dare-datasets 1.0.1

A quick and easy way to download datasets from the DARE lab.

Darelab Datasets Docs

A quick access library of datasets used in Darelab.

Installation

Install: pip install dare-datasets

Documentation: https://darelab.athenarc.gr/datasets-docs/add_dataset/

Datasets included:

• QR2T Benchmark from MikeXydas
• Iris from MikeXydas
• Spider from George Katsogiannis & Anna Mitsopoulou
• ToTTo from MikeXydas
• Wikitable from MikeXydas

Usage

from dare_datasets import QR2TBenchmark

qr2t_benchmark = QR2TBenchmark()
qr2t_data = qr2t_benchmark.get()

For each dataset, additional methods might exist. Check the documentation of each dataset for more details.

Dev Installation

For development purposes, additional libraries must be installed such as pytest and mkdocs.

Prerequisites:

• Python >=3.8
• Poetry
• PreCommit (Optional)

1. Clone the repository
2. poetry install
3. pre-commit install (Optional)
4. git branch new_dataset_name

After any contribution you should open a pull request.

Testing

pytest is the testing framework used for this project. Commands:

• pytest - Run all tests. Warning: This will download all datasets in a temp directory (~4GB).
• pytest -m "not download" - Run all tests except the ones that download datasets (suggested during development).