tenfingers/powershell-docs-rlat

PowerShell Docs — rlat知识模型是一个基于MicrosoftDocs/PowerShell-Docs构建的预建知识模型，专为检索增强的编码助手和LLM上下文注入而设计。目标用户是使用PowerShell 5.1或7.x进行脚本编写的Windows管理员、DevOps工程师和开发人员。数据集提供两种变体：远程版本在首次查询时从GitHub获取源文本并缓存本地；捆绑版本是完全自包含的，源文本打包在.rlat文件中，支持永久离线使用。两种变体在检索质量上完全相同，使用相同的编码器（BGE-large-en-v1.5）和压缩技术（zstd）。数据集包含使用说明、示例查询、语料来源、许可信息以及保持数据新鲜的更新方法。

The PowerShell Docs — rlat knowledge model is a prebuilt knowledge model based on MicrosoftDocs/PowerShell-Docs, designed for retrieval-augmented coding assistants and LLM context injection. The target audience includes Windows admins, DevOps engineers, and developers scripting with PowerShell 5.1 or 7.x. The dataset offers two variants: a remote version that fetches source text from GitHub on the first query and caches it locally, and a bundled version that is fully self-contained with source text packed inside the .rlat file for offline use. Both variants provide identical retrieval quality, using the same encoder (BGE-large-en-v1.5) and compression technique (zstd). The dataset includes usage instructions, sample queries, corpus provenance, licensing details, and methods to keep the dataset fresh.