Military Literature Search Engine Architecture And Realization

As one of the basic parts of Military information construction research, military information service system and creative researching have very significant means in finding information resource, obstacle-less transition, promoting technique evolution, assisting military research, cultivating modem military, and making up elite solders. In the current days, the development and utility of information source is very important for information construction. In the information construction of military, there is a needs for a stabile internet. In order to satisfy the needs of information sharing in military, we constructed the military Intranet. In the view of current structure of military Internet, LANs, which belong to different secret classes, compose the bone web, and the information was distributed by an outer-website.During the process of intranet searching, many literatures were generated. In order to use that information effectively, the former military system stores documents in the Oracle Database, and creates index and searches documents by Oracle Text technique. The use of Oracle Database searching technique can realize the literatures researching; however, it will cost lots of system resource, and may affect the operation of Database system. Further more, Oracle's index cannot be used by other Database systems, therefore, it cannot be searched in multi Database when the literatures were stored in different Database systems. Last, Oracle Text has a low efficiency in Chinese searching, and do not support Chinese-English crocs search.On the current bases, we construct the military literature searching structure. Although we still use the Oracle Database to store literature, we do not use Oracle's index structure; instead, we design an index-service module and a search-service module to help the searching.(1) Putting forward the architecture framework of retrieval system of military literature search engine. in this framework, database and index files to achieve loosely coupled, to solve low efficiency of database flows caused by index reconstruction in previous system.(2) To ensure sync of literature and index, design data collection services using a multithreading programming approach and introduce thread pool technology in order to improve the operating efficiency of data collection. Meanwhile, introduce the mechanism of the network reptiles in data collection services module. Using similar to the concept of vertical search expands collection method of military affairs information.(3) In C++ platform, design and implement text conversion framework with plug-and-play algorithm., using C++ virtual function mechanism to access to the multi-state acts of text-type conversion. The framework of a high degree of expandability, the algorithm can be integrated into framework without affecting framework itself and application.(4) On the basis of study of existing index technology, design and implement index services and enquiry services to meet system requirements. Index services not only created index for the body of literature, but also stored information of literature attributes and relevant information of document database to the index, which provides convenience for several enquiries ways of enquiry services.