Source code analytics platform using program analysis and information retrieval

摘要

In one embodiment, a code analytic platform may use a novel combination of information retrieval and program analysis techniques to develop a code relationship graph 514 to perform various code applications, such as intent based searches on a source code set, the documentation of undocumented code, risk analyses, natural language semantic searches, and others. A source code analytics platform may perform a code analysis of a source code set 410. The source code analytics platform may perform a metadata analysis of a code production data set 430 associated with the source code set 410. The source code analytics platform may create a code relationship graph 514 associating the source code set 410 with a descriptive metadata set.

主权项

1. A machine-implemented method, comprising:
applying an information retrieval module to a source code set and a code production data set associated with the source code set to retrieve a node artifact set for the source code set; executing a program analysis module to perform at least one of a code analysis of the source code set and a metadata analysis of the code production data set to refine the node artifact set; executing a machine learning module to correlate the node artifact set; creating a code relationship graph associating the source code set with a descriptive metadata set describing at least one representative element of the source code set; receiving a user natural language search entry; determining a user intent based on the user natural language search entry; searching the source code set based on the code relationship graph; and ranking a code result set based on the code relationship graph.