Water conservancy business dynamic response method and system based on large language model

Abstract

The invention provides a water conservancy business dynamic response method and system based on a large language model, and belongs to the technical field of water conservancy, and the method comprises the steps: collecting water conservancy business execution rules, building a water conservancy business process description standard according to the business execution rules, and employing the large language model to carry out the management of water conservancy assemblies, and the process standard and the component platform are in butt joint with the large language model, so that the large language model learns the information to master the construction rule of the water conservancy business. And after the big language model receives the business demand, generating a water conservancy business expression chain based on the learned rule, thereby realizing water conservancy business application generation based on the expression chain. According to the invention, the problem that the existing water conservancy application is difficult to adapt to dynamically changing water conservancy business requirements after delivery due to a fixed framework is solved, so that the flexibility and practicability of the application are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of water conservancy, and in particular relates to a method and system for dynamic response of water conservancy business based on a large language model. Background Art

[0002] As an important part of the national economy, the water conservancy industry covers key businesses such as flood control, irrigation area management and water resource dispatch. As global climate change, water shortage and other issues become increasingly prominent, the complexity and dynamism of water conservancy business are becoming increasingly apparent. How to efficiently respond to these challenges has become the focus of water conservancy information construction.

[0003] Although the existing dynamic response methods of water conservancy have adopted advanced technologies such as real-time monitoring and big data, they still face the problem of insufficient flexibility. The execution of existing water conservancy business is usually based on fixed rules and processes, and the preset logic can meet specific business needs in the early stage. However, the water conservancy industry has significant seasonal and regional differences. Different time periods and regions will put forward different requirements for the management and scheduling of water resources. For example, the early warning and disposal of flood disasters need to dynamically adjust the strategy according to real-time meteorological and hydrological data, while drought management requires timely assessment of water demand and adjustment of water supply plans. The needs of various water conservancy scenarios change with changes in the environment and time. The dynamic and complexity of the water conservancy industry make it difficult for such fixed-architecture systems to cope with real-time changes in the environment, resulting in the system's inability to adjust strategies in time when facing different hydrological conditions. This fixed logic development model makes the system slow to respond and lacks flexibility when dealing with complex scenarios, making it difficult to meet the needs of modern water resources management for rapid response and dynamic adjustment.

[0004] Therefore, in order to solve the problem of insufficient flexibility of the existing water conservancy business system in dealing with dynamic changes, a dynamic response method for water conservancy business based on a large language model is urgently needed. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a dynamic response method for water conservancy business based on a large language model, which aims to dynamically generate business expressions by constructing a large language model that can autonomously learn water conservancy business rules, and to construct specific business applications based on the business expression chain, so as to provide a technical solution to the problem of insufficient flexibility of existing water conservancy business systems when dealing with dynamic changes.

[0006] In order to achieve the above object, the present invention provides the following technical solutions: A method for dynamic response of water conservancy business based on a large language model comprises the following steps: Collect water conservancy business execution rules and establish water conservancy business process description standards based on the water conservancy business execution rules; Use the large language model to manage water conservancy components, provide a unified access address for component registration, integrate all component services, and build a component platform; Use the big language model to learn the water conservancy business description standard and understand the relationship between components and businesses in the component platform to obtain a learned big language model; The learned large language model generates a business expression chain based on the user's water conservancy business needs; Dynamically respond to user needs based on the business expression chain and generate business applications.

[0007] Preferably, the collection of water conservancy business execution rules specifically includes: collecting business description rules, business extension rules, component input and output rules, and the relationship between components and businesses; logical rules are divided according to business types, and subject domains of various business areas are constructed, and corresponding water conservancy business process description standards are formulated based on the subject domain types.

[0008] Preferably, the use of a large language model to manage water conservancy components also includes users independently adding components according to specific business needs, and the large language model semantically manages the components based on their input, output and application functions.

[0009] Preferably, the use of a large language model to learn the water conservancy business description standard and understand the relationship between components and businesses in the component platform specifically includes: The rules and relationships within each subject domain are sorted out and annotated according to the refined business needs. The annotated descriptions are fine-tuned and trained using the QLoRA method. In the process of generating business applications, the large language model uses few-shot and least-to-most strategies for typical business operation examples as reference information when executing business based on questions raised by users, and constructs the corresponding business rules into prompt templates as context information for the large language model, thus obtaining a learned large language model.

[0010] Preferably, the learned large language model generates a business expression chain according to the user's water conservancy business needs, specifically including: Through intent recognition, the business process description under the corresponding subject domain is found, and the business process is dynamically constructed based on the prompt project; according to the user intention category, the corresponding components are selected from the component platform to build a business component set, and the operating logic of the water conservancy business is generated into a business expression chain in the form of component series or parallel connection; the expression chain includes the called components, the connection relationship between components, input and output parameters, data sources, data types and system functions.

[0011] Preferably, dynamically responding to user needs according to the business expression chain to generate business applications specifically includes: Users modify the data nodes in the representation chain according to actual needs. The system background adjusts the node data according to the input and output constraints based on the content of the representation chain, and filters the database table of the business system according to the system's functions and component types, ultimately realizing the application construction based on business needs; when the water conservancy business needs change, additions, deletions or modifications are made based on the existing business representation chain, thereby realizing dynamic updates of the business system functions.

[0012] The present invention also provides a water conservancy business dynamic response system based on a large language model, comprising: A labeling establishment module is used to collect water conservancy business execution rules and establish water conservancy business process description standards based on the water conservancy business execution rules; The platform construction module is used to manage water conservancy components using a large language model, provide a unified access address for component registration, integrate all component services, and build a component platform; The learning module is used to learn the water conservancy business description standard using the large language model and understand the relationship between components and businesses in the component platform to obtain a learned large language model; The expression chain generation module is used to generate a business expression chain based on the user's water conservancy business needs after learning the large language model; The application generation module is used to dynamically respond to user needs according to the business expression chain and generate business applications.

[0013] The present invention also provides a computer device, including a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement any one of the steps in the method for dynamic response of water conservancy business based on a large language model.

[0014] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is loaded by a processor, it can execute any step of the method for dynamic response of water conservancy business based on a large language model.

[0015] The water conservancy business dynamic response method based on the large language model provided by the present invention has the following beneficial effects: The present invention introduces a large language model to autonomously learn water conservancy business processes, so that the system can dynamically and efficiently generate corresponding business expression chains when dealing with complex and changeable water conservancy business scenarios. Users can flexibly adjust each node in the business expression chain according to actual needs to ensure that the business process accurately matches the current situation. This mechanism significantly improves the ability of existing water conservancy applications to deal with emergencies and environmental changes, while bringing users higher decision-making efficiency and operational flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 This is a flow chart of a method for dynamic response of water conservancy services based on a large language model according to Embodiment 1 of the present invention; Figure 2 It is a schematic diagram of the business expression chain of the present invention. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0019] Example 1 The present invention provides a method for dynamic response of water conservancy business based on a large language model, specifically as follows Figure 1 As shown, the following steps are included: S1: Collect water conservancy business execution rules and establish water conservancy business process description standards based on the business execution rules.

[0020] Collect the execution rules related to water conservancy business, including business description rules, business extension rules, component input and output rules, and the relationship between components and business. For the convenience of management, these logical rules are classified by business type and corresponding subject domains are constructed, such as flood and drought disaster management, irrigation area water distribution, ecological compensation, prediction and forecasting, water rights trading and configuration regulation, etc., and corresponding water conservancy business process description standards are formulated based on the subject domain type.

[0021] a. Business description rules, including but not limited to the following aspects: the subject, content, type, scope of application, etc. of the business, clarifying the basic attributes and application scenarios of the business.

[0022] b. Business expansion rules mainly involve the relationship between the business and other related businesses, explaining its scalability and dependency to facilitate cross-business linkage.

[0023] c. Component input and output rules, which define in detail the purpose, scope of application, required input data type and output results of each component.

[0024] d. The relationship between components and services is used to describe how a specific service calls and schedules different components to implement service functions.

[0025] In order to make the large language model better understand, each subject domain contains detailed description text and keywords. These descriptions are vectorized and encoded through the embedding model to generate the corresponding subject domain vector. Then, the business process description under each subject domain is refined and annotated, and stored in the vector database through the same vectorization processing method to form a water conservancy business process description system.

[0026] S2: Use a large language model to manage water conservancy components, provide a unified access address for component registration, integrate all component services, and build a component platform.

[0027] Provide a unified management platform for the water conservancy business components dedicated to the large language model to integrate and manage all component services. The managed components include basic components and business components. The basic components mainly include process node components, database query components, routing selection components, etc. These components are mainly built-in by the system. The process node components are mainly input and output constraints at the beginning and end of the business chain. In more detail, they are what data the user needs to provide at the beginning of the business chain and what data the process needs to output at the end of the business chain; the function of the database query component is mainly to use the large language model to retrieve data according to the user's intention; the function of the routing selection component is mainly to control the direction of the business under different circumstances.

[0028] Water conservancy business components can be implemented by users uploading their own developed component interfaces or source codes. When uploading component interfaces or source codes, it is also necessary to upload component description information, including component purpose, function, input and output format, etc. After uploading, the system will parse the components submitted by the user, extract and translate these description information and input and output rules, convert them into JSON Schema format that can be understood by the large language model, and store them in the vector database according to component category. After parsing and conversion are completed, the business component will be encapsulated as a new RESTful API interface for the construction of subsequent business chains.

[0029] S3: Use the large language model to learn the water conservancy business description standard and understand the relationship between components and businesses in the component platform to obtain a learned large language model.

[0030] Using the water conservancy business process description standard constructed in the early stage (see S1 for specific data content) and the standard business description chain generated according to user operations in the later stage (see S4 for specific content), the large language model is optimized and trained in a targeted manner through fine-tuning technology. Specifically, the process description standard and the standard business description chain are first pre-processed. The pre-processing includes cleaning, word segmentation and annotation to ensure the quality and consistency of the data. After the data is pre-processed, fine-tuning algorithms such as Qlora are used for efficient parameter fine-tuning, which effectively saves computing resources while maintaining model performance. In this process, the large language model gradually establishes a comprehensive understanding of various water conservancy business fields by learning the subject domain vectors, business process descriptions and related annotations stored in the vector database, and makes intelligent dynamic responses based on specific business needs in practical applications.

[0031] On this basis, when specific business needs are met, prompt engineering strategies will be adopted to further enhance the intelligence level of the large language model in business understanding and component calling. Specifically, it includes: using running examples of related problems, building example data sets based on few-shot and Least-To-Most strategies, and using a small number of samples for rapid learning of user problem categories to enhance the accuracy of problem execution. In addition, the system task prompt template will be constructed, and the relevant water conservancy business rules will be used as the context information of this dialogue to constrain and inspire the behavior of the large language model, and obtain a learned large language model.

[0032] S4: The learned large language model generates a business expression chain based on the user’s water conservancy business needs.

[0033] After the user's needs are input, the user's question will first be vectorized and encoded, and the cosine similarity between it and the rule data, example data and component data in the vector library will be compared, and the part related to the user's question will be retrieved and dynamically constructed into the prompt set, example set and component set of the current dialogue. The large language model uses the prompt set as the context information of the current dialogue and the example set as the reference example of the task. Then, the large language model provides demand guidance to the user based on the user's needs, and guides the user to improve the business needs. For example, if the user does not mention the specific model algorithm used, the large language model will guide the user to supplement the part. When the user completes the demand content, the large language model will select the most appropriate component from the component set, splice the component information in the form of a business chain, and display the operating logic of the water conservancy business in the form of components in series or parallel, which is the business description chain.

[0034] like Figure 2As shown in the figure, the business description chain can be composed of a main chain or a combination of multiple sub-chains. The starting and stage parts of the chain are composed of fixed process component descriptions to control the input and output constraints of the entire link, such as controlling the necessary data that users need to provide when executing the business at the starting stage, and controlling the data type and data format that the business needs to output at the end to achieve the business goal. The rest is completed by splicing the description information of the database query component, routing component and business component. The description of this part of the component includes the basic information of the component, the connection relationship between the components, the input and output of the component, and the description information of the node. Each information is specific to the model method and data selected by the business.

[0035] According to the structure of the link, the business presentation chain can be divided into a sequence chain and a routing chain: a. Sequence chain: Organize independent functional components according to the logical relationship between them to form a clear business execution path, that is, each step in the business process is performed in a predetermined logical order.

[0036] b. Routing chain: Based on the specific needs of users, the routing component dynamically selects the appropriate sequence chain, that is, enters different sequence chains to execute business under different scenarios or needs.

[0037] The rules and logic in the business execution process are reflected through the sequence chain and routing chain, guiding the decision-making criteria and priority of process execution.

[0038] S5: Dynamically respond to user needs based on the business expression chain and generate business applications.

[0039] After the business expression chain is generated, the chain can be corrected and optimized based on expert experience. Users can add, delete, and modify components in the expression chain, such as modifying the source of the selected data, modifying the model category, and refining the model parameters, etc., to ensure the accuracy and rationality of the business process. When the user finally confirms the business expression chain, the backend system will organize the components in the form of workflow and the connection relationship according to the content of the expression chain to generate business applications.

[0040] When a user raises a business question to a business application, the user question will first be parsed through the starting component of the business expression chain to parse out the parameter information required for the business; after parsing out the parameter information, the business will be executed in the sequence chain. If there is no similar conditional judgment link in the business, the business components will be directly executed in sequence until the end, and the execution results of all business components will be summarized through the end to output standardized business result data. If there is a judgment link in the business, it will enter the routing chain, and the routing component will determine what business to perform under what circumstances, that is, go to different sequence chains. The business process can be composed of multiple sequence chains and routing chains until the end, and obtain standardized business data.

[0041] When water conservancy business needs change, the system provides two response strategies: a. Reconstruct the business expression chain.

[0042] If the current expression chain cannot meet the business needs under the changes in links, users can re-upload business components, or propose business construction requirements to the big language model based on existing components. The big language model will guide users to supplement their requirements based on their business needs, and then filter components, generate new business expressions, and finally generate new business applications to adapt to the current changing business logic.

[0043] b. Modification based on the existing expression chain.

[0044] If only some links of the current expression chain need to be adjusted, the component information can be modified based on the original expression chain, including replacing the type of model algorithm, modifying model parameters, adding business judgments, etc., to dynamically adjust the business process according to actual needs and update the application functions.

[0045] Generating water conservancy business applications in the form of expression chains greatly improves the flexibility of the system compared to traditional application architectures, ensuring that business applications can respond quickly to changes and achieve dynamic updating and optimization of functions.

[0046] In a specific engineering field, the user initially proposes an application construction requirement to the large language model: construct a flood evolution system for the XX river section. The process of the present invention is as follows: (1) Remind users to supplement application building requirements When the large language model receives user needs, it will first retrieve the example set, prompt set and component set related to the construction of the flood evolution system from the vector library. The example set includes samples of the business expression chain of the flood evolution system in some sections, the prompt set includes the construction rules for the flood evolution system, and the component set includes the components under the flood evolution annotation category. Then the retrieved example set and prompt set are filled into the prompt template to guide the large language model to parse the user's questions. After parsing the user's needs, it will determine whether additional demand construction information is needed, and then guide the user to supplement the information. For example, for the demand of "constructing a flood evolution system for XX river section", the user will be required to supplement the data range and the specific model algorithm selected.

[0047] (2) Component screening business chain construction When the user completes the requirements, the large language model selects components from the component set according to the user's requirements, queries the structural relationship of the key sections of the river section according to the business logic of flood evolution, and selects and assembles the data components of the starting point of the river section, the runoff component, and the flood evolution component according to the hydrological flow relationship, and sets the default parameter input and output format of each component. The format of the cross-section flood output is usually a time series format, such as [(time, value), (time, value)...]. After the information of each component is set, all components are connected in series and parallel to construct a sequence chain and a routing chain to generate a business description chain. For this problem, the sequence chain can be data query-runoff calculation-cross-section flood evolution, and the routing chain can be cross-section 1 using model a for flood evolution and cross-section 2 using model b for flood evolution.

[0048] (3) Business chain modification to generate business applications After the business expression chain is generated, the user can modify the generated expression chain again. After the modification, the components are combined according to the connection relationship of the expression chain to generate a business application. After the application is generated, the user can ask questions to the application, such as "obtain the flood process of section 2", and the system backend will calculate the results in sequence according to the chain and convert the results into the output format set by the user.

[0049] The present invention proposes a method for dynamic response of water conservancy business based on a large language model. The method aims to construct a large language model that can autonomously learn water conservancy business processes and logic, so as to dynamically generate water conservancy business expressions in the face of dynamically changing water conservancy environments, thereby achieving efficient and accurate matching of business applications, thereby improving the ability of existing water conservancy applications to deal with emergencies and environmental changes.

[0050] Based on the same inventive concept, the present invention also provides a water conservancy business dynamic response system based on a large language model, comprising: The annotation establishment module is used to collect water conservancy business execution rules and establish water conservancy business process description standards based on the water conservancy business execution rules.

[0051] The platform construction module is used to manage water conservancy components using a large language model, provide a unified access address for component registration, integrate all component services, and build a component platform.

[0052] The learning module is used to use the large language model to learn the water conservancy business description standard and understand the relationship between components and businesses in the component platform to obtain a learned large language model.

[0053] The expression chain generation module is used to generate a business expression chain based on the user's water conservancy business needs based on the learned large language model.

[0054] The application generation module is used to dynamically respond to user needs according to the business expression chain and generate business applications.

[0055] Each module in the above-mentioned water conservancy business dynamic response system based on a large language model can be implemented in whole or in part by software, hardware, and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0056] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps in the embodiment of the method for dynamic response of water conservancy business based on a large language model. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0057] Furthermore, the present invention also provides a non-temporary computer-readable storage medium containing instructions, and a computer program is stored on the storage medium. For example, a memory containing instructions, the above instructions can be executed by a processor of a computer device to complete the above method. For example, a non-temporary computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device. When the computer program is executed by the processor, the steps in the embodiment of the method for dynamic response of water conservancy business based on a large language model can be implemented. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0058] It should be understood by those skilled in the art that embodiments of the present invention may provide methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0062] It should be pointed out that the specific implementation methods described above can enable those skilled in the art to understand the invention more comprehensively, but do not limit the invention in any way. Therefore, although the invention has been described in detail in this specification and embodiments, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved. Any simple change or equivalent replacement of the technical solution that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belongs to the protection scope of the present invention.

Claims

1. A method for dynamic response of water conservancy business based on a large language model, characterized in that: include: Collect water conservancy business execution rules and establish water conservancy business process description standards based on the water conservancy business execution rules; Use the large language model to manage water conservancy components, provide a unified access address for component registration, integrate all component services, and build a component platform; Use the big language model to learn the water conservancy business description standard and understand the relationship between components and businesses in the component platform to obtain a learned big language model; The learned large language model generates a business expression chain based on the user's water conservancy business needs; Dynamically respond to user needs based on the business expression chain and generate business applications.

2. According to claim 1, a method for dynamic response of water conservancy business based on a large language model is characterized in that: The collection of water conservancy business execution rules specifically includes: collecting business description rules, business expansion rules, component input and output rules, and the relationship between components and businesses; logical rules are divided according to business types, and subject domains of various business areas are constructed, and corresponding water conservancy business process description standards are formulated based on the subject domain types.

3. A method for dynamic response of water conservancy business based on a large language model according to claim 2, characterized in that: The use of a large language model to manage water conservancy components also includes users independently adding components according to specific business needs, and the large language model semantically manages the components based on their input, output and application functions.

4. A method for dynamic response of water conservancy business based on a large language model according to claim 3, characterized in that: The use of a large language model to learn the water conservancy business description standard and understand the relationship between components and businesses in the component platform specifically includes: The rules and relationships within each subject domain are sorted out and annotated according to the refined business needs. The annotated descriptions are fine-tuned and trained using the QLoRA method. In the process of generating business applications, the large language model uses few-shot and least-to-most strategies for typical business operation examples as reference information when executing business based on questions raised by users, and constructs the corresponding business rules into prompt templates as context information for the large language model, thus obtaining a learned large language model.

5. A method for dynamic response of water conservancy business based on a large language model according to claim 4, characterized in that: The learned large language model generates a business expression chain according to the user's water conservancy business needs, specifically including: Through intent recognition, the business process description under the corresponding subject domain is found, and the business process is dynamically constructed based on the prompt project; according to the user intention category, the corresponding components are selected from the component platform to build a business component set, and the operating logic of the water conservancy business is generated into a business expression chain in the form of component series or parallel connection; the expression chain includes the called components, the connection relationship between components, input and output parameters, data sources, data types and system functions.

6. A method for dynamic response of water conservancy business based on a large language model according to claim 5, characterized in that: The dynamically responding to user needs according to the business expression chain and generating business applications specifically includes: Users modify the data nodes in the representation chain according to actual needs. The system background adjusts the node data according to the input and output constraints based on the content of the representation chain, and filters the database table of the business system according to the system's functions and component types, ultimately realizing the application construction based on business needs; when the water conservancy business needs change, additions, deletions or modifications are made based on the existing business representation chain, thereby realizing dynamic updates of the business system functions.

7. A water conservancy business dynamic response system based on a large language model, characterized in that: include: A labeling establishment module is used to collect water conservancy business execution rules and establish water conservancy business process description standards based on the water conservancy business execution rules; The platform construction module is used to manage water conservancy components using a large language model, provide a unified access address for component registration, integrate all component services, and build a component platform; The learning module is used to learn the water conservancy business description standard using the large language model and understand the relationship between components and businesses in the component platform to obtain a learned large language model; The expression chain generation module is used to generate a business expression chain based on the user's water conservancy business needs after learning the large language model; The application generation module is used to dynamically respond to user needs according to the business expression chain and generate business applications.

8. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is loaded into a processor, it can execute the steps of the method according to any one of claims 1 to 6.

Images