FOCUS: A Recommender System for Mining API Function Calls and Usage Patterns

Abstract

Software developers interact with APIs on a daily basis and, therefore, often face the need to learn how to use new APIs suitable for their purposes. Previous work has shown that recommending usage patterns to developers facilitates the learning process. Current approaches to usage pattern recommendation, however, still suffer from high redundancy and poor run-time performance. In this paper, we reformulate the problem of usage pattern recommendation in terms of a collaborativefiltering recommender system. We present a new tool, FOCUS, which mines open-source project repositories to recommend API method invocations and usage patterns by analyzing how APIs are used in projects similar to the current project. We evaluate FOCUS on a large number of Java projects extracted from GitHub and Maven Central and find that it outperforms the stateof-the-art approach PAM with regards to success rate, accuracy, and execution time. Results indicate the suitability of contextaware collaborative-filtering recommender systems to provide API usage patterns.

Democratizing the development of recommender systems by means of low-code platforms

“…Concerning the data preprocessing phase, we reported the most common ones used in software development, i.e., techniques that mine source code, documentation, and software projects. Such strategies have been excerpted both from existing RSSE as well as recommender systems that we have actually implemented [5,[16][17][18][19]25] in the context of the CROSSMINER project. According to the peculiar nature of data sources, one technique is more suitable rather than another.…”

“…Capturing context Producing recommendation Presenting recommendation Strathcona [7] AST parsing Keyword extraction Six heuristic functions IDE integration Sourcerer [12] AST parsing Indexing Custom ranking scheme (Heuristic) Web interface FaCoY [9] AST parsing Indexing Alternate query (Heuristic) Web interface PROMPTER [21] NLP Indexing Custom ranking model (Heuristic) IDE integration CrossSim [18] Graph representation Feature extraction Content-based filtering Web interface CrossRec [19] Graph representation Feature extraction Collaborative filtering IDE integration FOCUS [16] Tensor API calls extraction Context-aware collaborative filtering IDE integration AURORA [17] NLP Feature extraction Feed forward neural network Web interface MNB [5] Vectorization Feature extraction Bayesian Network Raw outcomes PostFinder [25] AST Parsing Indexing Heuristics IDE integration is that they reduce the development effort by avoiding complex data structures. Despite this, they may reach sub-optimal results compared to more sophisticated techniques and they should be carefully selected considering the context of the recommendations.…”

“…We have partially realized the paradigm presented in Fig. 1 in a series of recent work [16][17][18][19]. In particular, various components, which are listed in the gray rows of Table 1, have been properly implemented and evaluated using real-world datasets (marked as the bold edges in Fig.…”

“…FOCUS [16] supports the development activities by recommending API function calls. After the extraction of relevant elements from the developer's source code, the tool encodes these elements in a tensor and searches for possible matching by using a contextaware collaborative-filtering technique.…”

Detecting Java software similarities by using different clustering techniques

“…Having access to similar software projects helps developers speed up their development process. By looking at similar Open Source Software (OSS) projects, for example, developers are able to learn how relevant classes are implemented, and in some certain extent, to reuse useful source code [6,7,9].…”