Minor formatting changes (45009705) · Commits · 钟慕尧 / deepchem

deepchem/feat/molecule_featurizers/one_hot_featurizer.py

+15 −9

Original line number	Diff line number	Diff line
		@@ -49,9 +49,11 @@ class OneHotFeaturizer(MolecularFeaturizer):
		log_every_n: int = 1000) -> np.ndarray:
		datapoints = list(datapoints)
		if (len(datapoints) < 1):
		print("No datapoints are present in the parameter Iterable, so we return an empty array.")
		print(
		"No datapoints are present in the parameter Iterable, so we return an empty array."
		)
		return np.array([])
		ic(type(datapoints[0]))

		# Featurize Mol data
		if (type(datapoints[0]) == RDKitMol): # Mol
		return MolecularFeaturizer.featurize(self, datapoints, log_every_n)
		@@ -59,7 +61,9 @@ class OneHotFeaturizer(MolecularFeaturizer):
		elif (type(datapoints[0]) == str): # String
		return Featurizer.featurize(self, datapoints, log_every_n)
		else:
		print("One hot featurizer only supports strings and mols at this time, so returning {}")
		print(
		"One hot featurizer only supports strings and mols at this time, so returning {}"
		)
		return np.array([])

		def _featurize(self, datapoint: Any):
		@@ -69,7 +73,9 @@ class OneHotFeaturizer(MolecularFeaturizer):
		elif (type(datapoint) == RDKitMol):
		return self._featurizeMol(datapoint)
		else:
		print("One hot featurizer only supports strings and mols at this time, so returning {}")
		print(
		"One hot featurizer only supports strings and mols at this time, so returning {}"
		)
		return np.array([])

		def _featurizeString(self, string: str) -> np.ndarray:

+2 −8

Original line number	Diff line number	Diff line
		@@ -4,7 +4,7 @@ import numpy as np

		from deepchem.feat import OneHotFeaturizer
		from deepchem.feat.molecule_featurizers.one_hot_featurizer import ZINC_CHARSET
		from icecream import ic


		class TestOneHotFeaturizert(unittest.TestCase):
		"""
		@@ -20,9 +20,6 @@ class TestOneHotFeaturizert(unittest.TestCase):
		length = len(charset) + 1
		featurizer = OneHotFeaturizer(charset)
		feature = featurizer([string]) # Implicit call to featurize()
		ic(len(feature))
		ic(len(feature[0]))
		ic(len(feature[0][0]))
		assert feature.shape == (1, 100, length)
		# untransform
		undo_string = featurizer.untransform(feature[0])
		@@ -47,14 +44,11 @@ class TestOneHotFeaturizert(unittest.TestCase):
		"""
		Test one hot encoding with max_length.
		"""
		string = "abcdefghijklmnopqrstuvwxyz"
		string = "abcdefghijklmnopqrstuvwxyzvewqmc"
		charset = "abcdefghijklmnopqrstuvwxyz"
		length = len(charset) + 1
		featurizer = OneHotFeaturizer(charset, max_length=120)
		feature = featurizer([string])
		ic(len(feature))
		ic(len(feature[0]))
		ic(len(feature[0][0]))
		assert feature.shape == (1, 120, length)
		# untranform
		undo_string = featurizer.untransform(feature[0])