Design And Implemention Of Multifunctional Retrieval Surrogate Based On Linux

Information retrieval is an important application in our using compute works, it’sfunction and usage show the efficiency of computer workers. Referring to the baiduHard disk searcher, this thesis designs and implements the desktop information retrievalsurrogate based on linux. This surrogate implements searching native files based onfulltext retrieval and electronic dictionary based on WEB.Firstly, this paper gives a compare between existing desktop searching software andthe commonly used electronic dictionary under Linux, elaborates that the informationretrieval has the meaning to the Linux users, and all kinds of technology involved in thesystem are separately introduced in details. Then the thesis discusses the design andimplement of the native full-text retrieval, the WEB electronic dictionary, and thesurrogate server separately in details. The native full-text retrieval part achieved thecreation of native documents, segmentation, inverted index, snapshot structure andquick search arithmetic. These parts focus on how to search key word in mass dataefficiently, how to organize and search inverted index efficiently and how to organizeand search snapshot. According to the form of StarDic dictionary, the web dictionaryinterface part designed quickly searching arithmetic and combined the interface with thesystem by http protocol. And this part emphasized on full use of library web resourcesavailable online dictionary. And the last part builded the web server by usingBerkeley sockets to receive user request. The server deal with mass request by means ofthread pool technology and HTTP1.1is the application protocol in the server.