Loading .keras.jsondeleted 100644 → 0 +0 −6 Original line number Diff line number Diff line { "image_dim_ordering": "tf", "epsilon": 1e-07, "floatx": "float32", "backend": "tensorflow" } .travis.yml +1 −1 Original line number Diff line number Diff line Loading @@ -18,7 +18,7 @@ install: - source activate deepchem - pip install yapf==0.16.0 - python setup.py install - cp .keras.json ~/ - KERAS_BACKEND=tensorflow script: - nosetests -v deepchem --nologcapture - find ./deepchem | grep .py$ |xargs python -m doctest -v Loading deepchem/feat/one_hot.py +89 −2 Original line number Diff line number Diff line Loading @@ -15,35 +15,110 @@ class OneHotFeaturizer(Featurizer): """ def __init__(self, charset, padlength=120): """ Parameters ---------- charset: obj:`list` of obj:`str` Each string is length 1 padlength: int length to pad the smile strings to """ self.charset = charset self.pad_length = padlength def featurize(self, mols, verbose=True, log_every_n=1000): """ Parameters ---------- mols: obj List of rdkit Molecule Objects verbose: bool How much logging log_every_n: How often to log Returns ------- obj numpy array of features """ smiles = [Chem.MolToSmiles(mol) for mol in mols] if self.charset is None: self.charset = self._create_charset(mols) return np.array([self.one_hot_encoded(smile) for smile in smiles]) def one_hot_array(self, i): """ Create a one hot array with bit i set to 1 Parameters ---------- i: int bit to set to 1 Returns ------- obj:`list` of obj:`int` length len(self.charset) """ return [int(x) for x in [ix == i for ix in range(len(self.charset))]] def one_hot_index(self, c): """ TODO(LESWING) replace with map lookup vs linear scan :param charset: :return: Parameters ---------- c character whose index we want Returns ------- int index of c in self.charset """ return self.charset.index(c) def pad_smile(self, smile): """ Pad A Smile String to self.pad_length Parameters ---------- smile: str Returns ------- str smile string space padded to self.pad_length """ return smile.ljust(self.pad_length) def one_hot_encoded(self, smile): """ One Hot Encode an entire SMILE string Parameters ---------- smile: str smile string to encode Returns ------- object np.array of one hot encoded arrays for each character in smile """ return np.array([ self.one_hot_array(self.one_hot_index(x)) for x in self.pad_smile(smile) ]) def untransform(self, z): """ Convert from one hot representation back to SMILE Parameters ---------- z: obj:`list` list of one hot encoded features Returns ------- Smile Strings picking MAX for each one hot encoded array """ z1 = [] for i in range(len(z)): s = "" Loading @@ -54,6 +129,18 @@ class OneHotFeaturizer(Featurizer): return z1 def _create_charset(self, smiles): """ create the charset from smiles Parameters ---------- smiles: obj:`list` of obj:`str` list of smile strings Returns ------- obj:`list` of obj:`str` List of length one strings that are characters in smiles. No duplicates """ s = set() for smile in smiles: s.union(list(smile)) Loading deepchem/models/autoencoder_models/autoencoder.py +38 −10 Original line number Diff line number Diff line Loading @@ -19,11 +19,19 @@ class TensorflowMoleculeEncoder(Model): charset_length=len(zinc_charset), latent_rep_size=292): """ :param model_dir: Folder to store cached weights :weights_file: File to store cached weights in model_dir :param verbose: True for more logging :param charset_length: Length of one_hot_encoded vectors :param latent_rep_size: How large a 1D Vector for latent representation Parameters ---------- model_dir: str Folder to store cached weights weights_file: str File to store cached weights in model_dir verbose: bool True for more logging charset_length: int Length of one_hot_encoded vectors latent_rep_size: int How large a 1D Vector for latent representation """ super(TensorflowMoleculeEncoder, self).__init__( model_dir=model_dir, verbose=verbose) Loading @@ -38,6 +46,12 @@ class TensorflowMoleculeEncoder(Model): @staticmethod def zinc_encoder(): """ Returns ------- obj An Encoder with weights that were trained on the zinc dataset """ current_dir = os.path.dirname(os.path.realpath(__file__)) weights_filename = "zinc_model.h5" weights_file = os.path.join(current_dir, weights_filename) Loading Loading @@ -67,11 +81,19 @@ class TensorflowMoleculeDecoder(Model): charset_length=len(zinc_charset), latent_rep_size=292): """ :param model_dir: Folder To Store Cached Weights :weights_file: File to store cached weights in model_dir :param verbose: True for more logging :param charset_length: Length of one_hot_encoded vectors :param latent_rep_size: How large a 1D Vector for latent representation Parameters ---------- model_dir: str Folder to store cached weights weights_file: str File to store cached weights in model_dir verbose: bool True for more logging charset_length: int Length of one_hot_encoded vectors latent_rep_size: int How large a 1D Vector for latent representation """ super(TensorflowMoleculeDecoder, self).__init__( model_dir=model_dir, verbose=verbose) Loading @@ -89,6 +111,12 @@ class TensorflowMoleculeDecoder(Model): @staticmethod def zinc_decoder(): """ Returns ------- obj A Decoder with weights that were trained on the zinc dataset """ current_dir = os.path.dirname(os.path.realpath(__file__)) weights_filename = "zinc_model.h5" weights_file = os.path.join(current_dir, weights_filename) Loading deepchem/models/autoencoder_models/test_tensorflowEncoders.py +2 −4 Original line number Diff line number Diff line Loading @@ -12,9 +12,7 @@ from models.autoencoder_models.autoencoder import TensorflowMoleculeEncoder, Ten class TestTensorflowEncoders(TestCase): def test_fit(self): data_dir = "/home/leswing/Documents/data_sets/keras-molecule" tf_enc = TensorflowMoleculeEncoder(model_dir=data_dir) tf_enc = TensorflowMoleculeEncoder.zinc_encoder() smiles = [ "Cn1cnc2c1c(=O)n(C)c(=O)n2C", "O=C(O)[C@@H]1/C(=C/CO)O[C@@H]2CC(=O)N21", Loading @@ -33,7 +31,7 @@ class TestTensorflowEncoders(TestCase): dataset = DiskDataset.from_numpy(features, features) prediction = tf_enc.predict_on_batch(dataset.X) tf_de = TensorflowMoleculeDecoder(model_dir=data_dir) tf_de = TensorflowMoleculeDecoder.zinc_decoder() one_hot_decoded = tf_de.predict_on_batch(prediction) decoded_smiles = featurizer.untransform(one_hot_decoded) assert_equals(len(decoded_smiles), len(smiles)) Loading
.keras.jsondeleted 100644 → 0 +0 −6 Original line number Diff line number Diff line { "image_dim_ordering": "tf", "epsilon": 1e-07, "floatx": "float32", "backend": "tensorflow" }
.travis.yml +1 −1 Original line number Diff line number Diff line Loading @@ -18,7 +18,7 @@ install: - source activate deepchem - pip install yapf==0.16.0 - python setup.py install - cp .keras.json ~/ - KERAS_BACKEND=tensorflow script: - nosetests -v deepchem --nologcapture - find ./deepchem | grep .py$ |xargs python -m doctest -v Loading
deepchem/feat/one_hot.py +89 −2 Original line number Diff line number Diff line Loading @@ -15,35 +15,110 @@ class OneHotFeaturizer(Featurizer): """ def __init__(self, charset, padlength=120): """ Parameters ---------- charset: obj:`list` of obj:`str` Each string is length 1 padlength: int length to pad the smile strings to """ self.charset = charset self.pad_length = padlength def featurize(self, mols, verbose=True, log_every_n=1000): """ Parameters ---------- mols: obj List of rdkit Molecule Objects verbose: bool How much logging log_every_n: How often to log Returns ------- obj numpy array of features """ smiles = [Chem.MolToSmiles(mol) for mol in mols] if self.charset is None: self.charset = self._create_charset(mols) return np.array([self.one_hot_encoded(smile) for smile in smiles]) def one_hot_array(self, i): """ Create a one hot array with bit i set to 1 Parameters ---------- i: int bit to set to 1 Returns ------- obj:`list` of obj:`int` length len(self.charset) """ return [int(x) for x in [ix == i for ix in range(len(self.charset))]] def one_hot_index(self, c): """ TODO(LESWING) replace with map lookup vs linear scan :param charset: :return: Parameters ---------- c character whose index we want Returns ------- int index of c in self.charset """ return self.charset.index(c) def pad_smile(self, smile): """ Pad A Smile String to self.pad_length Parameters ---------- smile: str Returns ------- str smile string space padded to self.pad_length """ return smile.ljust(self.pad_length) def one_hot_encoded(self, smile): """ One Hot Encode an entire SMILE string Parameters ---------- smile: str smile string to encode Returns ------- object np.array of one hot encoded arrays for each character in smile """ return np.array([ self.one_hot_array(self.one_hot_index(x)) for x in self.pad_smile(smile) ]) def untransform(self, z): """ Convert from one hot representation back to SMILE Parameters ---------- z: obj:`list` list of one hot encoded features Returns ------- Smile Strings picking MAX for each one hot encoded array """ z1 = [] for i in range(len(z)): s = "" Loading @@ -54,6 +129,18 @@ class OneHotFeaturizer(Featurizer): return z1 def _create_charset(self, smiles): """ create the charset from smiles Parameters ---------- smiles: obj:`list` of obj:`str` list of smile strings Returns ------- obj:`list` of obj:`str` List of length one strings that are characters in smiles. No duplicates """ s = set() for smile in smiles: s.union(list(smile)) Loading
deepchem/models/autoencoder_models/autoencoder.py +38 −10 Original line number Diff line number Diff line Loading @@ -19,11 +19,19 @@ class TensorflowMoleculeEncoder(Model): charset_length=len(zinc_charset), latent_rep_size=292): """ :param model_dir: Folder to store cached weights :weights_file: File to store cached weights in model_dir :param verbose: True for more logging :param charset_length: Length of one_hot_encoded vectors :param latent_rep_size: How large a 1D Vector for latent representation Parameters ---------- model_dir: str Folder to store cached weights weights_file: str File to store cached weights in model_dir verbose: bool True for more logging charset_length: int Length of one_hot_encoded vectors latent_rep_size: int How large a 1D Vector for latent representation """ super(TensorflowMoleculeEncoder, self).__init__( model_dir=model_dir, verbose=verbose) Loading @@ -38,6 +46,12 @@ class TensorflowMoleculeEncoder(Model): @staticmethod def zinc_encoder(): """ Returns ------- obj An Encoder with weights that were trained on the zinc dataset """ current_dir = os.path.dirname(os.path.realpath(__file__)) weights_filename = "zinc_model.h5" weights_file = os.path.join(current_dir, weights_filename) Loading Loading @@ -67,11 +81,19 @@ class TensorflowMoleculeDecoder(Model): charset_length=len(zinc_charset), latent_rep_size=292): """ :param model_dir: Folder To Store Cached Weights :weights_file: File to store cached weights in model_dir :param verbose: True for more logging :param charset_length: Length of one_hot_encoded vectors :param latent_rep_size: How large a 1D Vector for latent representation Parameters ---------- model_dir: str Folder to store cached weights weights_file: str File to store cached weights in model_dir verbose: bool True for more logging charset_length: int Length of one_hot_encoded vectors latent_rep_size: int How large a 1D Vector for latent representation """ super(TensorflowMoleculeDecoder, self).__init__( model_dir=model_dir, verbose=verbose) Loading @@ -89,6 +111,12 @@ class TensorflowMoleculeDecoder(Model): @staticmethod def zinc_decoder(): """ Returns ------- obj A Decoder with weights that were trained on the zinc dataset """ current_dir = os.path.dirname(os.path.realpath(__file__)) weights_filename = "zinc_model.h5" weights_file = os.path.join(current_dir, weights_filename) Loading
deepchem/models/autoencoder_models/test_tensorflowEncoders.py +2 −4 Original line number Diff line number Diff line Loading @@ -12,9 +12,7 @@ from models.autoencoder_models.autoencoder import TensorflowMoleculeEncoder, Ten class TestTensorflowEncoders(TestCase): def test_fit(self): data_dir = "/home/leswing/Documents/data_sets/keras-molecule" tf_enc = TensorflowMoleculeEncoder(model_dir=data_dir) tf_enc = TensorflowMoleculeEncoder.zinc_encoder() smiles = [ "Cn1cnc2c1c(=O)n(C)c(=O)n2C", "O=C(O)[C@@H]1/C(=C/CO)O[C@@H]2CC(=O)N21", Loading @@ -33,7 +31,7 @@ class TestTensorflowEncoders(TestCase): dataset = DiskDataset.from_numpy(features, features) prediction = tf_enc.predict_on_batch(dataset.X) tf_de = TensorflowMoleculeDecoder(model_dir=data_dir) tf_de = TensorflowMoleculeDecoder.zinc_decoder() one_hot_decoded = tf_de.predict_on_batch(prediction) decoded_smiles = featurizer.untransform(one_hot_decoded) assert_equals(len(decoded_smiles), len(smiles))
