def test_gat_regression():
# load datasets
featurizer = MolGraphConvFeaturizer()
tasks, dataset, transformers, metric = get_dataset('regression',
featurizer=featurizer)
# initialize models
n_tasks = len(tasks)
model = GATModel(mode='regression',
n_tasks=n_tasks,
number_atom_features=30,
batch_size=10,
learning_rate=0.001)
# overfit test
model.fit(dataset, nb_epoch=500)
scores = model.evaluate(dataset, [metric], transformers)
assert scores['mean_absolute_error'] < 0.5
# test on a small MoleculeNet dataset
from deepchem.molnet import load_delaney
tasks, all_dataset, transformers = load_delaney(featurizer=featurizer)
train_set, _, _ = all_dataset
model = dc.models.GATModel(mode='regression',
n_tasks=len(tasks),
graph_attention_layers=[2],
n_attention_heads=1,
residual=False,
predictor_hidden_feats=2)
model.fit(train_set, nb_epoch=1)
def test_mpnn_regression():
# load datasets
featurizer = MolGraphConvFeaturizer(use_edges=True)
tasks, dataset, transformers, metric = get_dataset('regression',
featurizer=featurizer)
# initialize models
n_tasks = len(tasks)
model = MPNNModel(mode='regression', n_tasks=n_tasks, learning_rate=0.0005)
# overfit test
model.fit(dataset, nb_epoch=400)
scores = model.evaluate(dataset, [metric], transformers)
assert scores['mean_absolute_error'] < 0.5
# test on a small MoleculeNet dataset
from deepchem.molnet import load_delaney
tasks, all_dataset, transformers = load_delaney(featurizer=featurizer)
train_set, _, _ = all_dataset
model = MPNNModel(mode='regression',
n_tasks=len(tasks),
node_out_feats=2,
edge_hidden_feats=2,
num_step_message_passing=1,
num_step_set2set=1,
num_layer_set2set=1)
model.fit(train_set, nb_epoch=1)
def get_dataset(mode='classification', featurizer='GraphConv', num_tasks=2):
data_points = 20
if mode == 'classification':
tasks, all_dataset, transformers = load_bace_classification(
featurizer, reload=False)
else:
tasks, all_dataset, transformers = load_delaney(featurizer,
reload=False)
train, valid, test = all_dataset
for i in range(1, num_tasks):
tasks.append("random_task")
w = np.ones(shape=(data_points, len(tasks)))
if mode == 'classification':
y = np.random.randint(0, 2, size=(data_points, len(tasks)))
metric = dc.metrics.Metric(dc.metrics.roc_auc_score,
np.mean,
mode="classification")
else:
y = np.random.normal(size=(data_points, len(tasks)))
metric = dc.metrics.Metric(dc.metrics.mean_absolute_error,
mode="regression")
ds = NumpyDataset(train.X[:data_points], y, w, train.ids[:data_points])
return tasks, ds, transformers, metric
def get_dataset(self,
mode='classification',
featurizer='GraphConv',
num_tasks=2):
data_points = 10
if mode == 'classification':
tasks, all_dataset, transformers = load_bace_classification(featurizer)
else:
tasks, all_dataset, transformers = load_delaney(featurizer)
train, valid, test = all_dataset
for i in range(1, num_tasks):
tasks.append("random_task")
w = np.ones(shape=(data_points, len(tasks)))
if mode == 'classification':
y = np.random.randint(0, 2, size=(data_points, len(tasks)))
metric = dc.metrics.Metric(
dc.metrics.roc_auc_score, np.mean, mode="classification")
else:
y = np.random.normal(size=(data_points, len(tasks)))
metric = dc.metrics.Metric(
dc.metrics.mean_absolute_error, mode="regression")
ds = NumpyDataset(train.X[:data_points], y, w, train.ids[:data_points])
return tasks, ds, transformers, metric
def test_attentivefp_regression():
# load datasets
featurizer = MolGraphConvFeaturizer(use_edges=True)
tasks, dataset, transformers, metric = get_dataset(
'regression', featurizer=featurizer)
# initialize models
n_tasks = len(tasks)
model = AttentiveFPModel(mode='regression', n_tasks=n_tasks, batch_size=10)
# overfit test
model.fit(dataset, nb_epoch=100)
scores = model.evaluate(dataset, [metric], transformers)
assert scores['mean_absolute_error'] < 0.5
# test on a small MoleculeNet dataset
from deepchem.molnet import load_delaney
tasks, all_dataset, transformers = load_delaney(featurizer=featurizer)
train_set, _, _ = all_dataset
model = AttentiveFPModel(
mode='regression',
n_tasks=len(tasks),
num_layers=1,
num_timesteps=1,
graph_feat_size=2)
model.fit(train_set, nb_epoch=1)
def test_featurization_transformer():
fp_size = 2048
tasks, all_dataset, transformers = load_delaney('Raw')
train = all_dataset[0]
transformer = FeaturizationTransformer(
dataset=train, featurizer=dc.feat.CircularFingerprint(size=fp_size))
new_train = transformer.transform(train)
assert new_train.y.shape == train.y.shape
assert new_train.X.shape[-1] == fp_size
def test_featurization_transformer(self):
fp_size = 2048
tasks, all_dataset, transformers = load_delaney('Raw')
train = all_dataset[0]
transformer = FeaturizationTransformer(
transform_X=True,
dataset=train,
featurizer=dc.feat.CircularFingerprint(size=fp_size))
new_train = transformer.transform(train)
self.assertEqual(new_train.y.shape, train.y.shape)
self.assertEqual(new_train.X.shape[-1], fp_size)
def test_featurization_transformer(self):
fp_size = 2048
tasks, all_dataset, transformers = load_delaney('Raw')
train = all_dataset[0]
transformer = FeaturizationTransformer(
transform_X=True,
dataset=train,
featurizer=dc.feat.CircularFingerprint(size=fp_size))
new_train = transformer.transform(train)
self.assertEqual(new_train.y.shape, train.y.shape)
self.assertEqual(new_train.X.shape[-1], fp_size)
def load_dataset(args):
splitter = 'scaffold'
if args['featurizer'] == 'ECFP':
featurizer = 'ECFP'
elif args['featurizer'] == 'GC':
from deepchem.feat import MolGraphConvFeaturizer
featurizer = MolGraphConvFeaturizer()
if args['dataset'] == 'BACE_classification':
from deepchem.molnet import load_bace_classification
tasks, all_dataset, transformers = load_bace_classification(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'BBBP':
from deepchem.molnet import load_bbbp
tasks, all_dataset, transformers = load_bbbp(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'BACE_regression':
from deepchem.molnet import load_bace_regression
tasks, all_dataset, transformers = load_bace_regression(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'ClinTox':
from deepchem.molnet import load_clintox
tasks, all_dataset, transformers = load_clintox(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'Delaney':
from deepchem.molnet import load_delaney
tasks, all_dataset, transformers = load_delaney(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'HOPV':
from deepchem.molnet import load_hopv
tasks, all_dataset, transformers = load_hopv(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'SIDER':
from deepchem.molnet import load_sider
tasks, all_dataset, transformers = load_sider(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'Lipo':
from deepchem.molnet import load_lipo
tasks, all_dataset, transformers = load_lipo(
featurizer=featurizer, splitter=splitter, reload=False)
else:
raise ValueError('Unexpected dataset: {}'.format(args['dataset']))
return args, tasks, all_dataset, transformers
def test_pagtn_regression():
# load datasets
featurizer = PagtnMolGraphFeaturizer(max_length=5)
tasks, dataset, transformers, metric = get_dataset(
'regression', featurizer=featurizer)
# initialize models
n_tasks = len(tasks)
model = PagtnModel(mode='regression', n_tasks=n_tasks, batch_size=16)
# overfit test
model.fit(dataset, nb_epoch=150)
scores = model.evaluate(dataset, [metric], transformers)
assert scores['mean_absolute_error'] < 0.65
# test on a small MoleculeNet dataset
from deepchem.molnet import load_delaney
tasks, all_dataset, transformers = load_delaney(featurizer=featurizer)
train_set, _, _ = all_dataset
model = PagtnModel(mode='regression', n_tasks=n_tasks, batch_size=16)
model.fit(train_set, nb_epoch=1)
from __future__ import print_function
from __future__ import division
from __future__ import unicode_literals
import os
import shutil
import numpy as np
import deepchem as dc
from deepchem.molnet import load_delaney
# Only for debug!
np.random.seed(123)
# Load Delaney dataset
n_features = 1024
delaney_tasks, delaney_datasets, transformers = load_delaney()
train_dataset, valid_dataset, test_dataset = delaney_datasets
# Fit models
metric = dc.metrics.Metric(dc.metrics.pearson_r2_score, np.mean)
model = dc.models.MultitaskRegressor(
len(delaney_tasks),
n_features,
layer_sizes=[1000],
dropouts=[.25],
learning_rate=0.001,
batch_size=50,
verbosity="high")
# Fit trained model
from __future__ import division
from __future__ import unicode_literals
import numpy as np
from deepchem.models import GraphConvModel
np.random.seed(123)
import tensorflow as tf
tf.set_random_seed(123)
import deepchem as dc
from deepchem.molnet import load_delaney
# Load Delaney dataset
delaney_tasks, delaney_datasets, transformers = load_delaney(
featurizer='GraphConv', split='index')
train_dataset, valid_dataset, test_dataset = delaney_datasets
# Fit models
metric = dc.metrics.Metric(dc.metrics.pearson_r2_score, np.mean)
# Do setup required for tf/keras models
# Number of features on conv-mols
n_feat = 75
# Batch size of models
batch_size = 128
model = GraphConvModel(
len(delaney_tasks), batch_size=batch_size, mode='regression')
# Fit trained model
model.fit(train_dataset, nb_epoch=20)
def load_dataset(args):
splitter = 'scaffold'
if args['featurizer'] == 'ECFP':
featurizer = 'ECFP'
elif args['featurizer'] == 'GC':
from deepchem.feat import MolGraphConvFeaturizer
featurizer = MolGraphConvFeaturizer()
elif args['featurizer'] == 'AC':
from deepchem.feat import AtomicConvFeaturizer
featurizer = AtomicConvFeaturizer(frag1_num_atoms=100,
frag2_num_atoms=1000,
complex_num_atoms=1100,
max_num_neighbors=12,
neighbor_cutoff=4)
if args['dataset'] == 'BACE_classification':
from deepchem.molnet import load_bace_classification
tasks, all_dataset, transformers = load_bace_classification(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'BBBP':
from deepchem.molnet import load_bbbp
tasks, all_dataset, transformers = load_bbbp(featurizer=featurizer,
splitter=splitter,
reload=False)
elif args['dataset'] == 'BACE_regression':
from deepchem.molnet import load_bace_regression
tasks, all_dataset, transformers = load_bace_regression(
featurizer=featurizer, splitter=splitter, reload=False)
elif args['dataset'] == 'ClinTox':
from deepchem.molnet import load_clintox
tasks, all_dataset, transformers = load_clintox(featurizer=featurizer,
splitter=splitter,
reload=False)
elif args['dataset'] == 'Delaney':
from deepchem.molnet import load_delaney
tasks, all_dataset, transformers = load_delaney(featurizer=featurizer,
splitter=splitter,
reload=False)
elif args['dataset'] == 'HOPV':
from deepchem.molnet import load_hopv
tasks, all_dataset, transformers = load_hopv(featurizer=featurizer,
splitter=splitter,
reload=False)
elif args['dataset'] == 'SIDER':
from deepchem.molnet import load_sider
tasks, all_dataset, transformers = load_sider(featurizer=featurizer,
splitter=splitter,
reload=False)
elif args['dataset'] == 'Lipo':
from deepchem.molnet import load_lipo
tasks, all_dataset, transformers = load_lipo(featurizer=featurizer,
splitter=splitter,
reload=False)
elif args['dataset'] == 'PDBbind':
from deepchem.molnet import load_pdbbind
tasks, all_dataset, transformers = load_pdbbind(
featurizer=featurizer,
save_dir='.',
data_dir='.',
splitter='random',
pocket=True,
set_name='core', # refined
reload=False)
else:
raise ValueError('Unexpected dataset: {}'.format(args['dataset']))
return args, tasks, all_dataset, transformers