DeepTC-enhancer

Abstract

Automated test case generation tools have been successfully proposed to reduce the amount of human and infrastructure resources required to write and run test cases. However, recent studies demonstrate that the readability of generated tests is very limited due to (i) uninformative identifiers and (ii) lack of proper documentation. Prior studies proposed techniques to improve test readability by either generating natural language summaries or meaningful methods names. While these approaches are shown to improve test readability, they are also affected by two limitations: (1) generated summaries are often perceived as too verbose and redundant by developers, and (2) readable tests require both proper method names but also meaningful identifiers (within-method readability).In this work, we combine template based methods and Deep Learning (DL) approaches to automatically generate test case scenarios (elicited from natural language patterns of test case statements) as well as to train DL models on path-based representations of source code to generate meaningful identifier names. Our approach, called DeepTC-Enhancer, recommends documentation and identifier names with the ultimate goal of enhancing readability of automatically generated test cases.An empirical evaluation with 36 external and internal developers shows that (1) DeepTC-Enhancer outperforms significantly the baseline approach for generating summaries and performs equally with the baseline approach for test case renaming, (2) the transformation proposed by DeepTC-Enhancer results in a significant increase in readability of automatically generated test cases, and (3) there is a significant difference in the feature preferences between external and internal developers.

InterEvo-TR: Interactive Evolutionary Test Generation With Readability Assessment

“…Firstly, the number of participants (39) is aligned with previous studies in SBST [14], [15], and clearly superior to what is frequent in iSBSE [17]. In this sense, we consider that the sample size is adequate for the type of qualitative analysis presented.…”

“…For step 3, the actual interactive experiment, we chose ArrayIntList as the class under test. This is a well-known class of medium complexity used in other SBST experiments with participants related to test readability improvement [14], [15]. Due to the time required to complete the experiment and the cognitive burden of manually revising many tests, all participants used one class under test.…”

“…We extended the invitations to professional developers among our industrial contacts, as well as student collaborators. In total, we obtained positive answers from 39 participants (similar to other studies [14], [15]), divided into: 4 Bachelor's students, 4 Master's students, 7 PhD students, 10 faculty members with programming skills and 14 professional developers from industry. As part of the survey, we collected some basic information about their experience programming in Java and testing software in general.…”

Selectively Combining Multiple Coverage Goals in Search-Based Unit Test Generation

“…Braione et al [53] combined symbolic execution and SBST for programs with complex inputs; (3) readability of generated tests: Daka et al [54] proposed to assign names for tests by summarizing covered coverage goals. Roy et al [55] introduced deep learning approaches to generate test names; (4) fitness function design: Xu et al [56] proposed an adaptive fitness function for improving SBST. Rojas et al [4] proposed to combine multiple criteria to satisfy users' requirements.…”

Students' perception of medical professionalism at the University of Hong Kong

“…Zhou et al [4] proposed a method to select coverage goals from multiple criteria instead of combining all goals; (4) Readability of created tests: Daka et al [70] suggested naming tests by stating covered goals. Deep learning techniques were presented by Roy et al [71]; (5) Applying SBST to more software fields such as Machine Learning libraries [72], Android applications [73], Web APIs [74], and Deep Neural Networks [75].…”