Hypermodelling Live: OLAP for Code Clone Recommendation

Code bases contain often millions lines of code. Code recommendation systems ease programming by proposing developers mined and extracted use cases of a code base. Currently, recommender systems are based on hardcoded sets what makes it complicate to adapt them. Another research area is adaptable live detection of code clones. We advance clone detection and code recommender systems by presenting utilization of our Hypermodelling approach to realize an alternative technique. This approach uses Data Warehousing technology that scales for big data and allows for flexible and adaptable queries of source code. We present the generic idea to advance recommendation and clone detection based on queries and to evaluate our application with industry source code. Consequently, recommender systems and clone detection can be customized with flexible queries via Hypermodelling. This enables further research about more complex clone detection and context sensitive code recommendation.