Method and system for generating service fulfillment data communication configuration arrangement based on AI agent

Abstract

The invention discloses a method and a system for generating service fulfillment data communication configuration and arrangement based on an AI agent, and the method comprises the following steps: the AI agent analyzes a problem proposed by a user in combination with an LLM to extract keywords, assembles the keywords in combination with a user Prompt project, carries out RAG retrieval query in a local knowledge graph by using TEXT-to-Cypher, and if the query exists in the knowledge graph, carries out RAG retrieval query by using the TEXT-to-Cypher, carries out RAG retrieval query by using the TEXT-to-Cypher, and carries out RAG retrieval query by using the TEXT-to-Cypher; if yes, carrying out RAG (Random Access Gateway) enhanced system Prompt engineering; the AI agent is combined with workFlow to perform task decomposition according to the questions proposed by the user and the related information of the enhanced systemPrompt engineering template content, and answers are constructed; the AI agent calls an orchestrator according to the generated design process content and orchestrates an API in an atomic power library of configuration content required by service opening; and starting a dynamic parameter generation code and a jar packet of the API interface according to the configuration issued by each link of the service. According to the method and the system, when various services are opened, the data communication configuration arrangement content is automatically generated, the external API interface is automatically generated, and a developer only needs to finely adjust the code generated by the AI agent.

Description

Technical Field

[0001] The present invention relates to the field of service provisioning, and in particular to a method and system for generating service provisioning data communication configuration arrangement based on an AI agent. Background Art

[0002] With the development of the current communications industry and the diversification of business scenarios, when various services are activated, it is necessary to develop and arrange the configuration content of data communication equipment based on the scenario, and provide API interfaces to external systems to activate the overall business.

[0003] The existing service activation data communication configuration and arrangement process and the generation of API interfaces used by external systems have different configuration contents when activating services in the same service scenario due to the large number of equipment manufacturers and different equipment models. Designers are prone to incomplete considerations when outputting design documents, and the documents are not perfect; when developing, R&D personnel have more customized development content and a long development cycle, which affects customer satisfaction. Summary of the invention

[0004] In order to solve the above-mentioned problems existing in the prior art, the present invention provides a method and system for generating data communication configuration orchestration for service activation based on AIagent (AI intelligent body). When various services are activated, the data communication configuration orchestration content is automatically generated and the external API interface is automatically generated. Developers only need to fine-tune the code generated by the AI ​​intelligent body, which improves R&D efficiency, reduces human resource costs, achieves rapid response, automatic generation, and improves customer satisfaction.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] In one embodiment of the present invention, a method for generating service activation data communication configuration orchestration based on an AI agent is proposed, the method comprising:

[0007] The AI ​​agent uses LLM's NLP technology to parse the user's questions and extract keywords. It then uses the userPrompt project to assemble the extracted keywords and uses TEXT-to-Cypher to perform RAG search queries in the local knowledge graph. If the query exists in the knowledge graph, the RAG enhanced systemPrompt project is performed.

[0008] The AI ​​agent combines workFlow to decompose tasks and construct answers based on the questions raised by users and the relevant information of the enhanced systemPrompt project template content;

[0009] The AI ​​agent calls the orchestrator based on the generated design process content to orchestrate the APIs in the atomic capability library corresponding to the configuration content required for business activation; it generates code and external API interface jar packages based on the dynamic parameters configured and issued in each link of business activation.

[0010] Further, the steps to construct the answer are as follows:

[0011] Check whether the configuration content required for service activation already exists in the traditional database. If so, directly provide feedback in the answer content that it already exists, and prompt that the existing system can be checked. If not, query the knowledge graph and use the reasoning ability of its graph database to infer the relationship between the network, service, action, device type, device manufacturer, device model and configuration content, and find the path between the related relationships.

[0012] Analyze the configuration content found for service activation and confirm its accuracy, relevance, and timing;

[0013] Based on the relevance and timing of the analysis results, LLM is used to build the atomic capability process content required for configuration in service provisioning.

[0014] The AI ​​agent returns the design process content for confirmation by the questioner; after confirmation, the AI ​​agent stores the confirmation result in the graphical database of the knowledge graph.

[0015] Furthermore, LLM constructs the receipt content according to the configuration content required for service activation, and calls the flow chart that outputs the description chart type syntax.

[0016] In one embodiment of the present invention, a system for generating and arranging service provisioning data communication configuration based on an AI agent is also proposed, and the system includes:

[0017] Intent recognition module, used by AI agent to analyze and extract keywords from questions raised by users in combination with LLM's NLP technology, assemble the extracted keywords in combination with the userPrompt project, and use TEXT-to-Cypher to perform RAG search queries in the local knowledge graph. If the query exists in the knowledge graph, the enhanced systemPrompt project is performed;

[0018] Retrieve the enhanced generation module for RAG enhanced systemPrompt engineering;

[0019] The task decomposition module is used by the AI ​​agent to combine the workFlow to decompose tasks and construct answers based on the questions raised by the user and the relevant information of the enhanced systemPrompt project template content;

[0020] The answer building module is used to perform knowledge retrieval, information analysis, answer building, and generate design process content;

[0021] The output result module is used by the AI ​​agent to call the orchestrator according to the generated design process content, and to orchestrate the APIs in the atomic capability library corresponding to the configuration content required for business activation; it generates code and external API interface jar packages according to the dynamic parameters configured and issued in each link of business activation.

[0022] Furthermore, an answer module is constructed, which is specifically used for:

[0023] Check whether the configuration content required for service activation already exists in the traditional database. If so, directly provide feedback in the answer content that it already exists, and prompt that the existing system can be checked. If not, query the knowledge graph and use the reasoning ability of its graph database to infer the relationship between the network, service, action, device type, device manufacturer, device model and configuration content, and find the path between the related relationships.

[0024] Analyze the configuration content found for service activation and confirm its accuracy, relevance, and timing;

[0025] Based on the relevance and timing of the analysis results, LLM is used to build the atomic capability process content required for configuration in service provisioning.

[0026] The AI ​​agent returns the design process content for confirmation by the questioner; after confirmation, the AI ​​agent stores the confirmation result in the graphical database of the knowledge graph.

[0027] Furthermore, LLM constructs the receipt content according to the configuration content required for service activation, and calls the flow chart that outputs the description chart type syntax.

[0028] In one embodiment of the present invention, a computer device is also proposed, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the aforementioned AIagent-based service activation data communication configuration orchestration generation is implemented.

[0029] In one embodiment of the present invention, a computer-readable storage medium is further proposed, which stores a computer program for executing the configuration and arrangement generation of service activation data communication based on AI agent.

[0030] Beneficial effects:

[0031] After the designers input the service activation data communication configuration arrangement requirements in different combination dimensions such as business scenarios, equipment manufacturers, and equipment models, the AI ​​agent performs enhanced retrieval + autonomous learning, and then outputs the corresponding arrangement process content. After multiple rounds of manual conversations, the arrangement process can be improved. After the designers confirm the arrangement process, the AI ​​agent can generate the corresponding code and generate the corresponding package file. The R&D personnel can release the version for use after fine-tuning the code. The AI ​​agent combines LLM (large model), Prompt engineering (prompt word engineering), knowledge graph, RAG (search enhancement generation), and workflow (workflow) technologies to realize the service activation data communication configuration arrangement, ensuring the security, professionalism, accuracy, and standardization of enterprise-level requirements such as communication service data communication configuration content, and achieving rapid response to function development, automatic generation, and improving R&D efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of building a knowledge graph of the present invention;

[0033] Figure 2 It is a triplet schematic diagram of the multi-level knowledge system of business scenarios, physical resource classes and logical resource classes of the present invention;

[0034] Figure 3 It is a framework diagram of the execution process of the method for generating the configuration arrangement of service activation data communication based on AI agent of the present invention;

[0035] Figure 4 It is the framework diagram of the intention identification of the present invention;

[0036] Figure 5 It is a search enhancement generation framework diagram of the present invention;

[0037] Figure 6 It is the task decomposition framework diagram of the present invention;

[0038] Figure 7 It is a framework diagram for constructing an answer in the present invention;

[0039] Figure 8 This is the code generation and jar package framework diagram of the present invention;

[0040] Fig. 9 This is a schematic diagram of the overall steps of a method for generating service activation data communication configuration orchestration based on an AI agent according to an embodiment of the present invention;

[0041] Fig.10 This is a structural diagram after the knowledge graph of one embodiment of the present invention is constructed;

[0042] Fig.11 It is a schematic diagram of the contents of a new installation process of an IP metropolitan area network Internet dedicated line dedicated gateway according to an embodiment of the present invention;

[0043] Fig.12 It is a schematic diagram of the system structure generated by the service activation data communication configuration arrangement based on the AI ​​agent of the present invention;

[0044] Fig.13 It is a schematic diagram of the computer device structure of the present invention. DETAILED DESCRIPTION

[0045] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present disclosure more thorough and complete, and to fully convey the scope of the present disclosure to those skilled in the art.

[0046] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, system, device, method or computer program product. Therefore, the present disclosure may be implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0047] According to an implementation of the present invention, a method for generating data communication configuration orchestration for service activation based on an AI agent is proposed. When various services are activated, the data communication configuration orchestration content is automatically generated and an external API interface is automatically generated. Developers only need to fine-tune the code generated by the AI ​​agent, which improves R&D efficiency, reduces human resource costs, achieves rapid response, and automatically generates, thereby improving customer satisfaction.

[0048] The principle and spirit of the present invention are explained in detail below with reference to several representative embodiments of the present invention.

[0049] The present invention proposes a method for generating service activation data communication configuration arrangement based on AIagent. Since the system belongs to an enterprise-level AI agent, the enterprise-level AI agent must have the following characteristics:

[0050] Professionalism: Possess in-depth industry knowledge and be able to understand and handle complex issues in specific areas

[0051] Accuracy: When providing services or solutions, AI agents need to maintain high accuracy and reduce errors

[0052] Security: Data sources must be privatized and strict access control measures must be implemented to protect user privacy and sensitive information from being leaked.

[0053] Standardization: In the communications industry, input and output content must be input and output in accordance with strict content formats.

[0054] Scalability: It is scalable as business needs develop, and can be expanded at the function level, business level, and workflow level.

[0055] Therefore, this AI agent is built by combining LLM (large language model), Prompt engineering (prompt word engineering), knowledge graph, RAG (retrieval enhanced generation) and workflow technologies.

[0056] Prompt engineering: It can guide the generative model to understand user questions and must be combined with the retrieved relevant knowledge to make the answers more accurate and targeted.

[0057] RAG (Retrieval-Augmented Generation): RAG technology can leverage existing data resources without the need for large-scale pre-trained models to achieve efficient question-answering services and improve resource utilization. It combines traditional retrieval-based methods with modern neural network generation methods to provide more accurate and context-relevant answers.

[0058] Workflow: Workflow is a method of organizing tasks and steps in a logical order to ensure that each task can be completed efficiently and orderly. Through standardized workflow design, it ensures that each user request is processed according to the same process to provide a consistent service experience.

[0059] Knowledge Graph: The knowledge graph stores data in a structured manner in the form of entities, attributes, and relationships, which can be stored locally to ensure the accuracy and completeness of the information. Based on the relationships in the graph, the AI ​​Agent can perform logical reasoning and provide more in-depth and accurate answers.

[0060] The overall steps of this method can be broken down into preconditions and execution process:

[0061] 1. Prerequisites: Knowledge graph and prompt word project construction

[0062] 1. How to build a knowledge graph Figure 1 As shown:

[0063] (1) Knowledge modeling: Construct a multi-level knowledge system of business scenarios, physical resource categories, and logical resource categories, and determine the triples of each level of knowledge system, such as Figure 2 As shown in the figure, the basic unit of the knowledge graph is a triple consisting of "Entity-Relationship-Entity".

[0064] Entity: refers to something that is distinguishable and exists independently, such as: Internet dedicated line, port query, sub-interface query, and QOS template creation.

[0065] Relationship: refers to the connection between different entities, and refers to the connection between entities, such as port query, sub-interface query and QOS template creation required for Internet dedicated lines.

[0066] (2) Knowledge extraction: Convert data from different sources and structures (unstructured data, such as text on a website, requires the use of natural language processing (NLP) technology to identify entities and relationships; structured data, such as records in a database, can be directly converted into a knowledge graph through specific algorithms such as D2R) into triplets of knowledge graph data to ensure the validity and integrity of the data.

[0067] (3) Knowledge fusion: Fusion of multiple sources of duplicate knowledge information, including fusion calculation and manual operation fusion, to ensure the consistency and uniqueness of entities.

[0068] (4) Knowledge storage: Import the processed data into the graph database neo4j to build a knowledge graph.

[0069] 2. Prompt project construction:

[0070] (1) UserPrompt project: used to improve the efficiency of user question intention recognition and the standardization of question fields. The following is an example:

[0071] You are now a service provisioning expert in the communications industry. Please help me identify intent according to the following definition:

[0072] knowledge_name: the name of the knowledge system, such as business scenarios and physical resources;

[0073] service_name: business scenario name, such as Internet dedicated line;

[0074] process_name: the name of the opening step, such as port query;

[0075] action: abstract action behavior, such as: open, stop, close.

[0076] Here is the data index:

[0077] Knowledge: index of knowledge system;

[0078] service: index of business scenarios;

[0079] process: The index of the activation process.

[0080] Attributes are represented by =>, and multiple attributes are separated by &&.

[0081] Here are some examples:

[0082] How to open an Internet dedicated line? ? Expressed as: process|service_name:Internet dedicated line&&action:open;

[0083] What services are included in the business scenario knowledge system? ? Expressed as: service|knowledge_name:business scenario&&action:provide service;

[0084] What is the process for opening an Internet dedicated line in a business scenario? ? Expressed as: process|knowledge_name:business scenario&&service_name:Internet dedicated line&&action:open.

[0085] (2) SystemPrompt Engineering: Design and optimize AI agent instructions to ultimately generate the desired output.

[0086] role: indicates whether the agent is a design agent or an execution agent;

[0087] Background: describes the background of the AI ​​agent;

[0088] Goals: Describe the ultimate goal of the AI ​​agent, such as code generation, process generation, and intelligent question and answer.

[0089] Contrains: Constrain the scope of the AIagent's retrieval execution;

[0090] Skills List: describes the skills that the AI ​​agent possesses, such as knowledge retrieval, data analysis, and user interaction.

[0091] Workflow: describes the execution process of the entire AI AGENT;

[0092] Example: Provides examples of input and output.

[0093] Initialization: Initialization content. As a choreography generator, I have strong programming and scripting capabilities, strictly abide by the principle of simplicity and efficiency of scripts, and welcome users in a friendly manner. I am your choreography generator and I am happy to serve you. Please provide detailed requirements for the choreography you need to generate.

[0094] 2. Execution process Figure 3 As shown:

[0095] The execution process can be decomposed into: intent recognition, retrieval enhancement generation, task decomposition, answer construction and output results.

[0096] 1. Intent recognition Figure 4 As shown:

[0097] (1) Users initiate questions.

[0098] (2) The AI ​​agent combines the natural language processing (NLP) technology of the large language model (LLM) to parse and extract keywords from questions raised by users.

[0099] (3) Assemble the extracted keywords using the user prompt project (userPrompt);

[0100] (4) Parameter judgment: After the extracted keywords are assembled, use TEXT-to-Cypher (converting natural language text into Cypher query statements) to perform search queries (RAG search) in the constructed local knowledge graph. If the query does not exist in the knowledge graph, it will be fed back to the parameter judgment, indicating that the user input is not standardized; if the query exists, the enhanced systemPrompt project (RAG enhancement) will be performed, such as the service activation of XX service based on XX equipment manufacturer and XX equipment model in XX networking mode.

[0101] In this step, the knowledge graph is mainly used for constraints on questions raised by users and for the retrieval function of questions raised by users. The scope of questions raised by users must be within the knowledge graph.

[0102] 2. Retrieval Enhancement Generation (RAG) Figure 5 As shown:

[0103] After the AI ​​agent recognizes the intent, it performs a search query (RAG search) in the built local knowledge graph.

[0104] If the query exists in the knowledge graph, the systemPrompt project (RAG enhancement) is enhanced to improve the systemPrompt project so that LLM can generate accurate receipts more accurately when building receipts. For example: XX service of XX networking mode is based on the service activation of XX equipment manufacturer and XX equipment model.

[0105] 3. Task decomposition:

[0106] Based on the questions raised by the user and the relevant information of the enhanced systemPrompt project template content, workFlow performs AI agent task decomposition (such as knowledge retrieval, information analysis, answer construction, and generation of design process content), such as Figure 6 shown.

[0107] 4. Construct the answer:

[0108] (1) Knowledge retrieval: Query the traditional database to see whether the configuration content of the business scenario already exists. If so, the system will directly provide feedback in the answer content that it already exists, and prompt the user to check the existing system. If not, the system will query the knowledge graph and use the reasoning ability of the graph database to infer the relationship between network-business-action (activation, removal)-equipment type-equipment manufacturer-equipment model-configuration content, and find the path between the related relationships.

[0109] (2) Information analysis: Analyze the configuration content of the business scenario found to confirm its accuracy, relevance, and timing.

[0110] (3) Build answers: Based on the relevance and timing of the analysis results, build the atomic capability process content required to configure the content during service activation.

[0111] (4) Generate design process content: The AI ​​agent returns the design process content to the person who initiated the question for confirmation. After confirmation, the AI ​​agent will provide feedback and store the feedback result in the graph database of the knowledge graph. Figure 7 shown.

[0112] 5. Generate code and jar package:

[0113] (1) After manual fine-tuning or confirmation, the fine-tuning is fed back to the AI ​​agent, and the AI ​​agent uses the graph database to update the knowledge graph.

[0114] (2) After confirming the result, the AI ​​agent calls the orchestrator to automatically orchestrate the API in the atomic capability library corresponding to the existing configuration content (atomic capabilities of the configuration content). Generate code based on the dynamic parameters (dynamic parameters of atomic capabilities, such as the port number of port query) configured and issued in each link of the service activation, and generate a jar package of the external API interface based on the code, such as Figure 8 shown.

[0115] It should be noted that, although the operations of the method of the present invention are described in a specific order in the above embodiments and the accompanying drawings, this does not require or imply that the operations must be performed in the specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0116] In order to more clearly explain the above-mentioned method for generating service activation data communication configuration orchestration based on AI agent, a specific embodiment is used for illustration below. However, it should be noted that this embodiment is only for better illustrating the present invention and does not constitute an improper limitation to the present invention.

[0117] like Fig. 9 As shown in the figure, the process of implementing a method for orchestrating and generating service activation data communication configuration based on AI agent is as follows:

[0118] S1. Build a knowledge graph

[0119] The basic unit of the knowledge graph is the "entity-relationship-entity" triple.

[0120] The data required for the knowledge graph include: nodes, relationships, and attributes.

[0121] (1) Node: A node is used to store entity objects.

[0122] (2) Relationship: A relationship is used to connect nodes and represents the relationship or behavior between two nodes.

[0123] (3) Attribute: Attribute refers to information represented in the form of key-value pairs.

[0124] The nodes that need to be created based on the service provisioning configuration can be mainly divided into three categories:

[0125] (1) Business scenario (node): establishes a relationship with the configuration content, reflecting the association between the business scenario and the configuration content.

[0126] (2) Configuration content (node): When a service is activated, the configuration content delivered contains a certain order, which is used to establish the forward and backward link relationship between the configuration content and the association relationship between the configuration content and the input parameter variables.

[0127] The business scenario model is shown in Table 1:

[0128] Table 1

[0129]

[0130] The configuration contents are as shown in Table 2:

[0131] Table 2

[0132] Head Entity Device Type Equipment manufacturers relation Tail Entity Tail Entity Parameters Port Query MSE Huawei Open 1 Configuration Configure

[0133] Based on the above, the accumulated professional knowledge can be used to model the data of knowledge graph entities. The structure of the knowledge graph after construction is as follows: Fig.10 As shown:

[0134] S2. Prompt template construction

[0135] (1) UserPrompt project template:

[0136] ##Role: You are now a service provisioning expert in the communications industry. Please help me identify intent according to the following definition:

[0137] ##Goals:

[0138] knowledge_name: the name of the knowledge system, such as business scenarios and physical resources.

[0139] service_name: business scenario name, such as Internet dedicated line.

[0140] process_name: The name of the activation step, such as port query.

[0141] action: abstract action behavior, such as: open, stop, close.

[0142] Here is the data index:

[0143] knowledge: index of knowledge system.

[0144] service: index of business scenario.

[0145] process: The index of the activation process.

[0146] Attributes are represented by =>, and multiple attributes are separated by &&.

[0147] ##Skills:

[0148] -Ability to quickly retrieve relevant information from the knowledge base.

[0149] - Perform logical analysis on the information provided to ensure the accuracy of the answer.

[0150] ##Example:

[0151] How to open an Internet dedicated line? ? Expressed as: process|service_name:Internet dedicated line&&action:open.

[0152] What services are included in the business scenario knowledge system? ? Expressed as: service|knowledge_name:business scenario&&action:provide service.

[0153] What is the process for opening an Internet dedicated line in a business scenario? ? Expressed as: process|knowledge_name:business scenario&&service_name:Internet dedicated line&&action:open.

[0154] (2) SystemPrompt project template:

[0155] ##Role: Business activation data communication equipment configuration content automatic orchestration process generator.

[0156] ##Background: I am a professional orchestration process generator who can generate efficient automated orchestration process scripts according to user needs. I am familiar with multiple programming languages ​​and tools and can provide detailed documentation and scripts.

[0157] ##Preferences:

[0158] -The generated scripts should be concise, efficient and easy to maintain.

[0159] - Documentation should be detailed, clear, and easy for users to understand and execute.

[0160] -Supports multiple programming languages ​​and tools, such as Bash, Python, Java, etc.

[0161] ##Profile:

[0162] -author:liuhongyang.

[0163] -version:0.2.

[0164] -language: Chinese.

[0165] -description: Generate efficient automated orchestration processes based on user business requirements, generate final flow charts, interface documents and other requirement design documents, and generate the final jar package based on the code.

[0166] ##Goals:

[0167] -Generate efficient automated orchestration process scripts.

[0168] - Provide detailed documentation and comments.

[0169] -Supports multiple programming languages ​​and tools.

[0170] - Ensure script readability and maintainability.

[0171] ##Constrains:

[0172] -The script should be concise and efficient.

[0173] -The documentation should be detailed and clear.

[0174] -Supported programming languages ​​and tools should be within the user specified range.

[0175] - Don't add features that users haven't requested.

[0176] ##Skills:

[0177] -Ability to quickly retrieve relevant information from the knowledge base.

[0178] - Perform logical analysis on the information provided to ensure the accuracy of the answer.

[0179] -Strong programming and scripting skills.

[0180] -Familiar with multiple programming languages ​​and tools.

[0181] - Ability to generate detailed documentation and comments.

[0182] - Ensure script readability and maintainability.

[0183] ##Workflow:

[0184] - Receive inquiries: First receive and understand the user's questions.

[0185] -Knowledge retrieval: Find information related to the problem in the knowledge base.

[0186] -Information Analysis: Analyze the retrieved information to confirm its accuracy and relevance.

[0187] -Construct an answer: Based on the analysis results, construct an accurate and informative answer.

[0188] -Generate design process content: can generate detailed design process content.

[0189] -Generate code and jar package: Combine the code base to perform atomic capability orchestration and generate jar package.

[0190] ##Example:

[0191] ##Initialization:

[0192] As an orchestration process generator, I have strong programming and scripting skills, strictly abide by the principle of simplicity and efficiency of scripts, and welcome users in a friendly manner. I am your orchestration process generator and I am happy to serve you. Please provide the detailed requirements of the orchestration process you need to generate.

[0193] S3. Propose business activation requirements:

[0194] Input: New installation process of IP metropolitan area network Internet dedicated line dedicated gateway.

[0195] S4. Analysis of proposed content:

[0196] IP metropolitan area network, Internet dedicated line, exclusive gateway, new installation and process.

[0197] S5. Parameter assembly:

[0198] Parameters are assembled according to the configured keywords and specifications.

[0199] Major: IP metropolitan area network; Service: Internet dedicated line; Resource configuration: exclusive gateway; Scenario: new installation.

[0200] S6. Knowledge graph query:

[0201] Convert natural language text into Cypher query statements for knowledge graph retrieval query.

[0202] MATCH (n: business scenario content)

[0203] WHERE n.Professional = 'IP metropolitan area network' AND

[0204] n.Business = 'Internet dedicated line' AND

[0205] n. Resource configuration = 'exclusive gateway' AND

[0206] n. Scene = 'new clothes'

[0207] RETURN

[0208] S7. Enhanced prompt word engineering:

[0209] Improve the Prompt project, inform the large language model to query the professional: IP metropolitan area network; business: Internet dedicated line; resource configuration: exclusive gateway; scenario: newly installed related atomic capabilities, and perform orchestration.

[0210] ##Role: Business activation data communication equipment configuration content automatic orchestration process generator.

[0211] ##Goals:

[0212] -Query major: IP metropolitan area network; business: Internet dedicated line; resource configuration: exclusive gateway; scenario: newly installed related atomic capabilities, generate efficient automated orchestration process scripts.

[0213] - Provide detailed documentation and comments.

[0214] -Supports multiple programming languages ​​and tools.

[0215] - Ensure script readability and maintainability.

[0216] S8. Select the prompt word project:

[0217] After the large language model performs language understanding and intent recognition, it calls the appropriate prompt word project and queries whether it is necessary to perform the new installation process of the IP metropolitan area network Internet dedicated line exclusive gateway. The selected role is: IP metropolitan area network Internet dedicated line exclusive gateway new installation process orchestration jar package generator.

[0218] S9. Task decomposition:

[0219] According to the workflow in the S7 prompt word project, the tasks are decomposed and the overall task planning is as follows: (knowledge retrieval, information analysis, answer construction, and generation of design process content)

[0220] S10. Knowledge retrieval:

[0221] Need to obtain the configuration content related to the new installation of the IP metropolitan area network Internet dedicated line exclusive gateway,

[0222] MATCH(n:data communication configuration content)-[r]->(m)

[0223] WHERE n.Professional = 'IP metropolitan area network' AND

[0224] n.Business = 'Internet dedicated line' AND

[0225] n. Resource configuration = 'exclusive gateway' AND

[0226] n. Scene = 'new clothes'

[0227] RETURN n,r,m

[0228] S11. Information analysis:

[0229] The knowledge graph analysis shows that the configuration content related to the new installation of the IP metropolitan area network Internet dedicated line exclusive gateway includes multiple equipment models from manufacturers such as Huawei, ZTE, and XX. It is necessary to analyze the necessary steps and the order of steps according to different manufacturers, and analyze the exceptions.

[0230] Huawei: Port query--XX configuration 1---XX configuration 2.

[0231] ZTE: Port query--XX configuration N--XX configuration 2.

[0232] S12. Construct the answer:

[0233] The large language model is combined with the configuration content related to the new installation of the IP metropolitan area network Internet dedicated line exclusive gateway to construct the receipt content and call the output of the Mermaid (Mermaid is a tool for generating charts. It uses Markdown-style syntax to describe various chart types, including flowcharts, sequence diagrams, Gantt charts, etc.) syntax flowchart to facilitate designers to view it intuitively.

[0234] (1) Port check: First, check the port. If it exists, jump to the next step; if it does not exist, exit with an error.

[0235] (2) Huawei devices: XX configuration 1 is required. If the configuration is successful, proceed to the next step. If the configuration is unsuccessful, the system will exit with an error. ZTE devices: XX configuration N is required. If the configuration is successful, proceed to the next step. If the configuration is unsuccessful, the system will exit with an error.

[0236] (3) XX configuration 2: Perform XX configuration 2. If it exists, jump to the next step; if it does not exist, exit with an error.

[0237] Mermaid Contents:

[0238]

[0239]

[0240] S13. Generate design process content:

[0241] The contents of the new installation process of the IP metropolitan area network Internet dedicated line dedicated gateway are as follows Fig.11 As shown:

[0242] (1) Port check: First, check the port. If it exists, jump to the next step; if it does not exist, exit with an error.

[0243] (2) Huawei devices: XX configuration 1 is required. If the configuration is successful, proceed to the next step. If the configuration is unsuccessful, the system will exit with an error. ZTE devices: XX configuration N is required. If the configuration is successful, proceed to the next step. If the configuration is unsuccessful, the system will exit with an error.

[0244] (3) XX configuration 2: Perform XX configuration 2. If it exists, jump to the next step; if it does not exist, an error message is entered.

[0245] S14. Update the knowledge graph:

[0246] The AI ​​agent returns the design process content for the questioner to confirm. After confirmation, the AI ​​agent stores the feedback result in the constant knowledge base, pre-processes the data, including word segmentation, stop word removal, stemming, etc., and uses a suitable vectorization method (such as Word2Vec) to convert the text into a vector and store it in the vector knowledge base.

[0247] S15. Generate code and jar package:

[0248] According to the process content, the AI ​​agent calls the orchestrator to assemble and assemble the API of the atomic capability library, and generates the jar package and related interface documents of the external API interface according to the dynamic parameters sent by the configuration.

[0249] Based on the same inventive concept, the present invention also proposes a system for generating service activation data communication configuration orchestration based on AI agent. The implementation of the system can refer to the implementation of the above method, and the repeated parts will not be repeated. The term "module" used below can be a combination of software and / or hardware that implements a predetermined function. Although the system described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0250] Fig.12 This is a schematic diagram of the system structure generated by the service activation data communication configuration arrangement based on the AI ​​agent of the present invention. Fig.12 As shown, the system includes:

[0251] Intent recognition module 101 is used for AI agent to analyze and extract keywords from questions raised by users in combination with LLM's NLP technology, assemble the extracted keywords in combination with the userPrompt project, and use TEXT-to-Cypher to perform RAG search queries in the local knowledge graph. If the query exists in the knowledge graph, the enhanced systemPrompt project is performed;

[0252] The search enhancement generation module 102 is used to perform the RAG enhancement systemPrompt project;

[0253] The task decomposition module 103 is used for the AI ​​agent to decompose tasks and construct answers based on the questions raised by the user and the relevant information of the enhanced systemPrompt project template content in combination with the workFlow;

[0254] The answer building module 104 is used to perform knowledge retrieval, information analysis, answer building, and generate design process content; the details are as follows:

[0255] Check whether the configuration content required for service activation already exists in the traditional database. If so, directly provide feedback in the answer content that it already exists, and prompt that the existing system can be checked. If not, query the knowledge graph and use the reasoning ability of its graph database to infer the relationship between the network, service, action, device type, device manufacturer, device model and configuration content, and find the path between the related relationships.

[0256] Analyze the configuration content found for service activation and confirm its accuracy, relevance, and timing;

[0257] Based on the relevance and timing of the analysis results, LLM is used to build the atomic capability process content of the configuration content required for service activation; LLM builds the receipt content based on the configuration content required for service activation, and calls the flow chart that outputs the description of the chart type syntax;

[0258] The AI ​​agent returns the design process content for confirmation by the questioner; after confirmation, the AI ​​agent stores the confirmation result in the graphical database of the knowledge graph.

[0259] The output result module 105 is used for the AI ​​agent to call the orchestrator according to the generated design process content, and to orchestrate the API in the atomic capability library corresponding to the configuration content required for service activation; and to generate the code and the jar package of the external API interface according to the dynamic parameters configured and issued in each link of the service activation.

[0260] It should be noted that although several modules of the system for generating the configuration and orchestration of the service activation data communication based on the AI ​​agent are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more modules described above can be concretized in one module. Conversely, the features and functions of one module described above can be further divided into multiple modules for concretization.

[0261] Based on the above invention concept, Fig.13 As shown, the present invention also proposes a computer device 200, including a memory 210, a processor 220, and a computer program 230 stored in the memory 210 and executable on the processor 220. When the processor 220 executes the computer program 230, the aforementioned AI agent-based service activation data communication configuration orchestration generation is implemented.

[0262] Based on the aforementioned inventive concept, the present invention further proposes a computer-readable storage medium, which stores a computer program for executing the aforementioned AI agent-based service activation data communication configuration orchestration generation.

[0263] The method and system for generating the configuration arrangement of service activation data communication based on AI agent proposed in the present invention, after the designer inputs the configuration arrangement requirements of service activation data communication in different combination dimensions such as business scenarios, equipment manufacturers, equipment models, etc., the AI ​​agent performs enhanced retrieval + autonomous learning, and outputs the corresponding arrangement process content. After multiple rounds of manual conversations, the arrangement process can be improved. After the designer confirms the arrangement process, the AI ​​agent can generate the corresponding code and generate the corresponding package file. The R&D personnel can release the version for use after fine-tuning the code. The AI ​​agent combines LLM (large model), Prompt engineering (prompt word engineering), knowledge graph, RAG (search enhancement generation) and workflow (workflow) technologies to realize the configuration arrangement of service activation data communication, ensuring the security, professionalism, accuracy and standardization of enterprise-level requirements such as communication service data communication configuration content, achieving rapid response to function development, automatic generation, and improving R&D efficiency.

[0264] Although the spirit and principle of the present invention have been described with reference to several specific embodiments, it should be understood that the present invention is not limited to the disclosed specific embodiments, and the division of various aspects does not mean that the features in these aspects cannot be combined to benefit, and such division is only for the convenience of expression. The present invention is intended to cover various modifications and equivalent arrangements contained in the spirit and scope of the attached claims.

[0265] Regarding the limitation of the protection scope of the present invention, those skilled in the art should understand that, based on the technical solution of the present invention, various modifications or variations that can be made by those skilled in the art without creative work are still within the protection scope of the present invention.

Claims

1. A method for generating service activation data communication configuration arrangement based on AI agent, characterized in that: The method includes: The AI ​​agent uses LLM's NLP technology to parse the user's questions and extract keywords. It then uses the userPrompt project to assemble the extracted keywords and uses TEXT-to-Cypher to perform RAG search queries in the local knowledge graph. If the query exists in the knowledge graph, the RAG enhanced systemPrompt project is performed. The AI ​​agent combines workFlow to decompose tasks and construct answers based on the questions raised by users and the relevant information of the enhanced systemPrompt project template content; The AI ​​agent calls the orchestrator based on the generated design process content to orchestrate the APIs in the atomic capability library corresponding to the configuration content required for business activation; it generates code and external API interface jar packages based on the dynamic parameters configured and issued in each link of business activation.

2. The method for generating service activation data communication configuration arrangement based on AI agent according to claim 1, characterized in that: The steps to construct the answer are as follows: Check whether the configuration content required for service activation already exists in the traditional database. If so, directly provide feedback in the answer content that it already exists and prompt that the existing system can be checked; If it does not exist, the knowledge graph is queried, and the reasoning ability of its graph database is used to infer the relationship between the network, business, action, device type, device manufacturer, device model, and configuration content, and to find the path between the related relationships; Analyze the configuration content found for service activation and confirm its accuracy, relevance, and timing; Based on the relevance and timing of the analysis results, LLM is used to build the atomic capability process content required for configuration in service provisioning. The AI ​​agent returns the design process content for confirmation by the questioner; after confirmation, the AI ​​agent stores the confirmation result in the graphical database of the knowledge graph.

3. The method for generating service activation data communication configuration arrangement based on AI agent according to claim 2 is characterized in that: The LLM constructs the receipt content according to the configuration content required for service activation, and calls the flow chart that outputs the description chart type syntax.

4. A system for orchestrating and generating service activation data communication configuration based on AI agent, characterized in that: The system includes: Intent recognition module, used by AI agent to analyze and extract keywords from questions raised by users in combination with LLM's NLP technology, assemble the extracted keywords in combination with the userPrompt project, and use TEXT-to-Cypher to perform RAG search queries in the local knowledge graph. If the query exists in the knowledge graph, the enhanced systemPrompt project is performed; Retrieve the enhanced generation module for RAG enhanced systemPrompt engineering; The task decomposition module is used by the AI ​​agent to combine the workFlow to decompose tasks and construct answers based on the questions raised by the user and the relevant information of the enhanced systemPrompt project template content; The answer building module is used to perform knowledge retrieval, information analysis, answer building, and generate design process content; The output result module is used by the AI ​​agent to call the orchestrator according to the generated design process content, and to orchestrate the APIs in the atomic capability library corresponding to the configuration content required for business activation; it generates code and external API interface jar packages according to the dynamic parameters configured and issued in each link of business activation.

5. The system for generating service activation data communication configuration arrangement based on AI agent according to claim 1, characterized in that: The constructing answer module is specifically used for: Check whether the configuration content required for service activation already exists in the traditional database. If so, directly provide feedback in the answer content that it already exists and prompt that the existing system can be checked; If it does not exist, the knowledge graph is queried, and the reasoning ability of its graph database is used to infer the relationship between the network, business, action, device type, device manufacturer, device model, and configuration content, and to find the path between the related relationships; Analyze the configuration content found for service activation and confirm its accuracy, relevance, and timing; Based on the relevance and timing of the analysis results, LLM is used to build the atomic capability process content required for configuration in service provisioning. The AI ​​agent returns the design process content for confirmation by the questioner; after confirmation, the AI ​​agent stores the confirmation result in the graphical database of the knowledge graph.

6. The system for generating service activation data communication configuration arrangement based on AI agent according to claim 5, characterized in that: The LLM constructs the receipt content according to the configuration content required for service activation, and calls the flow chart that outputs the description chart type syntax.

7. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 3 is implemented.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program for executing the method according to any one of claims 1 to 3.

Images