datasetops 0.0.5

Fluent dataset operations, compatible with your favorite libraries

Dataset Ops

Friendly dataset operations for your data science needs

TL;DR

import datasetops as do

# prepare your data
train, val, test = (
    do.load_folder_class_data(path)
    .named("data", "label")
    .img_resize((240, 240))
    .one_hot("label")
    .shuffle(seed=42)
    .split([0.6, 0.2, 0.2])
)

# use your favorite framework
train_tf = trian.to_tensorflow() 
train_pt = trian.to_pytorch() #coming up!

# or do your own thing
for img, label in train:
    ...

Motivation

Collecting and preprocessing datasets is a tiresome and often underestimated part of the data science and machine learning lifecycle. While Tensorflow and PyTorch do have some useful datasets utilisites available, they are designed specifically with the respective frameworks in mind. Unsuprisingly, this makes it hard to switch between frameworks, and port training-ready dataset definitions.

Moreover, they do not aid you in standard scenarios where you want to:

• subsample your dataset, e.g with a fixed number of samples per class
• rescale, center, standardize, normalise you data
• combine multiple datasets, e.g. for parallel input in a multi-stream network
• create non-standard data splits

All of this is usually done by hand. Again and again and again...

Idea

In a nutshell, datasets for data science and machine learning are just a collection of samples that are often accompanied by a label.We should be able to read all these formats into a common representation, where most common operations can be performed.Subsequently, we should be able to transform these into the standard formats used in Tensorflow and PyTorch.

Implementation Status

The library is still under heavy development and the API may be subject to change.

What follows here is a list of implemented and planned features.

Loaders

• Loader (utility class used to define a dataset)
• load (load data from a path, automatically inferring type and structure)
• load_folder_data (load flat folder with data)
• load_folder_class_data (load nested folder with a folder for each class)
• load_folder_dataset_class_data (load nested folder with multiple datasets, each with a nested class folder structure )
• load_mat (load contents of a .mat file as a single dataaset)
• load_mat_single_mult_data (load contents of a .mat file as multiple dataasets)

Dataset information

• shape (get shape of a dataset item)
• counts (compute the counts of each unique item in the dataset by key)
• unique (get a list of unique items in the dataset by key)
• item_names (get a list of names for the elements in an item)
• named (supply names for the item elements)
• stats (provide an overview of the dataset statistics)
• origin (provide an description of how the dataset was made)

Sampling and splitting

• shuffle (shuffle the items in a dataset randomly)
• sample (sample data at random a dataset)
• filter (filter the dataset using a predicate)
• split (split a dataset randomly based on fractions)
• split_filter (split a dataset into two based on a predicate)
• allow_unique (handy predicate used for balanced classwise filtering/sampling)
• take (take the first items in dataset)
• repeat (repeat the items in a dataset, either itemwise or as a whole)

Item manipulation

• reorder (reorder the elements of the dataset items (e.g. flip label and data order))
• transform (transform function which takes other functions and applies them to the dataset items.)
• custom (function wrapper enabling user-defined function to be used as a transform)
• label (transforms an element into a integer encoded categorical label)
• one_hot (transforms an element into a one-hot encoded categorical label)
• numpy (transforms an element into a numpy.ndarray)
• reshape (reshapes numpy.ndarray elements)
• image (transforms a numpy array or path string into a PIL.Image.Image)
• image_resize (resizes PIL.Image.Image elements)
• center (modify each item according to dataset statistics)
• normalize (modify each item according to dataset statistics)
• standardize (modify each item according to dataset statistics)
• whiten (modify each item according to dataset statistics)

Dataset combinations

• concat (concatenate two datasets, placing the items of one after the other)
• zip (zip datasets itemwise, extending the size of each item)
• cartesian_product (create a dataset whose items are all combinations of items (zipped) of the originating datasets)

Converters

• to_tensorflow (convert Dataset into tensorflow.data.Dataset)
• to_pytroch (convert Dataset into torchvision.Dataset)