A code search engine for software ecosystems

Abstract

Searching and reusing source code play an increasingly significant role in the daily tasks of software developers. While code repositories, such as GitHub and Stackoverflow, may provide some results, a code search engine is generally considered most helpful when searching for code snippets as they typically crawl data from a wide range of code repositories. Code search engines enable software developers to search for code snippets using search terms. The accuracy of the search results can be increased if the searchers' intent can be modeled and predicted correctly. This study proposes a novel code search engine to model user intents through a dialogue system and then suggests a ranked list of code snippets that can meet user requirements.