A Pragmatic Decision Making Mechanism for Reusable Source Code Snippets Retrieval using Semantic Approach

Programmers frequently reuse exiting code available on internet sources like GitHub and Stack Overflow. Usually, programmers search for relevant/desired code by using natural language queries through search engines. However, the reusable code is very different from the normal textual searches. The main aim of code retrieval is to search for the most relevant snippet from a corpus of code snippets. The only way to efficiently retrieve the most relevant results is to eliminate the semantic gap between the code snippets and the user’s query (Natural language description) search in the repository. Bridging this gap is an immense challenge. The primary objective of the research is to retrieve code snippets that are semantically relevant to the user query and to provide programmers with the ability to locate source code that they can use when developing new applications. The proposed approach is implemented using a web platform through which users can search source code. In this research, a semantic model for code retrieval is used which generates meanings or synonyms of the words. The proposed model integrates Ontologies and Natural Language Processing. The proposed approach is evaluated on the dataset extracted from Stack Overflow. The performance measures and classification accuracy are computed using Mean Average Precision. The retrieved results are ranked using Levenshtein similarity, demonstrating that our approach outperforms a robust component matching significantly. Our evaluation shows that the use of semantic matching leads to improve retrieval effectiveness. This study marks a substantial advancement in the integration of programming expertise with code retrieval techniques. Moreover, this system enables users to know when and how it is used for successful semantic searching.