企业文档智能中的问题解析器：从用户输入提取关键词、范围、形状、分解与澄清
英文摘要
This tutorial presents a question parser for enterprise document intelligence that extracts five field families directly from a user’s query: keywords, scope, shape, decomposition, and clarification. The article provides code implementations for each extraction category. The parser is part of a larger system aimed at structuring user intent to improve document retrieval. By parsing these fields, the system can better interpret complex questions and guide downstream processes.
中文摘要
本教程介绍了一款面向企业文档智能的问题解析器，可从用户查询中直接提取五类字段：关键词、范围、形状、分解和澄清。文章给出了每类字段提取的代码实现。该解析器是一个更大系统的一部分，旨在结构化用户意图以提高文档检索效果。通过解析这些字段，系统能更好地理解复杂问题并指导下游处理。
关键要点
The parser extracts five field families: keywords, scope, shape, decomposition, and clarification.
解析器提取五类字段：关键词、范围、形状、分解和澄清。
Code snippets for each extraction family are included in the tutorial.
教程中包含了每类字段提取的代码示例。
It is part of a series on Enterprise Document Intelligence, focused on question parsing.
该内容是企业文档智能系列的一部分，专注于问题解析。