---
task_categories:
- text-classification
language:
- el
tags:
- legal
- et-gr
- greek
- legislation
size_categories:
- 1K<n<10K
---

# NOMOS_Greek_Legislation

**NOMOS_Greek_Legislation** is a Greek-language legal text corpus derived from the National Printing House (Ethniko Typografeio - ET.gr).
The dataset focuses on Greek national legislation (Laws, Presidential Decrees, Ministerial Decisions) and was developed in the context of the **+NOMOS project**.

It provides full-text legal documents enriched with metadata and thematic classification tags, suitable for Legal NLP tasks such as Text Classification and Language Modeling.

### File Organization

The dataset consists of a central metadata index and a folder containing the raw text files

### Metadata Index

| Field | Description |
| :--- | :--- |
| **id** | Unique internal ID |
| **filename** | The corresponding `.txt` filename in `files/` |
| **word_count** | Word count of the plain text |
| **title** | Short title of the act |
| **full_title** | Official full title (includes Law Number/Date) |
| **article_count** | Number of articles detected |
| **article_titles** | List of article titles (separated by `\|`) |
| **fek** | Government Gazette Issue (FEK) string |
| **fek_search_url** | Generated search URL for ET.gr |
| **fek_url** | Direct download link for the original PDF |
| **topics** | Thematic classification tags (separated by `\|`) |

# How to download the data

Use the provided download_script.py to pull the data to your local machine. This ensures the directory structure is preserved and only the necessary files are downloaded.

### Prerequisites

Ensure you have the huggingface_hub library installed:

  pip install huggingface_hub

### Usage

Run the following command in your terminal to download the files/ folder and the dataset_metadata.csv:

  python download_script.py

The script will create a local directory named greek_legal_data (or your chosen path) containing:
dataset_metadata.csv: The central index for all legal documents.
files/: A folder containing the raw .txt files corresponding to the metadata.