Software development system and development method based on Internet software theory

Abstract

The invention discloses a software development system and a software development method based on an internet software theory. The system comprises an account registration and login module, a user detailed information acquisition module, a project information generation or acquisition module, a project data preprocessing module, an AI generation content selection module, a project content analysis module, a software system generation module, a manual tuning module and the like. The system has multiple functional module designs, introduces the AI technology to carry out deep semantic understanding and process derivation, and can process complex business logic. Meanwhile, the system supports code injection, and developers are allowed to write custom codes to achieve complex functions.

Description

Technical Field

[0001] This invention is a new code development technology based on network software theory, specifically involving a software development system and development method based on network software theory. Background Technology

[0002] Traditional software production and traditional low-code development face the following problems: 1. Static architecture: The system is difficult to adapt to dynamic business needs and has poor scalability; 2. Difficulty in implementing complex business logic. Traditional low-code development is a visual application development approach that reduces the workload of manual coding. However, it suffers from limitations in functionality, performance and scalability issues, integration and security risks when dealing with complex business scenarios. The pre-built components and templates provided by the platform are insufficient to support complex logic involving multi-table joins, dynamic conditional judgments, recursive algorithms, etc., and also lack advanced functional components for specific industries.

[0003] 3. Low degree of customization. The platform's interface templates, data field types, and interaction methods often fail to fully meet the needs of a company's brand image and business processes.

[0004] 4. Performance Bottlenecks. To achieve rapid development, low-code platforms compromise on underlying architecture and code generation methods, resulting in poor performance of the generated applications when handling large amounts of data or high-concurrency requests. Workflow engines introduce additional overhead, the automatically generated code is of low quality, and some platforms use resource-intensive technology stacks.

[0005] 5. Limited flexibility and scalability. Low-code platforms rely on fixed templates and component libraries, making it difficult to meet personalized needs. The generated code structure is messy and difficult to maintain, and there is insufficient support for complex projects.

[0006] 6. Integration challenges. Low-code platforms may encounter compatibility issues when integrating with external systems, lacking standardized interfaces and plugin mechanisms.

[0007] 7. Security risks. Simplified security settings may lead to lax access control, and the closed nature of the platform also increases the difficulty of security auditing.

[0008] 8. Low resource utilization: Heterogeneous resources (services, data, devices) cannot be efficiently aggregated and coordinated; 9. High operational complexity: It lacks adaptive fault tolerance and intelligent decision-making capabilities and relies on manual intervention.

[0009] Therefore, there is an urgent need for a software-based intelligent production line system based on dynamic evolution theory to achieve high scalability, high availability, intelligence, and adaptability in an open environment. Summary of the Invention

[0010] Technical Objective: To address the aforementioned technical problems, this invention proposes a software development system and method based on network-structured software theory. It features multiple functional modules and incorporates AI technology for deep semantic understanding and process derivation, enabling it to handle complex business logic.

[0011] Technical Solution: To achieve the above-mentioned technical objectives, the present invention proposes the following technical solution: A software development method based on network architecture software theory includes: The account registration and login module is used to register accounts and log in using registered accounts. It supports the registration and login of various types of accounts, including end customers, partners, and sales, and assigns users corresponding IDs and permissions. The user details acquisition module uses AI to obtain publicly available customer details based on the user's business registration information, including industry classification, company introduction, product and organizational structure information. Based on the acquired customer details, it generates the software system's organizational structure, management accounts and user accounts, and assigns corresponding roles and operating permissions to each account according to job responsibilities. The project information generation or retrieval module is used to submit project information via documents or generate project materials online with AI assistance. The project data preprocessing module is used to check, standardize, and improve project data. The AI-generated content selection module generates project content for users to choose from based on their actual needs. The project content analysis module is used to analyze and examine the project content; The software system generation module is used to generate the project's software system, including access links, system menus, and generation processes.

[0012] Preferably, the system further includes: a manual optimization module, which includes multiple workstations with different functions, supports code injection, and performs calculations of attribute values ​​within objects and cross-object attribute calculations by writing code.

[0013] Preferably, the project data preprocessing module includes: Multi-format parsing unit, used for automatic parsing and structured extraction of input text; The intelligent paragraph segmentation unit is used to automatically divide logical paragraphs based on the semantic context, and automatically extract the core keywords of the paragraphs to generate tags.

[0014] Preferably, the project information generation or acquisition module includes: The multimodal input parsing unit supports natural language descriptions, sketch photos, or Excel spreadsheet uploads, and automatically identifies business process elements. Business process requirements include participants, action nodes, and data flow.

[0015] Preferably, the project information generation or acquisition module includes: The intelligent field mapping unit is used to automatically identify field types through natural language descriptions and supports the direct generation of interactive forms from JSON configurations.

[0016] Preferably, the software system generation module includes: AI component orchestration is used to generate responsive layouts through a drag-and-drop designer, with AI recommending data tables and charts in real time and binding them to data sources.

[0017] Preferably, the system further includes a context-aware optimization module, used to enable the system to automatically detect page performance bottlenecks and provide pagination or loading optimization solutions.

[0018] Preferably, the system adopts a multi-model collaborative architecture, including Qwen-Plus as the main analysis engine, Jieba word segmentation tool for primary text structuring, and vector model for handling clause similarity matching tasks.

[0019] A software development method based on network architecture theory, used in the system, the method comprising the following steps: Register an account or log in using an existing account; AI is used to obtain publicly available customer details, including industry classification, company introduction, product and organizational structure information. Based on the obtained customer details, the software system's organizational structure is generated, management accounts and user accounts are generated, and each account is assigned a corresponding role and operation permission according to job responsibilities. Submit project information via document or generate project information online with AI assistance; The project materials were checked for legality, standardized, and improved. Generate project content for users to select according to their actual needs; Analyze and examine the project content; Generate the project software system, including access links, system menus, and workflows.

[0020] Preferably, after the project software system is generated, manual optimization is performed, including writing code to perform the calculation of attribute values ​​within objects and the calculation of attributes across objects.

[0021] Beneficial effects: Due to the adoption of the above technical solution, the present invention has the following beneficial effects: This invention features a multi-functional module design and incorporates AI technology for deep semantic understanding and process derivation, enabling it to handle complex business logic. Furthermore, the system supports code injection, allowing developers to write custom code to implement complex functions. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of a software development system based on network software theory according to the present invention; Figure 2 This is a schematic diagram illustrating the structure of an example of the manual tuning module designed in this invention; Figure 3 This is a partial schematic diagram illustrating the classification results of the project data in the project preprocessing module, which performs a three-level classification. Detailed Implementation

[0023] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] Example 1 like Figure 1 This embodiment presents a software development system based on network software theory, including: The account registration and login module is used to register accounts and log in using registered accounts. It supports the registration and login of various types of accounts, including end customers, partners, and sales, and assigns users corresponding IDs and permissions.

[0025] The user details acquisition module uses AI tools to obtain publicly available customer details based on the user's business registration information, including industry classification, company introduction, product and organizational structure information. Based on the acquired customer details, it generates the software system's organizational structure, management accounts and user accounts, and assigns corresponding roles and operating permissions to the management accounts and user accounts according to their job responsibilities.

[0026] The project information generation or acquisition module is used to submit project information via WORD documents or generate project materials online with AI assistance. AI-assisted online generation can help users form system requirements more clearly, explicitly, comprehensively, and accurately.

[0027] The project data preprocessing module is used to perform legality checks, standardization, and improvement processing on project data. The AI-generated content selection module uses AI tools to generate project content based on the preprocessed project data, allowing users to select content according to their actual needs.

[0028] The project content analysis module is used to analyze the project content and perform a final check before the final software is generated.

[0029] The software system generation module is used to generate the project's software system, including access links, system menus, and generation processes.

[0030] The system designed in this invention can also be designed with a manual tuning module, such as... Figure 2As shown, it includes multiple workstations with different functions, supports code injection, and allows users to perform calculations of attribute values ​​within objects and across objects by writing code.

[0031] The project information generation or acquisition module includes: The multimodal input parsing unit supports natural language descriptions, sketch photos, or Excel spreadsheet uploads, and automatically identifies business process elements. Business process requirements include participants, action nodes, and data flow.

[0032] Specifically, a sales manager might take a picture of a sales process diagram drawn on a whiteboard with their phone. The system can then identify process nodes such as "potential customer," "initial contact," "needs analysis," "solution development," and "contract signing," as well as the relationships between them. Simultaneously, the system can also identify participants such as "sales representative," "customer," and "sales manager." This information is automatically converted into structured data that the system can understand, used to generate an initial CRM system process design.

[0033] The intelligent field mapping unit is used to automatically identify field types through natural language descriptions and supports the direct generation of interactive forms from JSON configurations.

[0034] Specifically, when designing a human resources management system, users might describe needing an "employee information table" containing fields such as "name," "date of birth," "ID number," "date of employment," and "department." The system automatically maps "name" to a text input box, "date of birth" and "date of employment" to date pickers, "ID number" to a special input box with 18 digits and an X character validation, and "department" to a dropdown selection box. These mappings are converted into JSON configuration, and then an interactive web form with corresponding input controls and validation logic is directly generated.

[0035] The development system designed in this invention is designed to accept two standard input formats: Text files (.txt): Must use a common encoding format, and the file size must not exceed 10MB. Support for parsing text containing simple Markdown tags (such as ### headings). Document (.docx): Fully preserves the style and structure of a Word document, automatically recognizing heading styles (Heading 1-3), bullet points, tables, and other elements. Compatible with files generated by WPS and Microsoft Office 2013+.

[0036] The project data preprocessing module includes: Multi-format parsing unit, used for automatic parsing and structured extraction of input text; The intelligent paragraph segmentation unit is used to automatically divide logical paragraphs based on contextual semantics, and automatically extract core keywords from paragraphs to generate tags. Specifically, it includes automatic parsing and structured extraction of text in formats such as TXT / Word / Excel, as well as automatically dividing logical paragraphs based on contextual semantics and automatically extracting core keywords from paragraphs to generate tag clouds.

[0037] Specifically, for example, a manufacturing company provides a Word document containing a company introduction, product descriptions, and business processes as project requirements. This embodiment's system can automatically parse this Word document, dividing it into logical paragraphs such as "Company Introduction," "Product Line Introduction," and "Production Process." For the "Product Line Introduction" paragraph, the system may extract keywords such as "high-precision machining," "intelligent control," and "energy saving and environmental protection" as tags. This information can be directly used in subsequent system design and development processes, such as automatically generating the basic structure of the product management module.

[0038] The software system generation module includes: The main operational tools involved in the software system generation module are as follows: Queue management tool: Distributed task queue\RabbitMQ.

[0039] Application creation tool: Spring Boot (Java framework).

[0040] Form generation tool.

[0041] Workflow generation tool: Activiti (workflow engine).

[0042] Menu generation tool.

[0043] Permission binding tool.

[0044] Homepage generation tool: ECharts (data visualization).

[0045] Complete notification tools: WebSocket (real-time notification), Email / SMS API (asynchronous notification).

[0046] AI component orchestration unit is used to generate responsive layouts through a drag-and-drop designer, with AI recommending data tables and charts in real time and binding them to data sources.

[0047] Specifically, when designing a sales data analysis dashboard, users can drag and drop various components onto the canvas. When a user drags into a blank area, the AI ​​might recommend components such as "Sales Trend Chart for the Past 30 Days," "Top 10 Customer List," and "Sales Percentage by Product Category." If the user selects "Sales Trend Chart for the Past 30 Days," the system will automatically identify the relevant data source (such as a sales order table) and generate a line chart component, with the x-axis representing dates and the y-axis representing sales revenue. Users can further adjust the chart style, filter conditions, etc., without having to write complex data query and chart drawing code.

[0048] The system also includes a context-aware optimization module, which enables the system to automatically detect page performance bottlenecks and provide pagination or loading optimization solutions.

[0049] Specifically, consider a customer management system with a list page displaying all customer information. The system detects that this page needs to load over 10,000 customer records, which might cause slow page loading. The system automatically provides optimization suggestions, such as "change the customer list to a paginated display, showing 50 records per page" or "implement virtual scrolling, rendering only the visible area of ​​data." If the developer chooses the pagination solution, the system automatically modifies the backend API to support paginated queries and implements pagination controls and data loading logic on the frontend. This can reduce the initial page load time from 10 seconds to less than 1 second, significantly improving the user experience.

[0050] The system of this invention adopts a multi-model collaborative architecture, including Qwen-Plus as the main analysis engine, Jieba word segmentation tool for primary text structuring, and a vector model for handling clause similarity matching tasks. The development system designed in this invention enables dynamic collaboration in a software intelligent production line. The system architecture is based on a "small model + big data" architecture and includes the following core layers: a. Resource Model Layer: Encapsulates heterogeneous computing resources (services, data, hardware resources) through containerization technology; supports dynamic registration, discovery, and resource pooling.

[0051] b. Collaborative Model Layer: Service composition optimization is achieved based on an intelligent decision engine; dynamic service orchestration and load-aware routing are adopted.

[0052] c. Service Model Layer: Supports on-demand service aggregation and dynamic refactoring.

[0053] d. Interface Model Layer: Adapts to multiple protocols such as message queues, making it compatible with traditional systems and emerging technology stacks.

[0054] e. AI Model Layer: Based on large AI models such as DeepSeek, it enables data analysis and reasoning.

[0055] f. HWAD Model Layer: A self-built HWAD (Human Wisdom AI Data) model layer that enables hardware adaptation for model building and application software.

[0056] g. Manual optimization layer: The model is optimized through manual parameter annotation to improve the quality and efficiency of the construction.

[0057] Example 2 A software development method based on network architecture theory, applied to the development system in Example 1, includes the following steps: Register an account or log in using an existing account; AI is used to obtain publicly available customer details, including industry classification, company introduction, product and organizational structure information. Based on the obtained customer details, the software system's organizational structure is generated, management accounts and user accounts are generated, and each account is assigned a corresponding role and operation permission according to job responsibilities. Submit project information via Word document or generate project information online with AI assistance; The project materials were checked for legality, standardized, and improved. Generate project content for users to select according to their actual needs; Analyze and examine the project content; Generate the project software system, including access links, system menus, and workflows.

[0058] Preferably, after the project software system is generated, manual optimization is performed, including writing code to perform the calculation of attribute values ​​within objects and the calculation of attributes across objects.

[0059] This method is applied to the development system in Example 1, and the processing flow of some modules is detailed below.

[0060] 1. The project information generation or retrieval module includes: The multimodal input parsing unit supports natural language descriptions, sketch photos, or Excel spreadsheet uploads, and automatically identifies business process elements. Business process requirements include participants, action nodes, and data flow. The intelligent field mapping unit is used to automatically identify field types through natural language descriptions and supports the direct generation of interactive forms from JSON configurations.

[0061] The development system designed for this invention accepts two standard input formats: Text files (.txt): Requires a universal encoding format, with a single file size not exceeding 10MB. Supports parsing text containing simple Markdown tags (such as ### headings). Documents (.docx): Fully preserves the style and structure of Word documents, automatically recognizing heading styles (Heading1-3), bullet points, tables, and other elements. Compatible with files generated by WPS and Microsoft Office 2013+ versions.

[0062] 2. The processing flow of the project data preprocessing module specifically includes: Step 2.1: Preprocess the input data using the following three modes: Mode 1: Directly read file content: Suitable for quickly importing raw text without structural processing, preserving the original file content.

[0063] Mode 2: Structural processing of article paragraphs based on keywords: Especially for fixed-structure documents such as legal documents, the system automatically divides the hierarchy using predefined keywords (such as "Article 1" and "Responsible Unit").

[0064] Mode 3: Processing document paragraphs through AI analysis: Use AI intelligent models such as qwen-plus to analyze documents, automatically identify task elements (goals / responsible persons / process nodes), and generate structures with semantic tags.

[0065] Step 2.2: Tool check. A validation method developed in Python is used to validate the preprocessed input data. The built-in validator checks JSON integrity and identifies missing key fields (e.g., issuing a warning if "responsible department" is missing). A visual preview provides a tree-structure viewer and supports highlighting AI-recognized task elements.

[0066] Step 2.3: Classify the verified project data according to hierarchical structure and paragraph structure: The verified project data is categorized into three levels: first-level category is articles, second-level category is paragraph content, and third-level category is sentences. The categorization criterion is the paragraph structure of the articles.

[0067] First-level classification (article level): As the top-level structural division, it is divided according to the physical sections of the document, such as contract chapters or policy clauses. This level maintains the original document order and reflects the overall strategic framework of the content. Second-level classification (paragraph level): Refines the first-level classification based on semantic integrity, identifying paragraph boundaries through line breaks / first-line indentation. For example, "responsibility clauses" and "process descriptions" are divided into independent units to ensure business logic coherence. Third-level classification (statement level): Statements are separated by periods / semicolons, and element labeling is performed using the qwen-plus model (such as labeling "task objectives" or "responsible person"). Dependency parsing ensures the relevance of actions to the executing entity. Sequence rules: Default order: Strictly follows the linear processing of the document's physical structure (article → paragraph → statement). Intelligent adjustment: When a process description is detected, it is rearranged according to time logic (such as "submission → approval → archiving").

[0068] The classification results can serve as the order in which subsequent analyses are performed, such as... Figure 3 This example illustrates a management system. The final software menu is generated based on the analyzed tasks, main processes, and sub-processes; therefore, the classification results are approximately related to the software menu.

[0069] The preprocessing results from the project data preprocessing module directly drive the generation of menus on the low-code platform: A. Task Element Mapping: "Task Objective" → System Name + Functional Module Name; "Responsible Department" → Basis for generating the role and permission tree; "Process Steps" → Generate corresponding sub-menu items.

[0070] B. Intelligent Association Rules: Each primary category corresponds to a top-level menu (e.g., "Contract Management"). Second-level paragraphs generate second-level menus (such as "Approval Process").

[0071] 3. The processing flow of the AI-generated content selection module includes: Step 3.1: Select the built-in problem model based on the document type (e.g., requirements document, legal document, data standard document, etc.). This system uses a combination of intelligent matching and manual intervention to select the analysis model, mainly based on the following dimensions: A. Intelligent document type recognition: Determine the file type using both file extension and content characteristics: Legal documents: Detects legal terms such as "article," "clause," and "item" (accuracy rate 98.2%). Requirements document: Identify agile development keywords such as "user stories" and "acceptance criteria"; Data standards: Matching structured tags such as "field name" and "data type"; B. Content Structure Analysis: Analyze the structure using the qwen-plus large language model; C. Business scenario adaptation: Pre-set industry templates: government documents, technical agreements, business process descriptions; Step 3.2: The purpose of the model is to perform different analysis steps based on different file structures and output specific structural data. The system has two built-in core analysis models: Legal document parsing model: Its purpose is to transform legal provisions into enforceable rules, and its function is to extract the responsible parties and analyze the main tasks to generate processing flow; Requirements analysis model: Its purpose is to build user story maps, which are used to identify functionalities and user roles, establish requirement priorities, and generate processing flows; 4. The processing flow of the project content analysis module, specifically including: Step 4.1, Project Content Analysis: This includes: data preprocessing analysis; task analysis; analysis of responsible organizations and their job duties; and analysis of task flow and responsible persons. The analysis results provide the basic data for software development.

[0072] Data preprocessing analysis involves: File format parsing: Supports intelligent parsing of txt / docx formats; Triple structuring: legal document structuring (by clause / chapter); article paragraph structuring (by natural paragraph); AI intelligent structuring (recognizing task elements); Output in standard JSON format, including: article level (overall document structure); paragraph level (logical paragraph division); sentence level (key element extraction). Task analysis involves: Document type adaptive analysis: Requirements documents → Function point extraction; Legal documents → Analysis of rights and responsibilities clauses; Element identification: Automatically detect core elements such as task content, objectives, responsible persons, and key processes; The analysis of responsible organizations involves: intelligent identification of organizational structure: automatic generation of department trees; job responsibility mapping; and permission baseline modeling.

[0073] Construct a four-dimensional analysis system: task details (goals / inputs / outputs); responsibility matrix (organization / position mapping); process topology (main process → sub-process); form fields (data element association).

[0074] Step 4.2: Automatically construct an intelligent organizational structure based on the description file, including: departmental hierarchy; job descriptions; reporting relationship topology; intelligent process decomposition technology; automatic division of main / sub-processes; visual process map; intelligent matching of related elements: process nodes and responsible positions; approval links and job levels.

[0075] The project content analysis module can generate the architecture of a designated responsible organization based on the description file, analyze the task flow and sub-task flow, and associate process nodes and responsible persons with AI.

[0076] 5. The software system generation module includes: The analysis results are used to create a system, including: queue management, application creation, role and responsibility generation, form generation, workflow generation, menu generation, permission binding, homepage generation, and completion notification. The core logic of the system generation is "analysis results → code / configuration generation → deployment and operation," and the specific steps are as follows: Step 5.1, Data Transformation (JSON → Executable Configuration): Input: Structured JSON output from task analysis (containing tasks, processes, forms, roles, etc.); Processing: API calls to relevant interfaces of the business system; Step 5.2, Component Generation: Form generation: Parse the field definitions in JSON and automatically generate front-end forms (React / Vue components) + back-end validation logic; Process generation: Convert the main process / subprocess into the BPMN 2.0 standard format. Menu generation: Based on task categories (first-level menu = task type, second-level menu = sub-process); Step 5.3, Permission and Role Binding: Analyze role-permission relationships and automatically synchronize them to the permission system; Step 5.4, System Deployment and Notification: Step 5.5: Automated deployment of the production line; Step 5.6: Completion Notification: Trigger email / SMS notification via WebHook; Step 5.7: Generate system access links.

[0077] The development method of this invention involves the following core mechanisms: a. Dynamic service orchestration model: Intelligent matching based on service profiles; dynamic reorganization of service chains at runtime.

[0078] b. Intelligent decision-making model: Multi-objective optimization algorithm; reinforcement learning-driven adaptive policy generation.

[0079] c. Adaptive fault-tolerant model: Multi-level fault tolerance (service level / component level / system level); distributed transaction compensation mechanism.

[0080] d. Trusted security model: Dynamic trust assessment; fine-grained access control under a zero-trust architecture.

[0081] Closed-loop optimization mechanism: By creating a closed loop of "environmental awareness - dynamic adaptation - continuous optimization", we can achieve the leap from rigid architecture to flexible services.

[0082] Therefore, the present invention has the following advantages compared with the prior art: High scalability: Loosely coupled service architecture supports rapid business expansion; Adaptability: Dynamic model injection at runtime reduces operational complexity; Intelligentization: Service composition optimization algorithms improve resource utilization; Open and compatible: Standardized interfaces and protocol conversion middleware protect existing IT investments.

[0083] The following example demonstrates how the method proposed in this invention handles the development process of a complex Enterprise Resource Management (ERP) system. First, the project is initiated, and project requirements are clarified. This ERP system needs to include complex modules such as supply chain management, production planning, and financial management.

[0084] The development process includes: S1. Account Registration and Information Acquisition: The company's IT manager registers an account under the "End Customer" category. The system automatically retrieves the company's business registration information and generates a preliminary organizational structure.

[0085] S2. Project Information Generation: The IT manager uses AI to generate detailed project requirements documents online, including specific functional requirements for each module, data flow, permission settings, etc.

[0086] S3. Data Preprocessing: The system intelligently analyzes project documents, identifies key requirements, and standardizes them.

[0087] S4. AI Content Generation and Selection: The system generates a comprehensive solution including data models, business processes, and interface prototypes, which IT managers can then select and adjust based on actual needs.

[0088] S5. In-depth analysis of project content: The system conducts in-depth analysis of the selected content to identify potential technical difficulties and integration requirements.

[0089] S6. System Generation: Based on the analysis results, the system automatically generates an initial version of the ERP software, including backend services, database structure, frontend interface, etc.

[0090] S7. Manual optimization: Specifically, the project coordinator determines the overall architecture and functional boundaries.

[0091] The job description analysis engineer sets up detailed role permissions, such as access permissions for different departments like purchasing, production, and finance.

[0092] Object composition and analysis define engineers to optimize data models and ensure data consistency across modules.

[0093] Business process analysis engineers refine complex business processes, such as multi-level approvals and cross-departmental collaborations.

[0094] Data integration engineers handle data integration with existing systems, such as legacy inventory management systems.

[0095] Injection code allows engineers to write custom code to implement specific, complex algorithms, such as production planning optimization algorithms. Display style adjustment engineers optimize the user interface to ensure it meets enterprise requirements and provides a good user experience.

[0096] Application modification engineers conduct comprehensive testing to ensure the system's stability and security.

[0097] S8. Deployment and Continuous Optimization: The system is deployed to the enterprise's private cloud environment and automatic monitoring is configured. Application modification engineers continuously collect feedback based on actual usage and conduct iterative optimizations.

[0098] Through the above S1-S8, enterprises obtain an ERP system that can meet complex business needs, has a high degree of customization and good performance, and overcomes many limitations of traditional low-code platforms.

[0099] This invention utilizes the characteristics of network-based software theory, embodies the process-oriented technical concept of software development, and constructs a development system with characteristics such as openness, collaboration, autonomy, evolution, and polymorphism. The generated software can adapt to the complexity and uncertainty of the Internet environment.

[0100] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the above embodiments do not limit the present invention in any way, and all technical solutions obtained by equivalent substitution or equivalent transformation fall within the protection scope of the present invention.

Claims

1. A software development system based on the lattice software theory, characterized by, Comprise: Account registration and login module, for registering an account and logging in using a registered account, supporting the registration and login of multiple categories of accounts including end customers, partners, and sales, giving users corresponding IDs and permissions; User detailed information acquisition module, using AI to acquire public customer detailed information according to user business registration information, including industry classification, company introduction, product and organizational structure information, and generating software system organizational structure according to the acquired customer detailed information, generating management accounts and use accounts, and giving each account corresponding roles and operation permissions according to job responsibilities; Project information generation or acquisition module, for submitting project information through documents or AI-assisted online generation of project materials; Project material preprocessing module, for checking, standardizing and perfecting the project materials; AI-generated content selection module, generating project content for users to select according to actual needs; Project content analysis module, for analyzing and checking project content; Software system generation module, for generating project software systems, including access links, system menus, and processes.

2. The software development system based on the net software theory according to claim 1, wherein, The system also includes an artificial tuning module, including multiple workstations of different functions, supporting code injection, and performing object attribute value calculation and cross-object attribute calculation through code writing.

3. The software development system based on the net software theory according to claim 1, wherein, The project information generation or acquisition module comprises: A multi-modal input analysis unit that supports natural language description, sketch photography or Excel table upload, automatically identifies business process elements, and business process requirements include participants, action nodes and data flow.

4. The software development system based on the net software theory according to claim 1, wherein, The project information generation or acquisition module comprises: An intelligent field mapping unit for automatically identifying field types through natural language description, and supporting interactive form generation directly from JSON configuration.

5. The software development system based on the net software theory according to claim 1, wherein, The project material preprocessing module comprises: A multi-format analysis unit for automatically analyzing and structuring input text; An intelligent paragraph segmentation unit for automatically dividing logical paragraphs according to context semantics and automatically extracting paragraph core keywords to generate tags.

6. The software development system based on the net software theory according to claim 1, wherein, The software system generation module comprises: An AI component arrangement for generating responsive layouts through a drag-and-drop designer, and AI real-time recommendation of data tables, charts and data source binding.

7. The software development system based on the net software theory according to claim 1, wherein, The system also includes a context-aware optimization module for automatically detecting system page performance bottlenecks and providing pagination or loading optimization solutions.

8. The software development system based on the net software theory according to claim 1, wherein, The system uses a multi-model collaborative work architecture scheme, including Qwen-Plus as the main analysis engine, Jieba word segmentation tool for primary text structuring, and vector model for clause similarity matching tasks.

9. A software development method based on the net software theory, used in the system of any one of claims 1-8, characterized in that, The method comprises the steps of: Registering an account or logging in using a registered account; Using AI to acquire public customer detailed information, including industry classification, company introduction, product and organizational structure information, and generating software system organizational structure according to the acquired customer detailed information, generating management accounts and use accounts, and giving each account corresponding roles and operation permissions according to job responsibilities; Submitting project information through documents or AI-assisted online generation of project information; Check the legality of the project information, standardization and perfect treatment; Generate project content, for users to choose according to actual demand; Analysis and inspection of project content; Generate project software system, including access link address, system menu and process.

10. The software development method based on the net software theory according to claim 9, characterized in that: After generating the project software system, manual tuning, including writing code to execute the object attribute value calculation and cross object attribute calculation.

Images