Loading deepchem/trans/transformers.py +64 −0 Original line number Diff line number Diff line Loading @@ -568,3 +568,67 @@ class CoulombFitTransformer(): raise NotImplementedError( "Cannot untransform datasets with FitTransformer.") class IRVTransformer(): """Performs randomization and binarization operations on batches of Coulomb Matrix features during fit.""" def __init__(self, K, dataset): """Initializes CoulombFitTransformer. Parameters: ---------- dataset: dc.data.Dataset object K: number of nearest neighbours that count """ self.X = dataset.X self.n_samples = X.shape[0] self.K = K self.y = dataset.y self.w = dataset.w @staticmethod def similarity(x, y): if x.shape != y.shape: raise ValueError( "Similarity measurement requires same size") return np.sum(np.min([x,y], axis=0))/np.sum(np.max([x,y], axis=0)) def realize(self, X_target): """Randomize features. """ def _realize_(x_target): similar_xs = [] similar_inds = [] threshold = 0 for count, x in enumerate(self.X): similar = IRVTransformer.similarity(x, x_target) if similar >= 1: continue if len(similar_xs) < self.K-1: similar_xs.append(similar) similar_inds.append(count) if len(similar_xs) == self.K-1: if self.w[count] similar_xs.append(similar) similar_inds.append(count) threshold = min(similar_xs) elif similar > threshold: ind = similar_xs.index(threshold) similar_xs.pop(ind) similar_inds.pop(ind) similar_xs.append(similar) similar_inds.append(count) threshold = min(similar_xs) return similar_xs.extend(similar_inds) return np.array([_realize_(z) for z in X_target]) def X_transform(self, X_target): X_target = self.realize(X_target) return X_target def transform(self, dataset): raise NotImplementedError( "Cannot transform datasets with FitTransformer") def untransform(self, z): raise NotImplementedError( "Cannot untransform datasets with FitTransformer.") Loading
deepchem/trans/transformers.py +64 −0 Original line number Diff line number Diff line Loading @@ -568,3 +568,67 @@ class CoulombFitTransformer(): raise NotImplementedError( "Cannot untransform datasets with FitTransformer.") class IRVTransformer(): """Performs randomization and binarization operations on batches of Coulomb Matrix features during fit.""" def __init__(self, K, dataset): """Initializes CoulombFitTransformer. Parameters: ---------- dataset: dc.data.Dataset object K: number of nearest neighbours that count """ self.X = dataset.X self.n_samples = X.shape[0] self.K = K self.y = dataset.y self.w = dataset.w @staticmethod def similarity(x, y): if x.shape != y.shape: raise ValueError( "Similarity measurement requires same size") return np.sum(np.min([x,y], axis=0))/np.sum(np.max([x,y], axis=0)) def realize(self, X_target): """Randomize features. """ def _realize_(x_target): similar_xs = [] similar_inds = [] threshold = 0 for count, x in enumerate(self.X): similar = IRVTransformer.similarity(x, x_target) if similar >= 1: continue if len(similar_xs) < self.K-1: similar_xs.append(similar) similar_inds.append(count) if len(similar_xs) == self.K-1: if self.w[count] similar_xs.append(similar) similar_inds.append(count) threshold = min(similar_xs) elif similar > threshold: ind = similar_xs.index(threshold) similar_xs.pop(ind) similar_inds.pop(ind) similar_xs.append(similar) similar_inds.append(count) threshold = min(similar_xs) return similar_xs.extend(similar_inds) return np.array([_realize_(z) for z in X_target]) def X_transform(self, X_target): X_target = self.realize(X_target) return X_target def transform(self, dataset): raise NotImplementedError( "Cannot transform datasets with FitTransformer") def untransform(self, z): raise NotImplementedError( "Cannot untransform datasets with FitTransformer.")
