call(self, inputs): # frag1_layer, frag2_layer, complex_layer, frag1_z, frag2_z, complex_z = inputs # atom_types = self.atom_types # num_layers = len(self.layer_sizes) # def atomnet(current_input, atomtype): # prev_layer = current_input # for i in range(num_layers): # #layer = tf.nn.bias_add( # # tf.matmul(prev_layer, self.type_weights[atomtype][i]), # # self.type_biases[atomtype][i]) # #layer = tf.nn.relu(layer) # layer = Dense(100)(prev_layer) # prev_layer = layer # #output_layer = tf.squeeze( # # tf.nn.bias_add( # # tf.matmul(prev_layer, self.output_weights[atomtype][0]), # # self.output_biases[atomtype][0])) # print("self.output_weights[atomtype][0].shape") # print(self.output_weights[atomtype][0].shape) # output_layer = Dense( # self.output_weights[atomtype][0].shape[0])(prev_layer) # return output_layer # frag1_zeros = tf.zeros_like(frag1_z, dtype=tf.float32) # frag2_zeros = tf.zeros_like(frag2_z, dtype=tf.float32) # complex_zeros = tf.zeros_like(complex_z, dtype=tf.float32) # frag1_atomtype_energy = [] # frag2_atomtype_energy = [] # complex_atomtype_energy = [] # for ind, atomtype in enumerate(atom_types): # frag1_outputs = tf.map_fn(lambda x: atomnet(x, ind), frag1_layer) # frag2_outputs = tf.map_fn(lambda x: atomnet(x, ind), frag2_layer) # complex_outputs = tf.map_fn(lambda x: atomnet(x, ind), complex_layer) # cond = tf.equal(frag1_z, atomtype) # frag1_atomtype_energy.append(tf.where(cond, frag1_outputs, frag1_zeros)) # cond = tf.equal(frag2_z, atomtype) # frag2_atomtype_energy.append(tf.where(cond, frag2_outputs, frag2_zeros)) # cond = tf.equal(complex_z, atomtype) # complex_atomtype_energy.append( # tf.where(cond, complex_outputs, complex_zeros)) # frag1_outputs = tf.add_n(frag1_atomtype_energy) # frag2_outputs = tf.add_n(frag2_atomtype_energy) # complex_outputs = tf.add_n(complex_atomtype_energy) # frag1_energy = tf.reduce_sum(frag1_outputs, 1) # frag2_energy = tf.reduce_sum(frag2_outputs, 1) # complex_energy = tf.reduce_sum(complex_outputs, 1) # binding_energy = complex_energy - (frag1_energy + frag2_energy) # return tf.expand_dims(binding_energy, axis=1) class AtomicConvModel(KerasModel): class AtomicConvModel(KerasModel): """Implements an Atomic Convolution Model. """Implements an Atomic Convolution Model. Loading Loading @@ -203,7 +107,7 @@ class AtomicConvModel(KerasModel): learning_rate: float learning_rate: float Learning rate for the model. Learning rate for the model. """ """ # TODO: Turning off queue for now. Safe to re-activate? self.complex_num_atoms = complex_num_atoms self.complex_num_atoms = complex_num_atoms self.frag1_num_atoms = frag1_num_atoms self.frag1_num_atoms = frag1_num_atoms self.frag2_num_atoms = frag2_num_atoms self.frag2_num_atoms = frag2_num_atoms Loading Loading @@ -260,16 +164,11 @@ class AtomicConvModel(KerasModel): regularizer = tf.keras.regularizers.l2(weight_decay_penalty) regularizer = tf.keras.regularizers.l2(weight_decay_penalty) else: else: regularizer = None regularizer = None # score = AtomicConvScore(self.atom_types, layer_sizes)([ # self._frag1_conv, self._frag2_conv, self._complex_conv, frag1_z, # frag2_z, complex_z # ]) # print("score") # print(score) prev_layer = concat prev_layer = concat # dropout_switch = Input(shape=tuple()) prev_size = concat.shape[0] prev_size = concat.shape[0] next_activation = None next_activation = None # Add the dense layers # Add the dense layers for size, weight_stddev, bias_const, dropout, activation_fn in zip( for size, weight_stddev, bias_const, dropout, activation_fn in zip( Loading @@ -278,9 +177,6 @@ class AtomicConvModel(KerasModel): layer = prev_layer layer = prev_layer if next_activation is not None: if next_activation is not None: layer = Activation(next_activation)(layer) layer = Activation(next_activation)(layer) print("size") print(size) # layer = Dense(100)(layer) layer = Dense( layer = Dense( size, size, kernel_initializer=tf.keras.initializers.TruncatedNormal( kernel_initializer=tf.keras.initializers.TruncatedNormal( Loading @@ -305,11 +201,6 @@ class AtomicConvModel(KerasModel): bias_initializer=tf.constant_initializer( bias_initializer=tf.constant_initializer( value=bias_init_consts[-1]))(prev_layer)) value=bias_init_consts[-1]))(prev_layer)) loss: Union[dc.models.losses.Loss, LossFn] loss: Union[dc.models.losses.Loss, LossFn] # prev_layer = Dense(100)(prev_layer) # output = Dense(1)(prev_layer) # print("output") # print(output) # loss = dc.models.losses.L2Loss() model = tf.keras.Model( model = tf.keras.Model( inputs=[ inputs=[ Loading @@ -317,7 +208,6 @@ class AtomicConvModel(KerasModel): frag2_nbrs_z, frag2_z, complex_X, complex_nbrs, complex_nbrs_z, frag2_nbrs_z, frag2_z, complex_X, complex_nbrs, complex_nbrs_z, complex_z complex_z ], ], # outputs=score) outputs=output) outputs=output) super(AtomicConvModel, self).__init__( super(AtomicConvModel, self).__init__( model, L2Loss(), batch_size=batch_size, **kwargs) model, L2Loss(), batch_size=batch_size, **kwargs) Loading
deepchem/models/tests/test_atomic_conv.py +36 −21 Original line number Original line Diff line number Diff line Loading @@ -6,14 +6,26 @@ import os import pytest import pytest import deepchem import numpy as np import tensorflow as tf import unittest import numpy as np import numpy as np from deepchem.models import atomic_conv from deepchem.models import atomic_conv from deepchem.data import NumpyDataset from deepchem.data import NumpyDataset from deepchem.feat import ComplexNeighborListFragmentAtomicCoordinates from deepchem.feat import AtomicConvFeaturizer def test_atomic_conv_initialize(): """Quick test of AtomicConv.""" acm = atomic_conv.AtomicConvModel( n_tasks=1, batch_size=1, layer_sizes=[ 1, ], frag1_num_atoms=5, frag2_num_atoms=5, complex_num_atoms=10) assert acm.complex_num_atoms == 10 assert len(acm.atom_types) == 15 @pytest.mark.slow @pytest.mark.slow Loading @@ -26,6 +38,7 @@ def test_atomic_conv(): atomic_convnet = atomic_conv.AtomicConvModel( atomic_convnet = atomic_conv.AtomicConvModel( n_tasks=1, n_tasks=1, batch_size=batch_size, batch_size=batch_size, layer_sizes=[10], frag1_num_atoms=5, frag1_num_atoms=5, frag2_num_atoms=5, frag2_num_atoms=5, complex_num_atoms=10, complex_num_atoms=10, Loading @@ -40,7 +53,6 @@ def test_atomic_conv(): frag1_z = np.random.randint(10, size=(N_atoms)) frag1_z = np.random.randint(10, size=(N_atoms)) frag2_coords = np.random.rand(N_atoms, 3) frag2_coords = np.random.rand(N_atoms, 3) frag2_nbr_list = {0: [], 1: [], 2: [], 3: [], 4: []} frag2_nbr_list = {0: [], 1: [], 2: [], 3: [], 4: []} #frag2_z = np.random.rand(N_atoms, 3) frag2_z = np.random.randint(10, size=(N_atoms)) frag2_z = np.random.randint(10, size=(N_atoms)) system_coords = np.random.rand(2 * N_atoms, 3) system_coords = np.random.rand(2 * N_atoms, 3) system_nbr_list = { system_nbr_list = { Loading @@ -63,20 +75,13 @@ def test_atomic_conv(): features = np.asarray(features) features = np.asarray(features) labels = np.random.rand(batch_size) labels = np.random.rand(batch_size) train = NumpyDataset(features, labels) train = NumpyDataset(features, labels) #atomic_convnet.fit(train, nb_epoch=300) atomic_convnet.fit(train, nb_epoch=150) atomic_convnet.fit(train, nb_epoch=150) print("labels") print(labels) print("atomic_convnet.predict(train)") print(atomic_convnet.predict(train)) assert np.allclose(labels, atomic_convnet.predict(train), atol=0.01) assert np.allclose(labels, atomic_convnet.predict(train), atol=0.01) @pytest.mark.slow @pytest.mark.slow def test_atomic_conv_variable(): def test_atomic_conv_variable(): """A simple test that initializes and fits an AtomicConvModel on variable input size.""" """A simple test that initializes and fits an AtomicConvModel on variable input size.""" # For simplicity, let's assume both molecules have same number of # atoms. frag1_num_atoms = 1000 frag1_num_atoms = 1000 frag2_num_atoms = 1200 frag2_num_atoms = 1200 complex_num_atoms = frag1_num_atoms + frag2_num_atoms complex_num_atoms = frag1_num_atoms + frag2_num_atoms Loading @@ -84,6 +89,9 @@ def test_atomic_conv_variable(): atomic_convnet = atomic_conv.AtomicConvModel( atomic_convnet = atomic_conv.AtomicConvModel( n_tasks=1, n_tasks=1, batch_size=batch_size, batch_size=batch_size, layer_sizes=[ 10, ], frag1_num_atoms=frag1_num_atoms, frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms) complex_num_atoms=complex_num_atoms) Loading @@ -108,6 +116,9 @@ def test_atomic_conv_variable(): labels = np.zeros(batch_size) labels = np.zeros(batch_size) train = NumpyDataset(features, labels) train = NumpyDataset(features, labels) atomic_convnet.fit(train, nb_epoch=1) atomic_convnet.fit(train, nb_epoch=1) preds = atomic_convnet.predict(train) assert preds.shape == (1, 1, 1) assert np.count_nonzero(preds) > 0 @pytest.mark.slow @pytest.mark.slow Loading @@ -117,28 +128,29 @@ def test_atomic_conv_with_feat(): ligand_file = os.path.join(dir_path, ligand_file = os.path.join(dir_path, "../../feat/tests/data/3zso_ligand_hyd.pdb") "../../feat/tests/data/3zso_ligand_hyd.pdb") protein_file = os.path.join(dir_path, protein_file = os.path.join(dir_path, "../../feat/tests/data/3zso_protein.pdb") "../../feat/tests/data/3zso_protein_noH.pdb") # Pulled from PDB files. For larger datasets with more PDBs, would use # Pulled from PDB files. For larger datasets with more PDBs, would use # max num atoms instead of exact. # max num atoms instead of exact. frag1_num_atoms = 44 # for ligand atoms frag1_num_atoms = 44 # for ligand atoms frag2_num_atoms = 2336 # for protein atoms frag2_num_atoms = 2334 # for protein atoms complex_num_atoms = 2380 # in total complex_num_atoms = 2378 # in total max_num_neighbors = 4 max_num_neighbors = 4 # Cutoff in angstroms # Cutoff in angstroms neighbor_cutoff = 4 neighbor_cutoff = 4 complex_featurizer = ComplexNeighborListFragmentAtomicCoordinates( complex_featurizer = AtomicConvFeaturizer(frag1_num_atoms, frag2_num_atoms, frag1_num_atoms, frag2_num_atoms, complex_num_atoms, max_num_neighbors, complex_num_atoms, neighbor_cutoff) max_num_neighbors, neighbor_cutoff) # arbitrary label # arbitrary label labels = np.array([0]) labels = np.array([0]) features, _ = complex_featurizer.featurize([ligand_file], [protein_file]) features = complex_featurizer.featurize([(ligand_file, protein_file)]) dataset = deepchem.data.DiskDataset.from_numpy(features, labels) dataset = NumpyDataset(features, labels) batch_size = 1 batch_size = 1 print("Constructing Atomic Conv model") print("Constructing Atomic Conv model") atomic_convnet = atomic_conv.AtomicConvModel( atomic_convnet = atomic_conv.AtomicConvModel( n_tasks=1, n_tasks=1, batch_size=batch_size, batch_size=batch_size, layer_sizes=[10], frag1_num_atoms=frag1_num_atoms, frag1_num_atoms=frag1_num_atoms, frag2_num_atoms=frag2_num_atoms, frag2_num_atoms=frag2_num_atoms, complex_num_atoms=complex_num_atoms) complex_num_atoms=complex_num_atoms) Loading @@ -146,3 +158,6 @@ def test_atomic_conv_with_feat(): print("About to call fit") print("About to call fit") # Run a fitting operation # Run a fitting operation atomic_convnet.fit(dataset) atomic_convnet.fit(dataset) preds = atomic_convnet.predict(dataset) assert preds.shape == (1, 1, 1) assert np.count_nonzero(preds) > 0
docs/source/api_reference/featurizers.rst +3 −3 Original line number Original line Diff line number Diff line Loading @@ -229,10 +229,10 @@ RdkitGridFeaturizer :members: :members: :inherited-members: :inherited-members: AtomConvFeaturizer AtomicConvFeaturizer ^^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^ .. autoclass:: deepchem.feat.NeighborListComplexAtomicCoordinates .. autoclass:: deepchem.feat.AtomicConvFeaturizer :members: :members: :inherited-members: :inherited-members: Loading
