.png){kind=icon}
Return
Git command recommendations using crowd-sourced knowledge
.png){kind=icon}
PRE
AI
DOI：10.1016/j.infsof.2023.107199.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Context: Git is a fast, scalable, distributed version control system with a rich command set that provides high-level operations and full access to the internals. It has been widely used by millions of developers worldwide. However, due to the flexibility of the usage of Git commands and the scarcity of Git documentation, many developers have experienced difficulties when using Git commands. Objective: This paper aims to propose an automatic approach to recommending Git commands for developers given a query described by natural language.Method: Our approach makes recommendations by mining the crowd-sourced knowledge related to Git on Stack Overflow. It first constructs a keyword-command mapping database from Git-related posts, then analyzes the similarity between the query given by the developer and the keywords in the database to retrieve the candidate commands, and proposes an algorithm to rank the candidate commands.Results: Our approach's recommendation results significantly outperform the baseline approaches in several metrics (e.g., Top-K accuracy). Meanwhile, the experimental results have shown that the favorable efficiency of our approach can promise its use by developers in real-world scenarios.Conclusion: The Git command recommendation approach proposed in this paper is effective and can be helpful for developers to use Git commands for more efficient development.
Keywords:
Git command
Recommendations
Stack overflow
Journal
IF:
4.3
Papers:
3.8K
Citations:
7.7K
Organization
Cited Papers
.png){kind=icon}
ShareSave
ShareSaveShareSaveShareSaveShareSave
View more