datalabs

DataLab is a unified platform that allows for NLP researchers to perform a number of data-related tasks in an efficient and easy-to-use manner. In particular, DataLab supports the following functionalities:

Data Diagnostics: DataLab allows for analysis and understanding of data to uncover undesirable traits such as hate speech, gender bias, or label imbalance.
Operation Standardization: DataLab provides and standardizes a large number of data processing operations, including aggregating, preprocessing, featurizing, editing and prompting operations.
Data Search: DataLab provides a semantic dataset search tool to help identify appropriate datasets given a textual description of an idea.
Global Analysis: DataLab provides tools to perform global analyses over a variety of datasets.

Installation

DataLab can be installed from PyPi

pip install --upgrade pip
pip install datalabs
python -m nltk.downloader omw-1.4 # to support more feature calculation

or from the source

# This is suitable for SDK developers
pip install --upgrade pip
git clone git@github.com:ExpressAI/DataLab.git
cd Datalab
pip install -e .[dev]
python -m nltk.downloader omw-1.4 # to support more feature calculation

By adding [dev], some extra libraries will be installed, such as pre-commit.

Code Quality Check?

If you would like to contribute to DataLab, checking the code style and quality before your pull request is highly recommended. In this project, three types of checks will be expected: (a) black (2) flake8 (3) isort

you could achieve this in two ways:

Manually (suitable for developers using Github Destop)

pre-commit install
git init .
pre-commit run --all-files or

where pre-commit run -all-files is equivalent to

pre-commit run black   # (this is also equivalent to python -m black .)
pre-commit run isort   # (this is also equivalent to isort .)
pre-commit run flake8  # (this is  also equivalent to flake8)

Notably, black and isort can help us fix code style automatically, while flake8 only provide hints with us, which means we need to fix these issues raised by flake8.

Automatically (suitable for developers using Git CLI)

pre-commit install
git init .
git commit -m "your update message"

The git commit will automatically activate the command pre-commit run -all-files

Using DataLab

Below we give several simple examples to showcase the usage of DataLab:

You can also view documentation:

# pip install datalabs
from datalabs import load_dataset
dataset = load_dataset("ag_news")


# Preprocessing operation
from preprocess import *
res=dataset["test"].apply(lower)
print(next(res))

# Featurizing operation
from featurize import *
res = dataset["test"].apply(get_text_length) # get length
print(next(res))

res = dataset["test"].apply(get_entities_spacy) # get entity
print(next(res))

# Editing/Transformation operation
from edit import *
res = dataset["test"].apply(change_person_name) #  change person name
print(next(res))

# Prompting operation
from prompt import *
res = dataset["test"].apply(template_tc1)
print(next(res))

# Aggregating operation
from aggregate.text_classification import *
res = dataset["test"].apply(get_statistics)

Acknowledgment

DataLab originated from a fork of the awesome Huggingface Datasets and TensorFlow Datasets. We highly thank the Huggingface/TensorFlow Datasets for building this amazing library. More details on the differences between DataLab and them can be found in the section. We thank Antonis Anastasopoulos for sharing the mapping data between countries and languages, and thank Alissa Ostapenko, Yulia Tsvetkov, Jie Fu, Ziyun Xu, Hiroaki Hayashi, and Zhengfu He for useful discussion and suggestions for the first version.

Metadata

Release files for datalabs 0.4.15

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for datalabs 0.4.15
File	Size	Uploaded
datalabs-0.4.15.tar.gz	340.9 kB	Dec 22, 2022	Details

Built distribution (wheel)

Table of built distributions (wheels) for datalabs 0.4.15
File	Interpreter	ABI	Platform	Reset
datalabs-0.4.15-py2.py3-none-any.whl 2.3 MB Dec 22, 2022	Python 3, Python 2	none	any	Details

Total release size: 2.6 MB

Release files / datalabs-0.4.15.tar.gz

Download URL	datalabs-0.4.15.tar.gz
Size	340.9 kB
Tags	Source
SHA-256 checksum How to use checksums	`020e88f01890f21614af117f51922e78eaf48d1003d2ff93df150101b0e5e9af`
BLAKE2b-256 checksum How to use checksums	`54b625da85ae5f1758a13dd3d1f3be17f28ffe6d190e0c05e218cf00af2ce2f4`
Upload date	Dec 22, 2022
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/4.0.2 CPython/3.9.16`

Release files / datalabs-0.4.15-py2.py3-none-any.whl

Download URL	datalabs-0.4.15-py2.py3-none-any.whl
Size	2.3 MB
Tags	Python 2 Python 3
SHA-256 checksum How to use checksums	`463d48186b98a70b1ce1d0abb10aa549f1cbaf65b0b8b7ed64d10b80969c483a`
BLAKE2b-256 checksum How to use checksums	`f21353c4f424079a8769a40cd0dda67ac1601e34455d2a3400ba9a4a68de0955`
Upload date	Dec 22, 2022
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/4.0.2 CPython/3.9.16`