---
dataset_info:
  features:
  - name: chapter
    dtype: int64
  - name: rom
    dtype: string
  - name: ro
    dtype: string
  - name: verse
    dtype: int64
  - name: book
    dtype: string
  splits:
  - name: train
    num_bytes: 7347749.210597389
    num_examples: 24881
  - name: test
    num_bytes: 1837158.7894026109
    num_examples: 6221
  download_size: 5593816
  dataset_size: 9184908
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
license: mit
task_categories:
- translation
language:
- ro
size_categories:
- 10K<n<100K
---
# Dataset Card for "ro-romani"

## Dataset Details

This dataset consists of parallel sentences extracted from online versions of the Bible (both Old and New Testaments) for the Romanian and [Kalderash Romani](https://en.wikipedia.org/wiki/Kalderash_Romani_language) languages.

## Dataset Structure

### Data Instances

An example looks like this:

```
{
 'book': 'GEN',
 'chapter': 38,
 'ro': 'A mai nascut iarasi un fiu, caruia i-a pus numele Sela; Iuda era la '
       'Czib cand a nascut ea.',
 'rom': 'Mai kărdea pale khă šeau, thai thodea lehkă anau Šela; o Iuda sas '
        'ando Czib, kana kărdeas oi.',
 'verse': 5
}
```

### Data Fields

- book: book name from where the text sequence was extracted
- chapter: book chapter number
- ro: sequence of text in Romanian
- rom: sequence of text in Kalderash Romani
- verse: verse number
  
## Dataset Creation

### Curation Rationale

<!-- Motivation for the creation of this dataset. -->

This dataset was built as part of a faculty project

### Source Data

<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->

The data was scraped from the following websites:

- https://www.ebible.ro/biblia/romana/cornilescu/ - for the verses in Romanian
- https://romani.global.bible/info - for the verses in Kalderash Romani


#### Data Collection and Processing

<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->

The verses were individiually scraped from the previously mentioned websites and aligned based on the book, chapter and verse order.

## Contributions

Thanks to [@cgrigoruta](https://huggingface.co/cgrigoruta) for extracting the Romanian verse translations.