摘   要

本文在古诗词的内容上会有简单的描述，然后对于古诗集句软件进行详尽分析，确定功能需求以及非功能需求，对于软件运作流程给以流程图的展示。

在需求分析基础上，进行软件的架构设计。进一步明确使用场景与功能点的关系。在查询模块主要任务是保证查询准确高效，而在集句生成部分，主要利用现有的可行手段，例如开源的中文分词工具对于待对偶诗句进行词性分析，在此基础上，给出相对较优的对偶句。最后，对于使用的开源工具运行机理进行介绍，并对系统，尤其是对偶生成算法的改进方向提出建议。

系统最突出的是将词句匹配以及对偶句诗句生成融合，将客户端以及服务器启动程序结合，增强了程序的易用性以及使用场景的多样性。

关键字：中文古诗词；诗词匹配；对偶句生成；中文分词使用

ABSTRACT

This paper will make a simple description of the ancient poetry. Then, there is an analysis of this ancient poetry software system in detail. After the function requirements and non-function requirements are curtained, there will take the flow chart about software operation process.

The software architecture will based on the demand analysis before. Then, the paper will clarify the relationship between scenes and function points. On the query module, the task is to ensure the query is accurate and efficient. And on the pair generation module, use existing open source project named HanLP, a segmentation tool, to generate sentence pairs with readability according to the word’s property.

At last, the paper will introduce some algorithms used in the project HanLP, and show the direction of system future improvement.  System put the client and server into one packet and put the function of pair generating and poetry matching into the other packet which makes the system more reliability and easy-using.

Keyword: Chinese poetry; poetry match; Generate pairs; the usage of segmentation