An AI-based test case automatic generation system

Abstract

The present application relates to the technical field of software testing, and particularly relates to an AI-based test case automatic generation system.The present application comprises a preprocessing module, which is used for exploratory execution of an input function, capture of implicit dependency objects accessed by the function during runtime, and construction of an implicit dependency graph;identification of test-inaccessible branches of the input function, calculation of an effective coverage benchmark with the number of full branches minus the number of test-inaccessible branches as the denominator;and a test generation module, which is used for input of the input function, the implicit dependency graph and the current uncovered branch, and calling of an AI model to generate a test case set.The present application positions a minimum modifiable node set with a failure line number, and performs subset checking on the changed node set and the minimum modifiable node set after repair, thereby forcibly limiting the repair range of the AI model within the minimum statement set directly related to the current defect, and preventing repair from introducing out-of-range changes or damaging the function signature.

Images