Enterprise application component building method and system based on large language model

Abstract

The invention provides an enterprise application component building method and system based on a large language model, and the method comprises the steps: S1, obtaining a user component demand, carrying out the analysis, generating a structured demand document, and determining whether to enter a step S2 or not according to a user demand confirmation result; s2, generating a core function module of a component through a component generation engine based on the structured demand document confirmed by the user; s3, matching a corresponding interface template from a preset universal interface library according to an adaptive system list in the structured demand document, and generating an interface adaptive code; and S4, deploying the core function module and the interface adaptation code to a preview environment for a user to preview, and selecting to automatically package and generate a component installation package or re-executing the steps S1, S2 or S3 according to a user preview confirmation result. According to the method, a full-link automation mechanism of'natural language demand-component generation-interface adaptation 'is constructed, so that the component building threshold is reduced, and rapid adaptation of the component and the existing system is realized.

Description

Technical Field

[0001] This invention relates to the field of enterprise application development technology, and in particular to a method and system for building enterprise application components based on a large language model. Background Technology

[0002] In the current process of enterprise digital transformation, building enterprise application components (such as customer management components, order processing components, and data analysis components) is a core step in constructing business systems. Traditional enterprise application component building models heavily rely on the coding skills of professional developers. Their processes typically include multiple stages such as requirements analysis, architecture design, code development, interface adaptation, and testing and debugging, and have the following significant drawbacks: 1. High development threshold: Component building requires mastery of multiple development languages ​​and frameworks such as Java, Python, and Vue. Business personnel without technical backgrounds cannot participate directly, resulting in communication barriers between requirements and development, and easy misunderstandings of requirements. 2. Long deployment cycle: A typical enterprise application component usually takes several weeks from requirement confirmation to deployment, with code development and interface adaptation accounting for more than 60% of the time, which cannot meet the needs of enterprises for rapid business iteration; 3. Difficulty in interface adaptation: Existing enterprise systems (such as ERP, CRM, OA, etc.) have diverse interface protocols (such as RESTful, SOAP, RPC, etc.). The integration of components with existing systems requires manual development of adaptation code, which results in low adaptation efficiency and difficulty in ensuring compatibility. 4. Cumbersome iterative adjustments: When business requirements change, code needs to be modified and interfaces need to be debugged, resulting in long adjustment cycles and poor flexibility.

[0003] In recent years, large language models have made breakthroughs in natural language understanding and code generation, enabling the automated construction of enterprise application components. However, currently, no technology can integrate natural language requirement parsing, core component function generation, automatic interface adaptation, and real-time preview adjustments into a fully automated mechanism, failing to fundamentally address the pain points of traditional component construction models. Furthermore, existing methods do not adequately consider the security, compliance, and integration with existing DevOps toolchains of the generated code, making them unsuitable for direct deployment in enterprise production environments. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for building enterprise application components based on a large language model, in order to solve the above-mentioned problems in the prior art.

[0005] This invention provides a method for building enterprise application components based on a large language model, including: S1: Obtain user component requirements, parse the component requirements, generate a structured requirement document, and select whether to proceed to step S2 based on the user requirement confirmation result; S2: Based on the structured requirements document confirmed by the user, the core functional modules of the component are generated through the component generation engine; S3: Match the corresponding interface template from the pre-set general interface library according to the adaptive system list in the structured requirements document, and generate interface adaptation code; S4: Deploy the core functional modules and interface adaptation code to the preview environment for user preview. Based on the user's preview confirmation, choose to automatically package and generate a component installation package or re-execute steps S1, S2, or S3.

[0006] This invention provides a system for building enterprise application components based on a large language model, comprising: The requirement parsing module is used to obtain user component requirements, parse the component requirements, generate a structured requirement document, and select whether to execute the component generation module based on the user requirement confirmation result. The component generation module is used to generate the core functional modules of components based on the structured requirements document confirmed by the user through the component generation engine. The interface adaptation module is used to match the corresponding interface template from the pre-set general interface library according to the adaptation system list in the structured requirements document, and generate interface adaptation code. The preview module is used to deploy the core functional modules and interface adaptation code to the preview environment for users to preview. Based on the user's preview confirmation, it can choose to automatically package and generate a component installation package or re-execute the requirement analysis module, component generation module, or interface adaptation module.

[0007] This invention also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the above-described method for building enterprise application components based on a large language model.

[0008] This invention also provides a computer-readable storage medium storing an information transmission implementation program, which, when executed by a processor, implements the steps of the above-described method for building enterprise application components based on a large language model.

[0009] The following advantages can be achieved by adopting embodiments of the present invention: Embodiments of the present invention provide a method for rapidly building enterprise application components based on a large model. By constructing a fully automated mechanism encompassing "natural language requirements - component generation - interface adaptation," it eliminates the need for manual code development, lowers the barrier to component building, shortens the building cycle from weeks to hours, and enables rapid adaptation of components to existing systems. Simultaneously, embodiments of the present invention address the security and compliance issues of generated code and ensure that the generated components can be seamlessly integrated into the enterprise's existing IT infrastructure and development processes. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in one or more embodiments of this specification or in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart of the enterprise application component building method based on a large language model according to an embodiment of the present invention; Figure 2 This is a schematic diagram of an enterprise application component building system based on a large language model, according to an embodiment of the present invention. Detailed Implementation

[0012] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this document.

[0013] Method Implementation Examples According to embodiments of the present invention, a method for building enterprise application components based on a large language model is provided. Figure 1 This is a flowchart of the enterprise application component building method based on a large language model according to an embodiment of the present invention, such as... Figure 1 As shown, the method for building enterprise application components based on a large language model according to an embodiment of the present invention specifically includes: Step S101: Obtain user component requirements, parse the component requirements, generate a structured requirements document, and select whether to proceed to step S102 based on the user requirements confirmation result. Specifically, this includes: Obtain component requirements input by the user in natural language form, parse the component requirements using a large language model, extract core elements and generate a structured requirement document, and feed the structured requirement document back to the user for confirmation; if the user confirms, proceed to step S102; if the user proposes modifications, update the structured requirement document according to the modifications and feed it back to the user for confirmation again. The large language model is fine-tuned based on the company's historical development documents and business specifications, and supports clarifying user needs through multi-round dialogues; the core elements include component types, functional modules, data dimensions, interaction requirements, business rules, and related systems; the structured requirements document includes a function list, data field definitions, interaction logic descriptions, and a list of compatible systems.

[0014] Step S102: Based on the structured requirements document confirmed by the user, generate the core functional modules of the component through the component generation engine, specifically including: Based on the structured requirements document confirmed by the user, the component generation engine is invoked to automatically generate the core functional modules of the component; The core functional modules include data structure, interaction logic, and visualization modules; the data structure includes database table structure, data entity classes, and data validation rules; the interaction logic includes front-end interaction code and back-end business logic code.

[0015] Step S103: Match the corresponding interface template from the pre-set general interface library according to the adaptation system list in the structured requirements document, and generate interface adaptation code, specifically including: Based on the list of compatible systems in the structured requirements document, the corresponding interface templates are matched from the pre-set general interface library to automatically generate interface adaptation code, and the interface compatibility is verified by the compatibility detection engine. The general interface library includes standardized interface templates for RESTful, SOAP, and RPC protocols, as well as customizable interfaces for common enterprise systems, and adopts a dynamic update mechanism. It also supports user-uploaded interface specification files to customize and extend interface templates.

[0016] Step S104: Deploy the core functional modules and interface adaptation code to the preview environment for user preview. Based on the user's preview confirmation, choose to automatically package and generate a component installation package or re-execute steps S101, S102, or S103. Specifically, this includes: The core functional modules and interface adaptation code are deployed to the preview environment for users to preview in real time, and user preview feedback is received. If the user confirms, the component installation package is automatically generated. If the user requests adjustments, at least one of steps S101, S102 or S103 is re-executed based on the adjustment requests. The preview environment is a Docker container sandbox environment; the adjustment requirements are input through natural language or visual configuration; the adjustment process supports version management and rollback.

[0017] The following describes in detail the above-mentioned technical solutions of the present invention with reference to the specific circumstances of the enterprise application component building method based on the large language model in the embodiments of the present invention.

[0018] This invention proposes an automated construction method for enterprise application components based on a large language model. This method overturns the traditional component development model that relies on manual coding, constructing a full-link automated mechanism of "natural language requirement input - intelligent component generation - automatic interface adaptation." Its core lies in: using a large language model to parse user business requirements described in natural language (such as "create a customer information management component, including name, phone number, and customer level fields"), automatically generating complete component code containing data entity structure, front-end UI interface, and interaction logic; automatically matching and generating integration interfaces with target business systems (such as ERP and CRM) through a built-in universal interface adaptation library, achieving seamless integration; and providing real-time component preview and parameter visualization adjustment functions, allowing users to instantly verify and fine-tune component functionality during the generation process. This invention shortens the development and construction cycle of enterprise-level application components from weeks to hours, significantly reducing the development threshold and system adaptation costs.

[0019] This invention proposes a method for rapidly building enterprise application components based on a large language model, comprising the following steps: 1. Natural Language Request Reception and Parsing ① Build a requirement input interface to support enterprise users to input component requirements in natural language (such as "Develop a customer management component, including customer information entry, query, modification and deletion functions, and need to statistically analyze customer source distribution"); ② Call a pre-trained large language model (such as an enterprise-customized model based on GPT-4 architecture) to perform semantic understanding and intent recognition of natural language requirements, and extract the core elements in the requirements, including component type, functional module, data dimension, interaction requirements, business rules and related systems; ③ The large language model generates a structured requirement document based on the extracted core elements, including a list of functions, data field definitions, interaction logic descriptions, and a list of compatible systems, and provides feedback to the user for confirmation. If the user provides modification suggestions, the model automatically iterates and optimizes the structured requirement document. ④ This large language model supports input in multiple languages, including Chinese and English, and guides users to clarify ambiguous needs through a multi-turn dialogue mechanism to ensure the accuracy of understanding the needs.

[0020] 2. Component core functional modules are automatically generated. Based on the confirmed structured requirements document, the large language model invokes its built-in component generation engine to automatically generate front-end user interface code, back-end data model definitions, and core business logic code that conform to the target technology stack, forming a runnable application component, specifically including: ① Data structure generation: Based on the data field definition, automatically generate database table structure (supports multiple database types such as MySQL, Oracle, MongoDB), data entity classes and data validation rules; ② Interaction Logic Generation: Based on the interaction logic description, generate front-end interaction code (supporting frameworks such as Vue and React) and back-end business logic code (supporting frameworks such as Spring Boot and Node.js) to implement the query, add, modify, and delete functions defined in the requirements; the generated code includes necessary unit test cases and API documentation comments, and follows the coding standards agreed upon by the enterprise; ③ Visualization module generation: Automatically generate chart components (such as bar charts, pie charts, and line charts) for data analysis needs, and support real-time data rendering; ④ The component generation engine has a built-in code optimization and security scanning module that automatically formats, removes redundancy, optimizes performance, and performs static security vulnerability scanning (SAST) on the generated code to identify and avoid common security risks (such as SQL injection, XSS attacks, etc.).

[0021] 3. Automatic interface adaptation ① Build a general interface library. This library has built-in standardized interface templates for various mainstream protocols (RESTful, SOAP, RPC) as well as dedicated adaptation interfaces for common enterprise systems (such as SAP, Yonyou, Kingdee ERP, Salesforce CRM). The library supports connection to enterprise private systems and allows customization and extension of interface templates by uploading WSDL, OpenAPI (Swagger) specification files or configuring connection parameters.

[0022] ② Based on the list of compatible systems specified by the user or in the structured requirements document, the interface adaptation module matches the corresponding interface template from the general interface library, and automatically generates interface adaptation code in combination with the core functional modules generated by the component, so as to realize the data interaction (data reading and writing, function calling) between the component and the existing system; for complex or non-standard interfaces, it supports the generation of interface mock code to facilitate parallel development and testing. ③ The interface adaptation module has a built-in compatibility detection engine that automatically detects the generated adapted interfaces, verifies the integrity and stability of data transmission, and automatically adjusts interface parameters or calls backup interface templates from the general interface library if compatibility issues are detected. This engine also performs data anonymization and compliance checks to ensure that sensitive information complies with data protection regulations such as GDPR.

[0023] 4. Real-time component preview and parameter adjustment ① Set up a component preview environment and deploy the generated core functional modules and interface adaptation code to the preview environment (supporting independent sandbox environments based on Docker), allowing users to preview the functional effects and interactive experience of the components in real time through a browser; ② Provides a parameter adjustment interface. Users can submit adjustment requests through natural language (such as "add a date range filter to the query conditions") or visual configuration methods (such as dragging and dropping component layouts and modifying field attributes). After receiving the adjustment requests, the large language model automatically updates the corresponding functional module code and interface adaptation code to form a closed-loop optimization and updates the preview effect simultaneously. The adjustment process supports version comparison, and users can view the change records and roll back to any historical version. ③ Once the user confirms that the preview effect meets the requirements, the component generation system automatically packages the component code, generates an installation package that can be directly deployed (supports Docker containerized deployment), and synchronously generates continuous integration / continuous deployment (CI / CD) pipeline configuration files (such as Jenkinsfile, GitLab CI Yaml), which facilitates automated integration into the enterprise's existing DevOps platform.

[0024] 5. System integration and scalability ① Provides a complete RESTful API, allowing component building capabilities to be embedded into third-party platforms or workflows (such as enterprise low-code platforms or project management software Jira); ② The system supports a component template marketplace mechanism, allowing users to publish verified components as templates for reuse in other projects, promoting internal sharing and asset accumulation within the enterprise; ③ The general interface library and component generation engine adopt a plug-in architecture, which supports the development of custom plug-ins to extend the supported databases, UI frameworks or business logic templates.

[0025] Furthermore, the large language model in Step 1 employs an enterprise-customized training method (specialized training on enterprise-level business scenario corpora and component design patterns, enabling it to understand common functional terms in enterprise-level applications such as "CRUD operations," "forms," ​​"dashboards," and "approval workflows"). It is fine-tuned by inputting historical enterprise component development documents, business process specifications, and interface protocol documents to improve the parsing accuracy for specific enterprise business needs. The training process prioritizes data privacy, employing either federated learning or localized deployment modes to ensure that sensitive enterprise data remains within its domain.

[0026] Furthermore, the general interface library adopts a dynamic update mechanism. By crawling the interface update logs of mainstream industry systems and user feedback on adaptation requirements, the interface templates are updated regularly to ensure the timeliness and compatibility of interface adaptation. At the same time, an interface compatibility test case library is established, and regression testing is automatically performed after each update to ensure stability.

[0027] Furthermore, the component generation engine has a built-in code optimization module that automatically formats the generated code, removes redundant code, and optimizes performance to improve component running efficiency. The optimization process includes, but is not limited to, database index suggestions, API response caching mechanism generation, and front-end component lazy loading strategies.

[0028] Furthermore, in this embodiment of the invention, the visual preview interface and the code generation engine are bidirectionally bound, so that the user's adjustments to the UI elements, layout, or attributes in the preview component can be synchronously mapped and modified in real time and automatically.

[0029] Example 1: Building a Customer Management Component 1. Natural Language Input Requirements: Users enter the following in the requirements input interface: "Develop a customer management component, including basic customer information (name, phone number, company, position) entry, fuzzy search by company name, and pie chart statistics by customer source (online promotion, offline exhibitions, old customer referrals); it needs to be integrated with the company's existing Yonyou ERP system to synchronize customer order data."

[0030] 2. Requirements Analysis: After parsing the large language model, a structured requirements document was generated, clearly defining the component type as "Customer Management Component." Functional modules include "Information Entry," "Fuzzy Search," and "Source Statistics." Data fields include Name (string), Phone Number (string), Company (string), Position (string), and Customer Source (enumeration type). The compatible system is "Yonyou ERP," and the user confirmed its accuracy. The model indicated that the "Phone Number" field might involve personal privacy and recommended encrypting its storage in the database and displaying it de-identified on the front end in non-essential scenarios. The user adopted this suggestion.

[0031] 3. Core Functionality Generation: The component generation engine automatically generates the MySQL database table "customer_info", including the aforementioned data fields, primary key, creation time, and other default fields, and creates an index for the 'Company' field; it generates Vue front-end code to implement the information entry form, query box, and pie chart rendering; and it generates Spring Boot back-end code to implement data CRUD and statistical logic. The code optimization module automatically generates parameterized query code to prevent SQL injection, and the security scanning module found no vulnerabilities.

[0032] 4. Interface Adaptation: The interface adaptation module matches the "Yonyou ERP RESTful Adaptation Interface" template from the general interface library and automatically generates interface adaptation code to achieve synchronization between customer information and Yonyou ERP order data; the compatibility detection engine passes the test, there are no data transmission problems, and the detection engine confirms that the transmitted order data does not contain sensitive monetary information, which complies with the enterprise's financial data security regulations.

[0033] 5. Preview and Adjustment: Users test the information entry and query functions in the preview interface and confirm that the pie chart data is displayed correctly; they suggest "changing the customer source statistics to a bar chart", and the large language model automatically updates the front-end chart code. After the preview effect is updated, the user confirms; the system generates a Docker installation package and the corresponding Kubernetes deployment description file, and the component is built. The whole process takes only 1.5 hours.

[0034] Example 2: Building an Order Processing Component 1. Natural Language Input Requirements: The user inputs "Develop an order processing component that supports order creation (associating customer ID, product information, and amount), modification of order status (pending payment, paid, shipped, completed), and counting the number of orders by status; integrate with Salesforce CRM to obtain customer IDs and with the logistics system to obtain tracking numbers."

[0035] 2. Requirements Analysis and Function Generation: After the large language model analyzes the requirements, it generates a structured document. The component generation engine automatically generates database tables and front-end and back-end code. For the 'amount' field, the back-end code automatically generates precision processing logic (Decimal type). The interface adaptation module matches the "Salesforce CRM SOAP interface" and "logistics system RPC interface" templates, generates adaptation code, and passes compatibility testing. Since the logistics system interface is an internal private protocol, the system guides users to create and learn a custom adaptation template by uploading interface documents.

[0036] 3. Preview and Adjustment: When the user previewed the app, they requested that "orders with an amount greater than 1,000 yuan be automatically marked as priority orders". The large language model updated the backend logic code, and the installation package was generated after the preview was confirmed. The whole process took 2 hours.

[0037] Example 3: Demonstration of Safety and Compliance A financial institution needs to build a "loan approval dashboard" component, which involves customer identity information. During the system's generation process: 1. Requirements Analysis Phase: Prompt users to include this component in the scope of the Enterprise Privacy Impact Assessment (PIA).

[0038] 2. Code generation stage: The database layer automatically encrypts fields such as "ID number" and "mobile phone number" using AES, and performs desensitization in the log output layer.

[0039] 3. Interface adaptation phase: When connecting with the internal risk control system, the compatibility detection engine verifies that the data transmission channel is an encrypted VPN and prompts for the authorization period.

[0040] The final generated component successfully passed the audit by the enterprise security team.

[0041] System Implementation Examples According to embodiments of the present invention, an enterprise application component building system based on a large language model is provided. Figure 2 This is a schematic diagram of an enterprise application component building system based on a large language model, as described in an embodiment of the present invention. Figure 2 As shown, the enterprise application component building system based on a large language model according to an embodiment of the present invention specifically includes: The requirement parsing module 20 is used to obtain user component requirements, parse the component requirements, generate a structured requirement document, and select whether to execute the component generation module based on the user requirement confirmation result. Specifically, it is used for: The system obtains component requirements input by the user in natural language, parses the component requirements using a large language model, extracts core elements, and generates a structured requirement document. The structured requirement document is then fed back to the user for confirmation. If the user confirms, the component generation module is executed. If the user suggests modifications, the structured requirement document is updated based on the modifications and fed back to the user for confirmation. The large language model is fine-tuned based on the company's historical development documents and business specifications, and supports clarifying user needs through multi-round dialogues; the core elements include component types, functional modules, data dimensions, interaction requirements, business rules, and related systems; the structured requirements document includes a function list, data field definitions, interaction logic descriptions, and a list of compatible systems.

[0042] Component generation module 22 is used to generate the core functional modules of components based on the structured requirements document confirmed by the user through the component generation engine. The interface adaptation module 24 is used to match the corresponding interface template from the pre-set general interface library according to the adaptation system list in the structured requirements document, and generate interface adaptation code. Preview module 26 is used to deploy the core functional modules and interface adaptation code to the preview environment for user preview. Based on the user's preview confirmation, it selects to automatically package and generate a component installation package or re-execute the requirement analysis module, component generation module, or interface adaptation module. Specifically, it is used for: The core functional modules and interface adaptation code are deployed to the preview environment for users to preview in real time, and user preview feedback is received. If the user confirms, the component installation package is automatically generated. If the user proposes an adjustment, at least one of the requirement parsing module, component generation module, or interface adaptation module is re-executed based on the adjustment requirement. The preview environment is a Docker container sandbox environment; the adjustment requirements are input through natural language or visual configuration; the adjustment process supports version management and rollback.

[0043] The embodiments of the present invention are system embodiments corresponding to the above method embodiments. The specific operation of each module can be understood by referring to the description of the method embodiments, and will not be repeated here.

[0044] In summary, the embodiments of the present invention have the following beneficial effects compared with the prior art: 1. Revolutionizing the traditional development model: Business personnel can directly input requirements through natural language without relying on manual code development, realizing the full-chain automated construction of components, greatly reducing the development threshold and eliminating communication barriers between requirements and development; 2. Significantly shorten the deployment cycle: Reduce the traditional component deployment cycle of several weeks to several hours, meeting the needs of enterprises for rapid business iteration; 3. Achieve automatic interface adaptation: Through a general interface library and compatibility detection engine, components can be seamlessly integrated with existing systems without the need for manual development of adaptation code, improving adaptation efficiency and compatibility; 4. Supports real-time adjustment and optimization: Users can intuitively view the component effect through the preview interface and quickly adjust parameters through natural language or visualization to achieve flexible optimization of the component; 5. Ensure security and compliance: Built-in code security scanning and data compliance checks mitigate security vulnerabilities and compliance risks from the source, generating high-quality code usable at the enterprise level; 6. Enhanced integration and scalability: Provides API integration, plug-in architecture, and template marketplace, making it easy to integrate with existing enterprise toolchains and supporting continuous expansion, thus protecting enterprise investments.

[0045] Device Example 1 This invention provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, performs the steps described in the method embodiment.

[0046] Device Example 2 This invention provides a computer-readable storage medium storing an information transmission implementation program, which, when executed by a processor, performs the steps described in the method embodiment.

[0047] The computer-readable storage media described in this embodiment include, but are not limited to, ROM, RAM, disk, or optical disk.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for building enterprise application components based on a large language model, characterized in that, include: S1: Obtain user component requirements, parse the component requirements, generate a structured requirement document, and select whether to proceed to step S2 based on the user requirement confirmation result; S2: Based on the structured requirements document confirmed by the user, the core functional modules of the component are generated through the component generation engine; S3: Match the corresponding interface template from the pre-set general interface library according to the adaptive system list in the structured requirements document, and generate interface adaptation code; S4: Deploy the core functional modules and interface adaptation code to the preview environment for user preview. Based on the user's preview confirmation, choose to automatically package and generate a component installation package or re-execute steps S1, S2, or S3.

2. The method according to claim 1, characterized in that, S1 specifically includes: Obtain component requirements input by the user in natural language form, parse the component requirements using a large language model, extract core elements and generate a structured requirement document, and feed the structured requirement document back to the user for confirmation; if the user confirms, proceed to step S2; if the user proposes modifications, update the structured requirement document according to the modifications and feed it back to the user for confirmation again. The large language model is fine-tuned based on the company's historical development documents and business specifications, and supports clarifying user needs through multi-round dialogues; the core elements include component types, functional modules, data dimensions, interaction requirements, business rules, and related systems; the structured requirements document includes a function list, data field definitions, interaction logic descriptions, and a list of compatible systems.

3. The method according to claim 2, characterized in that, S2 specifically includes: Based on the structured requirements document confirmed by the user, the component generation engine is invoked to automatically generate the core functional modules of the component; The core functional modules include data structure, interaction logic, and visualization modules; the data structure includes database table structure, data entity classes, and data validation rules; the interaction logic includes front-end interaction code and back-end business logic code.

4. The method according to claim 2, characterized in that, S3 specifically includes: Based on the list of compatible systems in the structured requirements document, the corresponding interface templates are matched from the pre-set general interface library to automatically generate interface adaptation code, and the interface compatibility is verified by the compatibility detection engine. The general interface library includes standardized interface templates for RESTful, SOAP, and RPC protocols, as well as customizable interfaces for common enterprise systems, and adopts a dynamic update mechanism. It also supports user-uploaded interface specification files to customize and extend interface templates.

5. The method according to claim 1, characterized in that, S4 specifically includes: The core functional modules and interface adaptation code are deployed to the preview environment for users to preview in real time, and user preview feedback is received. If the user confirms, the component installation package is automatically generated. If the user proposes an adjustment, at least one of steps S1, S2 or S3 is re-executed based on the adjustment request. The preview environment is a Docker container sandbox environment; the adjustment requirements are input through natural language or visual configuration; the adjustment process supports version management and rollback.

6. A system for building enterprise application components based on a large language model, characterized in that, include: The requirement parsing module is used to obtain user component requirements, parse the component requirements, generate a structured requirement document, and select whether to execute the component generation module based on the user requirement confirmation result. The component generation module is used to generate the core functional modules of components based on the structured requirements document confirmed by the user through the component generation engine. The interface adaptation module is used to match the corresponding interface template from the pre-set general interface library according to the adaptation system list in the structured requirements document, and generate interface adaptation code. The preview module is used to deploy the core functional modules and interface adaptation code to the preview environment for users to preview. Based on the user's preview confirmation, it can choose to automatically package and generate a component installation package or re-execute the requirement analysis module, component generation module, or interface adaptation module.

7. The system according to claim 6, characterized in that, The requirement analysis module is specifically used for: The system obtains component requirements input by the user in natural language, parses the component requirements using a large language model, extracts core elements, generates a structured requirement document, and sends the structured requirement document back to the user for confirmation. If the user confirms, the component generation module is executed. If the user suggests modifications, the structured requirements document will be updated according to the suggested modifications and the user will be notified again. The large language model is fine-tuned based on the company's historical development documents and business specifications, and supports clarifying user needs through multi-round dialogues; the core elements include component types, functional modules, data dimensions, interaction requirements, business rules, and related systems; the structured requirements document includes a function list, data field definitions, interaction logic descriptions, and a list of compatible systems.

8. The system according to claim 6, characterized in that, The preview module is specifically used for: The core functional modules and interface adaptation code are deployed to the preview environment for users to preview in real time, and user preview feedback is received. If the user confirms, the component installation package is automatically generated. If the user proposes an adjustment, at least one of the requirement parsing module, component generation module, or interface adaptation module is re-executed based on the adjustment requirement. The preview environment is a Docker container sandbox environment; the adjustment requirements are input through natural language or visual configuration; the adjustment process supports version management and rollback.

9. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored on the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method for building enterprise application components based on a large language model as described in any one of claims 1-5.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an information transmission implementation program, which, when executed by a processor, implements the steps of the enterprise application component building method based on a large language model as described in any one of claims 1-5.

Images