KESI: our first component for the ALERT project
As part of the work of URJC (LibreSoft) in the ALERT project, whose aim is to increase the efficiency of the developers in libre software projects, we are about to present the first iteration of the Knowledge Extractor for Structured Information, aka KESI. This component with a very complicated name has a simple mission, that is to gather information from source code repositories and from issue/bug tracking systems and to send it to the rest of the components of the ALERT platform.
The complexity of this component is the different structure of the information offered by the resources. Every code repository and tracker uses its own format to export information, then in order to include a new type of repo/tracker in the KESI it is necessary to create a new parser. As an example, a big percent of the information extracted from a remote issue tracker system needs to be parsed from HTML text, which is not the ideal format.
Once the information has been gathered by the KESI, it transforms it into a format known by the rest of the components of the platform and then publish it using a dedicated bus.
These are the main KESI features:
- supports CVS, Subversion and GIT
- supports Jira, Sourceforge and Bugzilla
- it performs an incremental analysis
- it publish the changes detected (events) to the rest of the platform
The information obtained by the KESI is critical in the following scenarios:
- recommend a developer which bug to solve
- detect duplicated bugs
- let the developer know about buggy parts of the code
- identifying inactive developers and orphaned parts of the code
We are looking forward to see it live with the rest of the platform!
[This entry is part of the work I do in LibreSoft and it is also available in my blog at libresoft.es]