README.md

---
license: cc-by-sa-4.0
---
```markdown
# RexTRO111/articles

Welcome to the **RexTRO111/articles** dataset. This repository contains a curated, multi-source text dataset specifically preprocessed and structured for large language model (LLM) pretraining, fine-tuning, and natural language processing pipelines. 🌐

The dataset features a clean, flat architecture optimized for high-performance streaming and distributed data loading. 📊

## 📂 Dataset Structure

The dataset is saved in the Apache Parquet format. Each shard contains a flat schema with the following fields:

*   `text` (string): The raw extracted text content, split exactly into **one line per paragraph** to preserve natural semantic context and chunking boundaries.
*   `url` (string): The exact source URL from which the text was extracted to maintain strict provenance and citation history.

### Repository Layout
```bash
.
├── README.md
├── articles-0000.parquet
└── ...

```

## 🛠️ Data Sources & Curation

This dataset aggregates high-quality, text-dense knowledge bases from across the web:

1. **Wikipedia**: Extracted main body content covering diverse domains, providing broad foundational knowledge. 📖
2. **arXiv**: Academic abstracts and structured paper content spanning machine learning, computer science, and related fields. 🔬
3. **Other Text Repositories**: Integrated text sequences designed to improve the general knowledge and reasoning capabilities of language models. 🧠

## 🚀 Quick Start / How to Use

You can easily stream or load this dataset using `pandas` or the Hugging Face `datasets` library.

### Using Pandas 🐼

```python
import pandas as pd

# Load a specific shard
df = pd.read_parquet("articles-0000.parquet")
print(df.head())

```

### Using Hugging Face Datasets 🤗

```python
from datasets import load_dataset

dataset = load_dataset("RexTRO111/articles")
print(dataset["train"][0])

```

## ⚖️ Licensing & Terms

This dataset is licensed under the **Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)** license.

* **Attribution**: You must give appropriate credit, provide a link to the license, and indicate if changes were made.
* **ShareAlike**: If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.
* *Disclaimer*: Content included in this dataset is gathered from public web sources. Users are responsible for ensuring downstream compliance regarding fair use and source-specific terms (such as individual author licenses on arXiv).

```