блочный тест, сгенерированный системой искусственного интеллекта(1) Code is parsed into our internal representation. 2) A semantic analysis constructs a formula which captures the behaviour of (part of) the program. 3) We automatically generate a series of queries about the code, such as which inputs are needed to reach line R, take branch B or to cause various kinds of exception. 4) The AI engine uses automated reasoning and learning to understand the code and find answers to the queries. 5) Answers from the AI engine are translated back into an execution trace of the program and these are used to build smart unit tests (with high coverage). Alex_Odeychuk)
