Source code generation method

Abstract

The invention relates to a source code generation method. The source code generation method comprises the steps of: S1, storing a special string sample, a general string sample corresponding to the special string sample, and a generate code segment sample corresponding to the general string sample in a database, wherein the special string sample is the subset of the general string sample; S2, retrieving the special string sample matching a string inputted by peripheral equipment in the database, executing step S3 is the special string sample is retrieved, otherwise, returning to step S1; and S3, retrieving the general string sample corresponding to the special string sample obtained in step S2 in the database, comparing the special string sample with the general string sample, and generating a first mapping relationship between the special string sample and the general string sample. According to the source code generation method provided by the invention, the natural language strings can automatically generate the code segments, thereby reducing the work loads of software development and meeting different use demands of users.

Description

technical field [0001] The invention relates to a source code generation method using natural language as input. Background technique [0002] At present, natural language processing technology is gradually beginning to be practical (for example: programming tools based on natural language, voice control system, etc.), but it is still at a relatively early stage. [0003] Existing natural language processing techniques usually just map a natural language string into a function. For example: map "send SMS" to a function in the mobile phone. Of course, it is also easy to map a natural language string into a sequence of functions, because this is equivalent to wrapping these functions into a new function. [0004] The essence of the above implementation methods is to establish a mapping table between "string" and "code segment", and when the "string" is input, execute the corresponding "code segment". Of course, this mapping can be many-to-many, for example: the same "string...

AI Technical Summary

Problems solved by technology

[0005] However, these methods have the following disadvantages: the "code segment" is pre-written by the engineer, which means that the engineer has to pre-determine which "strings" the user will input. Since a large number of users will have a large number of differentiated needs, so , these methods obviously do not meet the differentiation trend of demand