Recommending source code examples via API call usages and documentation
Collin McMillan, Denys Poshyvanyk, Mark Grechanik
Abstract
Collin McMillan, Denys Poshyvanyk, Mark Grechanik
Abstract
Online source code repositories contain software projects that already implement certain requirements that developers must fulfill. Programmers can reuse code from these existing projects if they can find relevant code without significant effort. We propose a new method to recommend source code examples to developers by querying against Application Programming Interface (API) calls and their documentations that are fused with structural information about the code. We conducted an empirical evaluation that suggests that our approach is lightweight and accurate.
OpenAlex reports 63 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Online source code repositories contain software projects that already implement certain requirements that developers must fulfill. Programmers can reuse code from these existing projects if they can find relevant code without significant effort. We propose a new method to recommend source code examples to developers by querying against Application Programming Interface (API) calls and their documentations that are fused with structural information about the code. We conducted an empirical evaluation that suggests that our approach is lightweight and accurate.
Key concepts: Computer science, Source code, KPI-driven code analysis, Application programming interface, Documentation, Internal documentation, Reuse, Code review