A software development method for internet and cloud computing

Abstract

The application discloses an internet and cloud computing software development method, relates to the field of internet and cloud computing, and specifically comprises the following steps: multi-modal demand analysis eliminates ambiguity and outputs a structured task tree, on the basis of which a cloud native hybrid architecture is built and low-code and high-code functions are divided; component plug-in collaboration is realized through visual development and bidirectional communication, and then code intelligent completion, detection and correction are used to improve quality, and finally version archives are built to support difference comparison and backtracking; the application eliminates demand ambiguity and unifies development standards through multi-modal demand analysis and hybrid architecture collaboration, and reduces the development threshold; then, code intelligent optimization and version intelligent management and control are used to improve code quality and running stability, realize multi-dimensional version comparison, flexible rollback and controllable risk, and comprehensively improve the collaboration efficiency, iteration safety and management and control capability of the whole software development life cycle, and adapt to the continuous delivery requirements of complex business scenarios.

Description

Technical Field

[0001] This invention relates to the fields of the Internet and cloud computing, specifically to a software development method for the Internet and cloud computing. Background Technology

[0002] With the deep integration and application of Internet technology and cloud computing architecture, the software development industry is moving towards a cloud-native, distributed, and intelligent development stage. The complexity of business application scenarios in various fields continues to increase, and software requirements are showing core characteristics of multimodality, specialization, and customization. The software development process needs to handle various forms of requirement information, such as natural language descriptions, visual graphics, and structured data. There are significant differences in business processes, standard systems, and functional requirements across different industries, which places stringent demands on the accuracy of requirement analysis, architectural adaptability, and standardization of development processes. Software in the cloud computing environment needs to have the ability to scale elastically, collaborate in a distributed manner, and be deployed in containers. Traditional development models cannot meet the development needs of complex businesses in multiple fields. The industry urgently needs to build a standardized software development solution that adapts to the characteristics of cloud computing, integrates intelligent technologies, and covers the entire development chain to improve development efficiency and the stability of software operation in the cloud environment.

[0003] Traditional software development methodologies suffer from numerous shortcomings in internet and cloud computing applications. The requirements analysis phase relies excessively on manual processing, failing to efficiently handle multimodal requirements and leading to ambiguity, misunderstandings, and disorganized business logic. It struggles to establish unified and standardized requirements execution guidelines. Architecture design lacks scientific quantitative support; the functional boundaries between low-code components and high-code plugins are arbitrarily defined, and their interaction standards, data transmission formats, and calling rules lack unified specifications, failing to meet the core requirements of distributed operation and elastic scaling in cloud computing. In the development phase, visual configuration and code development are disconnected, communication links between components and plugins are inefficient, code optimization relies on manual operations, resulting in low detection and correction efficiency and the potential for residual logical vulnerabilities, security risks, and performance issues. Version control lacks a full-process traceability mechanism, version comparison analysis and rollback schemes are inadequate, iteration risks are uncontrollable, the overall development cycle is long and costly, and software quality and operational stability cannot be effectively guaranteed. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a software development method for the Internet and cloud computing. This method eliminates requirement ambiguity through a domain-specific large model and outputs structured task tree quantification values; it builds a cloud-native hybrid architecture based on the quantification results, and rationally divides low-code components and high-code plugins; it supports visual component splicing and high-code plugin registration, and establishes bidirectional communication; it uses iterative adaptation algorithms and static analysis to achieve automatic code completion, detection, and correction; and finally, it constructs a traceable version archive, generates difference reports and rollback schemes, and supports flexible rollback.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a software development method for the Internet and cloud computing, the specific steps of which are as follows: S100, Multimodal Requirement Analysis: Collect multimodal requirement information, call domain-specific large models to perform requirement semantic analysis and feature integration, eliminate ambiguity in expression, sort out business process nodes and functional requirements, execute multimodal task quantification algorithms and output structured hybrid development task tree quantification values; S200, Hybrid Architecture Orchestration: Based on the quantization values ​​output by the multimodal task quantization algorithm, build a cloud-native hybrid development architecture, divide the functional scope of low-code components and high-code plugins, formulate interaction standards, data formats and calling rules, and clarify the cloud resource adaptation requirements and deployment methods of each module; S300 Component and Plugin Development: Through visual operations, low-code components are assembled and configured, component attributes, trigger events and data mapping relationships are set, standardized intermediate code adapted to the cloud computing environment is generated, high-code plugin development and registration are carried out simultaneously, and a two-way communication channel between low-code components and high-code plugins is established. S400, Intelligent Code Optimization: Retrieves the basic code of the developed high-code plugin, executes the code iteration and adaptation algorithm, combines business rules and code context, automatically identifies and performs code completion work through a domain-specific large model, improves logical branches and data verification content, and performs code detection and correction work through static analysis tools; S500 Intelligent Version Management: Collects development and modification information throughout the entire process and builds a traceable version archive, completes structured comparative analysis of different versions, generates version difference reports and rollback plans, and supports version preview and rollback operations.

[0006] Furthermore, the multimodal requirement information includes business requirement descriptions, functional boundary definitions, and business operation rules written by business personnel in natural language; business flow diagrams, functional module relationship diagrams, and interface layout diagrams drawn in visual graphics; business data fields, data relationships, and data processing rules compiled in structured table format; and data interaction logic, user permission allocation rules, and application deployment and operation environment requirements in business scenarios.

[0007] Furthermore, the domain-specific large models are pre-trained specifically for financial, medical, and industrial internet scenarios, respectively. They incorporate the corresponding domain's business terminology system, industry business process specifications, and development and implementation standards. When performing requirement semantic parsing, they perform categorical semantic decomposition and business intent recognition on various multimodal requirement information, extracting the business subject, execution process, constraints, and functional orientation information contained in the requirements. In the feature integration stage, the parsing results of different modalities are normalized and aligned, the requirement feature information corresponding to each modality is merged, and duplicate and conflicting content in the parsing results is removed to form a unified and standardized set of requirement features, providing the basic data required for the multimodal task quantization algorithm.

[0008] Furthermore, the mathematical expression of the multimodal task quantization algorithm is: in, The resulting structured hybrid development task tree is quantified and directly serves as the core basis for subsequent code development and iteration. Input feature values ​​for the i-th type of multimodal requirement, covering all requirement modalities such as natural language, flowcharts, and tables; The adaptation coefficients for domain-specific large models are dynamically assigned based on different business domains such as finance and healthcare. We analyze the weighting coefficients for multimodal requirements to characterize the impact of the accuracy of requirement understanding on development tasks. This serves as a quantification of the business complexity of functional modules, used to distinguish between basic functionalities and complex logic modules. This is a general compatibility coefficient for low-code components; the higher the value, the more suitable the corresponding function is for low-code implementation. Weighting coefficients are assigned to the hybrid development model to regulate the allocation ratio of low-code components and high-code plugins; n represents the total number of requirement input modalities.

[0009] Furthermore, the cloud-native hybrid development architecture is built on a containerized deployment model, comprising two core units: a low-code general component library and a high-code custom plugin container. The architecture supports elastic scaling and distributed collaborative operation in a cloud computing environment. When classifying the functional scope of low-code components and high-code plugins, based on the quantification values ​​output by the multimodal task quantification algorithm, page layout construction, form data display, basic information query, and general pop-up interaction functions are classified as low-code components, while business calculation processing, complex logic judgment, high-concurrency data operation, deep integration with third-party systems, and customized business logic execution functions are classified as high-code plugins. At the same time, a unified interaction standard, standardized data transmission format, and permission call rules are established between low-code components and high-code plugins, clarifying the resource adaptation requirements and specific deployment methods of each functional module in the cloud environment.

[0010] Furthermore, the interaction standard defines the communication interaction level between low-code components and high-code plugins, specifies the interaction sequence of components initiating requests to plugins and plugins returning results to components, defines data verification nodes and exception response mechanisms in the interaction process, and ensures the consistency and standardization of interaction behavior between the two types of development units. The data format adopts a standardized structured data format, unifies field naming rules, data type definitions, data length limits and encoding standards, supports cross-unit transmission of multiple types of business data, and adapts to the distributed data interaction needs in a cloud computing environment. The calling rules determine the calling permissions, calling frequency limits and calling triggering conditions of low-code components to high-code plugins, specify the scope of data returned by plugins to components and the timeliness requirements for return, standardize the calling chain and permission control logic between the two types of development units, and provide clear rule constraints for the collaborative operation of the two types of development units.

[0011] Furthermore, the standardized intermediate code is automatically generated based on the component configuration information from visual operations. The code adheres to containerized deployment specifications and distributed operation requirements in a cloud computing environment. The code structure is divided into a component layout definition section, an attribute parameter configuration section, a trigger event binding section, and a data mapping logic section. The component layout definition section determines the position coordinates, hierarchical relationships, and style configurations of each low-code component on the page; the attribute parameter configuration section stores the size, display rules, permission control, and validation logic of each component; the trigger event binding section records the trigger conditions and execution logic corresponding to each component; and the data mapping logic section defines the data source, transmission link, and synchronization rules between components. The code adopts a unified programming language specification, reserves standardized interfaces for interfacing with high-code plugins, supports containerized packaging and elastic scaling in cloud environments, and adapts to the collaborative operation requirements under a distributed deployment architecture.

[0012] Furthermore, the bidirectional communication channel between the low-code component and the high-code plugin is built upon a unified communication protocol based on a cloud-native hybrid development architecture, covering three types of interaction scenarios: data transmission, command interaction, and state synchronization. The channel clearly defines the interaction logic for low-code components to initiate calls to high-code plugins, including the call request initiation format, parameter encapsulation rules, permission verification nodes, and timeout determination mechanisms. It supports components transmitting business operation instructions, front-end interaction data, and user permission information to plugins. Simultaneously, it clearly defines the interaction logic for high-code plugins to send responses back to low-code components, including the response data encapsulation format, data verification rules, exception feedback mechanisms, and state synchronization rules. It supports plugins sending business processing results, data query results, and runtime status information back to components. The channel incorporates data encryption mechanisms, traffic control rules, and exception retry logic, adapts to the distributed deployment architecture of cloud computing environments, supports cross-container and cross-node communication interactions, is compatible with the elastic scaling and load scheduling mechanisms of cloud environments, and reserves extension interfaces to support the adaptation and access of subsequent new interaction scenarios and communication protocols.

[0013] Furthermore, the code iterative adaptation algorithm adopts the following mathematical expression: in, This represents the comprehensive quantitative value of intelligent version iteration; This represents the code generation adaptation weight coefficient, which adjusts the matching accuracy between development tasks and code implementation; This represents the quantized value of the structured hybrid development task tree, which is the output of the multimodal task quantization algorithm. These represent code context-aware feature values, characterizing the degree of context matching in large model code completion; This represents the version control optimization weight coefficient, which balances the effectiveness of code optimization with the stability of version iteration. This represents the quantitative value of static code analysis optimization, covering the results of code redundancy cleanup, performance improvement, and vulnerability remediation. This represents the version iteration risk coefficient; the higher the value, the greater the risk of failure and iteration in the current version.

[0014] Furthermore, the code detection and correction work relies on multi-dimensional static analysis tools. The detection work covers five dimensions: syntax standard verification, logical vulnerability investigation, performance bottleneck identification, security risk detection, and code redundancy analysis. It checks the code one by one for syntax correctness, boundary condition handling, loop complexity, memory leak risk, SQL injection vulnerability, and unauthorized access risk. At the same time, it compares with unified coding standards to identify issues such as non-standard naming, inconsistent formatting, and missing comments. The correction work categorizes and processes the detection results, automatically correcting syntax errors, formatting problems, and simple redundant code. It generates precise correction suggestions for logical vulnerabilities, performance bottlenecks, and security risks, marking the problem location, scope of impact, and remediation solutions. For complex problems, it generates multiple alternative correction solutions based on domain-specific large models for developers to choose from, ensuring code quality and operational stability.

[0015] Furthermore, the version difference report covers structured comparison results across four dimensions: code level, functional level, configuration level, and dependency level. Code-level differences mark newly added, modified, and deleted lines of code and associate them with corresponding development tasks and modification personnel. Functional-level differences outline the specific content and business impact of newly added, modified, and deprecated functions. Configuration-level differences compare changes to cloud resources, permissions, and interfaces. Dependency-level differences list version changes and compatibility risks of third-party dependency libraries. The report also marks the risk level corresponding to each difference item. The rollback plan is formulated based on the comprehensive quantitative value of intelligent version iteration and the risk level of the difference report, and is divided into three modes: full rollback, incremental rollback, and canary rollback, corresponding to high-risk major failures, partial functional failures, and medium-risk scenarios, respectively. The plan defines the execution steps, verification standards, and emergency response procedures for each mode.

[0016] Compared with existing technologies, this software development method for the Internet and cloud computing has the following advantages: I. This invention, through multimodal requirement analysis and hybrid architecture orchestration, comprehensively integrates various types of requirement information, including natural language, visual graphics, and structured tables. It relies on domain-specific pre-trained models to complete requirement semantic decomposition and intent recognition, eliminating ambiguity in requirement expressions and standardizing requirement feature sets. It accurately identifies business process nodes and functional requirements, and builds a cloud-native hybrid development architecture based on the quantified requirement results. It clearly defines the functional scope of low-code components and high-code plugins, establishes unified interaction standards, data formats, and calling rules, adapts to the containerized deployment and distributed operation requirements of cloud computing, completes low-code component configuration through visual operations and generates standardized intermediate code, establishes a two-way communication channel between low-code and high-code components, simplifies development processes, lowers the software development threshold, improves requirement transformation efficiency, ensures the adaptability of the development architecture to the cloud environment, and achieves efficient allocation and utilization of development resources.

[0017] II. This invention covers the entire development process through intelligent code optimization and intelligent version control. Relying on code iteration adaptation logic and multi-dimensional static analysis tools, it completes code completion, logic improvement, vulnerability detection and automatic correction. Combined with domain models, it provides correction solutions for complex problems, comprehensively improving code quality and operational stability. It collects development modification information throughout the process to build a traceable version archive, completes multi-dimensional version comparison of code, functions, configurations, and dependencies and generates difference reports. Based on risk levels, it formulates three rollback modes: full rollback, incremental rollback, and canary rollback, and clarifies execution steps and verification standards. This method enables version preview, rollback, and rapid repair, effectively reducing version iteration risks, ensuring the continuity of the development process, adapting to the continuous iteration needs of complex business scenarios, and improving the management efficiency and operational reliability of the entire software development lifecycle.

[0018] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0020] Figure 1 A flowchart illustrating a software development methodology for the Internet and cloud computing; Figure 2 A schematic diagram illustrating data transmission in a software development methodology based on the Internet and cloud computing; Figure 3 This is a schematic diagram illustrating the data transmission process of the intelligent code optimization steps in this invention. Detailed Implementation

[0021] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0022] Example 1: This embodiment targets core business scenarios in the financial sector. Relying on the underlying support environment of the internet and cloud computing, it employs the internet and cloud computing software development methods described in this invention to complete the full-process development and implementation of a financial cloud-native business management system. The entire development process strictly follows a standardized chain of requirements analysis, architecture orchestration, component and plugin development, code optimization, and version control. It fully aligns with the high security, high stability, and high concurrency characteristics of the financial industry, ensuring a deep fit between the development phase and financial business needs. This guarantees that the final delivered system can stably adapt to the distributed cloud computing environment, meeting the core requirements of continuous iteration and elastic scaling of financial businesses. Figure 1 As shown.

[0023] S100, Multimodal Requirements Analysis: Comprehensive collection of multimodal requirements information submitted by financial professionals, covering financial business requirements descriptions, functional boundary definitions, and business operation rules written in natural language; visual diagrams of financial business workflows, functional module relationship diagrams, and interface layout diagrams; and structured tables of financial business data fields, data relationships, and data processing rules. It also fully integrates data interaction logic, hierarchical user permission allocation rules, and cloud platform deployment and operation environment requirements in financial business scenarios, comprehensively covering all key requirements of financial business from front-end interaction to back-end operation. By invoking a domain-specific large-scale model pre-trained specifically for financial scenarios, and relying on the model's built-in financial industry business terminology system, industry business process specifications, and development and implementation standards, the system performs semantic decomposition and business intent identification on a categorized basis for various multimodal requirements. It accurately extracts the financial business entity, execution process, constraints, and functional orientation information contained in the requirements. During the feature integration stage, the parsing results of different modalities are normalized and aligned, merging the requirement feature information corresponding to each modality, eliminating duplicate and conflicting content in the parsing results, and forming a unified and standardized set of requirement features. This provides the basic data required for the multimodal task quantification algorithm. The algorithm is then executed, outputting a structured hybrid development task tree quantification value. This transforms vague descriptions of financial business requirements into precise and executable development guidelines, completely avoiding problems such as misunderstandings of requirements and chaotic business logic. It ensures that all subsequent development stages closely align with core financial business requirements, significantly improving the accuracy and efficiency of converting requirements into development tasks.

[0024] S200 Hybrid Architecture Orchestration: Based on the quantified values ​​output by the multimodal task quantification algorithm, a cloud-native hybrid development architecture adapted to financial scenarios is built. The architecture is constructed using a containerized deployment model and comprises two core units: a low-code general component library and a high-code custom plugin container. This perfectly supports elastic scaling and distributed collaborative operation in a cloud computing environment, giving the architecture the high availability and high scalability required for financial scenarios. The functional scope of low-code components and high-code plugins is clearly defined based on the quantified values. Functions such as page layout construction, form data display, basic information query, and general pop-up interaction are classified as low-code components, while functions such as financial business calculation and processing, complex logic judgment, high-concurrency data computation, deep integration with third-party financial systems, and customized financial business logic execution are classified as high-code plugins. This functional division aligns with the varying complexity of financial businesses, avoiding waste of development resources and redundancy in function implementation. Simultaneously, a unified interaction standard, standardized data transmission format, and permission call rules were established between low-code components and high-code plugins. The resource adaptation requirements and specific deployment methods of each functional module in the cloud environment were clarified. The interaction standard determined the communication interaction level, interaction sequence, data verification nodes, and exception response mechanism for the two types of development units. The data format adopted a standardized structured data format to unify various data specifications. The call rules clarified permissions, frequency, triggering conditions, and feedback requirements, making the collaborative operation of each module within the architecture more standardized and data interaction smoother. This fully adapts to the high concurrency and high security operation requirements of the financial system and ensures the stability and compatibility of the architecture operation in the cloud environment.

[0025] S300 Component Plugin Development: Enables the assembly and configuration of low-code components specifically designed for financial scenarios through visual operations. It precisely sets component attributes, trigger events, and data mapping relationships, automatically generating standardized intermediate code adapted to cloud computing environments. The code adheres to cloud computing containerization deployment specifications and distributed operation requirements, divided into component layout definition sections, attribute parameter configuration sections, trigger event binding sections, and data mapping logic sections. It can fully support all functional requirements of the financial front-end interaction, and reserves standardized interfaces for interfacing with high-code plugins, making low-code component development more efficient and configuration more flexible. Basic functionalities can be built without extensive manual coding. Simultaneously, we are developing and registering high-code plugins for financial business, writing customized code for core financial business logic, and building a bidirectional communication channel between low-code components and high-code plugins based on a unified communication protocol of a cloud-native hybrid development architecture. The channel covers three types of interaction scenarios: data transmission, command interaction, and state synchronization. It clearly defines the entire process logic of call requests and response feedback, and incorporates built-in data encryption mechanisms, traffic control rules, and exception retry logic. It can achieve stable communication and interaction across containers and nodes, and is compatible with cloud environment elastic scaling and load scheduling mechanisms. This allows low-code components and high-code plugins to work together seamlessly, and front-end interaction and back-end core business processing to be efficiently connected. It ensures the security of financial business data transmission and the accuracy of command execution, significantly improves overall development efficiency, and shortens the system development cycle.

[0026] S400 Intelligent Code Optimization: Retrieves the basic code of the high-code plugin for financial business that has been developed, executes the code iteration and adaptation algorithm, and combines the financial industry's exclusive business rules and code context information. Through a financial field-specific large model, it automatically identifies missing code content and completes the code, improves logical branches and financial data verification content, making the code logic more complete, the data processing more rigorous, and in line with the financial industry's data compliance and business standard requirements. Code inspection and correction are conducted using multi-dimensional static analysis tools. The inspection comprehensively covers five dimensions: syntax verification, logical vulnerability investigation, performance bottleneck identification, security risk detection, and code redundancy analysis. It meticulously checks the code's syntax correctness, boundary condition handling, loop complexity, memory leak risks, SQL injection vulnerabilities, and unauthorized access risks. Simultaneously, it compares the code against unified coding standards to identify formatting and comment issues. Correction is based on the categorized results, automatically correcting syntax errors, formatting issues, and simple redundant code. It generates precise correction suggestions for logical vulnerabilities, performance bottlenecks, and security risks. For complex problems, it combines domain-specific large-scale models to generate multiple alternative correction solutions. This continuously improves the quality of financial business code, eliminates system operational risks, ensures code stability and data security in high-concurrency financial scenarios, reduces the workload of manual code debugging and optimization, and minimizes human error during development. Figure 3 As shown.

[0027] S500 Intelligent Version Control: Collects development and modification information from the entire financial business management system, constructs a complete and traceable version archive, and accurately records and archives every development modification, enabling full-chain query and traceability of the development process. It completes structured comparative analysis of different versions, generating version difference reports covering four dimensions: code level, function level, configuration level, and dependency level. These reports clearly mark the specific content, business impact scope, and risk level of each difference, allowing developers to intuitively grasp the details of version iteration and potential risks. Based on the comprehensive quantitative value of intelligent version iteration and the risk level of the difference report, it formulates rollback plans for three modes: full rollback, incremental rollback, and canary rollback. It clarifies the execution steps, verification standards, and emergency response procedures for each mode, while supporting version preview and rollback operations. This ensures full control over the version iteration process, effectively reducing the operational risks brought by financial system version updates, ensuring uninterrupted business operations and no data loss during system iteration, achieving intelligent and standardized control over the entire lifecycle of development versions, and ensuring the continuous and stable delivery and iterative upgrades of the financial system.

[0028] This embodiment utilizes the method of the present invention to develop a cloud service platform for the medical field. Multimodal requirement analysis accurately matches medical diagnosis and management needs; the hybrid architecture orchestration conforms to medical industry standards and cloud environment requirements; component plugin development enables efficient functional construction; intelligent code optimization ensures medical data compliance and system stability; and version control reduces iteration risks. This method effectively solves the pain points and challenges of medical software development, improves development efficiency and platform security, enables the medical cloud platform to stably support the conduct of diagnosis and treatment services, and facilitates the efficient advancement of digital transformation in the medical industry.

[0029] Example 2: This embodiment is applied to the software development of a cloud service platform in the medical field. Based on the support of the Internet and cloud computing architecture, it utilizes the software development method proposed in this invention to complete the entire development process of the medical cloud service platform. The entire process aligns with the core requirements of the medical industry: data compliance, rigorous treatment procedures, and stable system operation. It covers the full-scenario functional development of medical services, from patient services and treatment operations to back-end management, ensuring the platform perfectly adapts to the operational needs of the medical cloud environment. This guarantees medical data security and the efficient operation of treatment services, providing stable and reliable technical support for the digital transformation of the medical industry. Figure 2 As shown.

[0030] S100 Multimodal Requirements Analysis: Comprehensive collection of multimodal requirements information across all categories in the medical field, including medical business requirements descriptions, functional boundary definitions, and business operation rules written by business personnel in natural language; visually illustrated flowcharts of diagnosis and treatment business processes, functional module relationship diagrams, and interface layout diagrams; medical business data fields, data relationships, and data processing rules compiled in structured tables; and complete integration of data interaction logic in medical scenarios, hierarchical user permission allocation rules for medical staff, and cloud platform deployment and operation environment requirements, comprehensively covering core medical needs such as outpatient diagnosis and treatment, inpatient management, data statistics, and system integration. By invoking a domain-specific large-scale model pre-trained specifically for medical scenarios, and relying on the model's built-in medical industry business terminology system, diagnosis and treatment process specifications, and development and implementation standards, the system performs categorized semantic decomposition and business intent recognition on various multimodal requirement information. It accurately extracts the medical business entities, diagnosis and treatment processes, data constraints, and functional orientation information contained in the requirements. During the feature integration stage, the parsing results of different modalities are normalized and aligned, and the requirement feature information corresponding to each modality is merged. Redundant and conflicting content in the parsing results is eliminated, forming a unified and standardized set of requirement features. This provides the basic data required for the multimodal task quantification algorithm. The algorithm is then executed, and a structured hybrid development task tree quantification value is output. This transforms complex medical requirements into standardized development tasks, completely eliminating ambiguity in requirement expression and ensuring that the development process accurately matches medical diagnosis and treatment processes and management specifications, making requirement implementation more efficient and accurate.

[0031] S200 Hybrid Architecture Orchestration: Based on the quantified values ​​output by the multimodal task quantification algorithm, a cloud-native hybrid development architecture specifically for medical scenarios is built. The architecture is constructed using a containerized deployment model and comprises two core units: a low-code general component library and a high-code custom plugin container. It can stably support elastic scaling and distributed collaborative operation in a cloud computing environment, meeting the expansion needs of the medical platform as its user base grows and business expands. The functional categories of low-code components and high-code plugins are scientifically divided according to the quantified values. Page layout construction, medical form data display, basic medical information query, and general pop-up interaction functions are classified as low-code components. Medical data processing, complex medical logic judgment, high-concurrency medical data processing, deep integration with third-party medical systems, and customized medical business logic execution functions are classified as high-code plugins. This functional division aligns with the differences in simplicity and complexity within medical businesses, optimizing the allocation of development resources. Simultaneously, unified interaction standards, data formats, and calling rules are established, clarifying the cloud resource adaptation requirements and deployment methods for each module. Interaction standards standardize communication levels, interaction sequences, and exception response mechanisms. Data formats unify medical data transmission specifications, and calling rules define permissions and execution logic. This ensures more orderly collaborative operation of modules within the architecture, more secure data interaction, and full compatibility with the distributed deployment and data compliance requirements of the medical platform, guaranteeing the stability of the medical cloud platform and the standardization of data interaction.

[0032] S300 Component Plugin Development: Enables the assembly and configuration of low-code components for medical scenarios through visual operations. It precisely sets component attributes, trigger events, and data mapping relationships, automatically generating standardized intermediate code adapted to cloud computing environments. The code adheres to containerized deployment and distributed operation specifications, divided into component layout definition sections, attribute parameter configuration sections, trigger event binding sections, and data mapping logic sections. This allows for rapid implementation of medical front-end interface construction and basic interactive functions. Standardized interfaces for interfacing with high-code plugins are reserved, making low-code development more convenient and significantly reducing the difficulty of front-end development. Simultaneously, we are developing and registering high-code plugins for medical services. We are writing customized code for core medical diagnosis and treatment logic and data processing rules. Based on a unified communication protocol of cloud-native hybrid development architecture, we are building a two-way communication channel between low-code components and high-code plugins. The channel covers three types of interaction scenarios: data transmission, command interaction, and status synchronization. It clearly defines the entire process logic of calls and feedback, and has built-in data encryption, traffic control, and abnormal retry mechanisms. It can achieve secure communication across containers and nodes, and is compatible with cloud environment elastic scaling and load scheduling. This allows for seamless connection between front-end interaction and back-end core medical business processing, ensuring the security of medical data transmission and the accuracy of diagnosis and treatment command execution, improving the development efficiency of the medical platform, and shortening the platform launch cycle.

[0033] S400 Intelligent Code Optimization: This system retrieves the base code of existing high-concurrency medical business plugins, executes an iterative code adaptation algorithm, and, combined with medical business rules and code context information, automatically identifies and completes code using a medical-specific large-scale model. This improves logical branches and medical data verification content, ensuring the code logic aligns with medical treatment standards and that data processing meets industry compliance requirements. Multi-dimensional static analysis tools are used for code detection and correction, covering five dimensions: syntax verification, logic vulnerability investigation, performance bottleneck identification, security risk detection, and code redundancy analysis. This comprehensive check addresses issues such as syntax, boundary handling, performance, security, and redundancy, strictly controlling medical code quality. Corrections are automatically handled for simple problems, while generating precise fixes for complex issues, making medical code more robust and secure. This eliminates system operational risks, ensures stable operation of the medical platform in high-concurrency treatment scenarios, reduces manual code optimization workload, and avoids code defects caused by human intervention.

[0034] S500 Intelligent Version Management: Collects development and modification information from the entire medical cloud service platform, constructs a traceable version archive, and fully records every development modification detail, achieving end-to-end retention of the development process. It completes structured comparative analysis of different versions, generating version difference reports across four dimensions: code level, function level, configuration level, and dependency level. These reports clearly mark the differences, business impact, and risk levels, allowing developers to accurately grasp the version iteration status. Based on the comprehensive quantitative value of intelligent version iteration and the risk level of the difference reports, it formulates three rollback schemes: full rollback, incremental rollback, and canary rollback, clearly defining the execution process and verification standards. It also supports version preview and rollback operations, ensuring full control over the medical platform's version iteration process. This effectively reduces the risk of system failures caused by version updates, ensuring uninterrupted medical services and no loss of medical data. It achieves intelligent management of the entire platform version lifecycle, ensuring the continuous and stable iteration and upgrade of the medical cloud service platform.

[0035] This embodiment utilizes the method of the present invention to develop a cloud service platform for the medical field. Multimodal requirement analysis accurately matches medical diagnosis and management needs; the hybrid architecture orchestration conforms to medical industry standards and cloud environment requirements; component plugin development enables efficient functional construction; intelligent code optimization ensures medical data compliance and system stability; and version control reduces iteration risks. This method effectively solves the pain points and challenges of medical software development, improves development efficiency and platform security, enables the medical cloud platform to stably support the conduct of diagnosis and treatment services, and facilitates the efficient advancement of digital transformation in the medical industry.

[0036] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A software development method for the Internet and cloud computing, characterized in that, The specific steps of this method are as follows: S100, Multimodal Requirement Analysis: Collect multimodal requirement information, call domain-specific large models to perform requirement semantic analysis and feature integration, execute multimodal task quantification algorithms and output structured hybrid development task tree quantification values; S200, Hybrid Architecture Orchestration: Based on the quantization values ​​output by the multimodal task quantization algorithm, a cloud-native hybrid development architecture is built, the functional scope of low-code components and high-code plugins is divided, and interaction standards, data formats and calling rules are defined; S300 Component and Plugin Development: Through visual operations, low-code components are assembled and configured, component attributes, trigger events and data mapping relationships are set, standardized intermediate code adapted to the cloud computing environment is generated, high-code plugin development and registration are carried out simultaneously, and a two-way communication channel between low-code components and high-code plugins is established. S400, Intelligent Code Optimization: Retrieves the basic code of the developed high-code plugin, executes the code iteration and adaptation algorithm, combines business rules and code context, automatically identifies and performs code completion work through a domain-specific large model, improves logical branches and data verification content, and performs code detection and correction work through static analysis tools; S500 Intelligent Version Management: Collects development and modification information throughout the entire process and builds a traceable version archive, completes structured comparative analysis of different versions, generates version difference reports and rollback plans, and supports version preview and rollback operations.

2. The software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S100, the multimodal requirement information includes business requirement descriptions, functional boundary definitions, and business operation rules written by business personnel in natural language; business flow flowcharts, functional module relationship diagrams, and interface layout diagrams drawn in visual graphics; business data fields, data relationships, and data processing rules compiled in structured table form; and data interaction logic, user permission allocation rules, and application deployment and operation environment requirements in the business scenario.

3. The software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S100, the domain-specific large model is a large model that has been pre-trained specifically for financial, medical and industrial Internet scenarios. It has built-in business terminology system, industry business process specifications and development and implementation standards for the corresponding domain. When performing demand semantic parsing, it performs class-by-class semantic decomposition and business intent recognition on various multimodal demand information to extract the business subject, execution process, constraints and functional orientation information contained in the demand. In the feature integration stage, the parsing results of different modalities are normalized and aligned, the requirement feature information corresponding to each modality is integrated, and duplicate and conflicting content in the parsing results is removed to form a unified and regular set of requirement features, providing the basic data required for the operation of multimodal task quantization algorithms.

4. The software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S100, the mathematical expression of the multimodal task quantization algorithm is: in, Quantify the final generated structured hybrid development task tree; Input feature values ​​for the i-th type of multimodal demand; For domain-specific large model fit coefficients; Analyze the weighting coefficients for multimodal demand; A quantifiable value for the business complexity of functional modules; A general adaptability coefficient for low-code components; Weighting coefficients are assigned to the hybrid development model; n represents the total number of requirement input modalities.

5. The software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S200, the cloud-native hybrid development architecture is built on a containerized deployment model and includes two core units: a low-code general component library and a high-code custom plugin container. The architecture supports elastic scaling and distributed collaborative operation in a cloud computing environment. When dividing the functional scope of low-code components and high-code plugins, based on the quantification value output by the multimodal task quantification algorithm, page layout construction, form data display, basic information query, and general pop-up interaction functions are classified as low-code components, while business calculation processing, complex logic judgment, high-concurrency data operation, deep integration with third-party systems, and customized business logic execution functions are classified as high-code plugins. At the same time, a unified interaction standard, standardized data transmission format, and permission call rules are established between low-code components and high-code plugins.

6. The software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S200, the interaction standard determines the communication interaction level between low-code components and high-code plugins, specifies the interaction sequence of components initiating requests to plugins and plugins returning results to components, and defines data verification nodes and exception response mechanisms in the interaction process. The data format adopts a standardized structured data format, with unified field naming rules, data type definitions, data length limits and encoding standards. It supports cross-unit transmission of multiple types of business data and adapts to the distributed data interaction needs in a cloud computing environment. The calling rules determine the calling permissions, calling frequency limits and calling trigger conditions of low-code components to high-code plugins, specify the scope of data returned by plugins to components and the timeliness requirements for return, and standardize the calling chain and permission control logic between the two types of development units.

7. The software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S300, the standardized intermediate code is automatically generated based on the component configuration information of the visual operation. The code follows the containerized deployment specifications and distributed operation requirements of the cloud computing environment. The code structure is divided into a component layout definition section, an attribute parameter configuration section, a trigger event binding section, and a data mapping logic section. The component layout definition section determines the position coordinates, hierarchical relationship, and style configuration of each low-code component on the page. The attribute parameter configuration section stores the size, display rules, permission control, and verification logic of each component. The trigger event binding section records the trigger conditions and execution logic corresponding to each component. The data mapping logic section defines the data source, transmission link, and synchronization rules between components.

8. The software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S400, the code iterative adaptation algorithm uses the following mathematical expression: in, This represents the comprehensive quantitative value of intelligent version iteration; This indicates the code generation adaptation weight coefficients; This represents the quantified value of the structured hybrid development task tree; This represents code context-aware feature values; This indicates the weighting coefficient for version control optimization. This represents the quantification value of code static analysis optimization. This indicates the risk factor for version iteration.

9. A software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S400, the code detection and correction work is performed using a multi-dimensional static analysis tool. The detection work covers five dimensions: syntax standard verification, logic vulnerability investigation, performance bottleneck identification, security risk detection, and code redundancy analysis. The code is checked one by one for syntax correctness, boundary condition handling, loop complexity, memory leak risk, SQL injection vulnerability, and unauthorized access risk. At the same time, it is compared with the unified coding standard to identify problems such as non-standard naming, inconsistent format, and missing comments. The correction process categorizes and handles the detection results, automatically correcting syntax errors, formatting issues, and simple, redundant code. It generates precise remediation suggestions for logical vulnerabilities, performance bottlenecks, and security risks, marking the location of the problem, the scope of impact, and the remediation plan; For complex problems, multiple alternative correction schemes are generated by combining domain-specific large models for developers to choose from.

10. A software development method for the Internet and cloud computing according to claim 1, characterized in that, In step S500, the version difference report covers structured comparison results across four dimensions: code level, function level, configuration level, and dependency level. Code-level differences mark the added, modified, and deleted lines of code and associate them with the corresponding development tasks and modification personnel. Function-level differences summarize the specific content and business impact of added, modified, and deprecated functions. Configuration-level differences compare changes in cloud resources, permissions, and interfaces. Dependency-level differences list version changes and compatibility risks of third-party dependency libraries. The report also marks the risk level corresponding to each difference item. The rollback plan is formulated based on the comprehensive quantitative value of version intelligent iteration and the risk level of difference report. It is divided into three modes: full rollback, incremental rollback and gray rollback, which correspond to high-risk major failures, partial functional failures and medium-risk scenarios, respectively. The plan determines the execution steps, verification standards and emergency response procedures for each mode.

Images