Project team: Mario Blažević, Jovana Brestovac, Nikola Domazet, Mario Dubravac, Marko Kujundžić, Teo Manojlović, Dea Marin, Nino Miškić-Pletenac, Igor Opačak, Nikola Pećarina, Nataša Prodić, Matej Raguzin, Emil Rubinić, Diego Sušanj, Ozren Šejić, Ivan Todorović, Armando Vega, Juraj Zorić
System Description: Data collection tool of open source projects for the purpose of software defect prediction (SDP). The tool collects the defect reports and the source code, offers a set of bugs-code linking techniques, calculates a number of software product metrics and generates reports for further analyses.
Project Motivation: There exists a number of approaches to data collection for SDP research, techniques to bug-code linking and software product metrics. Each variation in data collection procedure inevitably leads to data bias. This tool will offer a systematic data collection procedure applicable to a wide number of open source projects.
Project Goals: Make a tool that offers all the functionalities necessary to create the SDP datasets from the bug tracking and the source code management repositories. Allow the user to be unaware of all the underlying issues that might impose data bias and automate the process and make the otherwise time-consuming and difficult tasks less demanding.
Development of BuCo Analyzer tool was divided into several subprojects. The technical details of each are described in posters given below:
- Basic functionalities
- A set of bug-code linking techniques: Simple search, Authorship correspondence, Time correlation, Keyword inclusion/exclusion, Regular expression search
- Report generation
- Number of bugs per file calculation for each release
- Advanced functionalities
- ReLink interface for advanced technique of bug-code linking
- Bugzilla interface for easier access to the list of projects
- Automatic release tag recognition from GIT repository
- Automatic finding of GIT repository location for a project from Bugzilla