GSMA/3GPP-REL18

--- language: - en license: other license_name: 3gpp license_link: https://www.3gpp.org/specifications-technologies/legal-matters tags: - telecommunications - 3gpp - 5g - nr - lte - standards - release-18 pretty_name: 3GPP Release 18 Specifications configs: - config_name: raw default: true data_files: - split: 21_series path: data/raw/21_series-*.parquet - split: 22_series path: data/raw/22_series-*.parquet - split: 23_series path: data/raw/23_series-*.parquet - split: 24_series path: data/raw/24_series-*.parquet - split: 26_series path: data/raw/26_series-*.parquet - split: 27_series path: data/raw/27_series-*.parquet - split: 28_series path: data/raw/28_series-*.parquet - split: 29_series path: data/raw/29_series-*.parquet - split: 31_series path: data/raw/31_series-*.parquet - split: 32_series path: data/raw/32_series-*.parquet - split: 33_series path: data/raw/33_series-*.parquet - split: 36_series path: data/raw/36_series-*.parquet - split: 37_series path: data/raw/37_series-*.parquet - split: 38_series path: data/raw/38_series-*.parquet - config_name: images data_files: - split: 21_series path: data/images/21_series-*.parquet - split: 22_series path: data/images/22_series-*.parquet - split: 23_series path: data/images/23_series-*.parquet - split: 24_series path: data/images/24_series-*.parquet - split: 26_series path: data/images/26_series-*.parquet - split: 27_series path: data/images/27_series-*.parquet - split: 28_series path: data/images/28_series-*.parquet - split: 29_series path: data/images/29_series-*.parquet - split: 31_series path: data/images/31_series-*.parquet - split: 32_series path: data/images/32_series-*.parquet - split: 33_series path: data/images/33_series-*.parquet - split: 36_series path: data/images/36_series-*.parquet - split: 37_series path: data/images/37_series-*.parquet - split: 38_series path: data/images/38_series-*.parquet --- # 3GPP Release 18 Specifications 549 3GPP technical specifications and reports converted to markdown, organized by series. ## Usage ```python from datasets import load_dataset # Load all raw markdown for the 38-series raw = load_dataset("GSMA/3GPP", "raw", split="38_series") # Filter to one spec spec = raw.filter(lambda r: r["spec_id"] == "38331") print(spec[0]["text"][:500]) # Load all images for the 38-series imgs = load_dataset("GSMA/3GPP", "images", split="38_series") # Get images for one spec spec_imgs = imgs.filter(lambda r: r["spec_id"] == "38104") spec_imgs[0]["image"].show() ``` ## Structure Two **configs** (subsets): `raw` and `images`. Each has 14 **splits** (one per series). **`raw`** -- one row per specification: | spec_id | title | version | date | text | |---------|-------|---------|------|------| | 38104 | TS 38.104 | V18.4.0 | 2023-12 | # 3GPP TS 38.104... | **`images`** -- one row per extracted figure: | spec_id | filename | image | |---------|----------|-------| | 38104 | 64662465bba2..._img.jpg | PIL Image | ## Series (splits) | Split | Specs | Images | |-------|------:|-------:| | 21_series | 1 | 4 | | 22_series | 25 | 351 | | 23_series | 86 | 4539 | | 24_series | 55 | 1799 | | 26_series | 22 | 749 | | 27_series | 1 | 17 | | 28_series | 47 | 681 | | 29_series | 97 | 2269 | | 31_series | 5 | 208 | | 32_series | 20 | 498 | | 33_series | 64 | 891 | | 36_series | 25 | 904 | | 37_series | 17 | 313 | | 38_series | 84 | 2924 | ## Source Original DOCX files are in `original/`. Converted via Datalab API. ## License See [3GPP Legal Matters](https://www.3gpp.org/specifications-technologies/legal-matters).language: - 英语 license: - 其他 license_name: - 3GPP license_link: - https://www.3gpp.org/specifications-technologies/legal-matters tags: - 电信 - 3GPP - 5G - 新空口（NR） - 长期演进（LTE） - 标准 - Release 18（第18版） pretty_name: 3GPP第18版技术规范 configs: - config_name: raw default: true data_files: - split: 21_series path: data/raw/21_series-*.parquet - split: 22_series path: data/raw/22_series-*.parquet - split: 23_series path: data/raw/23_series-*.parquet - split: 24_series path: data/raw/24_series-*.parquet - split: 26_series path: data/raw/26_series-*.parquet - split: 27_series path: data/raw/27_series-*.parquet - split: 28_series path: data/raw/28_series-*.parquet - split: 29_series path: data/raw/29_series-*.parquet - split: 31_series path: data/raw/31_series-*.parquet - split: 32_series path: data/raw/32_series-*.parquet - split: 33_series path: data/raw/33_series-*.parquet - split: 36_series path: data/raw/36_series-*.parquet - split: 37_series path: data/raw/37_series-*.parquet - split: 38_series path: data/raw/38_series-*.parquet - config_name: images data_files: - split: 21_series path: data/images/21_series-*.parquet - split: 22_series path: data/images/22_series-*.parquet - split: 23_series path: data/images/23_series-*.parquet - split: 24_series path: data/images/24_series-*.parquet - split: 26_series path: data/images/26_series-*.parquet - split: 27_series path: data/images/27_series-*.parquet - split: 28_series path: data/images/28_series-*.parquet - split: 29_series path: data/images/29_series-*.parquet - split: 31_series path: data/images/31_series-*.parquet - split: 32_series path: data/images/32_series-*.parquet - split: 33_series path: data/images/33_series-*.parquet - split: 36_series path: data/images/36_series-*.parquet - split: 37_series path: data/images/37_series-*.parquet - split: 38_series path: data/images/38_series-*.parquet # 3GPP第18版技术规范 共收录549份3GPP技术规范与技术报告，已转换为Markdown格式，并按系列分组整理。 ## 使用方法 python from datasets import load_dataset # 加载38系列的所有原始Markdown内容 raw = load_dataset("GSMA/3GPP", "raw", split="38_series") # 筛选单份技术规范 spec = raw.filter(lambda r: r["spec_id"] == "38331") print(spec[0]["text"][:500]) # 加载38系列的所有插图 imgs = load_dataset("GSMA/3GPP", "images", split="38_series") # 获取单份规范对应的插图 spec_imgs = imgs.filter(lambda r: r["spec_id"] == "38104") spec_imgs[0]["image"].show() ## 数据集结构 本数据集包含两个**配置子集**：`raw`与`images`，每个配置子集下设14个**数据拆分**（每个拆分对应一个技术系列）。 **`raw` 配置**：每份技术规范对应一行数据： | 规范ID | 标题 | 版本 | 日期 | 文本 | |---------|-------|---------|------|------| | 38104 | TS 38.104 | V18.4.0 | 2023-12 | # 3GPP TS 38.104... | **`images` 配置**：每张提取的插图对应一行数据： | 规范ID | 文件名 | 图片 | |---------|----------|-------| | 38104 | 64662465bba2..._img.jpg | PIL 图像 | ## 技术系列（数据拆分） | 数据拆分 | 规范数量 | 插图数量 | |-------|------:|-------:| | 21_series | 1 | 4 | | 22_series | 25 | 351 | | 23_series | 86 | 4539 | | 24_series | 55 | 1799 | | 26_series | 22 | 749 | | 27_series | 1 | 17 | | 28_series | 47 | 681 | | 29_series | 97 | 2269 | | 31_series | 5 | 208 | | 32_series | 20 | 498 | | 33_series | 64 | 891 | | 36_series | 25 | 904 | | 37_series | 17 | 313 | | 38_series | 84 | 2924 | ## 来源说明 原始DOCX文件存储于`original/`目录中，转换工作通过Datalab API完成。 ## 许可协议 详见[3GPP法律事项](https://www.3gpp.org/specifications-technologies/legal-matters)。