Test Case Generation of Composite Web Services Based on Semantic Matching and Condition Recognition

With the rapid development of Internet technology, the application of Web service and their combinations is spreading widely as a vital role. However, due to the black box feature of Web services, test cases of Web service can be obtained only from the perspective of users, which is more challenging than traditional tests. We propose a method based on semantic matching and condition recognition, combined with document parsing to generate test cases of composite Web services. In this method, the first step is to parse the relevant XML documents, then to obtain the parameter types, keywords, as well as conditions and orders in the control flow. The following step is to match the parameter instances according to the conditions in the process or the semantic knowledge base. The last is to encapsulate them into test cases. Experimental results show that our method can generate test cases with low redundancy and high coverage, which can cover more than 85% nodes and paths in BPEL.