A Method for Automatically Generating Code Change Log Based on Nearest Neighbor Algorithm

Abstract

The invention discloses a code change log automatic generation method based on the nearest neighbor algorithm, which belongs to the field of code change log automatic generation. The method includes: preprocessing of input data, preprocessing of training set data, obtaining a set of word frequency vector pairs through bag-of-words model, calculating candidate intermediate results through KNN algorithm, calculating BLEU‑4 values, and finally obtaining output results. This method has the characteristics of simple model structure, strong interpretability, no training of the model, greatly reduced actual running time compared with NMT, insensitivity to noise, and strong robustness.

Description

technical field [0001] The invention belongs to the field of automatic generation of code change logs, in particular to a method for automatically generating code change logs based on the nearest neighbor algorithm. Background technique [0002] Currently, techniques for code changelogs have been extensively studied. DeltaDoc takes source code changes as input, obtains path predicates through symbolic execution, and then uses a set of predefined rules and transformations to generate commit messages. ChangeScribe first extracts the necessary information by analyzing the corresponding source code changes and the abstract syntax tree. It then populates a predefined template with the extracted information as a log of the commit. NMT is the first to apply deep neural networks to its field by using neural machine translation algorithms to automatically generate streamlined code change logs from source code. [0003] However, the following problems generally exist in the current...

AI Technical Summary

Problems solved by technology

[0004] (1) The model structure is complex and lacks interpretability;

[0005] (2) Model training and running time are expensive;

[0006] (3) Sensitive to noise and not robust

Images