# Copyright 2020 The TensorFlow Authors. All Rights Reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. # ============================================================================== """Keras timeseries dataset utilities.""" # pylint: disable=g-classes-have-attributes from __future__ import absolute_import from __future__ import division from __future__ import print_function import numpy as np from tensorflow.python.data.ops import dataset_ops from tensorflow.python.ops import array_ops from tensorflow.python.ops import math_ops from tensorflow.python.util.tf_export import keras_export @keras_export('keras.preprocessing.timeseries_dataset_from_array', v1=[]) def timeseries_dataset_from_array( data, targets, sequence_length, sequence_stride=1, sampling_rate=1, batch_size=128, shuffle=False, seed=None, start_index=None, end_index=None): """Creates a dataset of sliding windows over a timeseries provided as array. This function takes in a sequence of data-points gathered at equal intervals, along with time series parameters such as length of the sequences/windows, spacing between two sequence/windows, etc., to produce batches of timeseries inputs and targets. Arguments: data: Numpy array or eager tensor containing consecutive data points (timesteps). Axis 0 is expected to be the time dimension. targets: Targets corresponding to timesteps in `data`. It should have same length as `data`. `targets[i]` should be the target corresponding to the window that starts at index `i` (see example 2 below). Pass None if you don't have target data (in this case the dataset will only yield the input data). sequence_length: Length of the output sequences (in number of timesteps). sequence_stride: Period between successive output sequences. For stride `s`, output samples would start at index `data[i]`, `data[i + s]`, `data[i + 2 * s]`, etc. sampling_rate: Period between successive individual timesteps within sequences. For rate `r`, timesteps `data[i], data[i + r], ... data[i + sequence_length]` are used for create a sample sequence. batch_size: Number of timeseries samples in each batch (except maybe the last one). shuffle: Whether to shuffle output samples, or instead draw them in chronological order. seed: Optional int; random seed for shuffling. start_index: Optional int; data points earlier (exclusive) than `start_index` will not be used in the output sequences. This is useful to reserve part of the data for test or validation. end_index: Optional int; data points later (exclusive) than `end_index` will not be used in the output sequences. This is useful to reserve part of the data for test or validation. Returns: A tf.data.Dataset instance. If `targets` was passed, the dataset yields tuple `(batch_of_sequences, batch_of_targets)`. If not, the dataset yields only `batch_of_sequences`. Example 1: Consider indices `[0, 1, ... 99]`. With `sequence_length=10, sampling_rate=2, sequence_stride=3`, `shuffle=False`, the dataset will yield batches of sequences composed of the following indices: ``` First sequence: [0 2 4 6 8 10 12 14 16 18] Second sequence: [3 5 7 9 11 13 15 17 19 21] Third sequence: [6 8 10 12 14 16 18 20 22 24] ... Last sequence: [78 80 82 84 86 88 90 92 94 96] ``` In this case the last 3 data points are discarded since no full sequence can be generated to include them (the next sequence would have started at index 81, and thus its last step would have gone over 99). Example 2: temporal regression. Consider an array `data` of scalar values, of shape `(steps,)`. To generate a dataset that uses the past 10 timesteps to predict the next timestep, you would use: ```python input_data = data[:-10] targets = data[10:] dataset = tf.keras.preprocessing.timeseries_dataset_from_array( input_data, targets, sequence_length=10) for batch in dataset: inputs, targets = batch assert np.array_equal(inputs[0], data[:10]) # First sequence: steps [0-9] assert np.array_equal(targets[0], data[10]) # Corresponding target: step 10 break ``` """ # Validate the shape of data and targets if targets is not None and len(targets) != len(data): raise ValueError('Expected data and targets to have the same number of ' 'time steps (axis 0) but got ' 'shape(data) = %s; shape(targets) = %s.' % (data.shape, targets.shape)) if start_index and (start_index < 0 or start_index >= len(data)): raise ValueError('start_index must be higher than 0 and lower than the ' 'length of the data. Got: start_index=%s ' 'for data of length %s.' % (start_index, len(data))) if end_index: if start_index and end_index <= start_index: raise ValueError('end_index must be higher than start_index. Got: ' 'start_index=%s, end_index=%s.' % (start_index, end_index)) if end_index >= len(data): raise ValueError('end_index must be lower than the length of the data. ' 'Got: end_index=%s' % (end_index,)) if end_index <= 0: raise ValueError('end_index must be higher than 0. ' 'Got: end_index=%s' % (end_index,)) # Validate strides if sampling_rate <= 0 or sampling_rate >= len(data): raise ValueError( 'sampling_rate must be higher than 0 and lower than ' 'the length of the data. Got: ' 'sampling_rate=%s for data of length %s.' % (sampling_rate, len(data))) if sequence_stride <= 0 or sequence_stride >= len(data): raise ValueError( 'sequence_stride must be higher than 0 and lower than ' 'the length of the data. Got: sequence_stride=%s ' 'for data of length %s.' % (sequence_stride, len(data))) if start_index is None: start_index = 0 if end_index is None: end_index = len(data) # Determine the lowest dtype to store start positions (to lower memory usage). num_seqs = end_index - start_index - (sequence_length * sampling_rate) + 1 if num_seqs < 2147483647: index_dtype = 'int32' else: index_dtype = 'int64' # Generate start positions start_positions = np.arange(0, num_seqs, sequence_stride, dtype=index_dtype) if shuffle: if seed is None: seed = np.random.randint(1e6) rng = np.random.RandomState(seed) rng.shuffle(start_positions) sequence_length = math_ops.cast(sequence_length, dtype=index_dtype) sampling_rate = math_ops.cast(sampling_rate, dtype=index_dtype) positions_ds = dataset_ops.Dataset.from_tensors(start_positions).repeat() # For each initial window position, generates indices of the window elements indices = dataset_ops.Dataset.zip( (dataset_ops.Dataset.range(len(start_positions)), positions_ds)).map( lambda i, positions: math_ops.range( # pylint: disable=g-long-lambda positions[i], positions[i] + sequence_length * sampling_rate, sampling_rate), num_parallel_calls=dataset_ops.AUTOTUNE) dataset = sequences_from_indices(data, indices, start_index, end_index) if targets is not None: indices = dataset_ops.Dataset.zip( (dataset_ops.Dataset.range(len(start_positions)), positions_ds)).map( lambda i, positions: positions[i], num_parallel_calls=dataset_ops.AUTOTUNE) target_ds = sequences_from_indices( targets, indices, start_index, end_index) dataset = dataset_ops.Dataset.zip((dataset, target_ds)) if shuffle: # Shuffle locally at each iteration dataset = dataset.shuffle(buffer_size=batch_size * 8, seed=seed) dataset = dataset.batch(batch_size) return dataset def sequences_from_indices(array, indices_ds, start_index, end_index): dataset = dataset_ops.Dataset.from_tensors(array[start_index : end_index]) dataset = dataset_ops.Dataset.zip((dataset.repeat(), indices_ds)).map( lambda steps, inds: array_ops.gather(steps, inds), # pylint: disable=unnecessary-lambda num_parallel_calls=dataset_ops.AUTOTUNE) return dataset