Program source code automatic generation function descriptive annotation method

Abstract

The invention provides a program source code automatic generation function descriptive annotation method. The method comprises the steps of 1, extracting features of source codes; and 2, analyzing theextracted features, and expressing the features about functions automatically by using a natural language. According to the method, corresponding function descriptive annotations can be automaticallygenerated for various different program design languages; the effects and functions of the source codes are well expressed; and the maintenance and development of software can be effectively promoted. Compared with the prior art, the method has the following beneficial effects that 1, an analysis tree of the source codes is processed and the features are extracted, so that the features, not merely the local partial features, of the source codes can be represented more comprehensively in comparison with other methods; 2, the annotation generation no longer depends on a fixed template, so thatthe flexibility and diversity are higher in comparison with other methods, the generated annotation can express more functions, and the description is more comprehensive; and 3, the method is suitablefor various program design languages, and has universality.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method for automatically generating function descriptive annotations for program source codes. Background technique [0002] With the rapid development of the Internet, software development has become more and more popular and accompanied by a large number of open source code libraries. Reading and understanding other people's source code in the software development process is a time-consuming and labor-intensive task. If the source code can have corresponding functional descriptive comments, this process will be greatly simplified and the speed of software development will be accelerated. At the same time, perfect comments can also improve the maintainability of the software system. However, in the current situation, less than 20% of the code has corresponding comments, so the majority of program developers urgently need a method to automatically generate functional descrip...

AI Technical Summary

Problems solved by technology

Iyer (see Iyer, S.; Konstas, I.; Cheung, A.; and Zettlemoyer, L. 2016. Summarizing source code using ing a neural attention model. In ACL, 2073–2083) proposed a method using neural networks The method of training data to generate annotations, but this method still treats the source code as ordinary text, and does not extract any unique features and characteristics of programming languages, which exist in the source code feature extraction defect

Images