| --- |
| license: gpl-3.0 |
| dataset_info: |
| features: |
| - name: filename |
| dtype: string |
| - name: url |
| dtype: string |
| - name: content_text |
| dtype: string |
| - name: size_bytes |
| dtype: int64 |
| --- |
| # Intel Ark Frontend Assets (CSS & JS) |
|
|
| This dataset contains the enterprise frontend assets (Stylesheets and JavaScript files) extracted from Intel Ark (`ark.intel.com`). |
|
|
| ## 🎯 Primary Use Case |
| This dataset is specifically structured for **pre-training and fine-tuning AI coding assistants and web-navigating agents**. By analyzing production-grade code, models can learn how modern enterprise infrastructure (like Adobe Experience Manager) maps DOM elements to CSS rules, hydrates client-side state, and structures UI frameworks (like Bootstrap and custom Intel design systems). |
|
|
| ## 📊 Token Statistics |
|
|
| Asset counts and token volume were calculated using OpenAI's `tiktoken` (`cl100k_base` encoding). |
|
|
| | Asset Type | File Count | Raw Size | `cl100k_base` Tokens | |
| | :--- | :--- | :--- | :--- | |
| | **Stylesheets (.css)** | 7 | 215.18 KB | 424,625 | |
| | **Scripts (.js)** | 24 | 647.92 KB | 761,484 | |
| | **Total Corpus** | **31** | **863.10 KB** | **1,186,109** | |
|
|
| *Note: The exceptionally low file count compared to the page volume (3,700+ pages) is due to strict global deduplication. The entire Intel Ark domain is powered by these exact 31 centralized assets.* |
|
|
| ## 📁 Dataset Structure |
| - **`intel-ark_20260923_164720_css.parquet`**: Contains deduplicated `.css` files (minified text) and their source URLs. |
| - **`intel-ark_20260923_164720_js.parquet`**: Contains deduplicated `.js` files (minified text), including AEM hydration logic, analytics, and third-party trackers. |
|
|
| All files were normalized (bypassing strict WAF redirect loops on the `ark` subdomain) to fetch directly from Intel's global CDN (`www.intel.com`). |
|
|
| ## 📄 License |
| This dataset is published under the **GPL-3.0** License. |
|
|
