18 results about "Programming style" patented technology

The invention discloses a code reconstruction method, system and device based on function recognition and a medium, and the method specifically comprises the steps: scanning a project file topological structure, and constructing a knowledge graph based on a code import relation and a technology stack feature; performing context enhancement by analyzing the code function description and combining with the knowledge graph to generate a code modification instruction set; based on the code modification instruction set, in combination with data stream sensitivity marking and performance portrait analysis, evaluating the target code to obtain an evaluation result; based on an evaluation result, matching a historical optimal practice mode through a graph neural network, and generating an optimization scheme which retains an original design style; by comparing abstract syntax tree differences between original codes and verified optimization schemes, only functional logic nodes are replaced, and legal code style features are reserved. According to the method, the efficiency and quality of code reconstruction and optimization are improved, and a more convenient, efficient and intelligent code processing tool is provided for software developers.

The invention discloses a code generation model training method and electronic equipment, and relates to the technical field of computers, and the method comprises the following steps: inputting a sample cue word into a code generation model to obtain a sample code; based on code operation information of the sample code, determining a first reward parameter used for representing execution stability, and analyzing the code operation information, the sample code and the sample cue word by using a reference language model to obtain a second reward parameter used for representing a grammar error degree and a third reward parameter used for representing a code style; adjusting model parameters of a code generation model based on the first reward parameter, the second reward parameter and the third reward parameter; and under the condition that the adjusted code generation model reaches a convergence condition, determining a target code generation model according to the adjusted code generation model. The technical problem that the training efficiency of the code generation model is low is solved, and the technical effect of improving the training efficiency of the code generation model is achieved.

The invention discloses an integrated circuit design Verilog code generation method and device based on a large language model, equipment and a medium, and relates to the technical field of computers, and the method comprises the steps that semantic analysis is conducted on an input language of an integrated circuit user side through the large language model, and a target constraint set is determined based on an obtained performance index set and through a regular expression; a candidate architecture scheme is generated by utilizing thinking chain technology reasoning, the candidate architecture scheme is predicted by utilizing an XGBoost regression model, and a target architecture scheme is determined based on an index prediction result and the candidate architecture scheme by utilizing a non-dominated sorting genetic algorithm; and determining an initial Verilog code based on the target architecture scheme and a preset code template library, and performing code style conversion on the initial Verilog code based on a code style feature corresponding to the integrated circuit user side to obtain a target Verilog code. And the efficiency and availability of integrated circuit design are improved.

The invention provides an AI-based code review opinion automatic classification and priority ranking method. The method comprises the steps of collecting code review historical data; data preprocessing: preprocessing the collected data, and constructing a training data set; an AI model is selected and trained; inputting a new review opinion; automatic classification: the model automatically classifies the review suggestions, and classification categories include but are not limited to code errors, code style problems, performance optimization suggestions and potential safety hazard prompts; performing priority ranking: performing priority ranking on the review opinions based on a preset priority rule and the classification result; and generating result output. According to the code review method and device, automatic classification and priority ranking of the code review opinions are achieved through the AI model, a traditional manual processing mode is replaced, the problems that manual classification is low in efficiency and priority ranking is high in subjectivity are effectively solved, and the overall efficiency and result accuracy of code review are improved.

The invention discloses a task alignment Text-to-SQL (Structured Query Language) method and system based on a code style thinking chain. The method comprises the following steps: constructing a three-layer structured knowledge base comprising mode-level knowledge, instance-level knowledge and example-level knowledge; generating an initial query based on the target database and the user query; according to the list information in the initial query, constructing an instruction sequence represented by a similar SQL code, and forming a code style thinking chain; based on user query and initial query, acquiring multi-dimensional knowledge from the three-layer structured knowledge base and constructing an enhanced prompt context; and integrating the code style thinking chain and the enhanced prompt context based on the prompt template, and generating an executable final query through a multi-stage task alignment mechanism and a final prompt. According to the method, the problems of incoherence and inaccuracy of logical reasoning, insufficient knowledge generalization ability and multi-type model illusion existing in cross-domain, multi-table connection and complex query in the prior art are solved.

The invention discloses a method for dynamically constructing a UDF based on a large model. The method comprises the following steps: S1, constructing a multi-dimensional requirement cue word of a user-defined function according to characteristics and use requirements of a target platform; s2, the multi-dimensional requirement cue words are input into a pre-trained large language model for code generation, the large language model is subjected to multi-round iteration based on the multi-dimensional requirement cue words, and an optimal code sequence for achieving the user-defined function is output; s3, code style standardization is carried out on the optimal code sequence, and a compiler or an interpreter and a static analysis tool are respectively called to carry out grammar check and static analysis on the optimal code sequence, and constructing an optimization cue word according to an error log output by the compiler or the interpreter and a static analysis result output by the static analysis tool, and inputting the optimization cue word into the large language model to carry out iterative optimization on the optimal code sequence.

Some embodiments provide or utilize code-quality-focused replicated data type functionality for collaborative source code editing support. Some embodiments combine replicated data types with automated quality assessments of potential resolutions of source code update conflicts. The replicated data types support strong eventual consistency of copies of source code, even without a repository or other centralized facility. The quality assessments apply source code quality criteria, such as syntactic correctness, consistency with other source code, programming style criteria, and other measures of source code quality. Deterministic resolution rules are applied, such as last-write-wins, grow-only, or observed-remove, with various replicated data types. Some embodiments prompt a machine learning model for snippet quality assessments, for recommended snippets, or for both.

The invention discloses an interface testing method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining basic information of a to-be-tested interface and description information of a test case; on the basis of the basic information and the description information, constructing cue words; processing the cue word through a large language model, and generating a test script of the to-be-tested interface; and executing the test script to obtain test data of the to-be-tested interface, analyzing the test data through an artificial intelligence agent AI Agent to obtain an analysis result, and using the analysis result to optimize the test script and / or the test case. In this way, the test script can be quickly and automatically generated, the tedious process of traditional manual script writing is avoided, the generation efficiency of the test script is greatly improved, the problem of logic errors or inconsistent code styles possibly occurring during manual writing is avoided, and the quality and normalization of the test script are improved.

The invention provides an intelligent low-code development control method based on multi-modal interaction driving, and belongs to the technical field of development control, and the method comprises the steps: 1, constructing a front-end subsystem for low-code development control; and step 2, constructing a back-end subsystem for low-code development control. Aiming at the problems that a traditional low-code system component is low in dragging efficiency and difficult to adapt to a complex business scene, automatic generation of a prototype graph / voice / character-front-end interface and detailed design document-back-end code is realized through deep fusion of a visual large model, a code large model and a multi-modal interaction technology, and the method is high in practicability and high in practicability. The purposes of reducing the development time by more than 50%, reducing the code error rate and unifying the code style are achieved, and an efficient and high-quality solution is provided for enterprise software development.

The invention provides a code generation method and device, electronic equipment, a storage medium and a program product. The method comprises the following steps: acquiring a pre-generated standardized project structure; wherein the language adopted by the standardized project structure is a GO language; and on the basis of the standardized project structure, generating a code corresponding to each development component in the project in a declarative manner. According to the method, software development is carried out under the standardized project structure, and automatic generation is carried out in a declarative mode for each development component, so that on one hand, the code generation efficiency is improved, on the other hand, the generated codes are consistent in style, and the readability and maintainability of the codes are improved.

The application discloses a personalized code generation system and method based on residual learning. 1) Collecting code snippet data sets of programming languages, collecting style sequences corresponding to the code snippets through a code detection tool; 2) Setting residual terms based on the code of step 1) and the corresponding style sequence, and generating training data for residual learning; 3) Setting a learnable vector representation for each code style specification; 4) Building a residual learning framework through an LLM; 5) Obtaining the vector representation of each code style specification through residual learning of the residual learning framework of step 4); 6) Obtaining the style sequence of the code to be generated through a code detection tool, using the trained vector representation of the code style specification, and generating personalized code consistent with the historical code style. The application combines the code features learned through residual learning to form the code style of the developer, and guides the large language model to generate personalized code through the code style features.

The invention relates to the technical field of computers, and discloses an instruction optimization method and device based on an abstract syntax tree, equipment and a medium, according to the method, the abstract syntax tree is generated by analyzing a Vue template source code, code logic can be more accurately understood, and a list rendering instruction can be more accurately positioned. The abstract syntax tree is traversed to identify the statement parameters and the syntax structure of the list rendering instruction, deletion detection is performed based on the syntax structure, different deconstruction rules can be dynamically adapted, whether the instruction parameters are deleted or not can be accurately judged, and the code development accuracy and development efficiency are improved. When parameter missing is detected, missing parameters are complemented according to the grammar structure, optimization of a list rendering instruction is realized, development workload is reduced, consistency of code styles is guaranteed, maintainability of a front-end transaction system of a financial institution or an information management system of a medical institution is improved, system operation errors caused by parameter missing are avoided, and the system performance is improved. And the code quality and the development efficiency are obviously improved.

The invention discloses a code review method and system based on a local large language model. The method comprises the following steps: acquiring a code file submitted by a user; code detection rules are constructed, and the code detection rules comprise a code style detection rule, a sensitive information detection rule and a security vulnerability detection rule; constructing a local large language model; a code file and a code detection rule submitted by a user are input into the local large language model for code review, and a code review report in a text format is output; and performing format conversion on the code review report in the text format to obtain a code review report in a PDF or HTML format. According to the code review method and device, the problems that a current code review method mainly depends on a large language model, a user needs to upload a code to a cloud server for processing, once the cloud server crashes or is attacked, the risk of code leakage or loss possibly occurs, and the safety of user data is reduced are solved.

This application discloses a training method and electronic device for a code generation model, relating to the field of computer technology. The method includes: inputting sample prompts into a code generation model to obtain sample code; determining a first reward parameter to characterize execution stability based on the code execution information of the sample code, and parsing the code execution information, sample code, and sample prompts using a reference language model to obtain a second reward parameter to characterize the degree of syntax errors and a third reward parameter to characterize code style; adjusting the model parameters of the code generation model based on the first, second, and third reward parameters; and determining the target code generation model based on the adjusted code generation model when the adjusted model reaches convergence. This solves the technical problem of low training efficiency in code generation models and achieves the technical effect of improving the training efficiency of code generation models.

The invention provides a C language unit test generation method based on a reference test case, and belongs to the field of software engineering. Comprising the following steps: constructing a C project symbol index knowledge base; performing test case migration association retrieval; and generating a test case based on the reference test mode. According to the method, the pairwise resource'application / release 'mode is extracted and forcibly applied from the existing test, so that the problem that memory leakage may be introduced into the generated C language test case is solved; by injecting the test mode extracted from the real code, the generated test case is highly consistent with the tested item in data structure initialization, function call agreement and code style, and the readability and maintainability of the test code are improved; the expert knowledge and the generation capability of the large language model are deeply fused, and the test environment establishment and resource cleaning work are fully automated, so that the cost and time for compiling a high-quality test for a complex C language project are remarkably reduced.

At least one unit of a software application is identified. The at least one unit includes source code. The source code of the at least one unit is analyzed to determine a style of the source code. Metadata is extracted from the at least one unit based on the source code analysis. One or more features of the extracted metadata are classified. A template file is modified based on the extracted metadata and the classified features to create a modified template file.

The embodiment of the invention provides a training method, device and equipment for generating a large model of a unit test code. A specific embodiment of the method comprises the steps of obtaining a first unit code corresponding to a first program unit and a first test code for the first program unit; constructing a first cue indicating generation of a test code for the first program unit according to the first unit code and a first cue portion indicating a code style expected to be generated or prohibited from being generated; and taking the first prompt as a training sample, taking the first test code as a training label corresponding to the training sample, and training the pre-trained target large model again through the training sample and the training label. Through the method, the quality and the stability of the generated unit test code can be greatly improved, and the labor cost consumed by debugging the unit test code is reduced.

This disclosure provides a knowledge graph question answering method based on code style context learning. For an input code style learning question and a knowledge base, a large language model is first used to obtain a function call sequence through code style context learning. A dense retrieval system is then used to link entities extracted from the function calls and to perform relation matching on the relation names extracted from the function calls. Finally, a program interpreter executes the generated function call sequence to obtain the logical form, which is then further converted into an executable query to obtain the answer to the question. This method effectively solves the formatting error problem in logical form generation through metafunction design and retrieval enhancement strategies. It not only avoids secondary training of a large model but also significantly improves generalization performance in scenarios with few or no samples.