archit11/deepwiki2

DeepWiki数据集包含了从Hyperswitch支付路由项目文档中提取的章节性文档，这些文档被分块以便于使用。数据集共有68954个块，来自46个源文件，平均每个块包含798个字符。数据集按照标题级别分为三个层次：Level 1有5336个块，Level 2有22770个块，Level 3有40848个块。每个数据条目包含一个唯一标识符、源文件名、标题、标题级别、实际Markdown内容以及块索引。

The DeepWiki dataset consists of sectioned documentation extracted from the Hyperswitch payment router project, chunked for easy consumption. The dataset contains a total of 68954 chunks from 46 source files, with an average chunk length of 798 characters. It is categorized into three heading levels: Level 1 with 5336 chunks, Level 2 with 22770 chunks, and Level 3 with 40848 chunks. Each entry includes a unique identifier, source filename, heading, heading level, actual markdown content, and chunk index.