almanach/HALvest-Geometric

---
pretty_name: HALvest-Geometric

license: cc-by-4.0

configs:
- config_name: en
data_files: "en/*.gz"
- config_name: fr
data_files: "fr/*.gz"

language:
- en
- fr

size_categories:
- 100K<n<1M

task_categories:
- text-generation
- fill-mask

task_ids:
- language-modeling
- masked-language-modeling

tags:
- academia
- research
- graph

annotations_creators:
- no-annotation

multilinguality:
- multilingual

source_datasets:
- HALvest
---

<div align="center">
<h1> HALvest-Geometric </h1>
<h3> Citation Network of Open Scientific Papers Harvested from HAL </h3>
</div>

---

## Dataset Description

- **Repository:** [GitHub](https://github.com/Madjakul/HALvesting-Geometric)

## Dataset Summary

### overview:

French and English fulltexts from open papers found on [Hyper Articles en Ligne (HAL)](https://hal.science/) and its citation network.

You can download the dataset using Hugging Face datasets:
```py
from datasets import load_dataset

ds = load_dataset("Madjakul/HALvest-Geometric", "en")
```

### Details

#### Nodes

* Papers: 18,662,037
* Authors: 238,397
* Affiliations: 96,105
* Domains: 16

#### Edges

- Paper <-> Domain: 136,700
- Paper <-> Paper: 22,363,817
- Author <-> Paper: 238,397
- Author <-> Affiliation: 426,030

### Languages

ISO-639|Language|# Documents|# mT5 Tokens
-------|--------|-----------|--------
en|English|442,892|7,606,895,258
fr|French|193,437|8,728,722,255

## Considerations for Using the Data

The corpus is extracted from the [HAL's open archive](https://hal.science/) which distributes scientific publications following open access principles. The corpus is made up of both creative commons licensed and copyrighted documents (distribution authorized on HAL by the publisher). This must be considered prior to using this dataset for any purpose, other than training deep learning models, data mining etc. We do not own any of the text from which these data has been extracted.

## Dataset Copyright

The licence terms for HALvest strictly follows the one from HAL. Please refer to the below license when using this dataset.
- [HAL license](https://doc.archives-ouvertes.fr/en/legal-aspects/)

## Citation

```
@misc{kulumba2024harvestingtextualstructureddata,
title={Harvesting Textual and Structured Data from the HAL Publication Repository},
author={Francis Kulumba and Wissam Antoun and Guillaume Vimont and Laurent Romary},
year={2024},
eprint={2407.20595},
archivePrefix={arXiv},
primaryClass={cs.DL},
url={https://arxiv.org/abs/2407.20595},
}
```

pretty_name: HALvest-Geometric
license: cc-by-4.0
configs:
- config_name: en
data_files: "en/*.gz"
- config_name: fr
data_files: "fr/*.gz"
language:
- en
- fr
size_categories:
- 10万 < 样本数 < 100万
task_categories:
- 文本生成
- 掩码填充
task_ids:
- 语言建模
- 掩码语言建模
tags:
- 学术领域
- 研究
- 图谱
annotations_creators:
- 无注释
multilinguality:
- 多语言
source_datasets:
- HALvest

<div align="center">
<h1> HALvest-Geometric 数据集</h1>
<h3> 从HAL平台采集的开放科学论文引用网络</h3>
</div>

---

## 数据集说明

- **仓库地址：** [GitHub](https://github.com/Madjakul/HALvesting-Geometric)

## 数据集概览

### 总览：

本数据集包含从[Hyper Articles en Ligne（HAL）](https://hal.science/)平台获取的开放获取论文的英法双语全文及其对应的引用网络。

用户可通过Hugging Face datasets库下载本数据集：
py
from datasets import load_dataset

ds = load_dataset("Madjakul/HALvest-Geometric", "en")

### 详细信息

#### 节点类型

* 论文：18,662,037篇
* 作者：238,397位
* 所属机构：96,105个
* 研究领域：16个

#### 边关系

- 论文 <-> 研究领域：136,700条
- 论文 <-> 论文：22,363,817条
- 作者 <-> 论文：238,397条
- 作者 <-> 所属机构：426,030条

### 语言分布

|ISO-639代码|语言|文档数量|mT5 Token 数量|
|-------|--------|-----------|--------|
|en|英语|442,892|7,606,895,258|
|fr|法语|193,437|8,728,722,255|

## 数据使用注意事项

本数据集语料库源自[HAL开放存档库](https://hal.science/)，该平台遵循开放获取原则发布科学出版物。本语料库包含知识共享许可协议及受版权保护的文档（出版商已授权HAL平台进行分发）。在将本数据集用于深度学习模型训练、数据挖掘等任何用途前，需充分考虑上述版权情况。本团队不拥有提取自这些文本的任何所有权。

## 数据集版权

HALvest数据集的许可条款严格遵循HAL平台的相关规定。使用本数据集时，请参考以下许可协议：
- [HAL许可协议](https://doc.archives-ouvertes.fr/en/legal-aspects/)

## 引用格式

@misc{kulumba2024harvestingtextualstructureddata,
title={Harvesting Textual and Structured Data from the HAL Publication Repository},
author={Francis Kulumba and Wissam Antoun and Guillaume Vimont and Laurent Romary},
year={2024},
eprint={2407.20595},
archivePrefix={arXiv},
primaryClass={cs.DL},
url={https://arxiv.org/abs/2407.20595},
}