Grouping Test Results with the Common Root Cause Using String Similarity Algorithms

被引：0

作者：

Kramar, Vladimir T. ^{[1
]}

Nurminen, Jukka K. ^{[1
]}

Aalto, Tatu ^{[2
]}

机构：

[1] Univ Helsinki, Helsinki, Finland

[2] F Secure, Helsinki, Finland

来源：

PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON INNOVATIONS IN COMPUTING RESEARCH (ICR'22) | 2022年 / 1431卷

关键词：

String similarity; Pytest; Testing; Software logs; Test logs; Machine learning;

D O I：

10.1007/978-3-031-14054-9_21

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper presents an innovative concept of test log categorization that demonstrates how five different string similarity algorithms such as Pythons built-in diff library, Jaccard index, Jaro-Winkler distance, cosine similarity and Levenshtein ratio are applied on test logs from pytest - one of the popular assertion-based test frameworks. In order to minimize manual, error-prone work for software engineers of analyzing multiple test runs daily, these test logs are grouped into the following three distinctive categories; C1 - has similar failures in the same test, C2 - has a similar failure in two different tests and C3 - has a different failure in two same tests for easier root-cause analysis and fixes. The presented work demonstrates how efficient the string similarity algorithms can be.

引用

页码：214 / 224

页数：11