Loading deepchem/data/datasets.py +17 −3 Original line number Diff line number Diff line Loading @@ -420,6 +420,20 @@ class Dataset(object): """ raise NotImplementedError() def select(self, indices: Sequence[int], select_dir: Optional[str] = None): """Creates a new dataset from a selection of indices from self. Parameters ---------- indices: Sequence List of indices to select. select_dir: str, optional (default None) Path to new directory that the selected indices will be copied to. """ raise NotImplementedError() def get_statistics(self, X_stats: bool = True, y_stats: bool = True) -> Tuple[float, ...]: """Compute and return statistics of this dataset. Loading Loading @@ -1868,13 +1882,13 @@ class DiskDataset(Dataset): tasks=tasks) @staticmethod def merge(datasets: Iterable["DiskDataset"], def merge(datasets: Iterable["Dataset"], merge_dir: Optional[str] = None) -> "DiskDataset": """Merges provided datasets into a merged dataset. Parameters ---------- datasets: Iterable[DiskDataset] datasets: Iterable[Dataset] List of datasets to merge. merge_dir: str, optional (default None) The new directory path to store the merged DiskDataset. Loading @@ -1897,7 +1911,7 @@ class DiskDataset(Dataset): tasks = [] for dataset in datasets: try: tasks.append(dataset.tasks) tasks.append(dataset.tasks) # type: ignore except AttributeError: pass if tasks: Loading deepchem/splits/__init__.py +17 −5 Original line number Diff line number Diff line """ Gathers all splitters in one place for convenient imports """ # TODO(rbharath): Get rid of * import from deepchem.splits.splitters import * from deepchem.splits.splitters import ScaffoldSplitter from deepchem.splits.splitters import SpecifiedSplitter # flake8: noqa # basic splitter from deepchem.splits.splitters import Splitter from deepchem.splits.splitters import RandomSplitter from deepchem.splits.splitters import RandomGroupSplitter from deepchem.splits.splitters import SingletaskStratifiedSplitter from deepchem.splits.splitters import IndexSplitter from deepchem.splits.splitters import IndiceSplitter from deepchem.splits.splitters import RandomGroupSplitter # molecule splitter from deepchem.splits.splitters import ScaffoldSplitter from deepchem.splits.splitters import MolecularWeightSplitter from deepchem.splits.splitters import MaxMinSplitter from deepchem.splits.splitters import FingerprintSplitter from deepchem.splits.splitters import ButinaSplitter # other splitter from deepchem.splits.splitters import TimeSplitterPDBbind from deepchem.splits.task_splitter import merge_fold_datasets from deepchem.splits.task_splitter import TaskSplitter deepchem/splits/splitters.py +140 −130 Original line number Diff line number Diff line Loading @@ -6,14 +6,13 @@ import random import tempfile import itertools import logging from typing import List, Optional, Sequence, Tuple from typing import Any, Dict, List, Iterator, Optional, Sequence, Tuple, Union import numpy as np import pandas as pd import deepchem as dc from deepchem.data import Dataset, DiskDataset from deepchem.utils.save import load_data logger = logging.getLogger(__name__) Loading Loading @@ -76,6 +75,7 @@ class Splitter(object): else: rem_dataset = DiskDataset.from_numpy(dataset.X, dataset.y, dataset.w, dataset.ids) for fold in range(k): # Note starts as 1/k since fold starts at 0. Ends at 1 since fold goes up # to k-1. Loading @@ -89,16 +89,17 @@ class Splitter(object): **kwargs) cv_dataset = rem_dataset.select(fold_inds, select_dir=cv_dir) cv_datasets.append(cv_dataset) rem_dataset = rem_dataset.select(rem_inds) # FIXME: Incompatible types in assignment (expression has type "Dataset", variable has type "DiskDataset") rem_dataset = rem_dataset.select(rem_inds) # type: ignore train_ds_to_merge = filter(lambda x: x is not None, [train_ds_base, rem_dataset]) train_ds_to_merge: Iterator[Dataset] = filter( None, [train_ds_base, rem_dataset]) train_ds_to_merge = filter(lambda x: len(x) > 0, train_ds_to_merge) train_dataset = DiskDataset.merge(train_ds_to_merge, merge_dir=train_dir) train_datasets.append(train_dataset) update_train_base_merge = filter(lambda x: x is not None, [train_ds_base, cv_dataset]) update_train_base_merge: Iterator[Dataset] = filter( None, [train_ds_base, cv_dataset]) train_ds_base = DiskDataset.merge(update_train_base_merge) return list(zip(train_datasets, cv_datasets)) Loading Loading @@ -159,8 +160,7 @@ class Splitter(object): frac_test=frac_test, frac_valid=frac_valid, seed=seed, log_every_n=log_every_n, **kwargs) log_every_n=log_every_n) if train_dir is None: train_dir = tempfile.mkdtemp() if valid_dir is None: Loading Loading @@ -322,7 +322,7 @@ class RandomGroupSplitter(Splitter): caution if the number of elements per group varies significantly. """ def __init__(self, groups: Sequence, *args, **kwargs): def __init__(self, groups: Sequence): """Initialize this object. Parameters Loading @@ -342,7 +342,6 @@ class RandomGroupSplitter(Splitter): dataset.X : 0 1 2 3 4 5 6 7 8 """ self.groups = groups super(RandomGroupSplitter, self).__init__(*args, **kwargs) def split(self, dataset: Dataset, Loading Loading @@ -384,7 +383,7 @@ class RandomGroupSplitter(Splitter): # dict is needed in case groups aren't strictly flattened or # hashed by something non-integer like group_dict = {} group_dict: Dict[Any, List[int]] = {} for idx, g in enumerate(self.groups): if g not in group_dict: group_dict[g] = [] Loading Loading @@ -418,25 +417,21 @@ class RandomStratifiedSplitter(Splitter): For sparse multitask datasets, a standard split offers no guarantees that the splits will have any activate compounds. This class guarantees that each task will have a proportional split of the activates in a split. TO do this, a ragged split is performed with different numbers split. To do this, a ragged split is performed with different numbers of compounds taken from each task. Thus, the length of the split arrays may exceed the split of the original array. That said, no datapoint is copied to more than one split, so correctness is still ensured. Note that this splitter is only valid for boolean label data. TODO(rbharath): This splitter should be refactored to match style of other splitter classes. """ def __generate_required_hits(self, w, frac_split): # returns list of per column sum of non zero elements required_hits = (w != 0).sum(axis=0) for col_hits in required_hits: col_hits = int(frac_split * col_hits) return required_hits Notes ----- This splitter is only valid for boolean label data. """ def get_task_split_indices(self, y, w, frac_split): def get_task_split_indices(self, y: np.ndarray, w: np.ndarray, frac_split: float) -> List[int]: """Returns num datapoints needed per task to split properly.""" w_present = (w != 0) y_present = y * w_present Loading @@ -462,7 +457,12 @@ class RandomStratifiedSplitter(Splitter): # TODO(rbharath): Refactor this split method to match API of other # splits (or potentially refactor those to match this). def split(self, dataset, frac_split, split_dirs=None): def split( # type: ignore [override] self, dataset: Dataset, frac_split: float, split_dirs: Optional[List[str]] = None ) -> Tuple[Dataset, Optional[Dataset]]: """ Method that does bulk of splitting dataset. """ Loading Loading @@ -503,16 +503,20 @@ class RandomStratifiedSplitter(Splitter): return dataset_1, dataset_2 def train_valid_test_split(self, dataset, train_dir=None, valid_dir=None, test_dir=None, frac_train=.8, frac_valid=.1, frac_test=.1, seed=None, log_every_n=1000): # FIXME: Signature of "train_valid_test_split" incompatible with supertype "Splitter" def train_valid_test_split( # type: ignore [override] self, dataset: Dataset, train_dir: Optional[str] = None, valid_dir: Optional[str] = None, test_dir: Optional[str] = None, frac_train: float = 0.8, frac_valid: float = 0.1, frac_test: float = 0.1, seed: Optional[int] = None, log_every_n: int = 1000, **kwargs) -> Union[Tuple[Dataset, None, None], Tuple[Dataset, Dataset, Optional[Dataset]]]: """ Splits self into train/validation/test sets. Most splitters use the superclass implementation Loading @@ -521,20 +525,18 @@ class RandomStratifiedSplitter(Splitter): Parameters ---------- dataset: data like object. Dataset to be split. This should either be of type `dc.data.Dataset` or a type that `dc.utils.data.datasetify` can convert into a `Dataset`. train_dir: str, optional dataset: Dataset Dataset to be split. train_dir: str, optional (default None) If specified, the directory in which the generated training dataset should be stored. This is only considered if `isinstance(dataset, dc.data.DiskDataset)` valid_dir: str, optional valid_dir: str, optional (default None) If specified, the directory in which the generated valid dataset should be stored. This is only considered if `isinstance(dataset, dc.data.DiskDataset)` is True. test_dir: str, optional test_dir: str, optional (default None) If specified, the directory in which the generated test dataset should be stored. This is only considered if `isinstance(dataset, dc.data.DiskDataset)` Loading @@ -547,13 +549,15 @@ class RandomStratifiedSplitter(Splitter): The fraction of data to be used for the test split. seed: int, optional (default None) Random seed to use. log_every_n: int, optional log_every_n: int, optional (default 1000) Controls the logger by dictating how often logger outputs will be produced. Returns ------- Train and test datasets as dc.data.Dataset objects. Tuple[Dataset, Optional[Dataset], Optional[Dataset]] A tuple of train, valid and test datasets as dc.data.Dataset objects. In some cases, valid or test dataset is None. """ if train_dir is None: train_dir = tempfile.mkdtemp() Loading @@ -571,13 +575,38 @@ class RandomStratifiedSplitter(Splitter): else: return train_dataset, None, None # split remaining data into valid and test, treating sub test set also as sparse valid_dataset, test_dataset = self.split(rem_dataset, valid_percentage, # FIXME: Argument 1 to "split" of "RandomStratifiedSplitter" has incompatible type # "Optional[Dataset]"; expected "Dataset" valid_dataset, test_dataset = self.split( rem_dataset, # type: ignore valid_percentage, [valid_dir, test_dir]) return train_dataset, valid_dataset, test_dataset def k_fold_split(self, dataset, k, directories=None, **kwargs): """Needs custom implementation due to ragged splits for stratification.""" # FIXME: Signature of "k_fold_split" incompatible with supertype "Splitter" def k_fold_split( # type: ignore [override] self, dataset: Dataset, k: int, directories: Optional[List[str]] = None, **kwargs) -> List[Dataset]: """Needs custom implementation due to ragged splits for stratification. Parameters ---------- dataset: Dataset Dataset to be split. k: int Number of folds to split `dataset` into. directories: List[str], optional (default None) List of length k filepaths to save the result disk-datasets. Returns ------- fold_datasets: List[Dataset] List of dc.data.Dataset objects """ logger.info("Computing K-fold split") if directories is None: directories = [tempfile.mkdtemp() for _ in range(k)] Loading @@ -585,14 +614,18 @@ class RandomStratifiedSplitter(Splitter): assert len(directories) == k fold_datasets = [] # rem_dataset is remaining portion of dataset rem_dataset = dataset rem_dataset: Optional[Dataset] = dataset for fold in range(k): # Note starts as 1/k since fold starts at 0. Ends at 1 since fold goes up # to k-1. frac_fold = 1. / (k - fold) fold_dir = directories[fold] rem_dir = tempfile.mkdtemp() fold_dataset, rem_dataset = self.split(rem_dataset, frac_fold, # FIXME: Argument 1 to "split" of "RandomStratifiedSplitter" has incompatible type # "Optional[Dataset]"; expected "Dataset" fold_dataset, rem_dataset = self.split( rem_dataset, # type: ignore frac_fold, [fold_dir, rem_dir]) fold_datasets.append(fold_dataset) return fold_datasets Loading Loading @@ -625,7 +658,9 @@ class SingletaskStratifiedSplitter(Splitter): """ self.task_number = task_number def k_fold_split(self, # FIXME: Signature of "k_fold_split" incompatible with supertype "Splitter" def k_fold_split( # type: ignore [override] self, dataset: Dataset, k: int, directories: Optional[List[str]] = None, Loading @@ -643,7 +678,7 @@ class SingletaskStratifiedSplitter(Splitter): k: int Number of folds to split `dataset` into. directories: List[str], optional (default None) List of length 2*k filepaths to save the result disk-datasets. List of length k filepaths to save the result disk-datasets. seed: int, optional (default None) Random seed to use. log_every_n: int, optional (default None) Loading Loading @@ -930,8 +965,8 @@ class MolecularWeightSplitter(Splitter): mws = np.array(mws) sortidx = np.argsort(mws) train_cutoff = frac_train * len(sortidx) valid_cutoff = (frac_train + frac_valid) * len(sortidx) train_cutoff = int(frac_train * len(sortidx)) valid_cutoff = int((frac_train + frac_valid) * len(sortidx)) return (sortidx[:train_cutoff], sortidx[train_cutoff:valid_cutoff], sortidx[valid_cutoff:]) Loading Loading @@ -1081,7 +1116,7 @@ class ButinaSplitter(Splitter): log_every_n: int, optional (default None) Log every n examples (not currently used). cutoff: float, optional (default 0.18) The The cutoff value for similarity. Returns ------- Loading @@ -1101,7 +1136,7 @@ class ButinaSplitter(Splitter): except ModuleNotFoundError: raise ValueError("This function requires RDKit to be installed.") print("Performing butina clustering with cutoff of", cutoff) logger.info("Performing butina clustering with cutoff of", cutoff) mols = [] for ind, smiles in enumerate(dataset.ids): mols.append(Chem.MolFromSmiles(smiles)) Loading @@ -1128,8 +1163,8 @@ class ButinaSplitter(Splitter): # TODO (Ytz): for regression tasks we'd stop after only one cluster. active_populations = np.sum(ys[valid_inds], axis=0) if np.all(active_populations): print("# of actives per task in valid:", active_populations) print("Total # of validation points:", len(valid_inds)) logger.info("# of actives per task in valid:", active_populations) logger.info("Total # of validation points:", len(valid_inds)) break train_inds = list(itertools.chain.from_iterable(scaffold_sets[c_idx + 1:])) Loading Loading @@ -1189,7 +1224,8 @@ class ScaffoldSplitter(Splitter): frac_valid: float = 0.1, frac_test: float = 0.1, seed: Optional[int] = None, log_every_n: int = 1000) -> Tuple[List[int], List[int], List[int]]: log_every_n: Optional[int] = 1000 ) -> Tuple[List[int], List[int], List[int]]: """ Splits internal compounds into train/validation/test by scaffold. Loading Loading @@ -1220,7 +1256,9 @@ class ScaffoldSplitter(Splitter): train_cutoff = frac_train * len(dataset) valid_cutoff = (frac_train + frac_valid) * len(dataset) train_inds, valid_inds, test_inds = [], [], [] train_inds: List[int] = [] valid_inds: List[int] = [] test_inds: List[int] = [] logger.info("About to sort in scaffold sets") for scaffold_set in scaffold_sets: Loading Loading @@ -1387,79 +1425,51 @@ class FingerprintSplitter(Splitter): cur_distances[i] = new_dist class SpecifiedSplitter(Splitter): """ Class that splits data according to user specification. """ def __init__(self, input_file, split_field): """Provide input information for splits.""" raw_df = next(load_data([input_file], shard_size=None)) self.splits = raw_df[split_field].values def split(self, dataset, seed=None, frac_train=.8, frac_valid=.1, frac_test=.1, log_every_n=1000): """ Splits internal compounds into train/validation/test by user-specification. """ train_inds, valid_inds, test_inds = [], [], [] for ind, split in enumerate(self.splits): split = split.lower() if split == "train": train_inds.append(ind) elif split in ["valid", "validation"]: valid_inds.append(ind) elif split == "test": test_inds.append(ind) else: raise ValueError("Missing required split information.") return train_inds, valid_inds, test_inds class SpecifiedIndexSplitter(Splitter): """ Class that splits data according to user index specification """ def __init__(self, train_inds, valid_inds, test_inds): """Provide input information for splits.""" self.train_inds = train_inds self.valid_inds = valid_inds self.test_inds = test_inds class TimeSplitterPDBbind(Splitter): def split(self, dataset, seed=None, frac_train=.8, frac_valid=.1, frac_test=.1, log_every_n=1000): def __init__(self, ids, year_file: Optional[str] = None): """ Splits internal compounds into train/validation/test by user-specification. Parameters ---------- ids: .... WIP year_file: str, optional (default None) The file path of .... """ return self.train_inds, self.valid_inds, self.test_inds class TimeSplitterPDBbind(Splitter): def __init__(self, ids, year_file=None): self.ids = ids self.year_file = year_file def split(self, dataset, seed=None, frac_train=.8, frac_valid=.1, frac_test=.1, log_every_n=None): dataset: Dataset, frac_train: float = 0.8, frac_valid: float = 0.1, frac_test: float = 0.1, seed: Optional[int] = None, log_every_n: Optional[int] = None ) -> Tuple[List[int], List[int], List[int]]: """ Splits protein-ligand pairs in PDBbind into train/validation/test in time order. Parameters ---------- dataset: Dataset Dataset to be split. frac_train: float, optional (default 0.8) The fraction of data to be used for the training split. frac_valid: float, optional (default 0.1) The fraction of data to be used for the validation split. frac_test: float, optional (default 0.1) The fraction of data to be used for the test split. seed: int, optional (default None) Random seed to use. log_every_n: int, optional (default None) Log every n examples (not currently used). Returns ------- Tuple[List[int], List[int], List[int]] A tuple of train indices, valid indices, and test indices. Each indices is a list of integers. """ if self.year_file is None: try: Loading deepchem/splits/task_splitter.py +0 −1 Original line number Diff line number Diff line Loading @@ -71,7 +71,6 @@ class TaskSplitter(Splitter): n_tasks = len(dataset.get_task_names()) n_train = int(np.round(frac_train * n_tasks)) n_valid = int(np.round(frac_valid * n_tasks)) n_test = n_tasks - n_train - n_valid X, y, w, ids = dataset.X, dataset.y, dataset.w, dataset.ids Loading deepchem/splits/test_specified_index_splitter.pydeleted 100644 → 0 +0 −28 Original line number Diff line number Diff line from unittest import TestCase import deepchem import numpy as np from sklearn.model_selection import train_test_split from deepchem.splits import SpecifiedIndexSplitter class TestSpecifiedIndexSplitter(TestCase): def create_dataset(self): n_samples, n_features = 20, 10 X = np.random.random(size=(n_samples, n_features)) y = np.random.random(size=(n_samples, 1)) return deepchem.data.NumpyDataset(X, y) def test_split(self): ds = self.create_dataset() indexes = list(range(len(ds))) train, test = train_test_split(indexes) train, valid = train_test_split(train) splitter = SpecifiedIndexSplitter(train, valid, test) train_ds, valid_ds, test_ds = splitter.train_valid_test_split(ds) self.assertTrue(np.all(train_ds.X == ds.X[train])) self.assertTrue(np.all(valid_ds.X == ds.X[valid])) self.assertTrue(np.all(test_ds.X == ds.X[test])) Loading
deepchem/data/datasets.py +17 −3 Original line number Diff line number Diff line Loading @@ -420,6 +420,20 @@ class Dataset(object): """ raise NotImplementedError() def select(self, indices: Sequence[int], select_dir: Optional[str] = None): """Creates a new dataset from a selection of indices from self. Parameters ---------- indices: Sequence List of indices to select. select_dir: str, optional (default None) Path to new directory that the selected indices will be copied to. """ raise NotImplementedError() def get_statistics(self, X_stats: bool = True, y_stats: bool = True) -> Tuple[float, ...]: """Compute and return statistics of this dataset. Loading Loading @@ -1868,13 +1882,13 @@ class DiskDataset(Dataset): tasks=tasks) @staticmethod def merge(datasets: Iterable["DiskDataset"], def merge(datasets: Iterable["Dataset"], merge_dir: Optional[str] = None) -> "DiskDataset": """Merges provided datasets into a merged dataset. Parameters ---------- datasets: Iterable[DiskDataset] datasets: Iterable[Dataset] List of datasets to merge. merge_dir: str, optional (default None) The new directory path to store the merged DiskDataset. Loading @@ -1897,7 +1911,7 @@ class DiskDataset(Dataset): tasks = [] for dataset in datasets: try: tasks.append(dataset.tasks) tasks.append(dataset.tasks) # type: ignore except AttributeError: pass if tasks: Loading
deepchem/splits/__init__.py +17 −5 Original line number Diff line number Diff line """ Gathers all splitters in one place for convenient imports """ # TODO(rbharath): Get rid of * import from deepchem.splits.splitters import * from deepchem.splits.splitters import ScaffoldSplitter from deepchem.splits.splitters import SpecifiedSplitter # flake8: noqa # basic splitter from deepchem.splits.splitters import Splitter from deepchem.splits.splitters import RandomSplitter from deepchem.splits.splitters import RandomGroupSplitter from deepchem.splits.splitters import SingletaskStratifiedSplitter from deepchem.splits.splitters import IndexSplitter from deepchem.splits.splitters import IndiceSplitter from deepchem.splits.splitters import RandomGroupSplitter # molecule splitter from deepchem.splits.splitters import ScaffoldSplitter from deepchem.splits.splitters import MolecularWeightSplitter from deepchem.splits.splitters import MaxMinSplitter from deepchem.splits.splitters import FingerprintSplitter from deepchem.splits.splitters import ButinaSplitter # other splitter from deepchem.splits.splitters import TimeSplitterPDBbind from deepchem.splits.task_splitter import merge_fold_datasets from deepchem.splits.task_splitter import TaskSplitter
deepchem/splits/splitters.py +140 −130 Original line number Diff line number Diff line Loading @@ -6,14 +6,13 @@ import random import tempfile import itertools import logging from typing import List, Optional, Sequence, Tuple from typing import Any, Dict, List, Iterator, Optional, Sequence, Tuple, Union import numpy as np import pandas as pd import deepchem as dc from deepchem.data import Dataset, DiskDataset from deepchem.utils.save import load_data logger = logging.getLogger(__name__) Loading Loading @@ -76,6 +75,7 @@ class Splitter(object): else: rem_dataset = DiskDataset.from_numpy(dataset.X, dataset.y, dataset.w, dataset.ids) for fold in range(k): # Note starts as 1/k since fold starts at 0. Ends at 1 since fold goes up # to k-1. Loading @@ -89,16 +89,17 @@ class Splitter(object): **kwargs) cv_dataset = rem_dataset.select(fold_inds, select_dir=cv_dir) cv_datasets.append(cv_dataset) rem_dataset = rem_dataset.select(rem_inds) # FIXME: Incompatible types in assignment (expression has type "Dataset", variable has type "DiskDataset") rem_dataset = rem_dataset.select(rem_inds) # type: ignore train_ds_to_merge = filter(lambda x: x is not None, [train_ds_base, rem_dataset]) train_ds_to_merge: Iterator[Dataset] = filter( None, [train_ds_base, rem_dataset]) train_ds_to_merge = filter(lambda x: len(x) > 0, train_ds_to_merge) train_dataset = DiskDataset.merge(train_ds_to_merge, merge_dir=train_dir) train_datasets.append(train_dataset) update_train_base_merge = filter(lambda x: x is not None, [train_ds_base, cv_dataset]) update_train_base_merge: Iterator[Dataset] = filter( None, [train_ds_base, cv_dataset]) train_ds_base = DiskDataset.merge(update_train_base_merge) return list(zip(train_datasets, cv_datasets)) Loading Loading @@ -159,8 +160,7 @@ class Splitter(object): frac_test=frac_test, frac_valid=frac_valid, seed=seed, log_every_n=log_every_n, **kwargs) log_every_n=log_every_n) if train_dir is None: train_dir = tempfile.mkdtemp() if valid_dir is None: Loading Loading @@ -322,7 +322,7 @@ class RandomGroupSplitter(Splitter): caution if the number of elements per group varies significantly. """ def __init__(self, groups: Sequence, *args, **kwargs): def __init__(self, groups: Sequence): """Initialize this object. Parameters Loading @@ -342,7 +342,6 @@ class RandomGroupSplitter(Splitter): dataset.X : 0 1 2 3 4 5 6 7 8 """ self.groups = groups super(RandomGroupSplitter, self).__init__(*args, **kwargs) def split(self, dataset: Dataset, Loading Loading @@ -384,7 +383,7 @@ class RandomGroupSplitter(Splitter): # dict is needed in case groups aren't strictly flattened or # hashed by something non-integer like group_dict = {} group_dict: Dict[Any, List[int]] = {} for idx, g in enumerate(self.groups): if g not in group_dict: group_dict[g] = [] Loading Loading @@ -418,25 +417,21 @@ class RandomStratifiedSplitter(Splitter): For sparse multitask datasets, a standard split offers no guarantees that the splits will have any activate compounds. This class guarantees that each task will have a proportional split of the activates in a split. TO do this, a ragged split is performed with different numbers split. To do this, a ragged split is performed with different numbers of compounds taken from each task. Thus, the length of the split arrays may exceed the split of the original array. That said, no datapoint is copied to more than one split, so correctness is still ensured. Note that this splitter is only valid for boolean label data. TODO(rbharath): This splitter should be refactored to match style of other splitter classes. """ def __generate_required_hits(self, w, frac_split): # returns list of per column sum of non zero elements required_hits = (w != 0).sum(axis=0) for col_hits in required_hits: col_hits = int(frac_split * col_hits) return required_hits Notes ----- This splitter is only valid for boolean label data. """ def get_task_split_indices(self, y, w, frac_split): def get_task_split_indices(self, y: np.ndarray, w: np.ndarray, frac_split: float) -> List[int]: """Returns num datapoints needed per task to split properly.""" w_present = (w != 0) y_present = y * w_present Loading @@ -462,7 +457,12 @@ class RandomStratifiedSplitter(Splitter): # TODO(rbharath): Refactor this split method to match API of other # splits (or potentially refactor those to match this). def split(self, dataset, frac_split, split_dirs=None): def split( # type: ignore [override] self, dataset: Dataset, frac_split: float, split_dirs: Optional[List[str]] = None ) -> Tuple[Dataset, Optional[Dataset]]: """ Method that does bulk of splitting dataset. """ Loading Loading @@ -503,16 +503,20 @@ class RandomStratifiedSplitter(Splitter): return dataset_1, dataset_2 def train_valid_test_split(self, dataset, train_dir=None, valid_dir=None, test_dir=None, frac_train=.8, frac_valid=.1, frac_test=.1, seed=None, log_every_n=1000): # FIXME: Signature of "train_valid_test_split" incompatible with supertype "Splitter" def train_valid_test_split( # type: ignore [override] self, dataset: Dataset, train_dir: Optional[str] = None, valid_dir: Optional[str] = None, test_dir: Optional[str] = None, frac_train: float = 0.8, frac_valid: float = 0.1, frac_test: float = 0.1, seed: Optional[int] = None, log_every_n: int = 1000, **kwargs) -> Union[Tuple[Dataset, None, None], Tuple[Dataset, Dataset, Optional[Dataset]]]: """ Splits self into train/validation/test sets. Most splitters use the superclass implementation Loading @@ -521,20 +525,18 @@ class RandomStratifiedSplitter(Splitter): Parameters ---------- dataset: data like object. Dataset to be split. This should either be of type `dc.data.Dataset` or a type that `dc.utils.data.datasetify` can convert into a `Dataset`. train_dir: str, optional dataset: Dataset Dataset to be split. train_dir: str, optional (default None) If specified, the directory in which the generated training dataset should be stored. This is only considered if `isinstance(dataset, dc.data.DiskDataset)` valid_dir: str, optional valid_dir: str, optional (default None) If specified, the directory in which the generated valid dataset should be stored. This is only considered if `isinstance(dataset, dc.data.DiskDataset)` is True. test_dir: str, optional test_dir: str, optional (default None) If specified, the directory in which the generated test dataset should be stored. This is only considered if `isinstance(dataset, dc.data.DiskDataset)` Loading @@ -547,13 +549,15 @@ class RandomStratifiedSplitter(Splitter): The fraction of data to be used for the test split. seed: int, optional (default None) Random seed to use. log_every_n: int, optional log_every_n: int, optional (default 1000) Controls the logger by dictating how often logger outputs will be produced. Returns ------- Train and test datasets as dc.data.Dataset objects. Tuple[Dataset, Optional[Dataset], Optional[Dataset]] A tuple of train, valid and test datasets as dc.data.Dataset objects. In some cases, valid or test dataset is None. """ if train_dir is None: train_dir = tempfile.mkdtemp() Loading @@ -571,13 +575,38 @@ class RandomStratifiedSplitter(Splitter): else: return train_dataset, None, None # split remaining data into valid and test, treating sub test set also as sparse valid_dataset, test_dataset = self.split(rem_dataset, valid_percentage, # FIXME: Argument 1 to "split" of "RandomStratifiedSplitter" has incompatible type # "Optional[Dataset]"; expected "Dataset" valid_dataset, test_dataset = self.split( rem_dataset, # type: ignore valid_percentage, [valid_dir, test_dir]) return train_dataset, valid_dataset, test_dataset def k_fold_split(self, dataset, k, directories=None, **kwargs): """Needs custom implementation due to ragged splits for stratification.""" # FIXME: Signature of "k_fold_split" incompatible with supertype "Splitter" def k_fold_split( # type: ignore [override] self, dataset: Dataset, k: int, directories: Optional[List[str]] = None, **kwargs) -> List[Dataset]: """Needs custom implementation due to ragged splits for stratification. Parameters ---------- dataset: Dataset Dataset to be split. k: int Number of folds to split `dataset` into. directories: List[str], optional (default None) List of length k filepaths to save the result disk-datasets. Returns ------- fold_datasets: List[Dataset] List of dc.data.Dataset objects """ logger.info("Computing K-fold split") if directories is None: directories = [tempfile.mkdtemp() for _ in range(k)] Loading @@ -585,14 +614,18 @@ class RandomStratifiedSplitter(Splitter): assert len(directories) == k fold_datasets = [] # rem_dataset is remaining portion of dataset rem_dataset = dataset rem_dataset: Optional[Dataset] = dataset for fold in range(k): # Note starts as 1/k since fold starts at 0. Ends at 1 since fold goes up # to k-1. frac_fold = 1. / (k - fold) fold_dir = directories[fold] rem_dir = tempfile.mkdtemp() fold_dataset, rem_dataset = self.split(rem_dataset, frac_fold, # FIXME: Argument 1 to "split" of "RandomStratifiedSplitter" has incompatible type # "Optional[Dataset]"; expected "Dataset" fold_dataset, rem_dataset = self.split( rem_dataset, # type: ignore frac_fold, [fold_dir, rem_dir]) fold_datasets.append(fold_dataset) return fold_datasets Loading Loading @@ -625,7 +658,9 @@ class SingletaskStratifiedSplitter(Splitter): """ self.task_number = task_number def k_fold_split(self, # FIXME: Signature of "k_fold_split" incompatible with supertype "Splitter" def k_fold_split( # type: ignore [override] self, dataset: Dataset, k: int, directories: Optional[List[str]] = None, Loading @@ -643,7 +678,7 @@ class SingletaskStratifiedSplitter(Splitter): k: int Number of folds to split `dataset` into. directories: List[str], optional (default None) List of length 2*k filepaths to save the result disk-datasets. List of length k filepaths to save the result disk-datasets. seed: int, optional (default None) Random seed to use. log_every_n: int, optional (default None) Loading Loading @@ -930,8 +965,8 @@ class MolecularWeightSplitter(Splitter): mws = np.array(mws) sortidx = np.argsort(mws) train_cutoff = frac_train * len(sortidx) valid_cutoff = (frac_train + frac_valid) * len(sortidx) train_cutoff = int(frac_train * len(sortidx)) valid_cutoff = int((frac_train + frac_valid) * len(sortidx)) return (sortidx[:train_cutoff], sortidx[train_cutoff:valid_cutoff], sortidx[valid_cutoff:]) Loading Loading @@ -1081,7 +1116,7 @@ class ButinaSplitter(Splitter): log_every_n: int, optional (default None) Log every n examples (not currently used). cutoff: float, optional (default 0.18) The The cutoff value for similarity. Returns ------- Loading @@ -1101,7 +1136,7 @@ class ButinaSplitter(Splitter): except ModuleNotFoundError: raise ValueError("This function requires RDKit to be installed.") print("Performing butina clustering with cutoff of", cutoff) logger.info("Performing butina clustering with cutoff of", cutoff) mols = [] for ind, smiles in enumerate(dataset.ids): mols.append(Chem.MolFromSmiles(smiles)) Loading @@ -1128,8 +1163,8 @@ class ButinaSplitter(Splitter): # TODO (Ytz): for regression tasks we'd stop after only one cluster. active_populations = np.sum(ys[valid_inds], axis=0) if np.all(active_populations): print("# of actives per task in valid:", active_populations) print("Total # of validation points:", len(valid_inds)) logger.info("# of actives per task in valid:", active_populations) logger.info("Total # of validation points:", len(valid_inds)) break train_inds = list(itertools.chain.from_iterable(scaffold_sets[c_idx + 1:])) Loading Loading @@ -1189,7 +1224,8 @@ class ScaffoldSplitter(Splitter): frac_valid: float = 0.1, frac_test: float = 0.1, seed: Optional[int] = None, log_every_n: int = 1000) -> Tuple[List[int], List[int], List[int]]: log_every_n: Optional[int] = 1000 ) -> Tuple[List[int], List[int], List[int]]: """ Splits internal compounds into train/validation/test by scaffold. Loading Loading @@ -1220,7 +1256,9 @@ class ScaffoldSplitter(Splitter): train_cutoff = frac_train * len(dataset) valid_cutoff = (frac_train + frac_valid) * len(dataset) train_inds, valid_inds, test_inds = [], [], [] train_inds: List[int] = [] valid_inds: List[int] = [] test_inds: List[int] = [] logger.info("About to sort in scaffold sets") for scaffold_set in scaffold_sets: Loading Loading @@ -1387,79 +1425,51 @@ class FingerprintSplitter(Splitter): cur_distances[i] = new_dist class SpecifiedSplitter(Splitter): """ Class that splits data according to user specification. """ def __init__(self, input_file, split_field): """Provide input information for splits.""" raw_df = next(load_data([input_file], shard_size=None)) self.splits = raw_df[split_field].values def split(self, dataset, seed=None, frac_train=.8, frac_valid=.1, frac_test=.1, log_every_n=1000): """ Splits internal compounds into train/validation/test by user-specification. """ train_inds, valid_inds, test_inds = [], [], [] for ind, split in enumerate(self.splits): split = split.lower() if split == "train": train_inds.append(ind) elif split in ["valid", "validation"]: valid_inds.append(ind) elif split == "test": test_inds.append(ind) else: raise ValueError("Missing required split information.") return train_inds, valid_inds, test_inds class SpecifiedIndexSplitter(Splitter): """ Class that splits data according to user index specification """ def __init__(self, train_inds, valid_inds, test_inds): """Provide input information for splits.""" self.train_inds = train_inds self.valid_inds = valid_inds self.test_inds = test_inds class TimeSplitterPDBbind(Splitter): def split(self, dataset, seed=None, frac_train=.8, frac_valid=.1, frac_test=.1, log_every_n=1000): def __init__(self, ids, year_file: Optional[str] = None): """ Splits internal compounds into train/validation/test by user-specification. Parameters ---------- ids: .... WIP year_file: str, optional (default None) The file path of .... """ return self.train_inds, self.valid_inds, self.test_inds class TimeSplitterPDBbind(Splitter): def __init__(self, ids, year_file=None): self.ids = ids self.year_file = year_file def split(self, dataset, seed=None, frac_train=.8, frac_valid=.1, frac_test=.1, log_every_n=None): dataset: Dataset, frac_train: float = 0.8, frac_valid: float = 0.1, frac_test: float = 0.1, seed: Optional[int] = None, log_every_n: Optional[int] = None ) -> Tuple[List[int], List[int], List[int]]: """ Splits protein-ligand pairs in PDBbind into train/validation/test in time order. Parameters ---------- dataset: Dataset Dataset to be split. frac_train: float, optional (default 0.8) The fraction of data to be used for the training split. frac_valid: float, optional (default 0.1) The fraction of data to be used for the validation split. frac_test: float, optional (default 0.1) The fraction of data to be used for the test split. seed: int, optional (default None) Random seed to use. log_every_n: int, optional (default None) Log every n examples (not currently used). Returns ------- Tuple[List[int], List[int], List[int]] A tuple of train indices, valid indices, and test indices. Each indices is a list of integers. """ if self.year_file is None: try: Loading
deepchem/splits/task_splitter.py +0 −1 Original line number Diff line number Diff line Loading @@ -71,7 +71,6 @@ class TaskSplitter(Splitter): n_tasks = len(dataset.get_task_names()) n_train = int(np.round(frac_train * n_tasks)) n_valid = int(np.round(frac_valid * n_tasks)) n_test = n_tasks - n_train - n_valid X, y, w, ids = dataset.X, dataset.y, dataset.w, dataset.ids Loading
deepchem/splits/test_specified_index_splitter.pydeleted 100644 → 0 +0 −28 Original line number Diff line number Diff line from unittest import TestCase import deepchem import numpy as np from sklearn.model_selection import train_test_split from deepchem.splits import SpecifiedIndexSplitter class TestSpecifiedIndexSplitter(TestCase): def create_dataset(self): n_samples, n_features = 20, 10 X = np.random.random(size=(n_samples, n_features)) y = np.random.random(size=(n_samples, 1)) return deepchem.data.NumpyDataset(X, y) def test_split(self): ds = self.create_dataset() indexes = list(range(len(ds))) train, test = train_test_split(indexes) train, valid = train_test_split(train) splitter = SpecifiedIndexSplitter(train, valid, test) train_ds, valid_ds, test_ds = splitter.train_valid_test_split(ds) self.assertTrue(np.all(train_ds.X == ds.X[train])) self.assertTrue(np.all(valid_ds.X == ds.X[valid])) self.assertTrue(np.all(test_ds.X == ds.X[test]))
