README.md

---
language_creators:
  - expert-generated
language:
  - lij
license:
- cc-by-4.0
tags:
- text
configs:
- config_name: default
  data_files:
  - split: train
    path: "data/*.parquet"
---
# Ligurian National Corpus (LiNC)

## Dataset Description

Point of contact: [Council for Ligurian Linguistic Heritage (Conseggio pe-o patrimònio linguistico ligure)](https://www.conseggio-ligure.org/en/) <info@conseggio-ligure.org>

### Dataset Summary

The Ligurian National Corpus (LiNC) is a comprehensive text corpus of written Ligurian, with a primary focus on the Genoese variety, which represents the most widely spoken form of the Ligurian language. Compiled by the [Council for Ligurian Linguistic Heritage (Conseggio pe-o patrimònio linguistico ligure)](https://www.conseggio-ligure.org/en/), this corpus encompasses various genres.

### Supported Tasks

This dataset has been designed to support corpus linguistics research and the development of large language models.

### Languages

The corpus contains Genoese Ligurian (ISO 639-3 `lij`, Glottolog `geno1240`), the principal variety of the Ligurian language. It is spoken mainly in the central part of Liguria (in an area roughly between Noli and Moneglia on the coast and much of the relative hinterland) and is almost universally understood by speakers of other varieties.

## Dataset Structure

### Data Instances

Each instance in the dataset is structured as in the following example:

```json
{
    "id": "a08a13691119388e…",
    "content": "# O Zinâ\\n30/06/2023\\n## Liguria\\n### À Zena comensa i travaggi pe-o tunnel sotta o pòrto\\nDòppo trent'anni de discuscion l'é comensou i travaggi…",
    "doc_type": "webpage",
    "tokens": 1388,
    "citation_key": "zina",
    "year": 2023
}
```

### Data Fields

- `id`: A unique identifier for the entry.
- `content`: The textual content of the entry.
- `doc_type`: The document type (currently one of `webpage`, `magazine`, `dictionary`, `book`).
- `tokens`: The number of space-separated tokens in the current entry.
- `citation_key`: For webpages, the URL. For other materials, a citation key which references the bibliography (see the file `bibliography.yaml`).
- `year`:  The year the content was authored.

## Dataset Creation

### Curation Rationale

The Ligurian National Corpus was created to preserve and document written Ligurian language use across various genres and contexts. The corpus serves both as a linguistic resource for academic research and as training data for computational applications.

### Source Data

#### Initial Data Collection and Normalization

The corpus consists of publicly available written materials in Ligurian, collected and curated by the Council for Ligurian Linguistic Heritage.

#### Who are the source language producers?

The texts included in this corpus come from various authors and sources writing in Ligurian, with a focus on the Genoese variety. Provenance information for each entry is provided in the `citation_key` column of the dataset.

## Considerations for Using the Data

### Social Impact of Dataset

This dataset contributes to the preservation and computational processing of Ligurian, supporting efforts to maintain linguistic diversity and enable technological support for this language.

### Discussion of Biases

- The Ligurian National Corpus currently only focuses on the Genoese variety of Ligurian only.
- This release represents a subset of the complete LiNC corpus and includes only those documents that are publicly available.

## Additional Information

### Dataset Curators

The [Council for Ligurian Linguistic Heritage (Conseggio pe-o patrimònio linguistico ligure)](https://www.conseggio-ligure.org/en/) is responsible for the curation and maintenance of this corpus.

### Licensing Information

This public subset of the LiNC corpus is released under a Creative Commons Attribution 4.0 International License (CC-BY 4.0).