École de technolgie supérieure
Towards automatic context-aware summarization of code entities
Abstract
dc:descriptionSoftware developers are working with different methods and classes and in order to understand those that perplex them and–or that are part of their tasks, they need to tackle with a huge amount of information. Therefore, providing developers with high-quality summaries of code entities can help them during their maintenance and evolution tasks. To provide useful information about the purpose of code entities, informal documentation (Stack Overflow) has been shown to be an important source of information that can be leveraged. In this study, we investigate bug reports as a type of informal documentation and we apply machine learning to produce summaries of code entities (methods and classes) in bug reports. In the proposed approach, code entities are extracted using a technique in a form of an island parser that we implemented to identify code in bug reports. Additionally, we applied machine learning to select a set of useful sentences that will be part of the code entities’ summaries. We have used logistic regression as our machine learning technique to rank sentences based on their importance. To this aim, a corpus of sentences is built based on the occurrence of code entities in the sentences belonging to bug reports containing the code entities in question. In the last step, summaries have been evaluated using surveys to estimate the quality of produced summaries. The results show that the automatically produced summaries can reduce time and effort to understand the usage of code entities. Specifically, the majority of participants found summaries extremely helpful to decrease the understanding time (43.5%) and the effort to understand the code entities (39.1%). In the future, summaries can be produced by using other informal documentation such as mailing lists or stack overflow, etc. Additionally, the approach can be applied in practical settings. Consequently, it can be used within an IDE such as Eclipse to assist developers during their software maintenance and evolution tasks.
Degree
thesis:*- Name thesis:degree_name
- Mémoire de maîtrise électronique
- Level thesis:degree_level
- masters
- Grantor dc:publisher
- École de technolgie supérieure
- Year dc:date
- 2019
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Mohsenzadeh Korayem, Elmira
Subjects
dc:subject × 1Rights
- Language dc:language
- en
Identifiers
dc:identifier.*- Identifier
- INSERT-YOUR-LIBRARY-ID-HERE2433