Loading deepchem/models/tf_keras_models/__init__.py 0 → 100644 +0 −0 Empty file added. deepchem/models/tf_keras_models/containers.py 0 → 100644 +68 −0 Original line number Diff line number Diff line """Borrows Keras Container Abstraction for Models """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals from keras.engine.topology import Container class GraphContainer(Container): def __init__(self, sess, input, output, **kwargs): # Remove the keywords for the GraphModel self.graph_topology = kwargs.pop('graph_topology') kwargs["input"] = input kwargs["output"] = output super(GraphContainer, self).__init__(**kwargs) self.sess = sess # Save Tensorflow session self.op_mode_objects = {} def get_graph_topology(self): return self.graph_topology def get(self): return self.graph_topology.get_inputs() def get_batch_size(self): return self.graph_topology.get_batch_size() def get_num_output_features(self): # Gets the output shape of the featurization layers of the network return self.layers[-1].output_shape[1] def get_output(self): return self.output class SupportGraphContainer(Container): def __init__(self, sess, **kwargs): # Remove the keywords for the GraphModel self.graph_topology_test = kwargs.pop("graph_topology_test") self.graph_topology_support = kwargs.pop("graph_topology_support") super(SupportGraphContainer, self).__init__(**kwargs) self.sess = sess # Save Tensorflow session self.op_mode_objects = {} def get_test(self): return self.graph_topology_test.get_inputs() def get_support(self): return self.graph_topology_support.get_inputs() def get_test_batch_size(self): return self.graph_topology_test.get_batch_size() def get_support_batch_size(self): return self.graph_topology_support.get_batch_size() def get_featurization_dim(self): # Gets the output shape of the featurization layers of the network return self.layers[-1].output_shape[1] def get_test_output(self): return self.output[0] def get_support_output(self): return self.output[1] deepchem/models/tf_keras_models/graph_topology.py 0 → 100644 +147 −0 Original line number Diff line number Diff line """Manages Placeholders for Graph convolution networks. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals from keras.layers import Input from keras import backend as K from data_structures import ConvMol def merge_two_dicts(x, y): z = x.copy() z.update(y) return z def merge_dicts(l): """Convenience function to merge list of dictionaries.""" merged = {} for dict in l: merged = merge_two_dicts(merged, dict) return merged class GraphTopology(object): """Manages placeholders associated with batch of graphs and their topology""" def __init__(self, n_atoms, n_feat, batch_size, name='topology', max_deg=6, min_deg=0): """ Parameters ---------- n_atoms: int Number of atoms (max) in graphs. n_feat: int Number of features per atom. batch_size: int Number of molecules per batch. name: str, optional Name of this manager. max_deg: int, optional Maximum #bonds for atoms in molecules. min_deg: int, optional Minimum #bonds for atoms in molecules. """ self.n_atoms = n_atoms self.n_feat = n_feat self.batch_size = batch_size self.name = name self.max_deg = max_deg self.min_deg = min_deg self.init_keras_placeholders() def init_keras_placeholders(self): self.nodes_placeholder = Input( tensor=K.placeholder( shape=(None, self.n_feat), dtype='float32', name=self.name+'_nodes')) self.deg_adj_lists_placeholders = [ Input(tensor=K.placeholder( shape=(None, deg), dtype='int32', name=self.name+'_deg_adj'+str(deg))) for deg in range(1, self.max_deg+1)] self.deg_slice_placeholder = Input( tensor=K.placeholder( shape=(self.max_deg-self.min_deg+1,2), name="deg_slice", dtype='int32'), name=self.name+'_deg_slice') self.membership_placeholder = Input( tensor=K.placeholder(shape=(None,), dtype='int32', name="membership"), name=self.name+'_membership') # Define the list of tensors to be used as topology self.topology = [self.deg_slice_placeholder, self.membership_placeholder] self.topology.extend(self.deg_adj_lists_placeholders) self.inputs = [self.nodes_placeholder] self.inputs.extend(self.topology) def get_inputs(self): return self.inputs def get_topology(self): return self.topology def get_batch_size(self): return self.batch_size def get_nodes(self): return self.nodes_placeholder def get_deg_adj_lists(self): return self.deg_adj_lists_placeholders def get_deg_slice(self): return self.deg_slice_placeholder def get_membership(self): return self.membership_placeholder # TODO(rbharath): It's still not clear to me that this should live alone like # this... It's awkward to separate out part of the batch construction this # way. def batch_to_feed_dict(self, batch): """Converts the current batch into a feed_dict used by tensorflow. Assigns the graph information in batch to the placeholders tensors params ------ batch : np.ndarray Array of ConvMol objects returns ------- feed_dict : dict Can be merged with other feed_dicts for input into tensorflow """ # Merge mol conv objects batch = ConvMol.agglomerate_mols(batch) atoms = batch.nodes deg_adj_lists = [batch.deg_adj_lists[deg] for deg in range(1, self.max_deg+1)] # Generate dicts deg_adj_dict = dict(zip(self.deg_adj_lists_placeholders, deg_adj_lists)) atoms_dict = {self.nodes_placeholder : atoms, self.deg_slice_placeholder : batch.deg_slice, self.membership_placeholder : batch.membership} return merge_dicts([atoms_dict, deg_adj_dict]) def extract_topology(x): # Extracts the topology tensors from x topology = x[1::] # Extract parsed topology information deg_slice = topology[0] membership = topology[1] deg_adj_lists = topology[2::] return deg_slice, membership, deg_adj_lists def extract_nodes(x): # Extracts the nodes from x (just the first tensor in the list of tensors) return x[0] def extract_membership(x): return x[2] deepchem/models/tf_keras_models/tests/__init__.py 0 → 100644 +0 −0 Empty file added. deepchem/models/tf_keras_models/tests/test_containers.py 0 → 100644 +28 −0 Original line number Diff line number Diff line """ Simple Tests for Container objects. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals import unittest from deepchem.models.tf_keras_models.graph_topology import GraphTopology from tensorflow.python.framework import test_util class TestContainers(test_util.TensorFlowTestCase): """ Test Container usage." """ def setUp(self): super(TestContainers, self).setUp() self.root = '/tmp' def test_graph_containers(self): """Simple test that Graph Container can be initialized.""" n_atoms = 5 n_atom_feat = 10 batch_size = 3 with self.test_session() as sess: topology = GraphTopology(n_atoms, n_atom_feat, batch_size) container = GraphContainer(sess, input=topology.get_inputs(), output=topology.get_nodes()) Loading
deepchem/models/tf_keras_models/containers.py 0 → 100644 +68 −0 Original line number Diff line number Diff line """Borrows Keras Container Abstraction for Models """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals from keras.engine.topology import Container class GraphContainer(Container): def __init__(self, sess, input, output, **kwargs): # Remove the keywords for the GraphModel self.graph_topology = kwargs.pop('graph_topology') kwargs["input"] = input kwargs["output"] = output super(GraphContainer, self).__init__(**kwargs) self.sess = sess # Save Tensorflow session self.op_mode_objects = {} def get_graph_topology(self): return self.graph_topology def get(self): return self.graph_topology.get_inputs() def get_batch_size(self): return self.graph_topology.get_batch_size() def get_num_output_features(self): # Gets the output shape of the featurization layers of the network return self.layers[-1].output_shape[1] def get_output(self): return self.output class SupportGraphContainer(Container): def __init__(self, sess, **kwargs): # Remove the keywords for the GraphModel self.graph_topology_test = kwargs.pop("graph_topology_test") self.graph_topology_support = kwargs.pop("graph_topology_support") super(SupportGraphContainer, self).__init__(**kwargs) self.sess = sess # Save Tensorflow session self.op_mode_objects = {} def get_test(self): return self.graph_topology_test.get_inputs() def get_support(self): return self.graph_topology_support.get_inputs() def get_test_batch_size(self): return self.graph_topology_test.get_batch_size() def get_support_batch_size(self): return self.graph_topology_support.get_batch_size() def get_featurization_dim(self): # Gets the output shape of the featurization layers of the network return self.layers[-1].output_shape[1] def get_test_output(self): return self.output[0] def get_support_output(self): return self.output[1]
deepchem/models/tf_keras_models/graph_topology.py 0 → 100644 +147 −0 Original line number Diff line number Diff line """Manages Placeholders for Graph convolution networks. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals from keras.layers import Input from keras import backend as K from data_structures import ConvMol def merge_two_dicts(x, y): z = x.copy() z.update(y) return z def merge_dicts(l): """Convenience function to merge list of dictionaries.""" merged = {} for dict in l: merged = merge_two_dicts(merged, dict) return merged class GraphTopology(object): """Manages placeholders associated with batch of graphs and their topology""" def __init__(self, n_atoms, n_feat, batch_size, name='topology', max_deg=6, min_deg=0): """ Parameters ---------- n_atoms: int Number of atoms (max) in graphs. n_feat: int Number of features per atom. batch_size: int Number of molecules per batch. name: str, optional Name of this manager. max_deg: int, optional Maximum #bonds for atoms in molecules. min_deg: int, optional Minimum #bonds for atoms in molecules. """ self.n_atoms = n_atoms self.n_feat = n_feat self.batch_size = batch_size self.name = name self.max_deg = max_deg self.min_deg = min_deg self.init_keras_placeholders() def init_keras_placeholders(self): self.nodes_placeholder = Input( tensor=K.placeholder( shape=(None, self.n_feat), dtype='float32', name=self.name+'_nodes')) self.deg_adj_lists_placeholders = [ Input(tensor=K.placeholder( shape=(None, deg), dtype='int32', name=self.name+'_deg_adj'+str(deg))) for deg in range(1, self.max_deg+1)] self.deg_slice_placeholder = Input( tensor=K.placeholder( shape=(self.max_deg-self.min_deg+1,2), name="deg_slice", dtype='int32'), name=self.name+'_deg_slice') self.membership_placeholder = Input( tensor=K.placeholder(shape=(None,), dtype='int32', name="membership"), name=self.name+'_membership') # Define the list of tensors to be used as topology self.topology = [self.deg_slice_placeholder, self.membership_placeholder] self.topology.extend(self.deg_adj_lists_placeholders) self.inputs = [self.nodes_placeholder] self.inputs.extend(self.topology) def get_inputs(self): return self.inputs def get_topology(self): return self.topology def get_batch_size(self): return self.batch_size def get_nodes(self): return self.nodes_placeholder def get_deg_adj_lists(self): return self.deg_adj_lists_placeholders def get_deg_slice(self): return self.deg_slice_placeholder def get_membership(self): return self.membership_placeholder # TODO(rbharath): It's still not clear to me that this should live alone like # this... It's awkward to separate out part of the batch construction this # way. def batch_to_feed_dict(self, batch): """Converts the current batch into a feed_dict used by tensorflow. Assigns the graph information in batch to the placeholders tensors params ------ batch : np.ndarray Array of ConvMol objects returns ------- feed_dict : dict Can be merged with other feed_dicts for input into tensorflow """ # Merge mol conv objects batch = ConvMol.agglomerate_mols(batch) atoms = batch.nodes deg_adj_lists = [batch.deg_adj_lists[deg] for deg in range(1, self.max_deg+1)] # Generate dicts deg_adj_dict = dict(zip(self.deg_adj_lists_placeholders, deg_adj_lists)) atoms_dict = {self.nodes_placeholder : atoms, self.deg_slice_placeholder : batch.deg_slice, self.membership_placeholder : batch.membership} return merge_dicts([atoms_dict, deg_adj_dict]) def extract_topology(x): # Extracts the topology tensors from x topology = x[1::] # Extract parsed topology information deg_slice = topology[0] membership = topology[1] deg_adj_lists = topology[2::] return deg_slice, membership, deg_adj_lists def extract_nodes(x): # Extracts the nodes from x (just the first tensor in the list of tensors) return x[0] def extract_membership(x): return x[2]
deepchem/models/tf_keras_models/tests/test_containers.py 0 → 100644 +28 −0 Original line number Diff line number Diff line """ Simple Tests for Container objects. """ from __future__ import print_function from __future__ import division from __future__ import unicode_literals import unittest from deepchem.models.tf_keras_models.graph_topology import GraphTopology from tensorflow.python.framework import test_util class TestContainers(test_util.TensorFlowTestCase): """ Test Container usage." """ def setUp(self): super(TestContainers, self).setUp() self.root = '/tmp' def test_graph_containers(self): """Simple test that Graph Container can be initialized.""" n_atoms = 5 n_atom_feat = 10 batch_size = 3 with self.test_session() as sess: topology = GraphTopology(n_atoms, n_atom_feat, batch_size) container = GraphContainer(sess, input=topology.get_inputs(), output=topology.get_nodes())
