Design And Implementaion Of Semantic Analysis And Annotation Function For Restful Services

Recently, RESTful services grow rapidly. Because of their features of lightweight and simplicity, they have been used broadly on the Internet and have continually attracted attentions of big companies which used to employ WS*services as mainstream technique. From giants like Google, Amazon and Yahoo etc. to start-up companies most of them adopt RESTful services to implement open apis of their web sites. The statistic of ProgrammableWeb, the famous Web services portal site, shows that RESTful services have accounted for69%of the number of open Web services. Clearly, RESTful services have already become the mainstream of Web services techniques.Since RESTful services are published in the form of Web pages and their functions, interfaces and calling ways are described in natural language, it is hard to distinguish the description pages of RESTful services and general API documents from their appearance. Because of that, discovering and choosing fitting services from enormous non-relevant pages is a challenging work.To enable machine automatically recognizing and retrieving description of RESTful services, we propose an approach of semantic analysis and annotation for RESTful services. Our method can transform ordinary documents of RESTful services into semantic annotated and machine-readable description of services.In aspect of semantic analysis, we adopt techniques of IR and IE to analyze RESTful services pages semantically. The operations include preprocessing pages, segmenting pages, preprocessing text and semantic analyzing text. We adopt text categorization algorithm in IR as the basis of text semantic analysis algorithm, and acquire the features of RESTful services documents by observation to improve the basic algorithm. Finally, we introduce the semantic analysis algorithm of RESTful services. The experiment result shows that the precision and recall of our method both reach more than80%.In aspect of semantic annotation, we use hRESTS as our basic annotation language and add boundary information into hRESTS to meet the actual needs of annotating highly heterogeneous original documents.Based on the above method, we present the design and implementation of semantic analysis and annotation sub-system for RESTful services. First, we analyze the functional requirements of the system. Then we present the overall architecture and module interaction of the system, and detail the functional interfaces and procedure of every module. Finally, we demonstrate the system test and prove that our system can function normally.