Application development methods, devices, electronic devices and storage media

Abstract

This application discloses an application development method, apparatus, electronic device, and storage medium, belonging to the field of software development technology, to solve the problem that related application development technologies place high demands on developers. The method includes: obtaining target requirement information describing a target application through a low-code development platform; generating a corresponding DSL file for the low-code development platform based on the target requirement information and a trained application development model; wherein the DSL file includes one or more of a first DSL sub-file related to the interface of the target application, a second DSL sub-file related to the business logic of the target application, and a third DSL sub-file related to the workflow of the target application; and generating the target application through the low-code development platform based on the DSL file.

Description

Technical Field

[0001] This application belongs to the field of software development technology, specifically relating to an application development method, apparatus, electronic device, and storage medium. Background Technology

[0002] In software development, there are generally two approaches: high-code development using native programming languages ​​and low-code development based on low-code platforms. High-code development is a traditional software development model that relies on specialized programming languages ​​and tools to build applications. It offers strong flexibility and customizability, allowing developers to implement complex functions and business logic by writing intricate code. High-code development combined with AI has given rise to AI-assisted programming tools, such as Copilot, which directly generate code using large model technology to improve development efficiency. Low-code platforms typically provide a complete set of development tools and rich templates and components, enabling developers to quickly build customized applications based on specific needs.

[0003] High-code development involves complex programming languages ​​and development tools, requiring developers to have strong programming skills and extensive project experience. This technical requirement presents significant challenges in talent recruitment and training. Furthermore, the need for dedicated technical support and maintenance teams necessitates substantial investment in human resources. While AI can improve programming efficiency to some extent, its generated code is often confined to a single file due to limitations in the number of tokens, resulting in limited context. Humans need to finely break down tasks. However, even with fine-grained breakdown, the quality of AI-generated code is often unsatisfactory, making direct application difficult to maintain and limiting efficiency gains.

[0004] During low-code platform development, as the complexity of requirements increases, the complexity of using the low-code platform also rises. The final configuration complexity increases rapidly, requiring extremely high training costs to achieve proficiency. Ultimately, development costs are simply converted into training costs. Worse still, low-code platforms are not code-free; even with a highly complex platform, developers still need to write a large number of scripts to meet the requirements.

[0005] In other words, the relevant application development technologies have the problem of requiring high skill levels from developers. Summary of the Invention

[0006] This application provides an application development method, apparatus, electronic device, and storage medium, which can solve the problem that related application development technologies place high demands on developers.

[0007] In a first aspect, embodiments of this application provide an application development method, the method comprising: acquiring target requirement information describing a target application through a low-code development platform; generating a domain-specific language (DSL) file corresponding to the low-code development platform based on the target requirement information and a trained application development model; wherein the DSL file includes one or more of a first DSL sub-file related to the interface of the target application, a second DSL sub-file related to the business logic of the target application, and a third DSL sub-file related to the workflow of the target application; and generating the target application based on the DSL file through the low-code development platform.

[0008] Secondly, embodiments of this application provide an application development apparatus, the apparatus comprising: an acquisition module, configured to acquire target requirement information describing a target application through a low-code development platform; a first generation module, configured to generate a corresponding DSL file for the low-code development platform based on the target requirement information and a trained application development model; wherein the DSL file includes one or more of a first DSL sub-file related to the interface of the target application, a second DSL sub-file related to the business logic of the target application, and a third DSL sub-file related to the workflow of the target application; and a second generation module, configured to generate the target application based on the DSL file through the low-code development platform.

[0009] Thirdly, embodiments of this application provide an electronic device comprising: a processor; and a memory arranged to store computer-executable instructions configured to be executed by the processor, the executable instructions including instructions for performing the application development method as described in the first aspect.

[0010] Fourthly, embodiments of this application provide a storage medium for storing computer-executable instructions that cause a computer to perform the application development method as described in the first aspect.

[0011] Fifthly, embodiments of this application provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the application development method as described in the first aspect.

[0012] In a sixth aspect, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the application development method as described in the first aspect.

[0013] In this embodiment, a low-code development platform is used to obtain target requirement information describing the target application. Based on the target requirement information, a corresponding DSL file for the low-code development platform is generated using a trained application development model. The DSL file includes one or more of the following: a first DSL sub-file related to the target application's interface; a second DSL sub-file related to the target application's business logic; and a third DSL sub-file related to the target application's workflow. The low-code development platform then generates the target application based on the DSL file. Compared to related application development technologies, this application, through the natural language processing capabilities of the trained application development model, allows users to describe requirements in natural language and automatically generate the target application directly, eliminating the need for more complex coding processes. This reduces the requirements for developers, accelerates the development cycle, and makes application deployment faster; it solves the problem of related application development technologies having high requirements for developers. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating an application development method provided in an embodiment of this application; Figure 2 A flowchart illustrating another application development method provided in this application embodiment; Figure 3 This is a schematic diagram of the structure of an application development device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0017] The application development method, apparatus, electronic device, and storage medium provided in this application will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios.

[0018] Figure 1 This illustration shows an application development method provided by an embodiment of the present invention. The method can be executed by an electronic device, which may include a server and / or a terminal device, wherein the terminal device may be, for example, an in-vehicle terminal or a mobile phone terminal. The method includes the following steps: S102: Obtain target requirement information to describe the target application through a low-code development platform.

[0019] The target application is the application that the user wants to develop. Users include software developers, etc.

[0020] In practical applications, users can input target requirements information describing the intended application into the low-code development platform. This target requirements information can be in natural language format. The low-code development platform then receives this target requirements information; that is, it obtains the target requirements information through the low-code development platform.

[0021] S104: Based on the target requirements information, generate the corresponding Domain Specific Language (DSL) file for the low-code development platform through the trained application development model.

[0022] The DSL file includes one or more of the following: a first DSL sub-file related to the target application's interface, a second DSL sub-file related to the target application's business logic, and a third DSL sub-file related to the target application's workflow.

[0023] Specifically, the target requirement information can be input into the trained application development model, which will then generate one or more of the first, second, and third DSL sub-files. It should be noted that the first, second, and third DSL sub-files can be part of a single file, or they can be three independent files.

[0024] S106: Generate the target application based on the DSL file using a low-code development platform.

[0025] In practical applications, the target application can be generated from the DSL file using a low-code development platform.

[0026] Furthermore, during the process of generating the target application from the DSL file through the low-code development platform, a visual interface can be provided to users, allowing them to quickly build the target application by dragging and dropping components and configuring parameters, thus enabling non-professional developers to quickly build the target application.

[0027] The application development method provided in this invention uses a low-code development platform to obtain target requirement information describing the target application. Based on the target requirement information, a trained application development model generates a corresponding DSL file for the low-code development platform. The DSL file includes one or more of the following: a first DSL sub-file related to the target application's interface; a second DSL sub-file related to the target application's business logic; and a third DSL sub-file related to the target application's workflow. The low-code development platform then generates the target application based on the DSL file. Compared to related application development technologies, this application, through the natural language processing capabilities of the trained application development model, allows users to describe requirements in natural language and directly and automatically generate the target application without a more complex coding process. This reduces the requirements for developers, accelerates the development cycle, and makes application deployment faster; it solves the problem of related application development technologies having high requirements for developers.

[0028] In one implementation, before generating the corresponding domain-specific language DSL file (i.e., S104) for the low-code development platform based on the target requirement information and the trained application development model, steps A1 to A2 can be performed to obtain the trained application development model: Step A1: Obtain a predefined sample DSL file for generating the application, and obtain the requirement information for describing the application.

[0029] The sample DSL file includes one or more of the sample first DSL sub-file, sample second DSL sub-file, and sample third DSL sub-file.

[0030] Specifically, predefined sample DSL files can be stored in a lightweight data-interchange format (JavaScript Object Notation, JSON).

[0031] Step A2: Based on the sample DSL file and requirement information, train the pre-trained large language model to obtain the trained application development model.

[0032] In this embodiment, a set of sample DSL files is predefined for the low-code platform, including one or more of the sample first DSL sub-files related to the application's interface, sample second DSL sub-files related to the application's business logic, and sample third DSL sub-files related to the application's workflow, thereby improving the efficiency and accuracy of subsequent development.

[0033] In one implementation, based on the sample DSL file and requirement information, the pre-trained large language model is trained to obtain the trained application development model (i.e., step A2), and steps B1 to B2 can be executed as follows: Step B1: Based on the requirement information, obtain the corresponding historical DSL files and / or logs from the development and operations platform through a pre-trained large language model and Retrieval-augmented Generation (RAG).

[0034] Logs are related to tests performed on the application. Development and operations platforms, such as DevOps platforms, do not impose specific restrictions on this.

[0035] Step B2: Based on historical DSL files and / or logs, as well as sample DSL files and requirement information, train the pre-trained large language model to obtain the trained application development model.

[0036] Historical DSL files serve as a private knowledge context. Logs can also serve as a context.

[0037] In the actual training process, the requirement information can be input into the pre-trained large language model. The pre-trained large language model is generated through retrieval enhancement, automatically obtaining one or more of the corresponding historical DSL files and logs from the development and operation platform. Then, based on the historical DSL files and / or logs, as well as the sample DSL files and requirement information, a first prompt word can be generated. The first prompt word is then input into the pre-trained large language model, so that the pre-trained large language model can learn to obtain the trained application development model. Alternatively, steps C1 to C5 can be shown below.

[0038] In this embodiment, the corresponding historical DSL files are obtained from the development and operation platform using RAG, and combined with requirement information and logs, the pre-trained large language model is trained to obtain a more accurate application development model.

[0039] In one implementation, the pre-trained large language model is trained based on historical DSL files and / or logs, as well as sample DSL files and requirement information, to obtain the trained application development model (i.e., step B2). Steps C1 to C5 can also be performed as follows: Step C1: Based on the page layout-related DSL in the first DSL sub-file of the sample and the natural language description information corresponding to the page layout-related DSL in the requirement information, train the pre-trained large language model to obtain the first large language model.

[0040] In practical applications, a second cue word can be used to train a pre-trained large language model through single-sample learning. Specifically, the second cue word can include page layout-related DSLs and natural language descriptions of the requirements information corresponding to the page layout-related DSLs.

[0041] Considering that page layouts follow certain rules, in order to ensure the aesthetics of the pages corresponding to the final DSL files generated by the trained application development model, as well as the stability of colors, fonts, and tooltip page sizes, we can pre-set page layout-related DSLs for multiple pages, or pre-set second prompt words for multiple pages, and train the pre-trained large language model.

[0042] Step C2: Based on historical DSL files, demand information, logs, sample first DSL sub-files, and sample second DSL sub-files, train the first large language model to obtain the second large language model.

[0043] The second language model is used to perform cumulative reasoning (CR) based on historical DSL files, demand information, logs, and sample first DSL sub-files to generate sample second DSL sub-files.

[0044] Since the business logic enhances the page, the first sample DSL sub-file also needs to be input into the first large language model. The first large language model performs cumulative inference based on historical DSL files, requirement information, logs, and the first sample DSL sub-file to generate the business logic script corresponding to the requirement information. Specifically, a third prompt word can be generated based on historical DSL files, requirement information, logs, the first sample DSL sub-file, and the second sample DSL sub-file. This third prompt word is then input into the first large language model to perform cumulative inference and learning, thereby obtaining the second large language model.

[0045] Step C3: Based on the page type-related DSL, page navigation-related DSL, navigation element-related DSL in the first sample DSL sub-file, the second sample DSL sub-file, and the requirement information, train the second language model to obtain the third language model.

[0046] The second language model can already generate DSLs corresponding to several pages. However, in low-code development platforms, a complete application typically consists of multiple pages, each with its specific function and purpose. Therefore, in this step, it is also necessary to train the second language model to decompose and combine applications.

[0047] Specifically, based on the page type-related DSL, page navigation-related DSL, and navigation element-related DSL in the first sample DSL sub-file, the second sample DSL sub-file, and requirement information, a second large language model can be trained to recognize different page types, understand the role of different page types in the application and the navigation relationships between pages, as well as identify navigation elements, understand the navigation role of different navigation elements between pages, and combine multiple pages into a complete application according to business logic and requirement information; page types, such as homepage, detail page, list page, form page, etc.; navigation relationships between pages, such as how to jump from one page to another; navigation elements, such as buttons, links, or menu items, etc.

[0048] Step C4: Based on the workflow category-related DSL in the sample third DSL sub-file and the natural language description information corresponding to the workflow category-related DSL in the requirement information, train the third language model to obtain the trained application development model.

[0049] In low-code development platforms, workflow is a core component of the application, defining the interaction process between the user and the system. A workflow includes, but is not limited to, one or more of the following: triggering conditions, task execution, decision nodes, and termination conditions. Workflow types include, but are not limited to, linear workflows, parallel workflows, and conditional workflows.

[0050] The workflow category-related DSLs in the third DSL subfile of the sample, as well as the natural language description information corresponding to the workflow category-related DSLs in the requirement information, can correspond to various different workflows. This allows the trained application development model to learn workflows for different business scenarios corresponding to different requirement information. Furthermore, the trained application development model can understand data flow within the workflow, ensuring that it can correctly transmit and process data during task execution.

[0051] Of course, after training the third language model in step C4 above, a fourth model can be obtained. Then, the sample third DSL subfile can be combined with the sample second DSL subfile related to the application's business logic and the requirement information to train the fourth language model to understand how to translate the requirement information into specific workflow steps, generating dynamic workflows to adapt to different business scenarios, thus obtaining the trained application development model. Alternatively, the fourth language model can be trained based on the sample first DSL subfile, sample second DSL subfile, sample third DSL subfile, and requirement information to obtain the aforementioned trained application development model.

[0052] In addition, steps C2 to C4 above can also use prompt words to train the corresponding large language model through single-sample learning, similar to step C1 above, and will not be elaborated here.

[0053] In this embodiment, by training the pre-trained large language model in relation to the application's business logic, interface, and workflow, the DSL file automatically generated by the trained application development model is both efficient and meets user expectations, thereby improving the automation level and user experience of the electronic device executing this application (e.g., a low-code development platform).

[0054] In one implementation, based on the target requirement information, a corresponding domain-specific language (DSL) file (i.e., S104) for the low-code development platform is generated using the trained application development model, and the following steps D1 to D3 can be executed: Step D1: Input the target requirement information into the trained application development model, and generate the first DSL sub-file through the trained application development model.

[0055] Step D2: Using the trained application development model, generate the second DSL sub-file based on the first DSL sub-file and the target requirement information.

[0056] Step D3: Using the trained application development model, generate the third DSL sub-file based on the first DSL sub-file, the second DSL sub-file, and the target requirement information.

[0057] Of course, step D3 can generate a third DSL sub-file based on the target requirement information using the trained application development model. Alternatively, the trained application development model can combine the second DSL sub-file related to business logic with the target requirement information to generate the third DSL sub-file. Thus, the trained application development model can generate dynamic workflows to adapt to different business scenarios.

[0058] Optionally, the first DSL sub-file generated in step D1 can be output immediately after the trained application development model completes step D1, and the second DSL sub-file generated in step D2 can be output immediately after the trained application development model completes step D2; alternatively, the second DSL sub-file, the first DSL sub-file, and the third DSL sub-file can be output only after step D3 is executed.

[0059] In one implementation, steps E1 to E2 can also be performed to generate a test report: Step E1: Push the DSL file to the development and operations platform, and in response to the push of the DSL file, generate corresponding test cases based on the trained application development model.

[0060] The trained application development model ensures that the output DSL file syntax is correct and that the corresponding program can run normally. However, considering the possibility of business errors in the DSL file, it is necessary to adjust the DSL file to meet the target requirements. Specifically, before executing step E1, the following steps can also be performed: The system provides the DSL file to the user through a pre-defined visualization platform, receives the user's corrections to the DSL file, obtains the corrected DSL file, and generates the target application based on the corrected DSL file.

[0061] Accordingly, steps E1 and E2 can be performed on the revised DSL file.

[0062] Step E2: Based on test cases and DSL files, conduct testing and analysis through the development and operation platform and the trained application development model, generate test reports, and store the test reports in the development and operation platform.

[0063] The test report includes test cases and logs related to the tests performed on the target application.

[0064] Specifically, the development and operations platform can simulate different user operations and scenarios to execute these test cases, verifying the performance of the target application corresponding to the DSL file under various conditions. During test execution, the trained application development model monitors the target application's response and behavior, collects test results, and analyzes any errors or anomalies that occur. A detailed test report can then be generated. Optionally, the test report may include test cases that failed and their corresponding logs, and / or test cases that passed and their corresponding logs. This provides developers with clear feedback to quickly locate and correct problems.

[0065] The test report is then stored on the development and operations platform so that it can be retrieved in step B1 and added to the training set of the newly trained application development model. This continuously improves the accuracy of the application development model.

[0066] In addition, the above-mentioned trained application development model may also include a pre-trained test case generation model, through which the above steps E1 to E2 are executed.

[0067] In this embodiment, test cases are automatically generated through the trained application development model, and the quality and performance of the target application are ensured through automated testing. Feedback is collected after test execution to generate a test report of the target application. The logs in the test report are used as the training set for subsequent training, thereby continuously optimizing the performance of the application development model.

[0068] Figure 2 This is a flowchart illustrating another application development method provided in an embodiment of this application. For example... Figure 2 As shown, the method includes: Step 202: Obtain a predefined sample DSL file for generating the application, and obtain the requirement information for describing the application.

[0069] The sample DSL file includes one or more of the sample first DSL sub-file, sample second DSL sub-file, and sample third DSL sub-file.

[0070] Step 204: Based on the page layout-related DSL in the first DSL subfile of the sample and the natural language description information corresponding to the page layout-related DSL in the requirement information, train the pre-trained large language model to obtain the first large language model.

[0071] Step 206: Based on the requirement information, obtain the corresponding historical DSL files and logs from the development and operation platform through the pre-trained large language model and RAG.

[0072] The logs are related to the tests performed on the application. The historical DSL files serve as a private knowledge context.

[0073] Step 208: Based on historical DSL files, demand information, logs, sample first DSL sub-files, and sample second DSL sub-files, train the first large language model to obtain the second large language model.

[0074] The second language model is used to perform cumulative inference (CR) based on historical DSL files, demand information, logs, and sample first DSL sub-files to generate sample second DSL sub-files.

[0075] Step 210: Based on the page type-related DSL, page navigation-related DSL, navigation element-related DSL in the first sample DSL sub-file, the second sample DSL sub-file, and the requirement information, train the second language model to obtain the third language model.

[0076] Step 212: Based on the workflow category-related DSL in the sample third DSL sub-file and the natural language description information corresponding to the workflow category-related DSL in the requirement information, train the third language model to obtain the trained application development model.

[0077] Step 214: Obtain target requirement information to describe the target application through a low-code development platform.

[0078] Step 216: Based on the target requirement information, generate the corresponding DSL file for the low-code development platform using the trained application development model.

[0079] The DSL file includes one or more of the following: a first DSL sub-file related to the target application's interface, a second DSL sub-file related to the target application's business logic, and a third DSL sub-file related to the target application's workflow. Step 218: Push the DSL file to the development and operations platform, and in response to the push of the DSL file, generate corresponding test cases based on the trained application development model.

[0080] Step 220: Based on test cases and DSL files, conduct testing and analysis through the development and operation platform and the trained application development model, generate test reports, and store the test reports in the development and operation platform.

[0081] The test report includes test cases and logs related to the tests performed on the target application.

[0082] The specific processes of steps 202 to 220 above have been described in detail in the above embodiments, and will not be repeated here.

[0083] In this embodiment, a low-code development platform is used to obtain target requirement information describing the target application. Based on the target requirement information, a corresponding DSL file for the low-code development platform is generated using a trained application development model. The DSL file includes one or more of the following: a first DSL sub-file related to the target application's interface, a second DSL sub-file related to the target application's business logic, and a third DSL sub-file related to the target application's workflow. The target application is then generated using the low-code development platform based on the DSL file. Compared to related application development technologies, this application, through the natural language processing capabilities of the trained application development model, allows users to describe requirements in natural language and directly and automatically generate the target application without a more complex coding process. This reduces the requirements for developers, accelerates the development cycle, and makes application deployment faster; it solves the problem of related application development technologies having high requirements for developers.

[0084] Corresponding to the application development method provided in the above embodiments, based on the same technical concept, the present invention also provides an application development apparatus. Figure 3 This is a schematic diagram of an application development apparatus according to an embodiment of the present invention, which is used to execute... Figures 1 to 2 The application development method described, such as Figure 3 As shown, the application development device includes: an acquisition module 310, a first generation module 320, and a second generation module 330.

[0085] Module 310 is used to obtain target requirement information describing the target application through a low-code development platform. The first generation module 320 is used to generate a corresponding DSL file for a low-code development platform based on the target requirement information and through a trained application development model; wherein, the DSL file includes one or more of the following: a first DSL sub-file related to the interface of the target application, a second DSL sub-file related to the business logic of the target application, and a third DSL sub-file related to the workflow of the target application. The second generation module 330 is used to generate the target application based on the DSL file through a low-code development platform.

[0086] In one implementation, the application development apparatus further includes a training module. The training module includes: The acquisition unit is used to acquire a predefined sample DSL file for generating the application and to acquire requirement information for describing the application; wherein the sample DSL file includes one or more of the sample first DSL sub-file, sample second DSL sub-file, and sample third DSL sub-file; The training unit is used to train the pre-trained large language model based on sample DSL files and requirement information to obtain the trained application development model.

[0087] In one implementation, the training unit is specifically used for: Based on the requirements information, the corresponding historical DSL files and / or logs are obtained from the development and operations platform through pre-trained large language models and retrieval enhancement generation; among them, the logs are related to the tests performed on the application. Based on historical DSL files and / or logs, as well as sample DSL files and requirement information, a pre-trained large language model is trained to obtain a trained application development model; among them, historical DSL files serve as private knowledge context.

[0088] In one implementation, a pre-trained large language model is trained based on historical DSL files and / or logs, as well as sample DSL files and requirement information, to obtain a trained application development model, including: Based on the page layout-related DSL in the first DSL sub-file of the sample, and the natural language description information corresponding to the page layout-related DSL in the requirement information, the pre-trained large language model is trained to obtain the first large language model. Based on historical DSL files, demand information, logs, sample first DSL sub-files, and sample second DSL sub-files, the first large language model is trained to obtain the second large language model; the second large language model is used to perform cumulative inference CR based on historical DSL files, demand information, logs, and sample first DSL sub-files to generate sample second DSL sub-files. Based on the page type-related DSL, page navigation-related DSL, navigation element-related DSL in the first sample DSL sub-file, the second sample DSL sub-file, and the requirement information, the second language model is trained to obtain the third language model. Based on the workflow category-related DSLs in the sample third DSL sub-file and the natural language description information corresponding to the workflow category-related DSLs in the requirement information, the third language model is trained to obtain the trained application development model.

[0089] In one implementation, the first generation module 320 is specifically used for: Input the target requirement information into the trained application development model, and generate the first DSL sub-file through the trained application development model; Based on the first DSL sub-file and the target requirement information, the trained application development model generates the second DSL sub-file. The trained application development model generates a third DSL subfile based on the first DSL subfile, the second DSL subfile, and the target requirement information.

[0090] In one implementation, the application development apparatus further includes a testing module. The testing module is specifically used for: The DSL file is pushed to the development and operations platform, and in response to the push of the DSL file, corresponding test cases are generated based on the DSL file and the trained application development model. Based on test cases and DSL files, testing and analysis are performed through the development and operation platform and the trained application development model to generate test reports, which are then stored in the development and operation platform. The test reports include test cases and logs related to the tests performed on the target application.

[0091] In this embodiment, a low-code development platform is used to obtain target requirement information describing the target application. Based on the target requirement information, a corresponding DSL file for the low-code development platform is generated using a trained application development model. The DSL file includes one or more of the following: a first DSL sub-file related to the target application's interface, a second DSL sub-file related to the target application's business logic, and a third DSL sub-file related to the target application's workflow. The target application is then generated using the low-code development platform based on the DSL file. Compared to related application development technologies, this application, through the natural language processing capabilities of the trained application development model, allows users to describe requirements in natural language and directly and automatically generate the target application without a more complex coding process. This reduces the requirements for developers, accelerates the development cycle, and makes application deployment faster; it solves the problem of related application development technologies having high requirements for developers.

[0092] Those skilled in the art will understand that the above-described application development apparatus can be used to implement the application development method described above, and the detailed description therein should be similar to the method description in the preceding text. To avoid repetition, it will not be repeated here.

[0093] Based on the same technical concept, this application also provides an electronic device for executing the above-described application development method. Figure 4This is a schematic diagram of the structure of an electronic device to implement various embodiments of this application. The electronic device can vary significantly due to differences in configuration or performance, and may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440. The processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call a computer program stored in the memory 430 and executable on the processor 410 to perform the following steps: Use a low-code development platform to obtain target requirement information that describes the target application; Based on the target requirements information, the corresponding DSL file for the low-code development platform is generated through the trained application development model; wherein, the DSL file includes one or more of the following: a first DSL sub-file related to the interface of the target application, a second DSL sub-file related to the business logic of the target application, and a third DSL sub-file related to the workflow of the target application. The target application is generated based on the DSL file using a low-code development platform.

[0094] In this embodiment, a low-code development platform is used to obtain target requirement information describing the target application. Based on the target requirement information, a corresponding DSL file for the low-code development platform is generated using a trained application development model. The DSL file includes one or more of the following: a first DSL sub-file related to the target application's interface, a second DSL sub-file related to the target application's business logic, and a third DSL sub-file related to the target application's workflow. The target application is then generated using the low-code development platform based on the DSL file. Compared to related application development technologies, this application, through the natural language processing capabilities of the trained application development model, allows users to describe requirements in natural language and directly and automatically generate the target application without a more complex coding process. This reduces the requirements for developers, accelerates the development cycle, and makes application deployment faster; it solves the problem of related application development technologies having high requirements for developers.

[0095] The specific execution steps can be found in the various steps of the above application development method embodiments, and can achieve the same technical effect. To avoid repetition, they will not be repeated here.

[0096] It should be noted that the electronic devices in the embodiments of this application include: servers, terminals, or other devices besides terminals.

[0097] The above electronic device structure does not constitute a limitation on the electronic device. An electronic device may include more or fewer components than illustrated, or combine certain components, or arrange them differently. For example, an input unit may include a Graphics Processing Unit (GPU) and a microphone, and a display unit may use a liquid crystal display (LCD), organic light-emitting diode (OLED), or other similar display panels. User input units include at least one of a touch panel and other input devices. A touch panel is also called a touchscreen. Other input devices may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be elaborated further here.

[0098] Memory can be used to store software programs and various data. Memory can primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area can store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, memory can include volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0099] The processor may include one or more processing units; optionally, the processor integrates an application processor and a modem processor, wherein the application processor mainly handles operations related to the operating system, user interface, and applications, while the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into the processor.

[0100] This application also provides a storage medium storing computer-executable instructions. When these computer-executable instructions are executed by a processor, they implement the various processes of the above-described application development method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0101] The processor is the processor in the electronic device described in the above embodiments. The storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0102] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above application development method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0103] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0104] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the various processes of the above-described application development method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0105] It should be understood that the training and prediction processes of the AI ​​models involved in the various embodiments of this specification all adhere to multiple legal and compliant principles, including legal data sources, compliant data content, compliant data governance, compliant training objectives and schemes, compliant training processes, compliant training environments and tools, and compliant ethical verification of training results, and comply with the requirements of Article 5 of the Patent Law. Among them: Data source legitimacy: All datasets used for AI model training were obtained through legal means, covering three categories: publicly authorized data, data authorized by partners, and self-collected compliant data. Publicly authorized data comes from compliant data sources following open-source licenses such as Apache 2.0, with complete copyright attribution and authorization scope clearly marked, and no unauthorized open-source code or data reuse. Data authorized by partners has been subject to formal data usage agreements, clearly defining the scope, duration, and confidentiality obligations, and possessing a complete authorization chain. For self-collected data involving personal information, strict informed consent procedures have been followed, and anonymization processes (including but not limited to field masking, feature anonymization, and differential privacy technology applications) have been implemented to remove personally identifiable information, fully complying with the requirements of relevant laws and regulations such as the "Interim Measures for the Administration of Generative Artificial Intelligence Services" and the "Personal Information Protection Law."

[0106] Data content compliance: The AI ​​model's dataset undergoes multiple screenings and cleaning processes to remove all content that may violate social morality or harm public interests. It contains no obscene, pornographic, violent, discriminatory, or information that endangers national or public safety, nor does it involve the illegal acquisition or use of genetic resources. For data in sensitive fields (such as healthcare and finance), an additional privacy-preserving computation module (including federated learning and secure multi-party computation technologies) ensures that the data is "usable but not visible," avoiding compliance risks during the original data transmission process and ensuring that the data application scenarios and uses comply with public order and good morals and industry regulatory requirements.

[0107] Data governance norms: A complete data traceability system is established during the AI ​​model training process to automatically record the source, collection time, annotation process, cleaning rules, and permission allocation of training data, generating traceable compliance reports to ensure that the data is verifiable throughout its entire lifecycle. The dataset annotation process for AI models is completed by a professional human R&D team, clearly defining the proportion of human creative contributions and avoiding reliance on AI-generated data that has not undergone substantial human modification, thus meeting the examination requirements for "human main contributions" in AI patent applications.

[0108] Training objectives and plans are compliant: The AI ​​model training objective focuses on software development. The training scheme and the final output results do not violate any mandatory provisions of laws and administrative regulations, do not harm the public interest or the legitimate rights and interests of others, and do not pose any potential risks of being used for illegal activities, infringing on privacy, or disrupting public safety. The training strictly adheres to the ethical principle of "intelligent for good".

[0109] Training process compliance: A closed-loop training framework is adopted to ensure compliance and controllability of the training process. The specific process is as follows: First, training samples are obtained through compliant data sources. After the aforementioned data cleaning and desensitization, they are input into the neural network model to generate preliminary training results. Second, an expert system is introduced to verify the preliminary results. Based on preset rules and human expert experience, the feasibility of the results is evaluated, and outputs that may pose ethical risks or compliance hazards are corrected (such as removing decision-making logic that violates public order and good morals, and adjusting model parameters that do not comply with safety regulations). Finally, the loss function weights are dynamically optimized based on expert system feedback to strengthen the model's learning of compliant results, avoid overfitting errors or non-compliant labels, and form a closed-loop control of "data input - model training - expert verification - parameter optimization - result feedback" to ensure that the entire training process complies with A5 ethical review requirements.

[0110] Training environment and tool compliance: AI model training is implemented using nationally licensed chips and a compliant training platform. All open-source frameworks and components used in the training process have obtained their corresponding licenses, and copyright statements and patent citation information are fully retained, with no instances of infringement or reuse. The training environment is built using virtual devices (containers / virtual machines) with fixed random seeds and initial parameter configurations to ensure the reproducibility of the training process. Furthermore, through access control and operation log recording, risks such as data leakage and parameter tampering during training are prevented, ensuring the security and compliance of the training process.

[0111] Training results ethical verification compliance: After the model is trained, it undergoes additional third-party ethical compliance assessment and algorithm filing review to verify that the model output does not violate social morality or harm public interests. For potentially sensitive scenarios (such as public services and intelligent decision-making), a special result verification mechanism is established to ensure that the model always complies with Article 5 of the Patent Law and relevant laws and regulations in practical applications.

[0112] In summary, the data and training process used in the AI ​​model of this specification strictly comply with the relevant provisions of Article 5 of the Patent Law and the Patent Examination Guidelines (2023 Edition), and there are no violations of laws, social ethics, public interests, or illegal use of genetic resources. It fully meets the compliance requirements for patent authorization.

[0113] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include multitasking and parallel processing according to the functions involved, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0114] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods of the various embodiments of this application.

[0115] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. An application development method, characterized in that, The method includes: Use a low-code development platform to obtain target requirement information that describes the target application; Based on the target requirement information, a domain-specific language (DSL) file for the low-code development platform is generated through the trained application development model; wherein, the DSL file includes one or more of the following: a first DSL sub-file related to the interface of the target application, a second DSL sub-file related to the business logic of the target application, and a third DSL sub-file related to the workflow of the target application. The target application is generated based on the DSL file using a low-code development platform.

2. The method according to claim 1, characterized in that, Before generating the corresponding domain-specific language (DSL) file for the low-code development platform based on the target requirement information and the trained application development model, the method further includes: Obtain a predefined sample DSL file for generating the application, and obtain requirement information for describing the application; wherein, the sample DSL file includes one or more of a sample first DSL sub-file, a sample second DSL sub-file, and a sample third DSL sub-file; Based on the sample DSL file and the requirement information, the pre-trained large language model is trained to obtain the trained application development model.

3. The method according to claim 2, characterized in that, Based on the sample DSL file and the requirement information, the pre-trained large language model is trained to obtain the trained application development model, including: Based on the aforementioned requirements, a RAG is generated using the pre-trained large language model and retrieval enhancement, and corresponding historical DSL files and / or logs are obtained from the development and operation platform; wherein, the logs are related to the tests performed on the application; The pre-trained large language model is trained based on the historical DSL file and / or the log, as well as the sample DSL file and the requirement information, to obtain the trained application development model; wherein the historical DSL file serves as a private knowledge context.

4. The method according to claim 3, characterized in that, The step of training the pre-trained large language model based on the historical DSL file and / or the log, as well as the sample DSL file and the demand information, to obtain the trained application development model includes: Based on the page layout-related DSL in the first DSL sub-file of the sample, and the natural language description information in the requirement information corresponding to the page layout-related DSL, the pre-trained large language model is trained to obtain the first large language model. Based on the historical DSL file, the demand information, the log, the first sample DSL sub-file, and the second sample DSL sub-file, the first large language model is trained to obtain the second large language model; wherein, the second large language model is used to perform cumulative inference (CR) based on the historical DSL file, the demand information, the log, and the first sample DSL sub-file to generate the second sample DSL sub-file; Based on the page type-related DSL, page navigation-related DSL, navigation element-related DSL in the first DSL sub-file of the sample, the second DSL sub-file of the sample, and the requirement information, the second major language model is trained to obtain the third major language model; Based on the workflow category-related DSL in the sample third DSL sub-file and the natural language description information corresponding to the workflow category-related DSL in the requirement information, the third language model is trained to obtain the trained application development model.

5. The method according to claim 1, characterized in that, The step of generating the corresponding domain-specific language (DSL) file for the low-code development platform based on the target requirement information and through the trained application development model includes: The target requirement information is input into the trained application development model, and the first DSL sub-file is generated through the trained application development model. Based on the trained application development model, the second DSL sub-file is generated according to the first DSL sub-file and the target requirement information. The third DSL subfile is generated using the trained application development model based on the first DSL subfile, the second DSL subfile, and the target requirement information.

6. The method according to claim 1, characterized in that, The method further includes: The DSL file is pushed to the development and operation platform, and in response to the push of the DSL file, corresponding test cases are generated based on the DSL file and the trained application development model. Based on the test cases and the DSL file, testing and analysis are performed through the development and operation platform and the trained application development model to generate a test report, which is then stored in the development and operation platform. The test report includes the test cases and logs related to the testing of the target application.

7. An application development apparatus, characterized in that, The device includes: The acquisition module is used to acquire target requirement information describing the target application through a low-code development platform. The first generation module is used to generate a corresponding DSL file for the low-code development platform based on the target requirement information and through a trained application development model; wherein, the DSL file includes one or more of the following: a first DSL sub-file related to the interface of the target application, a second DSL sub-file related to the business logic of the target application, and a third DSL sub-file related to the workflow of the target application. The second generation module is used to generate the target application based on the DSL file using a low-code development platform.

8. An electronic device, characterized in that, include: processor; as well as A memory configured to store computer-executable instructions configured to be executed by the processor, the executable instructions including instructions for performing the application development method as described in any one of claims 1-6.

9. A storage medium, characterized in that, The storage medium is used to store computer-executable instructions that cause a computer to perform the application development method as described in any one of claims 1-6.

10. A computer program product, characterized in that, It includes a computer program that, when executed by a processor, implements the application development method as described in any one of claims 1-6.

Images