Abstract: An aggregate representation of a collection of source code examples is constructed. The collection includes positive examples that conform to a coding practice and negative examples do not conform to the coding practice. The aggregate representation includes nodes corresponding to source code elements, and edges representing relationships between code elements. Using an iterative analysis of the aggregate representation, a rule to automatically detect non-conformance is generated. The rule is used to provide an indication that a set of source code is non-conformant.

Abstract: Code may be analyzed according to natural language descriptions of coding practices. A practice for code written in a natural language description may be received. An embedding of the natural language description may be generated using a machine learning model trained to detect examples of practices. The embedding may be compared with embeddings of code portions stored in an index to detect one or more portions of code that satisfy a facet of the practice. The detected portions of code may be identified.