README.md

---
pretty_name: "IREKIER: An Easy Read Corpus for Basque and Spanish"
language:
  - es
  - eu
license: cc-by-nc-sa-4.0
multilinguality: multilingual
task_categories:
  - text-generation
task_ids:
  - text-simplification
size_categories:
  - 1K<n<10K
configs:
  - config_name: es-paragraph-almt
    default: true
    data_files:
      - split: train
        path: es/paragraph-almt/train.json
      - split: validation
        path: es/paragraph-almt/dev.json
      - split: test
        path: es/paragraph-almt/test.json

  - config_name: es-doc-almt
    data_files:
      - split: train
        path: es/doc-almt/documents.json

  - config_name: eu-paragraph-almt
    data_files:
      - split: train
        path: eu/paragraph-almt/train.json
      - split: validation
        path: eu/paragraph-almt/dev.json
      - split: test
        path: eu/paragraph-almt/test.json

  - config_name: eu-doc-almt
    data_files:
      - split: train
        path: eu/doc-almt/documents.json
---

# IREKIER: An Easy Read Corpus for Basque and Spanish

This dataset contains Spanish (`es`) and Basque (`eu`) Easy Read articles from the Basque Government's transparency portal (Irekia) in two alignment granularities:

- `doc-almt`: document-level aligned pairs
- `paragraph-almt`: paragraph-level aligned pairs (train/dev/test)

## Dataset structure

### Document-level alignments (`doc-almt`)

- `es/doc-almt/documents.json`
- `eu/doc-almt/documents.json`

Each entry contains:

- `id`: document id
- `src`: complex,original document text
- `tgt.seg`: adapted segmented version
- `tgt.no_seg`: adapted non-segmented version

### Paragraph-level alignments (`paragraph-almt`)

Spanish:
- `es/paragraph-almt/train.json`
- `es/paragraph-almt/dev.json`
- `es/paragraph-almt/test.json`

Basque:
- `eu/paragraph-almt/train.json`
- `eu/paragraph-almt/dev.json`
- `eu/paragraph-almt/test.json`

All paragraph files are JSON arrays of documents:

- `id` (string): document id
- `alignments` (array): list of aligned paragraph pairs for that document

Each item in `alignments` contains:

- `src` (string): source,complex paragraph text
- `tgt` (string): adapted paragraph text

Split-specific note:

- `train.json` alignments also include `score` (float similarity score)
- `dev.json` and `test.json` alignments contain only `src` and `tgt`

These files provide paragraph alignment data for training, development, and evaluation.

Full details of alignment process can be found in the Citation reference.

## Languages

- Spanish (`es`)
- Basque (`eu`)

## License

This dataset is distributed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license (CC BY-NC-SA 4.0).

## Citation

If you use this dataset, please cite:


```bibtex
@inproceedings{calleja2026irekier,
  title = {{IREKIER: An Easy Read Corpus for Basque and Spanish}},
  author = {Calleja, Jesús and Etchegoyhen, Thierry},
  booktitle = {Proceedings of the 15th edition of the Language Resources and Evaluation Conference (LREC 2026)},
  note = {to appear}
  year = {2026}
}
```