Most recent version of code (4744652d) · Commits · 钟慕尧 / deepchem

.travis.yml

+1 −0

Original line number	Diff line number	Diff line
		@@ -22,6 +22,7 @@ install:
		- conda install six
		- conda install dill
		- conda install runipy
		- conda install ipyparallel
		- pip install runipy
		- pip install nglview
		- conda install -c omnia mdtraj

deepchem/featurizers/nnscore.py

+7 −0

Original line number	Diff line number	Diff line
		@@ -714,32 +714,39 @@ class NNScoreComplexFeaturizer(ComplexFeaturizer):
		"""
		Compute Binana fingerprint for complex.
		"""
		print("In _featurize_complex")
		mol_pdb_file = tempfile.NamedTemporaryFile(suffix="pdb")
		with open(mol_pdb_file.name, "w") as mol_f:
		mol_f.writelines(mol_pdb)
		protein_pdb_file = tempfile.NamedTemporaryFile(suffix="pdb")
		with open(protein_pdb_file.name, "w") as protein_f:
		protein_f.writelines(protein_pdb)
		print("Written temp pdb files")

		mol_hyd_file = tempfile.NamedTemporaryFile(suffix="pdb")
		mol_pdbqt_file = tempfile.NamedTemporaryFile(suffix="pdbqt")
		hydrogenate_and_compute_partial_charges(
		mol_pdb_file.name, "pdb", mol_hyd_file.name,
		mol_pdbqt_file.name)
		print("Hydrogenated mol file")

		protein_hyd_file = tempfile.NamedTemporaryFile(suffix="pdb")
		protein_pdbqt_file = tempfile.NamedTemporaryFile(suffix="pdbqt")
		hydrogenate_and_compute_partial_charges(
		protein_pdb_file.name, "pdb", protein_hyd_file.name,
		protein_pdbqt_file.name)
		print("Hydrogenated protein file")

		mol_pdb_obj = PDB()
		mol_pdb_obj.load_from_files(mol_pdb_file.name, mol_pdbqt_file.name)
		print("Loaded mol pdb object")

		protein_pdb_obj = PDB()
		protein_pdb_obj.load_from_files(
		protein_pdb_file.name, protein_pdbqt_file.name)
		print("Loaded protein pdb object")

		features = self.binana.compute_input_vector(mol_pdb_obj, protein_pdb_obj)
		print("Computed binana features.")

		return features

deepchem/featurizers/tests/test_nnscore.py

+2 −0

Original line number	Diff line number	Diff line
		@@ -19,6 +19,7 @@ class TestNNScoreComplexFeaturizer(unittest.TestCase):
		"""
		Create internal featurizer.
		"""
		print("Doing setup")
		self.nnscore_featurizer = NNScoreComplexFeaturizer()
		### 3zso comes from PDBBind-CN
		_3zso_protein_pdb_file = os.path.join(data_dir(), "3zso_protein.pdb")
		@@ -31,6 +32,7 @@ class TestNNScoreComplexFeaturizer(unittest.TestCase):
		_3zso_ligand_pdb = f.readlines()

		self.test_cases = [("3zso", _3zso_ligand_pdb, _3zso_protein_pdb)]
		print("Done with setup")

		def testNNScore(self):
		"""

deepchem/hyperparameters/init.py

+11 −3

Original line number	Diff line number	Diff line
		@@ -19,7 +19,7 @@ class HyperparamOpt(object):
		self.task_types = task_types

		def hyperparam_search(self, params_dict, train_dataset, valid_dataset,
		output_transformers, metric):
		output_transformers, metric, verbosity=None):
		"""Perform hyperparams search according to params_dict.

		Each key to hyperparams_dict is a model_param. The values should be a list
		@@ -44,7 +44,7 @@ class HyperparamOpt(object):
		model_params[hyperparam] = hyperparam_val

		model_dir = tempfile.mkdtemp()
		model = self.model_class(self.task_types, model_params, verbosity=None)
		model = self.model_class(self.task_types, model_params, verbosity=verbosity)
		model.fit(train_dataset)
		model.save(model_dir)

		@@ -54,6 +54,7 @@ class HyperparamOpt(object):
		valid_score = score.iloc[0][metric]
		print("Model %d/%d, Metric %s, Validation set %s: %f" %
		(ind, number_combinations, metric, ind, valid_score))
		print("\tbest_validation_score so far: %f" % best_validation_score)
		all_scores[hyperparameter_tuple] = valid_score

		if valid_score > best_validation_score:
		@@ -66,6 +67,13 @@ class HyperparamOpt(object):
		else:
		shutil.rmtree(model_dir)

		train_csv_out = tempfile.NamedTemporaryFile()
		train_stats_out = tempfile.NamedTemporaryFile()
		train_evaluator = Evaluator(best_model, train_dataset, output_transformers)
		train_df, train_score = train_evaluator.compute_model_performance(
		train_csv_out, train_stats_out)
		train_score = train_score.iloc[0][metric]
		print("Best hyperparameters: %s" % str(zip(hyperparams, best_hyperparams)))
		print("best_validation_score: %f" % best_validation_score)
		print("train_score: %f" % train_score)
		print("validation_score: %f" % best_validation_score)
		return best_model, best_hyperparams, all_scores

deepchem/models/init.py

+7 −1

Original line number	Diff line number	Diff line
		@@ -127,7 +127,10 @@ class Model(object):
		task_names = dataset.get_task_names()
		pred_task_names = ["%s_pred" % task_name for task_name in task_names]
		w_task_names = ["%s_weight" % task_name for task_name in task_names]
		column_names = (['ids'] + task_names + pred_task_names + w_task_names
		raw_task_names = [task_name+"_raw" for task_name in task_names]
		raw_pred_task_names = [pred_task_name+"_raw" for pred_task_name in pred_task_names]
		column_names = (['ids'] + raw_task_names + task_names
		+ raw_pred_task_names + pred_task_names + w_task_names
		+ ["y_means", "y_stds"])
		pred_y_df = pd.DataFrame(columns=column_names)

		@@ -147,12 +150,15 @@ class Model(object):
		y_pred = np.reshape(y_pred, np.shape(y))

		# Now undo transformations on y, y_pred
		y_raw, y_pred_raw = y, y_pred
		y = undo_transforms(y, transformers)
		y_pred = undo_transforms(y_pred, transformers)

		shard_df = pd.DataFrame(columns=column_names)
		shard_df['ids'] = ids
		shard_df[raw_task_names] = y_raw
		shard_df[task_names] = y
		shard_df[raw_pred_task_names] = y_pred_raw
		shard_df[pred_task_names] = y_pred
		shard_df[w_task_names] = w
		pred_y_df = pd.concat([pred_y_df, shard_df])

Admin message