Merge pull request #370 from CCXD/dev (8b004c53) · Commits · 钟慕尧 / deepchem

examples/chembl/chembl_datasets.py

+3 −13

Original line number	Diff line number	Diff line
		@@ -67,8 +67,9 @@ def load_chembl(shard_size=2000, featurizer="ECFP", set="5thresh", split="random
		transformers = [
		dc.trans.NormalizationTransformer(transform_y=True, dataset=train_dataset)]
		for transformer in transformers:
		for dataset in [train_dataset, valid_dataset, test_dataset]:
		transformer.transform(dataset)
		train = transformer.transform(train_dataset)
		valid = transformer.transform(valid_dataset)
		test = transformer.transform(test_dataset)
		else:
		transformers = [
		dc.trans.NormalizationTransformer(transform_y=True, dataset=dataset)]
		@@ -82,17 +83,6 @@ def load_chembl(shard_size=2000, featurizer="ECFP", set="5thresh", split="random
		splitter = splitters[split]
		print("Performing new split.")
		train, valid, test = splitter.train_valid_test_split(dataset)
		elif split == "year":
		print("Featurizing train datasets")
		train = loader.featurize(
		train_files, shard_size=shard_size)

		print("Featurizing valid datasets")
		valid = loader.featurize(
		valid_files, shard_size=shard_size)

		print("Featurizing test datasets")
		test = loader.featurize(
		test_files, shard_size=shard_size)

		return chembl_tasks, (train, valid, test), transformers

+3 −4

Original line number	Diff line number	Diff line
		@@ -78,13 +78,12 @@ def load_kaggle(shard_size=2000, featurizer=None):
		dc.trans.NormalizationTransformer(transform_y=True,
		dataset=train_dataset)]

		# TODO(rbharath): Is this a bug in the Kaggle data transformation?
		for transformer in transformers:
		print("Performing transformations with %s"
		% transformer.__class__.__name__)
		for dataset in [train_dataset, valid_dataset, test_dataset]:
		print("Transforming dataset")
		transformer.transform(dataset)
		train_dataset = transformer.transform(train_dataset)
		valid_dataset = transformer.transform(valid_dataset)
		test_dataset = transformer.transform(test_dataset)

		print("Shuffling order of train dataset.")
		train_dataset.sparse_shuffle()