| With the extremely rapid development of the Internet, the construction of the information infrastructure in the enterprise is also speeding up. Handling information in the enterprise is becoming more and more difficult, traditional ways of accessing information in the enterprise are out of date. General search engine can help people get what they want from the Internet whereas the information inner an enterprise is kind of special, so we cannot use a general search engine directly in an enterprise. Generally speaking, information search in an enterprise has the following difficulties, first are the organize and store of multi-source and heterogeneous data, second is handling the comprehensiveness and precision of the search content, third is the personalized search and finally is the security.In this paper, we put forward a framework of enterprise search engine, then we analysis and implement a prototype of a distributed enterprise search engine based on solr, the prototype can offer some help for the deeper study of the enterprise search engine. Besides, we analysis the navigation of a certain kind of documents in reality, then implement an important function of the enterprise search- the navigation of the search result. In general, the main work of this paper are as follows:1. Introduce the principle conception of the search engine and graph database and then study the core technology of search engine and graph database deeply, we also analysis the system architecture of the enterprise search engine.2. Put forward a framework of enterprise search engine, analysis and implement a distributed enterprise search engine prototype based on solr, define and implement every component of the system.3. Analysis and define the navigation of a certain kind of documents formally and proposed two different ways to solve the subset of this problem-the result navigation of the enterprise search, this can be an implementation reference of this kind of applications.At last, we test the search time and navigation precision of the system use some text dataset, the experiments prove that the system achieve the design goal. |
