System architecture reconstruction method, device and equipment and readable storage medium

Abstract

The invention discloses a system architecture reconstruction method, device and equipment and a readable storage medium, and is applied to the technical field of computers, and the system architecture reconstruction method comprises the following steps: obtaining multi-modal data of a to-be-migrated old system; analyzing the multi-modal data by using a system function analysis agent to obtain analyzed old system architecture data; mapping by using a data model agent according to the analyzed old system architecture data to obtain new system architecture data corresponding to the old system architecture data; and generating a new application system by utilizing a low-code business development platform according to the new system architecture data. Compared with the current manual reproduction of the existing system function, the method has the advantages that the new system architecture data corresponding to the old system architecture data are obtained through the multi-agent collaborative automatic analysis system function, so that the new application system is generated by utilizing the low-code service development platform based on the system architecture data, the old system functions do not need to be manually connected one by one, and the system efficiency is improved. The upgrading period is obviously shortened, and the upgrading efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a system architecture reconstruction method, apparatus, device, and readable storage medium. Background Technology

[0002] As enterprises continue to deepen their digital transformation, many companies' outdated MES (Enterprise Resource Planning) and ERP (Manufacturing Execution System) systems are no longer fully adapted to current business needs. When upgrading and transforming existing systems while retaining their functionality, customers often face the following problems: Outdated system architecture and difficulty in upgrading: The traditional architecture used in older systems cannot meet the current performance requirements of the enterprise even after upgrading; some systems have even lost the basic conditions for development and upgrade due to the dissolution of the original development vendor and the loss of core development personnel.

[0003] It is evident that improving the efficiency of upgrading old systems is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a system architecture reconstruction method, apparatus, device and readable storage medium, which solves the technical problem of low efficiency in upgrading old systems in the prior art.

[0005] To address the aforementioned technical problems, this invention provides a system architecture reconstruction method, comprising:

[0006] Obtain multimodal data of the legacy system to be migrated;

[0007] The system function analyzer is used to analyze the multimodal data to obtain the analyzed old system architecture data;

[0008] Based on the parsed old system architecture data, a data model agent is used to map the data to obtain new system architecture data corresponding to the old system architecture data.

[0009] Based on the new system architecture data, a new application system is generated using a low-code business development platform.

[0010] Optionally, before generating a new application system using a low-code business development platform based on the new system architecture data, the following steps are also included:

[0011] Based on the new system architecture data, the process engine intelligent agent is used to generate business process configuration information for the new system.

[0012] Accordingly, based on the new system architecture data, a new application system is generated using a low-code business development platform, including:

[0013] Based on the new system architecture data and the new system's business process configuration information, the new application system is generated using the low-code business development platform.

[0014] Optionally, the system function parsing agent is used to parse the multimodal data to obtain the parsed old system architecture data, including:

[0015] The system uses the built-in webpage parsing tool, image parsing tool, and document parsing tool in the intelligent agent to parse the multimodal data and obtain initial parsed data;

[0016] The initial parsed data is processed into a structured form to obtain the parsed old system architecture data.

[0017] Optionally, after acquiring the multimodal data of the legacy system to be migrated, the following may also be included:

[0018] The intelligent agent collaboration component is used to invoke all intelligent agents in a set order; the intelligent agents include at least the system function parsing intelligent agent and the data model intelligent agent.

[0019] The output data and process markers of each agent are recorded in chronological order using the agent collaboration component; wherein the process markers are markers that determine the progress of the task.

[0020] Optionally, after generating a new application system using a low-code business development platform based on the new system architecture data, the method further includes:

[0021] Obtain the error information returned by the low-code business development platform;

[0022] Based on the error information, the correction command agent performs error analysis to obtain error analysis logic, suggested modifications, and suggested correction agents.

[0023] Optionally, based on the error information, a correction command agent performs error analysis to obtain error analysis logic, suggested modifications, and a suggested correction agent, including:

[0024] Based on the error information, the error analysis logic and the suggested modifications are determined using the large language model in the corrective command agent.

[0025] Using a machine learning model to classify errors based on error information, the agent with the highest probability corresponding to each error is determined, and the agent with the highest probability is used as the agent for the proposed correction.

[0026] The rule engine determines the number of times the corrective command agent and the machine learning model are called. When the number of calls exceeds a set number, a prompt message is sent.

[0027] Optionally, it can be applied to the old raw material procurement system that needs to be upgraded.

[0028] The present invention also provides a system architecture reconstruction apparatus, comprising:

[0029] The multimodal data acquisition module is used to acquire multimodal data from the legacy system to be migrated.

[0030] The parsing module is used to parse the multimodal data by the intelligent agent using system functions to obtain the parsed old system architecture data;

[0031] The mapping module is used to map the parsed old system architecture data using a data model agent to obtain new system architecture data corresponding to the old system architecture data.

[0032] The system architecture reconstruction module is used to generate a new application system based on the new system architecture data using a low-code business development platform.

[0033] The present invention also provides a system architecture reconstruction device, comprising:

[0034] Memory, used to store computer programs;

[0035] A processor for executing the computer program to implement the steps of the above-described system architecture reconstruction method.

[0036] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described system architecture reconstruction method.

[0037] The present invention also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the above-described system architecture reconstruction method.

[0038] As can be seen, this invention acquires multimodal data of the legacy system to be migrated; uses a system function analysis agent to parse the multimodal data, obtaining parsed legacy system architecture data; uses a data model agent to map the parsed legacy system architecture data, obtaining new system architecture data corresponding to the legacy system architecture data; and uses a low-code business development platform to generate a new application system based on the new system architecture data. This invention automates the parsing of system functions and the construction of data models through the collaborative automation of a system function analysis agent and a data model agent, replacing the traditional manual reproduction mode. It eliminates the need for manual integration of legacy system functions, significantly reducing customization development and debugging time, shortening the upgrade cycle, and improving upgrade efficiency.

[0039] In addition, the present invention also provides a system architecture reconstruction apparatus, device and readable storage medium, which also have the above-mentioned beneficial effects. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0041] Figure 1 A flowchart of a system architecture reconstruction method provided in an embodiment of the present invention;

[0042] Figure 2 An architectural framework diagram of a system reconstruction method provided in an embodiment of the present invention;

[0043] Figure 3 A flowchart example of a system reconstruction method provided in an embodiment of the present invention;

[0044] Figure 4 This is a schematic diagram of a system function analysis intelligent agent provided in an embodiment of the present invention;

[0045] Figure 5 This is a schematic diagram of a system architecture reconstruction device provided in an embodiment of the present invention;

[0046] Figure 6 This is a structural schematic diagram of a system architecture reconstruction device provided in an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] Some terms that appear in the description of the embodiments of this application are subject to the following interpretation:

[0049] Intelligent agent: A smart system capable of perceiving environmental information, making autonomous decisions, and executing actions to achieve goals. In this invention, it refers to a large-scale artificial intelligence model system that incorporates knowledge base technology.

[0050] Low-code business development platform: A software development tool or environment that allows users to quickly build applications through a graphical interface, drag-and-drop components, and configuration logic, significantly reducing the workload of traditional manual coding. It makes development simpler, faster, and more accessible, enabling both professional developers and business personnel (such as HR, administration, operations, and other non-programmers) to participate in application building.

[0051] Implementation is a crucial stage in software development and IT projects, encompassing the entire management process from requirement implementation to system usability. This invention focuses on the process of building a low-code platform application based on user business information. The user ultimately receives the application generated by the platform, not the low-code platform itself.

[0052] Please refer to Figure 1 , Figure 1 A flowchart illustrating a system architecture reconstruction method provided in an embodiment of the present invention. The method may include:

[0053] S101, Obtain multimodal data of the old system to be migrated.

[0054] Each step in this embodiment can be executed by a designated electronic device, which can specifically be a server, a portable terminal, or other forms. This embodiment does not limit the specific legacy system to be migrated. For example, the legacy system to be migrated in this embodiment can be an outdated MES (Enterprise Resource Planning); or the legacy system to be migrated in this embodiment can also be an ERP (Manufacturing Execution System) applied to an old raw material procurement system to be upgraded. Multimodal data in this embodiment refers to a data set containing multiple different types or forms (i.e., modalities), which can include text, images, audio, video, etc. Here, "modality" refers to the form of data existence, information source, or perception method. For example, multimodal data in this embodiment can include images, web pages, document data, etc. The multimodal data acquired in this embodiment can include third-party system webpage links, system screenshots, and user manual documents. Third-party system webpage links provide real-time system interaction data (such as page elements, interface requests / responses, and operation-triggered process jumps), directly reflecting the actual business chain during system operation and compensating for dynamic interaction logic that static documents cannot reflect. System screenshots visually present the interface layout, function buttons, data display format, and operation paths, enabling rapid location of core functional modules and assisting the parsing agent in identifying "implicit operation logic" (such as pop-up trigger conditions and sequential constraints of multi-step operations). User manual documents contain explicit information such as the system design intent, core business processes, functional module descriptions, and parameter configuration rules, forming the basis for the parsing agent to obtain "officially defined logic" and avoiding rule deviations caused by inferring solely from interaction data. The combination of these three elements covers "dynamic interaction + visual scenarios + explicit rules," addressing the pain points of missing documentation and implicit logic in older systems, and providing comprehensive data support for subsequent function reproduction. Furthermore, API (interface documentation), database dictionaries, and system operation logs can also be acquired and processed.

[0055] It should be further noted that, based on any of the above embodiments, after obtaining the multimodal data of the legacy system to be migrated, the following may also be included:

[0056] Step 1: Use the agent collaboration component to call all agents in a set order; the agents must include at least the system function analysis agent and the data model agent;

[0057] Step 2: Use the agent collaboration component to record the output data and process markers of each agent in chronological order; where the process markers are markers that determine the progress of the task.

[0058] The agent collaboration component in this embodiment enables agent invocation and output coordination. It records the output and process markers of each agent in chronological order, and invokes agents according to the scheme flow. The agents in this embodiment may include system function analysis agents, data model agents, process engine agents, and correction command agents, etc. The process markers in this embodiment are metadata added to achieve process control and traceability, and may include: Task ID (identifier): uniquely identifies a complete migration task, linking all records in a single task; Timestamp: records the precise time each agent is invoked and terminated, used for performance analysis and fault diagnosis; Agent identifier: indicates which agent generated this record; Status markers: such as running, success, error; Debug count marker: specifically used to record the number of times the correction command agent has attempted to resolve the same problem. This is the key basis for triggering the "maximum 3 retries before manual intervention" rule.

[0059] S102, using the system function parsing agent to parse the multimodal data, obtain the parsed old system architecture data.

[0060] The system function analysis intelligence in this embodiment is a dedicated AI agent designed for reverse engineering and analyzing existing software systems. It performs multi-dimensional, in-depth analysis of legacy systems in an automated manner to obtain the analyzed legacy system architecture data. This system function analysis intelligence agent is a collaborative platform integrating multiple analysis tools, capable of processing different types of input data and determining the business logic of the analyzed information.

[0061] It should be further noted that, based on any of the above embodiments, the above-described method of using a system function parsing agent to parse multimodal data and obtain parsed old system architecture data may include:

[0062] S1021, the built-in webpage parsing tool, image parsing tool, and document parsing tool in the system function parsing agent are used to parse the multimodal data to obtain the initial parsing data;

[0063] S1022, perform structured processing on the initial parsed data to obtain the parsed old system architecture data.

[0064] The system function analysis agent in this embodiment incorporates tools such as webpage parsing, image parsing, and document parsing. It identifies the interface and functional logic of systems through third-party system links, system screenshots, and user manuals, and uses this preprocessed data as core input. The structured processing of the initial parsed data in this embodiment aims to achieve information unification, thereby obtaining the parsed old system architecture data. The system function analysis agent in this embodiment implements multimodal fusion, business logic extraction, and system function modeling.

[0065] S103, based on the parsed old system architecture data, the data model agent is used to perform mapping to obtain the new system architecture data corresponding to the old system architecture data.

[0066] In this embodiment, the data model agent is a dedicated AI agent responsible for data layer design and mapping. It transforms or "translates" the business function descriptions of the old system provided by the "system function analysis agent" into the business logic of the new system architecture. This business logic is a standardized, structured data model that the low-code business development platform can directly recognize and execute. The knowledge base built into the data model agent includes common business flow data models from common industries and data model information from the low-code business development platform.

[0067] S104: Based on the new system architecture data, use a low-code business development platform to generate a new application system.

[0068] The low-code development platform in this embodiment is a software development environment that allows developers and general business users to build applications through a graphical user interface, drag-and-drop components, and model-driven logic. This low-code platform can automatically generate runnable applications based on new system architecture data through its internal engine. It should be noted that the challenge in combining the system function parsing agent and the low-code business development platform lies in the fact that the low-code platform relies on standardized component libraries and configuration rules, while the functions of the old system are mostly customized (such as dedicated business processes and personalized computational logic). The "non-standardized function descriptions" extracted by the parsing agent cannot be directly mapped to the standardized components and configuration items of the low-code platform. This embodiment decomposes the extracted customized functions of the old system into "atomic functional units" through the parsing agent and establishes multi-dimensional feature tags (such as data processing, process approval, etc.). On the other hand, an extensible standardized component mapping rule library is built on the low-code platform side, supporting automatic matching and adaptation of components through feature tags. For personalized logic that cannot be directly matched, a visual flexible configuration interface is provided, enabling flexible combination of "atomic functional units" and low-code components.

[0069] It should be further explained that, based on any of the above embodiments, before generating a new application system using a low-code business development platform based on the new system architecture data, the process may further include: generating business process configuration information for the new system using a process engine intelligent agent based on the new system architecture data; correspondingly, generating a new application system using a low-code business development platform based on the new system architecture data includes: generating a new application system using a low-code business development platform based on the new system architecture data and the new system's business process configuration information. In this embodiment, the process engine intelligent agent's built-in knowledge base includes department and job settings for common industries and form process information from the low-code business development platform. This embodiment can generate process configurations based on the process engine intelligent agent. In addition to basic business requirements, the new system generated by this embodiment can upgrade approval processes to the new application system when approval processes exist, improving the comprehensiveness of the new application system's functions.

[0070] It should be further noted that, based on any of the above embodiments, after generating a new application system using a low-code business development platform based on the new system architecture data, it may further include:

[0071] Step 1: Obtain the error information returned by the low-code business development platform;

[0072] Step 2: Based on the error information, use the correction command agent to perform error analysis to obtain the error analysis logic, suggested modifications, and the agent to be corrected.

[0073] In this embodiment, the correction command agent has a built-in knowledge base that includes common error information from low-code business development platforms. Based on this error information, the correction command agent can provide error analysis logic, suggested modifications, and suggested agents for correction. The four agents in this embodiment (system function parsing, data model, process engine, and correction command) collaborate in a closed-loop process of "parsing-modeling-building-verification" under the scheduling of the collaboration component: After receiving preprocessed data, the agent collaboration component first triggers the system function parsing agent to extract old system information, then sequentially triggers the data model agent to generate data mapping, the process engine agent to generate process configuration, and finally the correction command agent to detect and correct errors. Collaboration is ensured through a shared bus and process recording. Configuring four agents is optimal for adapting to various scenarios because it fully covers the core links of old system upgrades, accurately solves technical difficulties at each stage, and avoids overloading or redundancy if added or removed.

[0074] It should be further explained that, based on any of the above embodiments, the above-mentioned error analysis using a correction command agent based on error information to obtain error analysis logic, suggested modification content, and suggested correction agent may include: determining the error analysis logic and suggested modification content based on the large language model in the correction command agent using error information; classifying errors based on error information using a machine learning model to determine the agent with the highest probability for each error, and using the agent with the highest probability as the agent for suggested correction; determining the number of calls to the correction command agent and the machine learning model based on a rule engine, and sending a prompt message when the number of calls exceeds a set number.

[0075] The error analysis logic in this embodiment is a logical reasoning and explanation of the root cause of the error. It answers why the error occurred. The suggested modifications are specific and actionable steps proposed to fix the error; it answers how to fix the error. The suggested agent for correction is the responsible agent designated to perform the correction task based on the error type and root cause. It answers who should perform this correction task and how to optimize this agent. This embodiment can determine the number of calls to the correction command agent and machine learning model based on a rule engine. When the number of calls exceeds a set number, a prompt message is sent. This embodiment uses a machine learning model to classify errors based on the error information, determine the agent with the highest probability for each error, and use the agent with the highest probability as the suggested agent for correction. For example, the error information is fed into a pre-trained machine learning classification model (such as a text classification model). This model analyzes the features of the error information (such as keywords and phrases) and quickly calculates the probability that the error belongs to the responsibility area of ​​each agent (such as a data model agent or a process engine agent). It outputs the agent with the highest probability as the "suggested agent for correction". In this embodiment, the large language model can perform contextual analysis based on error information and related logs to determine error analysis logic and suggested modifications. This embodiment assigns different technical tasks to different agents, improving the accuracy of determining suggested modifications and the agents responsible for correction. It classifies low-code platform errors using a rule engine and machine learning model, generates correction suggestions based on historical data, and iteratively corrects the target agent, ensuring efficient adaptation.

[0076] This invention provides a system architecture reconstruction method, which may include: S101, acquiring multimodal data of the old system to be migrated; S102, using a system function parsing agent to parse the multimodal data to obtain parsed old system architecture data; S103, using a data model agent to map the parsed old system architecture data to obtain new system architecture data corresponding to the old system architecture data; S104, using a low-code business development platform to generate a new application system based on the new system architecture data. This invention automates the parsing of system functions and the construction of data models and business processes through multi-agent collaborative automation, replacing the traditional manual reproduction mode. It eliminates the need for manual integration of old system functions, significantly reducing customization development and debugging time, shortening the upgrade cycle, and improving upgrade efficiency.

[0077] Some technologies have attempted to address these issues, such as manually reproducing existing system functions or using intelligent agent collaboration and low-code platforms to automatically build application systems through requirement descriptions. However, these technologies still generally suffer from pain points such as long implementation cycles, omission of key information, and incomplete reproduction of functions.

[0078] For a clearer understanding of this invention, please refer to the following details. Figure 2 , Figure 2 An architectural framework diagram of a system reconstruction method provided in an embodiment of the present invention may specifically include:

[0079] In this architecture framework, based on the input layer data, the system function parsing agent, data model agent, process engine agent, and correction command agent sequentially transmit their processing results to the agent collaboration component. Based on the collaborative results from the agent collaboration component, these results are then transmitted to the integration layer. Based on the processing at the integration layer, the system interfaces with the low-code business development platform's form design, process template, and other modules to build adapted system functions. Finally, based on the function reproduction and adaptation completed on the low-code platform, the upgrade and transformation of legacy systems is achieved. The integration layer is responsible for the data communication integration required between the system and the low-code business platform.

[0080] Figure 2 Please refer to the corresponding flowchart for details. Figure 3 , Figure 3 A flowchart example of a system reconstruction method provided in this embodiment of the invention may specifically include:

[0081] S201. Obtain the webpage link, system screenshots, and user manual document of the third-party system, perform preprocessing, and obtain the preprocessed data.

[0082] S202. Based on the preprocessed data, the system function parsing agent extracts old system information to obtain old system architecture information.

[0083] S203. Based on the old system architecture information, use the data model agent to perform mapping to obtain the new system architecture information corresponding to the old system architecture information.

[0084] S204. Based on the new system architecture information, process configuration information is generated using the process engine intelligent agent.

[0085] S205. Generate a new application system using a low-code business development platform based on the new system architecture information and process configuration information.

[0086] S206. Obtain error information based on the new application system, and use the correction command agent to correct the agent based on the error information.

[0087] The beneficial effects of the embodiments of the present invention may include:

[0088] Significantly shortens the system upgrade implementation cycle: This invention utilizes a multi-agent collaborative automated analysis system.

[0089] It streamlines functionality, builds data models and business processes, replacing the traditional manual reproduction method. It eliminates the need for manual integration of legacy system functions, significantly reducing customization development and debugging time, and shortening the upgrade cycle.

[0090] Significantly reduces overall upgrade and transformation costs: The solution eliminates the need for complete reconstruction or replacement of the old system. Core functions are accurately reproduced through intelligent agents, reducing costs associated with hardware replacement, custom development, and cross-departmental coordination. It avoids the high manpower investment and redundant development issues inherent in traditional upgrades, significantly reducing transformation costs and alleviating financial pressure on enterprises.

[0091] Ensuring completeness and accuracy of function reproduction: The system function analysis intelligent agent integrates multi-dimensional analysis tools and is supported by an industry knowledge base to comprehensively capture the interface logic, data relationships, and process details of the old system. This effectively avoids the loss of key information and omission of functions caused by manual operation, ensuring the stable availability of the upgraded system.

[0092] Lowering the technical barrier to upgrading legacy systems: No reliance on the original development team or system architecture details is required; functionality can be analyzed and built simply by inputting screenshots and documentation. Adaptable to various legacy MES and ERP systems, solving upgrade challenges due to outdated architecture and lack of development capabilities, enabling SMEs to efficiently advance their digital transformation.

[0093] Taking the scenario of reproducing and constructing the business function of raw material procurement in the manufacturing industry as an example, the steps of the present invention are explained in detail.

[0094] Step 1: Knowledge Base Initialization. This step is performed when the system is first deployed or when new industry support is added to build the built-in knowledge base for each agent.

[0095] It should be noted that the knowledge base content required for adding new industry support can be obtained from successfully run tasks provided by the low-code business development platform.

[0096] Low-code business platforms require that the data model include application type, form structure, form data model, and form process information.

[0097] Step 1.1: Prepare the data model intelligent agent knowledge base.

[0098] Based on the data model requirements of the low-code business platform, a predefined industry-standard data model template library is imported into the system. Each template contains a complete form structure and form data model (field definitions, relationships, and calculation rules). In this embodiment, the industry-standard data model and process knowledge base are built using a dual-path approach: "top-down (standard import) + bottom-up (case extraction)". Dynamic updates are achieved through "automatic pattern extraction + manual review", and a full lifecycle management system of "version control + industry-adaptive templates + incremental learning" is implemented to ensure the accuracy, scalability, and timeliness of the knowledge base.

[0099] Step 1.2: Prepare the knowledge base of the process engine intelligent agent.

[0100] Low-code business development platforms require form process data to include application type and form process information.

[0101] Step 2: Requirements Input and Structured Analysis

[0102] Collect access links, screenshots of core functional modules, and operation manuals for the outdated raw material procurement system to be upgraded.

[0103] Define a clear and unambiguous output format to form instructions for the large model, enabling it to correctly complete the tasks of information extraction and transformation. The output format and instructions include "identifying the core business entities in the file", "extracting key business processes", "organizing all forms required for each process", and "listing the key fields contained in each form (including field name, type, and whether it is required)".

[0104] Upon completion of this step, you will obtain a core business entity, N key business processes, M forms for each business process, and O form fields for each form.

[0105] Add a task start marker to the above output and output it to the collaboration component.

[0106] Step 3: System function analysis of the intelligent agent.

[0107] The system utilizes built-in parsing tools to analyze input source information. For example, webpage parsing tools crawl menu levels, interface interaction logic, and API call records through system links; image parsing tools identify the element layout and data display format of each functional interface; document parsing tools extract business rules such as "the raw material procurement list must be associated with material numbers"; and video parsing tools extract keyframes from videos, capturing user actions (clicking input boxes, entering text, clicking buttons, etc.) through frame difference analysis. It compares the text / value changes of UI (interface design) elements in adjacent keyframes (e.g., an input box changes from "empty" to "100," or "total price" changes from "0" to "500") to identify the causal chain of "operation A → field B change → field C change." For easier understanding, please refer to [link / reference needed]. Figure 4 , Figure 4 This is a schematic diagram of a system function parsing intelligent agent provided in an embodiment of the present invention. The intermediate presentation layer unifies the heterogeneous output of multiple source parsing tools such as web pages, screenshots, documents, and videos, standardizes UI elements, semantics, interactions, and time-series data, generates a "Raw Material Procurement System Functional Parsing Report," clearly defines core functional modules, field details, and business logic constraints, and synchronously pushes it to the intelligent agent collaboration component for archiving. The web page parsing tool uses dynamic crawling and mining technologies; the image parsing tool uses OCR (Optical Character Recognition) and YOLOv8 (Object Detection Algorithm) technologies; the document parsing tool uses NLP (Natural Language Processing) technologies; and the video parsing tool uses OpenCV (Open Source Computer Vision Library), OCR, and YOLOv8 technologies.

[0108] Step 4: Data Model Agent: Define a clear and unambiguous output format to form a large model instruction, so that the agent can build the forms required for the corresponding process based on common sense, knowledge base and input content (functional analysis report).

[0109] The collaboration component receives the content from the previous step and calls the data model agent multiple times. Each time, it passes in a business entity, a business process, a form, and the documentation collected in the previous step, for a total of N*M calls.

[0110] Upon completion of the call, N*M detailed form data entries with form fields will be obtained. The collaborative component receives the output and determines whether it has completed the design of all forms required for a certain process; if so, it proceeds to the next step; otherwise, it waits.

[0111] Step 5: Process Engine Intelligent Agent.

[0112] The collaboration component receives multiple form data models from the previous step. If all the necessary forms for a key business process are collected, this step begins. A clear and unambiguous output format is defined to form a large model instruction, allowing the agent to construct the corresponding business process for the business entity based on common sense, the knowledge base, and the input content. The process engine agent receives the business description and other documents from the previous step, along with the multiple form data models output from the previous step. Combining this with department and job description information from its own knowledge base, it synthesizes the process information required by the low-code business platform and sends it to the collaboration component. The collaboration component receives the output and determines whether it has completed all process information; if so, it proceeds to the next step; otherwise, it waits.

[0113] Step 6: Low-code development platform.

[0114] After collecting all necessary process information, the collaboration component sends all process information, form designs, business entities, and other necessary user information to the integration layer. The integration layer then performs operations such as logging in, opening or creating business entities according to the low-code platform requirements, and finally submits the process information and form designs. Based on this data, the low-code business development platform generates the application or outputs error messages back to the integration layer, which then sends them back to the collaboration component.

[0115] Step 7: Modify the command agent.

[0116] Upon receiving the error message, the collaboration component identifies all inputs and outputs for the current task based on the task start marker and potential debugging count markers. If the debugging count is greater than or equal to 3, the process terminates and manual correction is initiated. If the debugging count does not meet the termination condition, the error message, the current task's inputs and outputs, and the debugging count marker (incremented if it already exists) are passed to the correction command agent. A clear and unambiguous output format is defined to form a large model instruction, allowing the agent to construct error analysis, suggested modifications, suggested correction agents, and debugging count markers based on common sense, the knowledge base, and the input content. The correction command agent sends the above output to the collaboration component. The collaboration component receives the error analysis, suggested modifications, and suggested correction agents. Based on the suggested correction agent data, it decides whether to overlay the error analysis and suggested modifications onto the data model agent input or the process engine agent input. Then, it executes the steps accordingly.

[0117] Experimental verification shows that the system upgrade (architecture reconstruction) implementation cycle based on the method of this invention is shortened by 83%, the cost is reduced by 75%, and the accuracy of function reproduction is 93%. In a manufacturing procurement system migration example, multi-system integration was completed in 15 days, and the error debugging efficiency was improved by 82%. Please refer to Table 1, which is a comparison table of reconstruction efficiency provided by an embodiment of this invention, and please refer to Table 2, which is a comparison table of debugging times provided by an embodiment of this invention.

[0118] Table 1. Comparison of Reconstruction Efficiency

[0119]

[0120] Table 2. Comparison of Reconstruction Efficiency

[0121]

[0122] The system architecture reconstruction apparatus provided in the embodiments of the present invention will be described below. The system architecture reconstruction apparatus described below can be referred to in correspondence with the system architecture reconstruction method described above.

[0123] Please refer to the details. Figure 5 , Figure 5 A schematic diagram of a system architecture reconstruction device provided in an embodiment of the present invention may include:

[0124] The multimodal data acquisition module 100 is used to acquire multimodal data of the old system to be migrated;

[0125] The parsing module 200 is used to parse the multimodal data by the intelligent agent using system functions to obtain the parsed old system architecture data;

[0126] The mapping module 300 is used to map the parsed old system architecture data using a data model agent to obtain new system architecture data corresponding to the old system architecture data.

[0127] The system architecture reconstruction module 400 is used to generate a new application system based on the new system architecture data using a low-code business development platform.

[0128] Furthermore, based on any of the above embodiments, the above system architecture reconstruction method may further include:

[0129] The business process configuration information determination module is used to generate business process configuration information for the new system based on the new system architecture data and using the process engine intelligent agent.

[0130] Correspondingly, the system architecture reconstruction module 400 may include:

[0131] The system architecture reconstruction unit is used to generate the new application system based on the new system architecture data and the business process configuration information of the new system, using the low-code business development platform.

[0132] Furthermore, based on any of the above embodiments, the parsing module 200 may include:

[0133] The initial parsing data determination unit is used to parse the multimodal data using the built-in webpage parsing tool, image parsing tool, and document parsing tool in the system's parsing agent to obtain initial parsing data;

[0134] The parsing unit is used to perform structured processing on the initial parsed data to obtain the parsed old system architecture data.

[0135] Furthermore, based on any of the above embodiments, the above system architecture reconstruction method may further include:

[0136] The agent invocation module is used to invoke all agents in a set order using the agent collaboration component; the agents include at least the system function parsing agent and the data model agent.

[0137] The process marking module is used to record the output data and process markings of each agent in chronological order using the agent collaboration component; wherein the process markings are markings that determine the progress of the task.

[0138] Furthermore, based on any of the above embodiments, the above system architecture reconstruction method may further include:

[0139] The error information acquisition module is used to acquire error information returned by the low-code business development platform;

[0140] The correction module is used to perform error analysis using the correction command agent based on the error information, and to obtain error analysis logic, suggested modifications, and suggested correction agents.

[0141] Furthermore, based on any of the above embodiments, the above-mentioned correction module may include:

[0142] The error analysis unit is used to determine the error analysis logic and the suggested modification content based on the error information using the large language model in the correction command agent;

[0143] The agent determination unit for proposed correction is used to classify errors based on error information using a machine learning model, determine the agent with the highest probability corresponding to each error, and use the agent with the highest probability as the agent for proposed correction.

[0144] The prompt message sending unit is used to determine the number of calls to the correction command agent and the machine learning model based on the rule engine, and to send a prompt message when the number of calls exceeds a set number.

[0145] Furthermore, based on any of the above embodiments, the above system architecture reconstruction device is applied to the old raw material procurement system to be upgraded.

[0146] It should be noted that the order of the modules and units in the above-mentioned system architecture reconstruction device can be changed without affecting the logic.

[0147] This invention provides a system architecture reconstruction apparatus, which may include: a multimodal data acquisition module 100, used to acquire multimodal data of the old system to be migrated; a parsing module 200, used to parse the multimodal data using a system function parsing agent to obtain parsed old system architecture data; a mapping module 300, used to map the parsed old system architecture data using a data model agent to obtain new system architecture data corresponding to the old system architecture data; and a system architecture reconstruction module 400, used to generate a new application system using a low-code business development platform based on the new system architecture data. This invention replaces the traditional manual reproduction mode by automating the parsing of system functions and the construction of data models and business processes through multi-agent collaborative automation. It eliminates the need for manual integration of old system functions one by one, significantly reducing the time spent on customized development and debugging, shortening the upgrade cycle, and improving upgrade efficiency.

[0148] The following describes a system architecture reconstruction device provided by an embodiment of the present invention. The system architecture reconstruction device described below and the system architecture reconstruction method described above can be referred to in correspondence.

[0149] Please refer to Figure 6 , Figure 6 A schematic diagram of a system architecture reconstruction device provided in an embodiment of the present invention may include:

[0150] Memory 10 is used to store computer programs;

[0151] Processor 20 is used to execute computer programs to implement the system architecture reconstruction method described above.

[0152] The memory 10, processor 20, and communication interface 30 all communicate with each other through the communication bus 40.

[0153] In this embodiment of the invention, the memory 10 is used to store one or more programs. The programs may include program code, which includes computer operation instructions. In this embodiment of the invention, the memory 10 may store programs for implementing the following functions:

[0154] Obtain multimodal data of the legacy system to be migrated;

[0155] The system functions are used to analyze the intelligent agent to parse the multimodal data and obtain the parsed old system architecture data;

[0156] Based on the parsed old system architecture data, a data model agent is used to perform mapping to obtain new system architecture data corresponding to the old system architecture data.

[0157] Based on the new system architecture data, a new application system is generated using a low-code business development platform.

[0158] In one possible implementation, the memory 10 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; and the data storage area may store data created during use.

[0159] Furthermore, memory 10 may include read-only memory and random access memory, providing instructions and data to the processor. A portion of the memory may also include NVRAM. The memory stores operating systems and operating instructions, executable modules, or data structures, or subsets thereof, or extended sets thereof, wherein the operating instructions may include various operating instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and handling hardware-based tasks.

[0160] Processor 20 can be a central processing unit (CPU), an application-specific integrated circuit, a digital signal processor, a field-programmable gate array, or other programmable logic device. Processor 20 can be a microprocessor or any conventional processor. Processor 20 can call programs stored in memory 10.

[0161] The communication interface 30 can be an interface for the communication module, used to connect with other devices or systems.

[0162] Of course, it should be noted that, Figure 6 The structure shown does not constitute a limitation on the system architecture reconstruction device in the embodiments of the present invention. In practical applications, the system architecture reconstruction device may include more than Figure 6 More or fewer components as shown, or combinations of certain components.

[0163] The following describes the readable storage medium (i.e., computer-readable storage medium) provided in the embodiments of the present invention. The computer-readable storage medium described below can be referred to in correspondence with the system architecture reconstruction method described above.

[0164] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described system architecture reconstruction method.

[0165] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0166] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0167] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0168] Finally, it should be noted that in this document, relationships such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0169] The above provides a detailed description of a system architecture reconstruction method, apparatus, device, and readable storage medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A system architecture reconstruction method, characterized in that, include: Obtain multimodal data of the legacy system to be migrated; The system function analyzer is used to analyze the multimodal data to obtain the analyzed old system architecture data; Based on the parsed old system architecture data, a data model agent is used to map the data to obtain new system architecture data corresponding to the old system architecture data. Based on the new system architecture data, a new application system is generated using a low-code business development platform.

2. The system architecture reconstruction method according to claim 1, characterized in that, Before generating a new application system using a low-code business development platform based on the new system architecture data, the process also includes: Based on the new system architecture data, the process engine intelligent agent is used to generate business process configuration information for the new system. Accordingly, based on the new system architecture data, a new application system is generated using a low-code business development platform, including: Based on the new system architecture data and the new system's business process configuration information, the new application system is generated using the low-code business development platform.

3. The system architecture reconstruction method according to claim 1, characterized in that, The system functional analysis agent parses the multimodal data to obtain the parsed old system architecture data, including: The system uses the built-in webpage parsing tool, image parsing tool, and document parsing tool in the intelligent agent to parse the multimodal data and obtain initial parsed data. The initial parsed data is processed into a structured form to obtain the parsed old system architecture data.

4. The system architecture reconstruction method according to any one of claims 1 to 3, characterized in that, After acquiring the multimodal data of the legacy system to be migrated, the following steps are also included: The intelligent agent collaboration component is used to invoke all intelligent agents in a set order; the intelligent agents include at least the system function parsing intelligent agent and the data model intelligent agent. The output data and process markers of each agent are recorded in chronological order using the agent collaboration component; wherein the process markers are markers that determine the progress of the task.

5. The system architecture reconstruction method according to claim 1, characterized in that, After generating a new application system using a low-code business development platform based on the new system architecture data, the process also includes: Obtain the error information returned by the low-code business development platform; Based on the error information, the correction command agent performs error analysis to obtain error analysis logic, suggested modifications, and suggested correction agents.

6. The system architecture reconstruction method according to claim 5, characterized in that, Based on the error information, an error analysis is performed using a correction command agent to obtain error analysis logic, suggested modifications, and a suggested correction agent, including: Based on the error information, the error analysis logic and the suggested modifications are determined using the large language model in the corrective command agent. The machine learning model is used to classify errors based on error information, and the agent with the highest probability corresponding to each error is determined. The agent with the highest probability is then used as the agent for the proposed correction. The rule engine determines the number of times the corrective command agent and the machine learning model are called. When the number of calls exceeds a set number, a prompt message is sent.

7. The system architecture reconstruction method according to claim 1, characterized in that, It is applied to the old raw material procurement system that needs to be upgraded.

8. A system architecture reconstruction device, characterized in that, include: The multimodal data acquisition module is used to acquire multimodal data from the legacy system to be migrated. The parsing module is used to parse the multimodal data by the intelligent agent using system functions to obtain the parsed old system architecture data; The mapping module is used to map the parsed old system architecture data using a data model agent to obtain new system architecture data corresponding to the old system architecture data. The system architecture reconstruction module is used to generate a new application system based on the new system architecture data using a low-code business development platform.

9. A system architecture reconstruction device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the steps of the system architecture reconstruction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the system architecture reconstruction method as described in any one of claims 1 to 7.

Images