Search

However, searching for software components on GitHub is currently inefficient, as searches are based on text strings. This paper introduces OntoGitHubSearch as a search model that allows the retrieval and classification of software architectures stored on GitHub. This proposed model retrieves all repositories that match the searched architectural concepts; to provide the search with semantics and context, it implements the domainspecific ontology Architecture Ontology Version 2.0, as well as a natural language processing module to analyze the text in the repository. The repositories are automatically identified and classified according to the text and concepts found in their descriptions. To evaluate the proposed model, we developed a web application called WebOntoGitHubSearch, which allows interaction with users during the search process. The evaluation of the model was carried out with the participation of software developers and architects from several colombian businesses. Information retrieval metrics such as Precision at k. The results obtained during the evaluation process are promising and allow verifying the effectiveness of the proposed model. [ABSTRACT FROM AUTHOR]