Nickmancol/etymology

这是一个结构化、全面的、多语言的词源数据集，通过解析Wiktionary的词源部分而创建。它包含超过420万个词源关系，涉及超过200万个术语，涵盖3300多种语言或方言。数据集区分了31种不同的词源关系类型，如继承、借用等，并保留了关系的层级结构，例如术语在语言中的演变。数据集的所有信息都是直接通过半结构化文本解析从Wiktionary获取的，没有进行结果验证。

This is a structured, comprehensive, and multilingual etymology dataset created by parsing Wiktionarys etymology sections. It contains over 4.2 million etymological relationships involving more than 2.0 million terms across over 3300 languages/dialects. The dataset distinguishes between 31 types of etymological relations, such as inheritance and borrowing, and preserves hierarchical relationship structures, such as the evolution of a term across languages.