AIStudioDelta/Eurovoc_2025_by_language

🇪🇺 🏷️ EuroVoc dataset (by language)

This is the EuropeanParliament/Eurovoc_2025 dataset, but split up by language, not by period.

The original is split up into periods (1996-03 through 2025-11), with documents in different languages mixed together. For ease of training this dataset splits the data by language instead, with documents in different periods put together.

License

This dataset is redistributed under the original European Union Public License 1.2. When using this version of the dataset, please give the appropriate credit to the original dataset.

Code

The dataset was created with the following script:

from datasets import concatenate_datasets, get_dataset_config_names, load_dataset
from tqdm import tqdm


def main(dataset_name, languages, output_dataset, private=True):
    periods = sorted(get_dataset_config_names(dataset_name))

    for language in languages:
        print(language)
        subsets = []
        for period in tqdm(periods):
            dataset = load_dataset(dataset_name, period, split='train')
            dataset = dataset.filter(lambda lang: lang == language, input_columns=['lang'])

            if len(dataset) > 0:
                dataset = dataset.add_column('subset', [period] * len(dataset))

                subsets.append(dataset)

        lang_dataset = concatenate_datasets(subsets)

        lang_dataset.push_to_hub(output_dataset, config_name=language, private=private)


if __name__ == '__main__':
    main(dataset_name='EuropeanParliament/Eurovoc_2025',
         languages=['bul', 'ces', 'dan', 'deu', 'ell', 'eng', 'est', 'fin', 'fra', 'gle', 'hrv', 'hun', 'isl', 'ita', 'lav', 'lit', 'mkd', 'mlt', 'nld', 'nor', 'pol', 'por', 'ron', 'slk', 'slv', 'spa', 'srp', 'swe', 'tur'],
         output_dataset='AIStudioDelta/Eurovoc_2025_by_language',
         private=True)