Application generation method and device based on large language model, equipment and storage medium

Abstract

The invention provides an application generation method and device based on a large language model, equipment and a storage medium, and relates to the field of artificial intelligence, in particular to the field of intelligent search. According to the specific implementation scheme, in response to a received application generation request, according to application description information in the application generation request, industry application knowledge matched with the application field of the application description information is determined; the industry application knowledge and the application description information are input into the large language model, application development information is output, and the application development information comprises design information used for indicating an application control and development information used for achieving the function of the application control; and according to the application development information, generating a target application matched with the application description information.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence, in particular to the field of intelligent search, and specifically to an application generation method, device, equipment and storage medium based on a large language model. Background Art

[0002] With the development of artificial intelligence, application development is also developing continuously. As business scenarios become more complex and business needs become more diversified, applications are currently usually developed in an intelligent generation way. For example, by building the functional modules required for application development in a visual building platform, applications that meet user needs can be generated. Summary of the invention

[0003] The present disclosure provides an application generation method, device, electronic device, storage medium and program product based on a large language model.

[0004] According to one aspect of the present disclosure, a method for generating an application based on a large language model is provided, comprising:

[0005] In response to the received application generation request, determining industry application knowledge matching the application field of the application description information according to the application description information in the application generation request;

[0006] Inputting the above-mentioned industry application knowledge and the above-mentioned application description information into a large language model, and outputting application development information, wherein the above-mentioned application development information includes design information for indicating application controls and development information for realizing the functions of the above-mentioned application controls;

[0007] According to the application development information, a target application matching the application description information is generated.

[0008] According to another aspect of the present disclosure, there is provided a device for generating an application based on a large language model, comprising:

[0009] A knowledge determination module, configured to respond to a received application generation request and determine, based on the application description information in the application generation request, industry application knowledge matching the application field of the application description information;

[0010] An information processing module, used to input the above-mentioned industry application knowledge and the above-mentioned application description information into a large language model, and output application development information, wherein the above-mentioned application development information includes design information for indicating application controls and development information for realizing the functions of the above-mentioned application controls;

[0011] The application generation module is used to generate a target application matching the application description information according to the application development information.

[0012] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a method as disclosed in the present disclosure.

[0013] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method of the present disclosure.

[0014] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the method of the present disclosure is implemented.

[0015] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0017] Figure 1 Schematically illustrates an exemplary system architecture to which a method and device for generating an application based on a large language model can be applied according to an embodiment of the present disclosure;

[0018] Figure 2 A flowchart of a method for generating an application based on a large language model according to an embodiment of the present disclosure is schematically shown;

[0019] Figure 3 A schematic diagram of the architecture of application expansion information according to an embodiment of the present disclosure is schematically shown;

[0020] Figure 4 A schematic diagram schematically shows the generation of control development information for an application control according to an embodiment of the present disclosure;

[0021] Figure 5 A schematic diagram schematically shows an interactive interface according to an embodiment of the present disclosure;

[0022] Figure 6 A schematic diagram schematically shows the association of form components according to an embodiment of the present disclosure;

[0023] Figure 7 A schematic diagram schematically shows a preview of a target application according to an embodiment of the present disclosure;

[0024] Figure 8Schematically shows an application generation architecture diagram based on a large language model according to an embodiment of the present disclosure;

[0025] Fig. 9 A flowchart of a method for generating an application based on a large language model according to another embodiment of the present disclosure is schematically shown;

[0026] Fig.10 A block diagram of an application generation device based on a large language model according to an embodiment of the present disclosure is schematically shown;

[0027] Fig.11 A block diagram of an electronic device suitable for implementing an application generation method based on a large language model according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0028] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0029] Figure 1 An exemplary system architecture to which the method and apparatus for generating an application based on a large language model according to an embodiment of the present disclosure can be applied is schematically shown.

[0030] It should be noted that Figure 1 The examples shown are only examples of system architectures to which the embodiments of the present disclosure can be applied, in order to help those skilled in the art understand the technical content of the present disclosure, but do not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios. For example, in another embodiment, an exemplary system architecture to which the method and apparatus for generating an application based on a large language model can be applied may include a terminal device, but the terminal device may implement the method and apparatus for generating an application based on a large language model provided by the embodiments of the present disclosure without interacting with the server.

[0031] like Figure 1 As shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0032] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as knowledge reading applications, web browser applications, search applications, instant messaging tools, email clients and / or social platform software, etc. (only as examples).

[0033] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0034] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for the content browsed by the user using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to the user request) to the terminal device.

[0035] It should be noted that the application generation method based on the large language model provided in the embodiment of the present disclosure can generally be executed by the terminal device 101, 102, or 103. Accordingly, the application generation apparatus based on the large language model provided in the embodiment of the present disclosure can also be set in the terminal device 101, 102, or 103.

[0036] Alternatively, the application generation method based on the large language model provided in the embodiment of the present disclosure may also be generally executed by the server 105. Accordingly, the application generation device based on the large language model provided in the embodiment of the present disclosure may generally be arranged in the server 105. The application generation method based on the large language model provided in the embodiment of the present disclosure may also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the application generation device based on the large language model provided in the embodiment of the present disclosure may also be arranged in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0037] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0038] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision, disclosure and application of user personal information involved comply with the provisions of relevant laws and regulations, take necessary confidentiality measures, and do not violate public order and good morals.

[0039] In the technical solution of the present disclosure, the user's authorization or consent is obtained before obtaining or collecting the user's personal information.

[0040] Figure 2 The flowchart of the application generation method based on the large language model according to the embodiment of the present disclosure is schematically shown.

[0041] like Figure 2 As shown, the method includes operations S210 to S230.

[0042] In operation S210, in response to the received application generation request, industry application knowledge matching the application field of the application description information is determined according to the application description information in the application generation request.

[0043] In operation S220, industry application knowledge and application description information are input into a large language model, and application development information is output, wherein the application development information includes design information for indicating an application control and development information for realizing a function of the application control.

[0044] In operation S230, a target application matching the application description information is generated according to the application development information.

[0045] According to the embodiments of the present disclosure, in the process of application development, due to the complexity of business scenarios and the diversity of business needs, applications are usually developed in an intelligent generation manner. Specifically, when intelligently generating applications, users can enter application description information in the development platform to express their needs for application generation. The application description information can be a text description composed of natural language. For example: generate an asset management system. After the user enters the application description information, an application generation request is triggered, and the target application matching the application description information is generated.

[0046] According to the embodiments of the present disclosure, when using a large language model to analyze application description information, since the basic large language model does not have the knowledge reserve related to the industry application, it often answers content that does not meet the actual business needs during the application generation process. Therefore, in order to improve the processing effect of generating a large language model for the needs, the industry application knowledge matching the application field of the application description information is obtained, and the industry application knowledge and application description information are input into the large language model for analysis and processing.

[0047] According to the embodiments of the present disclosure, the large language model analyzes the industry application knowledge, and since the industry application knowledge includes the design information of each industry application, specifically including the page design, hierarchical distribution and process creation of the application, when processing the application description information, the design information indicating the application control when developing the target application and the development information for realizing the application control function are determined by referring to the application industry knowledge.

[0048] According to the embodiments of the present disclosure, the application control is a development engine deployed in the development platform to generate the target application, and each application control implements different functions. Specifically, it may include a page engine, an entity engine, a process engine, a report engine, etc. The large language model generates design information for indicating the application control by analyzing industry application knowledge, such as: the distribution of the design page, the statistical chart style used when designing data statistics in the page, etc. The large language model will also generate development information for implementing the application control function, that is, how to implement the corresponding function according to the above-mentioned design information.

[0049] According to the embodiments of the present disclosure, when the large language model processes the application description information, the application description information is used as the demand and the industry application knowledge is used as the context corresponding to the demand. The prompt word template is constructed according to the demand and the context, and then the key domain-specific language (DSL) description corresponding to each type of application control is generated according to the prompt word template, that is, the application development information.

[0050] According to an embodiment of the present disclosure, the function of each application control is implemented in a development platform according to the application development information, and the functions of multiple application controls are integrated to generate a target application that matches the application description information.

[0051] According to an embodiment of the present disclosure, industry application knowledge is acquired based on application description information, and both the application description information and the industry application knowledge are used as inputs of a large language model, so that the large language model can better generate application development information that conforms to the application description information, thereby effectively improving application generation efficiency.

[0052] Reference below Figure 3~Figure 9 , combined with specific embodiments Figure 2 The method shown is further explained.

[0053] Figure 3 The schematic diagram of the architecture of application expansion information according to an embodiment of the present disclosure is schematically shown.

[0054] According to an embodiment of the present disclosure, industry application knowledge and application description information are input into a large language model, and application development information is output, including: inputting industry application knowledge, application description information and an application expansion prompt template into a large language model, and outputting application expansion information, wherein the application expansion information includes identification information of each of a plurality of application controls and function information of each of the plurality of application controls; when it is determined that the application expansion information meets predetermined conditions, inputting the application expansion information and an application framework prompt template into the large language model, and outputting application development information.

[0055] According to an embodiment of the present disclosure, industry application knowledge, application description information and application expansion prompt template are input into a large language model to expand the demand. Specifically, the large language model generates application expansion information by splicing industry application knowledge application description information with application expansion template.

[0056] like Figure 3 As shown, the large language model first vertically expands the application description information to obtain multiple application expansion information 310, and then expands each application expansion information horizontally to obtain multiple functions 320. After the expansion is completed, the multiple application expansion information obtained by the expansion and the function corresponding to each application expansion information are associated to form an overall application framework.

[0057] For example: the application description information entered by the user is asset management, and the multiple application expansion information obtained by expansion are: asset information management, asset warehousing, asset collection, asset return, asset scrapping and asset maintenance. Specifically, the functions corresponding to asset information management are asset information addition and asset information management. The functions corresponding to asset warehousing application are asset warehousing application and asset warehousing record. The functions corresponding to asset collection are asset collection application and asset collection record. The functions corresponding to asset return are asset return application and asset return record. The functions corresponding to asset scrapping are asset scrapping application and asset scrapping record. The functions corresponding to asset scrapping are asset maintenance application and asset maintenance record.

[0058] According to an embodiment of the present disclosure, after the application expansion information is generated, Figure 3 The information input by the user in the user interaction box 330 is used to determine whether the expanded application expansion information meets the predetermined conditions. The information input by the user can be a text description or instruction used to indicate whether to confirm the application expansion information. If it is determined that the predetermined conditions are met, the application expansion information and the application framework prompt template are input into the large language model for further processing, thereby generating application development information.

[0059] According to the embodiments of the present disclosure, by expanding the application description information based on industry application knowledge, various aspects of the requirements can be clearly defined and described, which helps to efficiently generate the target application.

[0060] According to an embodiment of the present disclosure, the application generation method based on a large language model also includes: displaying application expansion information on an interactive interface; in response to a determination instruction for the application expansion information, determining that the application expansion information meets predetermined conditions; in response to a modification instruction for the application expansion information, modifying the application expansion information based on modification information input into the interactive interface to obtain application expansion information that meets the predetermined conditions.

[0061] According to an embodiment of the present disclosure, the application expansion information is displayed on the interactive interface on the development platform, and can be specifically displayed in the form of a frame. After receiving an instruction sent by the user through the user interaction frame to indicate the determination of the application expansion information, it is determined that the application expansion information meets the predetermined conditions, and the next step of application development can be performed based on the current application expansion information.

[0062] According to an embodiment of the present disclosure, if an instruction to modify the application expansion information sent by the user is received, the application expansion information is modified according to the modification information input by the user, so that the user can adjust his or her needs in a timely manner. If the user wants to modify his or her needs again after the application expansion information is generated, the needs can also be expanded again using the large language model according to the modification information input by the user, thereby obtaining new application expansion information.

[0063] According to the embodiments of the present disclosure, in the process of demand expansion, by interacting with the user, the application expansion information finally obtained is more in line with the user's demand, and the overall application generation process is made more intelligent.

[0064] Figure 4 A schematic diagram schematically illustrates a method for generating control development information for an application control according to an embodiment of the present disclosure.

[0065] According to an embodiment of the present disclosure, application expansion information and an application framework prompt template are input into a large language model, and application development information is output, including: determining multiple control task prompt templates corresponding to the application control from the application framework prompt template according to the functional information of the application control; repeating the following operations in the order of task execution for generating the application control until the control development information of the application control is obtained: inputting the current control task prompt template, the identification information of the application control, and the output result of the previous task into the large language model to obtain the output result of the current task, wherein the current control task prompt template is determined from multiple control task prompt templates according to the order of task execution, and the output result of the previous task is the result of the large language model executing the output of the previous task; based on the control development information of each of the multiple application controls, the application development information of the target application is obtained.

[0066] According to an embodiment of the present disclosure, the functional information of an application control is matched with an application framework prompt template, thereby obtaining a plurality of control task prompt templates corresponding to the application control. Specifically, since there are multiple application controls, during the matching process, the control task prompt template of the application control can be determined in a one-to-one matching manner, or the functional information of multiple application controls can be processed simultaneously in a many-to-many manner. For each application control, since the application control includes multiple sub-functions, during the matching process, the corresponding control task prompt template can be determined from the application framework prompt template according to the functional information of each sub-function.

[0067] According to an embodiment of the present disclosure, multiple control task prompt templates are input into a large language model, and each control task prompt template corresponds to a task. According to the task execution order of generating application controls, multiple tasks are processed in sequence. Among them, the task execution order can be set by the development platform according to the development process of the application, and specifically can be developed in the order from the back end to the front end. For example: first configure the back end process and interface, and finally generate sample data for front end display.

[0068] According to an embodiment of the present disclosure, after the large language model initially receives the control task prompt template and the identification information of the application control, the control task prompt template and the identification information are spliced ​​to prompt the large language model to generate a task output result corresponding to the application control function.

[0069] According to an embodiment of the present disclosure, when the large language model processes the next task, the current control task prompt template, the identification information of the application control and the output result of the previous task are input into the large language model, and the three are spliced ​​to prompt the large language model to generate the output result of the current task.

[0070] According to an embodiment of the present disclosure, after the large language model completes processing of multiple control task prompt templates corresponding to the current application control, the multiple task output results are combined to obtain control development information of the application control.

[0071] like Figure 4 As shown, the large language model processes multiple application controls 410 in sequence. Specifically, the control development information of application control 1, application control 2, application control 3, and application control 4 is generated at one time. For application control 2, application control 2 includes multiple sub-functions 420. According to the function information of each sub-function, the corresponding control task prompt template is determined from the application framework prompt template. Specifically, the three control task prompt templates correspond to task one, task two, and task three, respectively.

[0072] According to an embodiment of the present disclosure, each task is processed in turn according to the task execution order of generating the application control 2: Task 1-Task 2-Task 3. During the processing of the large language model, for Task 1, the input of the large language model is the identification information of the application control 2 corresponding to the task, and the control task prompt template corresponding to the task. After the large language model completes the processing of Task 1, the task output result output by Task 1, the identification information of the application control 2, and the control task prompt template corresponding to Task 2 are input into the large language model together to process Task 2. Similarly, Task 3 is also processed according to the task output result output by Task 2. Finally, the task output results of Task 1, the task output results of Task 2, and the task output results of Task 3 are respectively integrated to obtain the control development information 430 of the application control 2.

[0073] According to an embodiment of the present disclosure, in the process of processing Task 1, the large language model may also refer to the control development information of Application Control 1 for processing. Specifically, the identification information of Application Control 2, the control task prompt template corresponding to Task 1, and the control development information of Application Control 1 are input into the large language model, thereby generating the task output result of Task 1.

[0074] According to an embodiment of the present disclosure, after Task 1, Task 2 and Task 3 are processed, since the large language model processes the data step by step and in association, the task output result of Task 3 can also be directly determined as the control development information 430 of the application control 2.

[0075] According to an embodiment of the present disclosure, after the control development information of application control 1, application control 2, application control 3, and application control 4 is generated, the multiple control development information are integrated and associated to obtain the application development information of the target application.

[0076] According to the embodiments of the present disclosure, when generating control development information, by using the output of the previous task as the input of the next task, the final output of the large language model is more accurate and meets user needs. The large language is learned layer by layer, which effectively improves the processing efficiency of the overall task.

[0077] Figure 5 A schematic diagram of an interactive interface according to an embodiment of the present disclosure is schematically shown.

[0078] According to an embodiment of the present disclosure, a current control task prompt template, identification information of an application control, and an output result of a previous task are input into a large language model to obtain a current task output result, including: inputting a current control task prompt template, identification information of an application control, and an output result of a previous task into a large language model to obtain an initial current task output result; displaying preview information corresponding to the initial current task output result on an interactive interface; and obtaining a current task output result based on the initial current task output result in response to a determination instruction for the preview information.

[0079] According to an embodiment of the present disclosure, after the large language model processes each task of the application control, the task output result is displayed through an interactive interface. Among them, the task output result can be a section of code including fields and field descriptions. By executing the code in the task output result, each field is rendered to generate a preview result. Specifically, the preview result can be a form interface, such as a table including fields such as warehouse entry number, label, specification and quantity; it can also be a flowchart including the relationship between the various nodes of the approval process.

[0080] like Figure 5 As shown, the task output result 510 and the preview information 520 are displayed simultaneously in the interactive interface. If the user is not satisfied with the current preview result, the task output result can be modified through the user interaction box 530 to adjust the preview result. If the user is satisfied with the current preview result, the user interaction box 530 can input a confirmation instruction to indicate that the currently generated task output result meets the requirements.

[0081] According to the embodiments of the present disclosure, by displaying the preview results in real time, users can modify and adjust the task output results in an intuitive and easy-to-understand operation interface. This effectively improves the ability of human-computer interaction during the application creation process, and also makes the final generated task output results more accurate, reducing the workload of users for subsequent modifications and improving task processing efficiency.

[0082] According to an embodiment of the present disclosure, the application generation method based on the large language model also includes: inputting the application development information into a development conversion tool, and mapping the elements in the application development information to the form components and functions of the zero-code platform according to the conversion rules in the development conversion tool, so as to obtain the target application resources required to generate the target application.

[0083] According to the embodiments of the present disclosure, since the generated application development information is described in DSL, it cannot be directly called and run to implement the corresponding functions. Therefore, the application development information described in DSL is input into the development conversion tool (DSL Builder) to convert the DSL description into zero-code application resources. Specifically, during the conversion process, the development conversion tool maps the elements in the application development information to the form components and functions of the zero-code platform one by one, thereby obtaining zero-code resources. Among them, the pages displayed by the zero-code resources are more in line with the actual business demands, and are convenient for business personnel without any coding managers to view and adjust them.

[0084] According to an embodiment of the present disclosure, the development conversion tool can perform fault tolerance and correction on the application development information before or during the conversion process to eliminate the wrong answers output by the large language model. Specifically, garbled codes can be eliminated, and the logic of unexecutable codes can be modified.

[0085] According to the embodiments of the present disclosure, a DSL description is generated using a large language model, and further converted using a development conversion tool, which effectively reduces the difficulty of generating the large language model. The large language model no longer needs to learn the generation logic of the code during the training process, which effectively improves the processing efficiency.

[0086] Figure 6 The figure schematically shows the association of form components according to an embodiment of the present disclosure.

[0087] According to an embodiment of the present disclosure, the application generation method based on the large language model also includes: determining the association relationship between multiple form components based on the field description of each form component in the target application resource; establishing a hierarchical relationship between the multiple form components based on the association relationship to associate the multiple form components to obtain a data structure.

[0088] According to an embodiment of the present disclosure, the field description of each form component is determined, specifically including a header and multiple attributes of the form component. The header may be the name of the form component itself. Based on the association relationship between the header in the form component and multiple attributes in other form components, the hierarchical relationship between the form components is determined to associate multiple form components.

[0089] For example, there is an association between the attribute "Asset Name" in the Asset Entry Application Form component and the header "New Asset Information" in the Add Asset Form component. Since the information is filled in according to the drop-down list selection when the asset is entered into the warehouse, if the Asset Entry Application Form component is not associated with the Add Asset Form component, the drop-down box will not have the corresponding information, which will cause the Asset Entry Application Form component to not work properly. Therefore, the Asset Entry Application Form component is associated with the Add Asset Form component.

[0090] According to the embodiment of the present disclosure, multiple form components are associated in the database to establish a hierarchical relationship between the multiple form components. Specifically, the hierarchical relationship is established layer by layer according to the field description, which can be established from the bottom to the top or from the top to the bottom.

[0091] like Figure 6 As shown, when facing form component 1, form component 2 and form component 3, the corresponding field descriptions are determined respectively. When it is known that field description 2 of form component 3 is associated with form component 2, and field description 1 in form component 2 is associated with form component 1, according to the association relationship between the field descriptions, the hierarchical relationship between the three form components is determined to be form component 1-form component 2-form component 3. That is, the information in form component 2 can be obtained in form component 3, and similarly, form component 2 can obtain the information in form component 1.

[0092] According to the embodiment of the present disclosure, multiple form components are associated in the database to establish a complete data structure, thereby forming a complete framework structure of the target application, making the process of generating the target application more convenient and efficient.

[0093] According to an embodiment of the present disclosure, a target application matching the application description information is generated based on the application development information, including: parsing the application development information to generate calling information for calling target application resources; based on the calling information, calling the target application resources to generate the target application.

[0094] According to an embodiment of the present disclosure, in the process of generating a target application by a development platform, the corresponding function is realized by calling the target application resource corresponding to each application control. Before calling, the application development information needs to be parsed to generate calling information for calling the target application resource. Specifically, the corresponding resource can be determined from the target application resource according to the identification information and function information of each application control, and the calling information is generated according to the storage address of the corresponding resource.

[0095] According to the embodiments of the present disclosure, during the development process, the relevant target application resources are called into the development platform based on the call information, and the resources are run to generate corresponding forms and pages, etc. By using the resource call method to generate the target application, the functions of each application control are not interfered with each other during the application generation process, and the application development can be efficiently completed in the development platform.

[0096] According to an embodiment of the present disclosure, based on the calling information, the target application resources are called to generate the target application, including: for the application control, the corresponding resource data is obtained from the target application resources based on the calling information to build multiple form pages in the development platform; according to the hierarchical relationship between the multiple form components in the data structure, the constructed multiple form pages are associated to generate functional modules corresponding to the application control; when it is determined that the functional modules for the multiple application controls have been generated, the multiple functional modules are integrated to generate the target application.

[0097] According to an embodiment of the present disclosure, after the application control obtains the corresponding resource data, a form page is constructed according to the content in the resource data. Specifically, the resource data can be a section of code including a table header and table name. The form page is obtained by extracting relevant information from the code and rendering the page. Among them, the form page is an empty page that does not include data and is constructed on the development platform. The data structure is imported into the development platform, and multiple form pages are associated according to the hierarchical relationship between the form components to enrich the multiple empty pages constructed.

[0098] According to the embodiments of the present disclosure, the development platform can process multiple application controls at the same time during the process of processing resource data. After generating a form page corresponding to each application control, the multiple form pages are uniformly associated to build multiple functional modules. By uniformly associating multiple form pages, the optionality of the association relationship between form pages is effectively enhanced, thereby enabling the establishment of functional modules with more complex functions.

[0099] According to the embodiments of the present disclosure, in the process of processing resource data, the development platform can also process each application control one by one according to the development order of multiple application controls in the development platform. After the form page association of each application control is completed, the constructed multiple functional modules are integrated. By processing the application controls one by one according to the development order, the development process is more stable and the occurrence of form page association errors is effectively reduced.

[0100] According to an embodiment of the present disclosure, in the process of associating multiple form pages, the association of pages and sub-pages is specifically included. For example, when operating one of the form pages, it can automatically jump to the corresponding form page in response to the process adding instruction or page information viewing instruction clicked by the user. After associating multiple form pages, the functions corresponding to the application control can all be used normally, thereby generating corresponding functional modules.

[0101] According to the embodiments of the present disclosure, after each functional module is generated, the multiple functional modules are integrated in the development platform to generate a target application. In addition, after each functional module is generated, it can also be tested first, and then integrated after the test passes, so that the generated target application is more secure. By creating the target application in the manner of generating the functional modules, the generation process of the target application is more reasonable and orderly.

[0102] Figure 7 A preview diagram of a target application according to an embodiment of the present disclosure is schematically shown.

[0103] According to an embodiment of the present disclosure, the application generation method based on a large language model also includes: importing sample data generated by using a data generation control in a development platform into a target application, so as to display an application page embedded with the sample data when an application preview request sent by a user is received.

[0104] According to the embodiment of the present disclosure, the data generation control generates sample data that conforms to the field type according to the field type and field name of each form component in the target resource data. For example, if the field type is a date type, the corresponding sample data "20××-×-× 22:55:55" is generated according to the corresponding field name.

[0105] According to an embodiment of the present disclosure, sample data is stored in a target application according to field names, wherein the sample data are all real data determined according to industry application knowledge. When an application preview request is received by a user by clicking a "preview" button in an interactive interface, the application page embedded with the sample data is displayed.

[0106] According to an embodiment of the present disclosure, there is also an "Edit" button in the interactive interface. If the user is not satisfied with the application page after previewing the target application, the user can click the "Edit" button to edit and change the application page by dragging and dropping.

[0107] like Figure 7As shown, the displayed application page includes a navigation bar, and the specific navigation bar includes multiple functions 710, specifically including function 1, function 2, function 3 and function 4. The user can click on the corresponding function logo to implement the corresponding function. The application page also includes embedded sample data 720, specifically including sample data 1, sample data 2, sample data 3 and sample data 4. Among them, the sample data can be data corresponding to a certain function in the application page. The application page also includes a pie chart 730 and a bar chart 740 for sample data statistics. The sample data is statistically analyzed and then displayed, so that the user can see the data distribution effect more intuitively.

[0108] According to the embodiments of the present disclosure, by generating sample data and importing it into a target application, a user can more intuitively see the preview effect of the target application, making the generation of the application more intelligent and improving the user experience.

[0109] According to an embodiment of the present disclosure, the application generation method based on a large language model also includes: when the generation of the target application is completed, the target application resources are converted into low-code resources, so that when an editing request for a form component in the low-code resource is received from a user, the generated target application is modified according to the acquired edited low-code resources.

[0110] According to the embodiments of the present disclosure, the development platform provides application building capabilities for zero-code development and low-code development for different user roles. Zero-code is suitable for business personnel who do not have any coding managers, and low-code is suitable for business personnel with a small amount of programming experience. Low-code capabilities can cover zero-code. In order to reduce the difficulty of generating large language models, choose zero-code to build applications. Zero-code pages are more in line with real business demands. After the target application is built using zero code, the zero-code application resources can be converted into low-code resources through data transmission on the development platform.

[0111] According to an embodiment of the present disclosure, since low-code resources are user-editable resources, users can adjust variables in form components by rewriting codes as needed. Therefore, when an edit request for a form component in a low-code resource is received from a user, the edited low-code resource is obtained. Based on the obtained edited low-code resource, the form component is modified, thereby updating the generated target application.

[0112] According to an embodiment of the present disclosure, after the target application is created, the zero-code resources are converted into low-code resources, which effectively increases the development operability of the user and facilitates the user to modify and adjust the target application.

[0113] According to an embodiment of the present disclosure, industry application knowledge matching the application field of the application description information is determined based on the application description information in the application generation request, including: determining the industry application knowledge from an application knowledge base based on the application field and the application mapping relationship of the application description information, wherein the application mapping relationship represents the correspondence between the application field and the industry application knowledge, and the industry application knowledge in the application knowledge base is obtained by preprocessing a set of industry application knowledge in the same application field.

[0114] According to the embodiments of the present disclosure, before acquiring industry application knowledge, in order to make the data more usable, it is necessary to collect, clean and annotate the industry application system data to complete the data preprocessing. In order to reduce labor costs, a visual data collection, cleaning and annotation tool is provided. After the data is preprocessed, the application description information is matched in the application knowledge base based on the application field and application mapping relationship of the application description information to obtain the industry application knowledge. Specifically, in order to enable the large language model to better generate application development information that meets real business needs, the knowledge retrieval model adopts hybrid retrieval.

[0115] According to the embodiments of the present disclosure, by acquiring industry application knowledge, the large language model can better generate application development information that conforms to the application description information, thereby effectively improving the application generation efficiency.

[0116] Figure 8 The following schematically shows an architecture diagram of application generation based on a large language model according to an embodiment of the present disclosure.

[0117] like Figure 8 As shown, in the process of generating the target application, after the user inputs the application description information, the knowledge determination module 810 is used to extract industry application knowledge from the application knowledge base 860. The application knowledge base is maintained through the industry application data engineering, and the data is specifically produced, cleaned and annotated before being stored in the application knowledge base 860.

[0118] After the application knowledge base 860 returns the industry application knowledge, the application description information, the industry application knowledge and the prompt word template generated by the prompt word template generation module 820 are input into the large language model to generate application development information and obtain the output of the large language model.

[0119] The application development information output by the large language model is input into the development conversion tool 840 for processing, so as to convert the application development information into zero-code resources and obtain the target application resources required for generating the target application. In the process of generating the target application resources, the knowledge determination module 810, the prompt word template generation module 820 and the development conversion tool 840 are all uniformly scheduled by the integration scheduling module 830.

[0120] The integrated scheduling module 830 is responsible for the logical arrangement of the entire generation process, and is composed of multiple levels of intelligent agents, including application function modules, page navigation, form body, automated process, report, test data, and intelligent assistant. Among them, the function module supports user intervention, that is, to adjust the requirements through multiple rounds of dialogue, and all interactions are streamed returns.

[0121] The generated application development information is uploaded to the development platform 870 to generate corresponding functions using the application controls in the development platform 870. Finally, the functional modules generated by each application control are integrated to obtain the target application, which is then fed back to the user end.

[0122] According to the embodiments of the present disclosure, in the process of generating a target application, the entire process of target application development from the front-end to the back-end, to system configuration and business data testing is covered. Users do not need to worry about complex functional module design, page layout, logic design, database design, business process design, system role design, etc., making application development more intelligent.

[0123] According to an embodiment of the present disclosure, the application generation method based on a large language model also includes: performing a data structure query in a database based on application description information and industry application knowledge to obtain a query result; when the query result indicates that there is a data structure corresponding to the application description information in the database, generating a target application in a development platform according to the hierarchical relationship between multiple form components in the data structure.

[0124] According to an embodiment of the present disclosure, in response to a received application generation request, a data structure query is performed in the database based on the application description information in the application generation request and the acquired industry application knowledge. If the large language model has historically generated applications similar to the application description information, the database will store relevant data structures. Therefore, if the query result indicates that there is a data structure corresponding to the application description information in the database, the target application is directly generated in the development platform based on the hierarchical relationship between multiple form components in the data structure, without the need for the large language model to generate the application development information, which effectively improves the efficiency of application development.

[0125] Fig. 9 The flowchart of the method for generating an application based on a large language model according to another embodiment of the present disclosure is schematically shown.

[0126] like Fig. 9 As shown, the method includes operations S901 to S923.

[0127] In operation S901, application description information is acquired.

[0128] In operation S902, it is determined whether the application description information is modified.

[0129] According to the embodiments of the present disclosure, during the application development process, the user can modify the application description information at any time to adjust the requirements.

[0130] In operation S903, if the user has not modified the application description information, the industry application knowledge is determined from the application knowledge base according to the application description information.

[0131] In operation S904, a data structure query is performed in the database according to the application description information and the industry application knowledge to determine whether the query result of the data result is a hit.

[0132] In operation S905 , when the characterization hits, relevant information is acquired from the data.

[0133] Operation S910 is performed respectively to obtain control development information of the application control. Operation S911 is performed to obtain page generation information. Operation S912 is performed to obtain table body generation information. Operation S913 is performed to obtain process generation information.

[0134] In the case of a representation miss, the large language model is used to generate corresponding application development information.

[0135] Operation S906 is performed respectively to generate application loss information based on industry application knowledge and application description information. Operation S907 is performed to generate a page. Operation S908 is performed to generate a table body. Operation S909 is performed to generate a process.

[0136] In operation S914, after it is determined that the application development information is obtained, a target application is created.

[0137] In operation S915, a form page is constructed.

[0138] In operation S916 , entity modeling is performed.

[0139] According to an embodiment of the present disclosure, after constructing an empty form page in a development platform, data results are determined in a database and imported to perform entity modeling to enrich the empty form page and implement corresponding functions.

[0140] After executing operation S916, execute operation S917 in parallel to generate a report page. Operation S918 to generate an intelligent assistant. Operation S919 to generate a home page. Operation S920 to generate a list page. Operation S921 to generate an approval process. Operation S922 to generate sample data.

[0141] According to an embodiment of the present disclosure, when generating an intelligent assistant, a business assistant module is obtained by integrating the business logic in multiple functional modules. An introduction text corresponding to the business assistant module is generated according to the application description information and industry application knowledge. The business assistant module and the introduction text are configured in the development platform so that when a process generation requirement is received from a user according to the introduction text, the corresponding business process is automatically initiated.

[0142] According to the embodiments of the present disclosure, in the process of generating each function of the page, a corresponding interface will be configured to implement adding, deleting, checking and modifying the data in the page. By rendering the front-end interface, integrating multiple functional modules, and embedding sample data, the target application is generated.

[0143] In operation S923, the zero-code resources are converted into low-code resources.

[0144] According to an embodiment of the present disclosure, after the target application is generated, the zero-code resources are converted into low-code resources, which effectively increases the development operability of the user and facilitates the user to modify and adjust the target application.

[0145] According to the embodiments of the present disclosure, in the process of generating the target application, the native low-code reconstruction of the development platform is realized, and the traditional visual development platform is upgraded to an enterprise application development platform for intelligent development and intelligent operation, which effectively improves the application development efficiency.

[0146] Fig.10 A block diagram of an application generation device based on a large language model according to an embodiment of the present disclosure is schematically shown.

[0147] like Fig.10 As shown, the device 1000 for generating an application based on a large language model in this embodiment includes a knowledge determination module 1010 , an information processing module 1020 and an application generation module 1030 .

[0148] The knowledge determination module 1010 is used to respond to the received application generation request and determine the industry application knowledge matching the application field of the application description information according to the application description information in the application generation request.

[0149] The information processing module 1020 is used to input industry application knowledge and application description information into the large language model, and output application development information, wherein the application development information includes design information for indicating application controls and development information for realizing application control functions.

[0150] The application generation module 1030 is used to generate a target application matching the application description information according to the application development information.

[0151] According to an embodiment of the present disclosure, the information processing module 1020 includes a prompt expansion submodule.

[0152] The prompt expansion submodule is used to input industry application knowledge, application description information and application expansion prompt template into the large language model, and output application expansion information, wherein the application expansion information includes identification information of each of the multiple application controls and function information of each of the multiple application controls.

[0153] The information processing submodule is used to input the application expansion information and the application framework prompt template into the large language model and output the application development information when it is determined that the application expansion information meets the predetermined conditions.

[0154] According to an embodiment of the present disclosure, the device 1000 for application generation based on a large language model further includes an expansion display module, a condition determination module, and an information modification module.

[0155] The expansion display module is used to display application expansion information on the interactive interface.

[0156] The condition determination module is used to determine whether the application extension information meets a predetermined condition in response to a determination instruction for the application extension information.

[0157] The information modification module is used to respond to the modification instruction for the application extension information and modify the application extension information based on the modification information inputted in the interactive interface to obtain the application extension information meeting the predetermined condition.

[0158] According to an embodiment of the present disclosure, the information processing submodule includes a template determination unit, a task execution unit, and an information generation unit.

[0159] The template determining unit is used to determine a plurality of control task prompt templates corresponding to the application control from the application framework prompt template according to the function information of the application control.

[0160] The task execution unit is used to repeatedly perform the following operations according to the task execution sequence of generating the application control until the control development information of the application control is obtained:

[0161] The current control task prompt template, the identification information of the application control, and the output result of the previous task are input into the large language model to obtain the current task output result, wherein the current control task prompt template is determined from multiple control task prompt templates according to the task execution order, and the output result of the previous task is the result of the large language model executing the output of the previous task.

[0162] The information generating unit is used to obtain the application development information of the target application based on the control development information of each of the multiple application controls.

[0163] According to an embodiment of the present disclosure, the task execution unit includes an initial determination subunit, a preview display subunit and a task output subunit.

[0164] The initial determination subunit is used to input the current control task prompt template, the identification information of the application control, and the output result of the previous task into the large language model to obtain the initial current task output result.

[0165] The preview display subunit is used to display the preview information corresponding to the initial current task output result on the interactive interface.

[0166] The task output subunit is used to obtain the current task output result based on the initial current task output result in response to the determination instruction for the preview information.

[0167] According to an embodiment of the present disclosure, the device 1000 for application generation based on a large language model further includes an information mapping module.

[0168] The information mapping module is used to input application development information into the development conversion tool, and according to the conversion rules in the development conversion tool, map the elements in the application development information to the form components and functions of the zero-code platform to obtain the target application resources required to generate the target application.

[0169] According to an embodiment of the present disclosure, the apparatus 1000 for application generation based on a large language model further includes a field association module and a component association module.

[0170] The field association module is used to determine the association relationship between multiple form components according to the field description of each form component in the target application resource.

[0171] The component association module is used to establish a hierarchical relationship between multiple form components based on the association relationship, so as to associate multiple form components to obtain a data structure.

[0172] According to an embodiment of the present disclosure, the application generation module 1030 includes an information parsing submodule and a resource calling submodule.

[0173] The information parsing submodule is used to parse the application development information and generate calling information for calling target application resources.

[0174] The resource calling submodule is used to call the target application resources based on the calling information and generate the target application.

[0175] According to an embodiment of the present disclosure, the resource calling submodule includes a page establishing unit, a page associating unit, and an application generating unit.

[0176] The page building unit is used to obtain corresponding resource data from the target application resources based on the call information for the application control, so as to build multiple form pages in the development platform.

[0177] The page association unit is used to associate multiple constructed form pages according to the hierarchical relationship between multiple form components in the data structure, and generate functional modules corresponding to the application controls.

[0178] The application generation unit is used to integrate the multiple functional modules to generate a target application when it is determined that the functional modules of the multiple application controls have been generated.

[0179] According to an embodiment of the present disclosure, the device 1000 for application generation based on a large language model further includes a data import module.

[0180] The data import module is used to import sample data generated by the data generation control in the development platform into the target application, so as to display the application page embedded with the sample data when receiving an application preview request sent by the user.

[0181] According to an embodiment of the present disclosure, the apparatus 1000 for application generation based on a large language model further includes a resource conversion module.

[0182] The resource conversion module is used to convert the target application resources into low-code resources after the target application is generated, so that when an editing request for a form component in the low-code resource is received from a user, the generated target application can be modified according to the edited low-code resources obtained.

[0183] According to an embodiment of the present disclosure, the knowledge determination module 1010 includes a knowledge determination submodule.

[0184] The knowledge determination submodule is used to determine the industry application knowledge from the application knowledge base based on the application domain and application mapping relationship of the application description information, wherein the application mapping relationship represents the correspondence between the application domain and the industry application knowledge, and the industry application knowledge in the application knowledge base is obtained by preprocessing the industry application knowledge set of the same application domain.

[0185] According to an embodiment of the present disclosure, the device 1000 for application generation based on a large language model further includes a structure query module and a target generation module.

[0186] The structure query module is used to perform data structure query in the database based on application description information and industry application knowledge to obtain query results.

[0187] The target generation module is used to generate the target application in the development platform according to the hierarchical relationship between multiple form components in the data structure when there is a data structure corresponding to the application description information in the query result representation database.

[0188] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0189] According to an embodiment of the present disclosure, an electronic device includes: at least one processor, and a memory connected to the at least one processor in communication, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above method.

[0190] According to an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the above method.

[0191] According to an embodiment of the present disclosure, a computer program product includes a computer program, and the computer program implements the above method when executed by a processor.

[0192] Fig.11 A block diagram of an electronic device suitable for implementing an application generation method based on a large language model according to an embodiment of the present disclosure is schematically shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0193] like Fig.11 As shown, the device 1100 includes a computing unit 1101, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1102 or a computer program loaded from a storage unit 1108 to a random access memory (RAM) 1103. In the RAM 1103, various programs and data required for the operation of the device 1100 can also be stored. The computing unit 1101, the ROM 1102, and the RAM 1103 are connected to each other via a bus 1104. An input / output (I / O) interface 1105 is also connected to the bus 1104.

[0194] A number of components in the device 1100 are connected to the I / O interface 1105, including: an input unit 1106, such as a keyboard, a mouse, etc.; an output unit 1107, such as various types of displays, speakers, etc.; a storage unit 1108, such as a disk, an optical disk, etc.; and a communication unit 1109, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 1109 allows the device 1100 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0195] The computing unit 1101 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 1101 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 1101 performs the various methods and processes described above, such as an application generation method based on a large language model. For example, in some embodiments, the application generation method based on a large language model may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 1108. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 1100 via the ROM 1102 and / or the communication unit 1109. When the computer program is loaded into the RAM 1103 and executed by the computing unit 1101, one or more steps of the application generation method based on the large language model described above may be performed. Alternatively, in other embodiments, the computing unit 1101 may be configured to execute the application generation method based on the large language model in any other appropriate manner (for example, by means of firmware).

[0196] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0197] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0198] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0199] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0200] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0201] A computer system may include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises through computer programs running on respective computers and having a client-server relationship to each other. The server may be a cloud server, a server in a distributed system, or a server combined with a blockchain.

[0202] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution provided by this disclosure can be achieved, and this document does not limit this.

[0203] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for generating applications based on a large language model, include: In response to the received application generation request, determining industry application knowledge matching the application field of the application description information according to the application description information in the application generation request; Inputting the industry application knowledge and the application description information into a large language model, and outputting application development information, wherein the application development information includes design information for indicating an application control and development information for realizing a function of the application control; A target application matching the application description information is generated according to the application development information.

2. The method according to claim 1, in, The step of inputting the industry application knowledge and the application description information into a large language model and outputting application development information includes: Inputting the industry application knowledge, the application description information and the application expansion prompt template into the large language model, and outputting application expansion information, wherein the application expansion information includes identification information of each of the plurality of application controls and function information of each of the plurality of application controls; When it is determined that the application expansion information meets a predetermined condition, the application expansion information and an application framework prompt template are input into the large language model, and the application development information is output.

3. The method according to claim 2, further comprising: include: Displaying the application expansion information on the interactive interface; In response to a determination instruction for the application extension information, determining that the application extension information satisfies the predetermined condition; In response to a modification instruction for the application extension information, the application extension information is modified based on the modification information inputted in the interactive interface to obtain the application extension information that meets the predetermined condition.

4. The method according to claim 2, wherein the application expansion information and the application framework prompt template are input into the large language model, and the application development information is output. include: Determining, according to the function information of the application control, a plurality of control task prompt templates corresponding to the application control from the application framework prompt template; According to the task execution order of generating the application control, repeatedly perform the following operations until the control development information of the application control is obtained: Inputting a current control task prompt template, identification information of the application control, and a previous task output result into the large language model to obtain a current task output result, wherein the current control task prompt template is determined from a plurality of control task prompt templates according to a task execution order, and the previous task output result is a result of the large language model executing the previous task output; Based on the control development information of each of the plurality of application controls, the application development information of the target application is obtained.

5. The method according to claim 4, wherein the current control task prompt template, the identification information of the application control, and the output result of the previous task are input into the large language model to obtain the output result of the current task, include: Inputting the current control task prompt template, the identification information of the application control, and the output result of the previous task into the large language model to obtain the initial current task output result; Displaying preview information corresponding to the initial current task output result on the interactive interface; In response to a determination instruction for the preview information, the current task output result is obtained based on the initial current task output result.

6. The method according to claim 1, further comprising: include: The application development information is input into a development conversion tool, and according to the conversion rules in the development conversion tool, the elements in the application development information are mapped to the form components and functions of the zero-code platform to obtain the target application resources required to generate the target application.

7. The method according to claim 6, further comprising: include: Determine the association relationship between the plurality of form components according to the field description of each form component in the target application resource; According to the association relationship, a hierarchical relationship between the multiple form components is established to associate the multiple form components to obtain a data structure.

8. The method according to claim 7, wherein the target application matching the application description information is generated according to the application development information, include: Parsing the application development information to generate calling information for calling target application resources; Based on the calling information, the target application resources are called to generate the target application.

9. The method according to claim 8, in, The calling the target application resource based on the calling information to generate the target application includes: For the application control, obtaining corresponding resource data from the target application resource based on the call information, so as to construct multiple form pages in the development platform; According to the hierarchical relationship between the multiple form components in the data structure, the multiple constructed form pages are associated to generate a functional module corresponding to the application control; When it is determined that the functional modules of the plurality of application controls are all generated, the plurality of functional modules are integrated to generate the target application.

10. The method according to claim 9, further comprising: include: The sample data generated by the data generation control in the development platform is imported into the target application, so that when an application preview request sent by a user is received, the application page embedded with the sample data is displayed.

11. The method according to claim 6, further comprising: include: When the generation of the target application is completed, the target application resources are converted into low-code resources, so that when an editing request for the form component in the low-code resource is received from the user, the generated target application can be modified according to the acquired edited low-code resources.

12. The method according to claim 1, wherein the application description information in the application generation request is used to determine the industry application knowledge that matches the application field of the application description information. include: Based on the application field and application mapping relationship of the application description information, the industry application knowledge is determined from an application knowledge base, wherein the application mapping relationship represents the correspondence between the application field and the industry application knowledge, and the industry application knowledge in the application knowledge base is obtained by preprocessing the industry application knowledge set of the same application field.

13. The method according to claim 1, further comprising: include: Performing a data structure query in a database based on the application description information and the industry application knowledge to obtain a query result; In a case where the query result indicates that a data structure corresponding to the application description information exists in the database, the target application is generated in the development platform according to the hierarchical relationship between multiple form components in the data structure.

14. A device for application generation based on a large language model, include: A knowledge determination module, configured to respond to a received application generation request and determine, based on application description information in the application generation request, industry application knowledge matching the application field of the application description information; An information processing module, used for inputting the industry application knowledge and the application description information into a large language model, and outputting application development information, wherein the application development information includes design information for indicating application controls and development information for realizing functions of the application controls; The application generation module is used to generate a target application matching the application description information according to the application development information.

15. An electronic device, include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 13.

16. A non-transitory computer-readable storage medium storing computer instructions, in, The computer instructions are used to cause the computer to execute the method according to any one of claims 1-13.

17. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 13.

Images