Operation and maintenance service report generation method and equipment based on large model scheduling and arrangement
By using large-scale model scheduling and orchestration technology, and leveraging intent analysis, problem classification, and parameter extraction to generate visualized operation and maintenance service reports, the problem of low efficiency and accuracy in operation and maintenance report generation is solved, and operation and maintenance development costs are reduced.
Patent Information
- Application Number
- CN202510972247.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-14
- Publication Date
- 2025-11-11
AI Technical Summary
In the field of security equipment operation and maintenance, as the types and quantities of equipment increase, the efficiency and accuracy of operation and maintenance service report generation are affected, leading to increased operation and maintenance development costs.
A large-scale model-based scheduling and orchestration approach is adopted to generate a visualized operation and maintenance service report through intent analysis, problem classification, parameter extraction, and data acquisition. The report is generated by constructing a rule chain using large-scale model components for intent analysis, problem classification, parameter extraction, and data.
It improves the efficiency and accuracy of generating operation and maintenance service reports, reduces operation and maintenance development costs, and adapts to complex and personalized operation and maintenance business scenarios.
Smart Images
Figure CN120929070A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of large model technology, and in particular to a method and device for generating operation and maintenance service reports based on large model scheduling and orchestration. Background Technology
[0002] In the field of security equipment operation and maintenance, operation and maintenance teams usually need to conduct statistical analysis of operation and maintenance services and generate operation and maintenance service reports to have a clear understanding of the status of the entire operation and maintenance project, or to report and track it regularly.
[0003] However, as the types and number of devices managed in projects increase, the complexity of operation and maintenance service reports also increases, often requiring a lot of custom development, which increases operation and maintenance development costs.
[0004] Improving the efficiency and accuracy of operation and maintenance service report generation has become a pressing technical problem. Summary of the Invention
[0005] In view of this, this application provides a method and device for generating operation and maintenance service reports based on large model scheduling and orchestration.
[0006] According to a first aspect of the embodiments of this application, a method for generating operation and maintenance service reports based on large-scale model scheduling and orchestration is provided, including:
[0007] Obtain a natural language requirement description for generating operation and maintenance service reports;
[0008] Using large-scale intent analysis model components, intent analysis is performed on the natural language demand description to obtain intent analysis results;
[0009] Based on the intent analysis results, the corresponding target problem scenario is determined using the components of the problem classification model.
[0010] Based on the target problem scenario, the operation and maintenance parameters are extracted from the natural language requirement description using the parameter extraction large model components to obtain the target operation and maintenance parameters;
[0011] Based on the target operation and maintenance parameters, a corresponding target operation and maintenance parameter instruction set is generated, and based on the target operation and maintenance parameter instruction set, the target operation and maintenance data is obtained using the corresponding target data element;
[0012] Based on the target problem scenario, target code is generated using code large model components for the target operation and maintenance data, and a visual operation and maintenance service report page is generated based on the target code.
[0013] According to a second aspect of the embodiments of this application, an operation and maintenance service report generation device based on large-scale model scheduling and orchestration is provided, comprising:
[0014] The acquisition unit is configured to acquire a natural language requirement description used to generate an operation and maintenance service report;
[0015] The controller unit is configured to perform intent analysis on the natural language demand description using intent analysis big model elements to obtain intent analysis results;
[0016] The controller unit is also configured to determine the corresponding target problem scenario based on the intent analysis results and using the problem classification big model components;
[0017] The controller unit is also configured to extract operation and maintenance parameters from the natural language requirement description based on the target problem scenario and the parameter extraction large model component to obtain the target operation and maintenance parameters.
[0018] The controller unit is also configured to generate a corresponding target operation and maintenance parameter instruction set based on the target operation and maintenance parameters, and to acquire target operation and maintenance data using the corresponding target data element based on the target operation and maintenance parameter instruction set.
[0019] The controller unit is also configured to generate target code based on the target problem scenario, using code big model components to generate target code for the target operation and maintenance data, and generate a visual operation and maintenance service report page based on the target code.
[0020] According to a third aspect of the embodiments of this application, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;
[0021] Memory, used to store computer programs;
[0022] The processor, when executing a program stored in memory, implements the method provided in the first aspect.
[0023] According to a fourth aspect of the embodiments of this application, a non-transitory computer-readable storage medium is provided, wherein a computer program is stored in the non-transitory computer-readable storage medium, and the computer program implements the method provided in the first aspect when executed by a processor.
[0024] The operation and maintenance service report generation method based on large model scheduling and orchestration in this application embodiment obtains a natural language requirement description for generating the operation and maintenance service report. It then uses intent analysis large model components to perform intent analysis on the obtained natural language requirement description. Based on the intent analysis results, it uses a problem classification large model component to determine the corresponding target problem scenario. According to the target problem scenario, it uses a parameter extraction large model component to extract operation and maintenance parameters from the natural language requirement description. The extracted target operation and maintenance parameters generate a corresponding target operation and maintenance parameter instruction set. Based on the target operation and maintenance parameter instruction set, it uses a corresponding target tree component to obtain target operation and maintenance data. Furthermore, based on the target problem scenario, it uses a code large model component to generate target code for the target operation and maintenance data, and generates a visual operation and maintenance service report page based on the target code. By encapsulating the capabilities of the large model into large model components, the flexibility of large model function calls is improved. Different large model components can be scheduled and orchestrated according to actual needs to construct rule chains. The generated operation and maintenance service report is achieved based on the constructed rule chains, improving the accuracy and efficiency of operation and maintenance service report generation. Attached Figure Description
[0025] Figure 1 This is a flowchart illustrating a method for generating operation and maintenance service reports based on large-scale model scheduling and orchestration, as provided in an embodiment of this application.
[0026] Figure 2 This is a flowchart illustrating an operation and maintenance data integrity analysis provided in an embodiment of this application;
[0027] Figure 3 This is a flowchart illustrating a method for generating operation and maintenance service reports based on large-scale model scheduling and orchestration, as provided in an embodiment of this application.
[0028] Figure 4 This is a schematic diagram of the structure of an operation and maintenance service report generation device based on large model scheduling and orchestration provided in an embodiment of this application;
[0029] Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0030] To enable those skilled in the art to better understand the technical solutions provided in the embodiments of this application, some technical terms involved in the embodiments of this application will be briefly explained below.
[0031] Prompt Engineering refers to the techniques and methods used to design and optimize natural language prompts to guide large-scale artificial intelligence models in generating content that meets expectations.
[0032] AIGC (AI Generated Content) large-scale models refer to large-scale machine learning models that utilize artificial intelligence technology to autonomously generate various forms of content, including text, images, audio, and video. Through learning and training on massive amounts of data, they master the rules and patterns of content generation, thereby achieving automated, high-quality content creation.
[0033] Wensheng SQL (Structured Query Language) Big Model: This refers to a model based on big language modeling technology that can convert natural language into SQL query statements. It aims to allow non-technical users to easily query databases using natural language, without needing to master complex SQL syntax, thereby improving the ease of data access and the efficiency of data analysis and maintenance.
[0034] To make the above-mentioned objectives, features and advantages of the embodiments of this application more apparent and understandable, the technical solutions of the embodiments of this application will be further described in detail below with reference to the accompanying drawings.
[0035] It should be noted that the sequence number of each step in the embodiments of this application does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0036] Please see Figure 1 The above is a flowchart illustrating a method for generating operation and maintenance service reports based on large-scale model scheduling and orchestration, as provided in an embodiment of this application. Figure 1 As shown, the method for generating operation and maintenance service reports based on large-scale model scheduling and orchestration may include the following steps:
[0037] Step S100: Obtain the natural language requirement description used to generate the operation and maintenance service report.
[0038] In this embodiment of the application, relevant personnel can trigger the system to generate operation and maintenance service reports by inputting a natural language description describing the requirements for generating operation and maintenance service reports into the operation and maintenance service report generation system based on large model scheduling and orchestration (hereinafter referred to as the system).
[0039] The system can obtain a natural language requirement description for generating an operation and maintenance service report, so as to determine the generation requirements of the operation and maintenance service report based on the obtained natural language requirement description, and perform operation and maintenance service report generation processing according to the generation requirements.
[0040] Step S110: Using the intent analysis large model components, perform intent analysis on the obtained natural language demand description to obtain the intent analysis results.
[0041] In this embodiment of the application, in order to achieve more flexible large model scheduling, large model components can be encapsulated based on large models with different capabilities.
[0042] Among them, through the large model component, the processing capabilities of the associated large model (such as the AIGC large model) can be invoked to realize some functional processing in the process of generating operation and maintenance service reports.
[0043] For example, any large model component encapsulates component identification information (used to uniquely identify the large model component), component description information, package information, and parameter configuration information.
[0044] The component description information can be used to describe the functionality of large model components.
[0045] For example, the functions of large model components may include, but are not limited to, code generation, intent analysis, question classification, or parameter extraction.
[0046] The package can be used to run and invoke the functionality of the associated large model components.
[0047] The parameter configuration information can include the large model information associated with the large model components and the prompt words corresponding to the problem scenario.
[0048] For example, for any large model component, the associated large model can be called by running the package encapsulated in the large model component, based on the large model information included in the parameter configuration information, and the associated large model can be used for corresponding processing based on the prompt words included in the parameter configuration information.
[0049] In one example, for any type of large model element, different problem scenarios correspond to different large model elements.
[0050] In another example, different problem scenarios correspond to the same large model element, which encapsulates the prompt word projects corresponding to different problem scenarios.
[0051] In this embodiment of the application, the obtained natural language requirement description can be analyzed using intent analysis large model components to obtain intent analysis results.
[0052] For example, for intent analysis large model components, the corresponding prompt word engineering is used to prompt the large model to perform intent analysis on natural language demand descriptions.
[0053] By utilizing large-scale intent analysis components, associated large-scale models can be invoked to perform intent analysis on natural language demand descriptions.
[0054] Step S120: Based on the intent analysis results, use the problem classification model components to determine the corresponding target problem scenario.
[0055] In this embodiment of the application, after obtaining the intent analysis result, the problem scenario (which can be called the target problem scenario) corresponding to the intent analysis result can be determined by using the problem classification large model components.
[0056] For example, for a large model component for problem classification, the corresponding prompt word engineering is used to prompt the large model to classify problems, and includes candidate problem scenario options.
[0057] By utilizing the large problem classification model components, associated large models can be invoked, and the target problem scenario can be determined from the candidate problem scenario options based on the intent analysis results.
[0058] Step S130: Based on the target problem scenario, use the parameter extraction large model components to extract operation and maintenance parameters from the natural language requirement description to obtain the target operation and maintenance parameters.
[0059] In this embodiment of the application, the operation and maintenance parameters may include, but are not limited to, time, region, device type, indicators, scenarios, and some or all of the data in the report.
[0060] For example, the types of operation and maintenance parameters required to generate operation and maintenance service reports can be different or the same (i.e., different types are allowed) in different problem scenarios.
[0061] Once the target problem scenario is identified, the parameter extraction large model components can be used to further parse the natural language requirement description, extract the operation and maintenance parameters contained therein, and obtain the operation and maintenance parameters used to generate this operation and maintenance service report (which can be called target operation and maintenance parameters).
[0062] In one example, based on the target problem scenario, the operation and maintenance parameters are extracted from the natural language requirement description using large model components to obtain the target operation and maintenance parameters, which may include:
[0063] Based on the target problem scenario, determine the target parameter extraction large model components corresponding to the target problem scenario; among them, the target parameter extraction large model components encapsulate the prompt word project that matches the target problem scenario;
[0064] By using parameter extraction of large model components, operation and maintenance parameters are extracted from natural language requirement descriptions to obtain target operation and maintenance parameters.
[0065] For example, corresponding parameters can be pre-encapsulated to extract large model components for different problem scenarios.
[0066] Each parameter extraction module contains a prompt word project that matches the corresponding problem scenario within its large model components.
[0067] It should be noted that when new problem scenarios emerge, new parameter extraction large model components can be encapsulated for parameter extraction in response to the new problem scenarios.
[0068] Accordingly, once the target problem scenario is determined, the parameter extraction large model components corresponding to the target problem scenario can be determined (which can be called the target parameter extraction large model proximity). Using the target parameter extraction large model components, based on the encapsulated prompt word project that matches the target problem scenario, the obtained natural language requirement description is used to extract operation and maintenance parameters to obtain the target operation and maintenance parameters.
[0069] In another example, based on the target problem scenario, the parameter extraction large model components are used to extract operation and maintenance parameters from the natural language requirement description to obtain the target operation and maintenance parameters, which may include:
[0070] Based on the target problem scenario, the corresponding target prompt word project is determined from the prompt word project of the large model component encapsulation extracted from the parameters;
[0071] Based on the target prompt word project, the operation and maintenance parameters are extracted from the natural language requirement description by using the parameter extraction large model components to obtain the target operation and maintenance parameters.
[0072] For example, different problem scenarios can use a unified parameter extraction method to extract large model components for operation and maintenance parameters.
[0073] Among them, the parameter extracts large model components encapsulated with prompt word projects corresponding to different problem scenarios.
[0074] It should be noted that if a new problem scenario is added, a prompt word project corresponding to the new problem scenario can be added to the prompt word project for extracting large model components using this parameter.
[0075] Accordingly, once the target problem scenario is determined, the corresponding prompt word project (which can be called the target prompt word project) can be determined from the prompt word project encapsulated by the parameter extraction large model components. Based on the target prompt word project, the operation and maintenance parameters are extracted from the natural language requirement description using the parameter extraction large model components to obtain the target operation and maintenance parameters.
[0076] Step S140: Generate the corresponding target operation and maintenance parameter instruction set based on the target operation and maintenance parameters, and obtain the target operation and maintenance data using the corresponding target data element based on the target operation and maintenance parameter instruction set.
[0077] In this embodiment of the application, in order to eliminate semantic ambiguity, unify technical cooperation standards, and adapt to machine processing logic, the determined target operation and maintenance parameters can be used to generate a corresponding operation and maintenance parameter instruction set (which can be called the target operation and maintenance parameter instruction set).
[0078] For example, operational parameters can be translated and combined into a set of operational parameter instructions.
[0079] For example, the corresponding operation and maintenance parameter instruction set for "encoder device" can be "encoderdevice"; the corresponding operation and maintenance parameter instruction set for "online status" can be "onlinestatus".
[0080] In this embodiment of the application, the corresponding operation and maintenance data (which can be called target operation and maintenance data) can be obtained by using the corresponding data element (which can be called target) based on the target operation and maintenance parameter instruction set.
[0081] For example, if the operation and maintenance parameter instruction set includes "encoderdevice" and "onlinestatus", then the corresponding target operation and maintenance parameter data can include online status details of the encoding device (including the online status of different encoding devices).
[0082] For example, the data element includes data elements corresponding to different data sources, used to acquire and transmit data from the corresponding data sources respectively.
[0083] The data sources may include, but are not limited to, interfaces, databases, caches, or message queues.
[0084] For example, data elements may include HTTP (Hypertext Transfer Protocol) interface elements for acquiring and transmitting data that conforms to the HTTP interface specification.
[0085] Data elements may include PG database (an open-source relational database) elements that can be used to acquire and pass through data that conforms to the PG database specification.
[0086] The Redis (Remote Dictionary Server) component can be used to retrieve and pass through data that conforms to the Redis caching specification.
[0087] Kafka (a distributed stream processing platform) components can be used to retrieve and pass through data that conforms to the Kafka message queue specification.
[0088] It should be noted that, in the embodiments of this application, in addition to acquiring and transmitting data, the data element can also perform some general conversions and other processing on the acquired data through scripts or other means. The embodiments of this application do not limit this.
[0089] Step S150: Based on the target problem scenario, use the code big model components to generate target code for the target operation and maintenance data, and generate a visual operation and maintenance service report page based on the target code.
[0090] In this embodiment of the application, when the target operation and maintenance data is obtained, an operation and maintenance service report can be generated based on the obtained target operation and maintenance data.
[0091] For example, code-based large model components can be used to generate code (which can be called code) for target operation and maintenance data. For instance, HTML (HyperText Markup Language) code (such as H5 code) can be generated, and page rendering can be performed based on the target code to generate a visual operation and maintenance service report.
[0092] For example, in the process of generating target code using code big model components, the corresponding prompt word project can be determined according to the target problem scenario, and the big model associated with the code big model component can be called according to the prompt word project to generate code for the target operation and maintenance data.
[0093] It can be seen that, in Figure 1 In the illustrated method flow, a natural language requirement description for generating an operations and maintenance service report is obtained. Intent analysis is performed on the obtained natural language requirement description using intent analysis large model components. Based on the intent analysis results, a problem classification large model component is used to determine the corresponding target problem scenario. Based on the target problem scenario, an operation and maintenance parameter extraction large model component is used to extract operation and maintenance parameters from the natural language requirement description. The extracted target operation and maintenance parameters are used to generate a corresponding target operation and maintenance parameter instruction set. Based on the target operation and maintenance parameter instruction set, a target operation and maintenance data is obtained using the corresponding target tree component. Then, based on the target problem scenario, a target code can be generated using a code large model component for the target operation and maintenance data. A visual operation and maintenance service report page is generated based on the target code. By encapsulating the capabilities of the large model into large model components, the flexibility of large model function calls is improved. Different large model components can be scheduled and orchestrated according to actual needs to build rule chains. The operation and maintenance service report is generated based on the constructed rule chains, improving the accuracy and efficiency of operation and maintenance service report generation.
[0094] In some embodiments, such as Figure 2As shown, based on the target problem scenario, the above-mentioned method uses large model components for parameter extraction to extract operational parameters from the natural language requirement description. After obtaining the target operational parameters, it may also include:
[0095] Based on the parameter types corresponding to the target problem scenario, a completeness analysis of the target operation and maintenance parameters is performed;
[0096] If the target operation and maintenance parameters are found to be incomplete, a prompt message will be output to remind relevant personnel to supplement the requirements.
[0097] Supplement natural language descriptions based on input requirements, extract large model components using parameters to determine supplementary operation and maintenance parameters, and update target operation and maintenance parameters based on supplementary operation and maintenance parameters;
[0098] The above-mentioned set of instructions for generating corresponding target operation and maintenance parameters based on target operation and maintenance parameters may include:
[0099] If the current target operation and maintenance parameters are complete, generate the corresponding target operation and maintenance parameter instruction set based on the current target operation and maintenance parameters.
[0100] For example, considering that in real-world scenarios, the natural language description of requirements given by relevant personnel in a single instance may be incomplete and may contain missing key information, which may result in incomplete operation and maintenance parameters and affect the generation of operation and maintenance service reports.
[0101] Accordingly, once the target operation and maintenance parameters have been extracted in the manner described above, an integrity analysis can be performed on the current target operation and maintenance parameters.
[0102] For example, based on the parameter type corresponding to the target problem scenario (which can be called the target parameter type), it can be determined whether the current target operation and maintenance parameters include operation and maintenance parameters of all target parameter types, so as to achieve a completeness analysis of the target operation and maintenance parameters.
[0103] Specifically, for any target parameter type, if the current target operation and maintenance parameters do not contain operation and maintenance parameters of that type, the extracted operation and maintenance parameter of that type is empty, or the extracted operation and maintenance parameter of that type is an abnormal parameter value, then it can be determined that the current target operation and maintenance parameters are incomplete.
[0104] For example, if it is determined that the current target operation and maintenance parameters are incomplete, a prompt message can be output to remind relevant personnel to supplement the requirements.
[0105] For example, the initial natural language requirement description is "generate an online rate report", but for the generation of the online rate maintenance service report, the device type also needs to be specified. The prompt message could be "Please specify the online rate of what type of device".
[0106] When the user receives a natural language description of the required response prompt, the user can use parameter extraction to determine the supplementary operation and maintenance parameters for large model components. The user can then update the target operation and maintenance parameters based on the supplementary parameters and perform an integrity analysis on the current target operation and maintenance parameters again.
[0107] If the current target operation and maintenance parameters are still incomplete, they can be supplemented again.
[0108] If the current target operation and maintenance parameters are complete, the corresponding target operation and maintenance parameter instruction set can be generated based on the current target operation and maintenance parameters.
[0109] In some embodiments, the above-mentioned generation of target code based on the target problem scenario and utilizing code large model components for target operation and maintenance data may include:
[0110] Based on the target problem scenario, the target operation and maintenance data is statistically analyzed using the corresponding target operation and maintenance components to obtain the operation and maintenance statistics results;
[0111] Based on the target problem scenario, target code is generated using code large model components based on operation and maintenance statistics results.
[0112] For example, in order to implement operation and maintenance statistics, different types of operation and maintenance components can be built as the smallest business processing unit for generating operation and maintenance service reports, thereby improving the accuracy of generation.
[0113] For example, operation and maintenance components may include, but are not limited to, organizational percentage components, organizational ranking components, and operation and maintenance detail components.
[0114] Among them, "organization" refers to a security management entity or institution with hierarchical management.
[0115] Taking the security field as an example, the organization tree in the security field refers to the core data structure that hierarchically models physical security equipment.
[0116] Under the national standard, organizations can correspond to different levels of administrative divisions, such as province-city-district.
[0117] For example, organizations can also be divided into other hierarchical structures, such as company-branch; company-department; park-office building, etc.
[0118] Taking the organization percentage element as an example, the organization percentage element can be used to calculate the equipment percentage of a certain indicator for a certain equipment type in different organizations based on the number of devices of a certain indicator for different equipment types obtained by the data element according to the organization, as well as the total number of devices of that equipment type.
[0119] For example, assuming the device type is a camera and the metric is online, the online rate of cameras in different organizations can be determined by using the organization percentage element based on the number of online cameras in different organizations and the total number of cameras obtained from the data element.
[0120] For example, for target operation and maintenance data obtained through target data elements, the required target operation and maintenance elements can be determined based on the target problem scenario.
[0121] For example, if the target problem scenario requires percentage statistics, the target operation and maintenance components may include organizational percentage components.
[0122] For example, if the target problem scenario requires ranking statistics, the target operation and maintenance components may include organizational ranking components.
[0123] For example, the target operation and maintenance component corresponding to the target problem scenario may include one or more operation and maintenance components.
[0124] For example, if the target problem scenario requires a percentage ranking, the target operation and maintenance components may include organizational percentage components and organizational ranking components.
[0125] For example, the target maintenance component can be used to perform maintenance statistics on the target maintenance data to obtain maintenance statistics results.
[0126] In one example, the above-mentioned use of the corresponding target operation and maintenance components to perform operation and maintenance statistics on target operation and maintenance data may include:
[0127] The target operation and maintenance data obtained from multiple different target data elements are used to perform operation and maintenance statistics.
[0128] For example, for any target operation and maintenance element, the target operation and maintenance data obtained through multiple different target data elements can be used to perform operation and maintenance statistics. Thus, in the process of operation and maintenance statistics, operation and maintenance statistics across sources, across tables, and in complex operation and maintenance scenarios can be achieved.
[0129] In some embodiments, the above-mentioned generation of target code based on the target problem scenario and utilizing code large model components for target operation and maintenance data may include:
[0130] Based on the target operation and maintenance data cached in the cache element, operation and maintenance statistics are performed using the statistical large model element to obtain the operation and maintenance statistics results;
[0131] Based on the target problem scenario, target code is generated using code large model components based on operation and maintenance statistics results.
[0132] For example, large models can be used to perform statistical processing of operational data.
[0133] Correspondingly, statistical large model components can be pre-encapsulated and cache components can be set up. The cache components can be used to cache target operation and maintenance data, and the statistical large model components can be used to perform operation and maintenance statistical processing on the target operation and maintenance data cached in the cache components.
[0134] It should be noted that since large model components process data by calling associated large models, and the data processing of large models has limitations on the context length, in practical applications, it can be determined whether to perform operation and maintenance statistical processing through operation and maintenance components or through statistical large model components, depending on the size of the target operation and maintenance data to be processed.
[0135] In one example, if the amount of target maintenance data is less than a preset data volume threshold, a large statistical model can be used to perform maintenance statistics on the target maintenance data; if the amount of target maintenance data is greater than or equal to the preset data volume threshold, the target maintenance components can be used to perform maintenance statistics on the target maintenance data.
[0136] In some embodiments, after generating the visual operation and maintenance service report page based on the target code, the following may also be included:
[0137] Export the operation and maintenance components from the report, and export a snapshot PDF of the visual operation and maintenance service report page.
[0138] For example, after generating a visual operation and maintenance service report page in the manner described above, a snapshot PDF of the visual operation and maintenance report page can also be exported.
[0139] In some embodiments, after generating the visual operation and maintenance service report page based on the target code, the following may also be included:
[0140] Export maintenance components and detailed maintenance data using the details.
[0141] For example, after generating a visual operation and maintenance service report page in the manner described above, the operation and maintenance details data can also be exported, such as exporting operation and maintenance details CSV (Comma-Separated Values) data.
[0142] It should be noted that, in this embodiment of the application, when an operation and maintenance service report is generated, analysis suggestions can also be given based on the large inference model.
[0143] To enable those skilled in the art to better understand the technical solutions provided in the embodiments of this application, the technical solutions provided in the embodiments of this application are described below with reference to specific examples.
[0144] In this embodiment, an intelligent operation and maintenance service reporting system based on large model scheduling and orchestration is proposed to solve the problems of low accuracy and high operation and maintenance cost in the field of operation and maintenance.
[0145] Considering that the traditional method of generating operation and maintenance service reports through the Wensheng SQL large model has low accuracy and does not support complex scenarios such as cross-data source and cross-table multi-table join queries, it is necessary to consider that the traditional method of generating operation and maintenance service reports through the Wensheng SQL large model has low accuracy and does not support complex scenarios such as cross-data source and cross-table multi-table join queries.
[0146] Based on this, the embodiments of this application improve the accuracy of operation and maintenance service report generation while ensuring the efficiency of operation and maintenance service report generation by intelligently arranging components such as large model components, operation and maintenance components, and data components.
[0147] In this embodiment, an intelligent orchestration engine is built, which encapsulates the functionality of the large model into large model components and references them in the rule chain.
[0148] The rule chain is used to schedule and orchestrate large model components with different capabilities and parameters, such as language large model components, code large model components, intent analysis large model components, parameter extraction large model components, etc.
[0149] For example, any large model component encapsulates component identification information, component description information, package, and parameter configuration information.
[0150] For example, various types of operation and maintenance components can be built to define the minimum business processing unit for generating operation and maintenance service reports, thereby improving the accuracy of generation. These can include components such as organizational percentage, organizational ranking, and operation and maintenance details. Through these operation and maintenance components, the system can be dynamically expanded to adapt to different operation and maintenance business scenarios and solve personalized problems in complex scenarios. Compared to repeatedly developing a complete operation and maintenance report page, the cost is greatly reduced.
[0151] For example, the implementation details of some operation and maintenance components can be shown in Table 1:
[0152] Table 1
[0153]
[0154]
[0155]
[0156] It should be noted that the "equipment type" and "indicator" in Table 1 above can refer to a single equipment type or indicator, or they can be sets of data. For example, a "equipment type" can include a set of equipment types (including multiple equipment types).
[0157] For example, various types of data elements can be built to match with operation and maintenance elements to handle various operation and maintenance service report scenarios, providing accurate data collection and processing capabilities, such as interface elements, database elements, etc., which, together with operation and maintenance elements, can solve cross-source, cross-table, and complex operation and maintenance report scenarios, and provide convenient scalable element capabilities.
[0158] For example, the implementation details of some data elements are shown in Table 2:
[0159] Table 2
[0160]
[0161]
[0162] It should be noted that, in the embodiments of this application, the main functions of the data element include data acquisition and transparent transmission. In addition, the data element can also be used to perform some general data processing (such as data conversion) on the acquired data. The data element can perform data conversion and other processing on the acquired data through scripts or other means, which is not limited in this embodiment of the application.
[0163] In this embodiment, a business rule chain can be formed by connecting large model components, operation and maintenance components, and data components based on an intelligent orchestration engine, which can be used to generate intelligent operation and maintenance service reports.
[0164] The following describes the implementation process for generating intelligent operation and maintenance service reports.
[0165] like Figure 3 As shown, in this embodiment, generating an intelligent operation and maintenance service report may include the following steps:
[0166] S1. Obtain the operation and maintenance service report in natural language (i.e., the natural language requirement description above).
[0167] S2. Utilize the large-scale intent analysis model components to perform report requirement intent analysis and obtain the intent analysis results.
[0168] S3. Based on the intent analysis results, use the components of the problem classification model to determine the corresponding problem scenario (i.e., the target problem scenario).
[0169] For example, problem scenarios may include, but are not limited to, operational statistics, comprehensive statistics, service reports, or work order statistics.
[0170] For example, different problem scenarios can correspond to different rule chain branches. Among them, the problem classifier includes problem scenarios such as operation statistics, comprehensive statistics, service reports, and work order statistics, and is scalable.
[0171] S4. Based on the target problem scenario, extract large model components using parameters, extract operation and maintenance parameters from the natural language requirement description, and obtain the target operation and maintenance parameters.
[0172] In the example, parameters can be used to extract large model components for further analysis of user intent, and natural language can be used for induction and understanding to extract and convert them into operation and maintenance parameters.
[0173] For example, operation and maintenance parameters may include: T time, A region, B device type, C indicator, D scenario, and E report.
[0174] Among them, T, A, B, C, D, and E can be parameter sets (that is, any type of operation and maintenance parameter can include multiple sets).
[0175] For example, suppose the natural language requirement is described as "Help me generate a regional ranking report of the VQD anomalies, offline duration, and upgrade failures of all GNSS (Global Navigation Satellite System) displacement monitors, hyperspectral water quality monitors, and other devices in Zhejiang Province for the past month," then T represents the past month; A represents Zhejiang Province; B includes GNSS displacement monitors and hyperspectral water quality monitors; C includes VQD anomalies, offline duration, and upgrade failures; D includes regional percentages and percentage rankings; and E is the report.
[0176] For example, for the obtained target operation and maintenance parameters, an integrity analysis can be performed to determine whether there are any types of parameters that have not been extracted, the extracted parameter values are empty, or the extracted parameters are abnormal parameter values, etc.
[0177] For example, if it is determined that the current target operation and maintenance parameters are incomplete, a prompt message can be output to allow the target operation and maintenance parameters to be supplemented.
[0178] If the current target operation and maintenance parameters are complete, an operation and maintenance parameter instruction set can be constructed based on the current target operation and maintenance parameters.
[0179] S5. Obtain the operation and maintenance parameter instruction set, and have the operation and maintenance report controller element schedule and execute the specific instruction set.
[0180] S6. Under the scheduling of the operation and maintenance report controller element, the target operation and maintenance data is obtained by using the corresponding data element according to the operation and maintenance parameter instruction set.
[0181] For example, a database component can be used to connect to an operations and maintenance (O&M) metric library, or an interface component can be used to connect to an O&M OpenAPI interface to obtain detailed datasets of corresponding O&M parameters. For instance, a detailed dataset of metric C for region A, device type B.
[0182] S7. Based on the target maintenance data, use the corresponding target maintenance components to perform maintenance statistics on the target maintenance data.
[0183] For example, based on the detailed dataset of specified indicators for specified equipment types in different organizations, the organization share element can be used to determine the equipment share of specified indicators for specified equipment types in different organizations, and the organization ranking element can be used to determine the ranking of the equipment share of specified indicators for specified equipment types in different organizations.
[0184] For example, through rule chain orchestration, operation and maintenance components can be connected with corresponding large model components and data components to generate operation and maintenance service reports.
[0185] S8. Import the operation and maintenance processing results of the operation and maintenance components into the code big model component through the operation and maintenance report controller component to obtain the corresponding target code (such as H5 code).
[0186] S9. Render the H5 code on the platform's web page and display a visual report of the page to the user.
[0187] S10. Export maintenance components via report, export H5 page snapshot PDF; or, export maintenance components via details, export maintenance detail CSV data.
[0188] The rule chain will be explained below with specific examples.
[0189] 1. Begin.
[0190] 2. User question: Please generate a report page showing the online rate of encoding devices.
[0191] 3. Obtain users' natural language requirements.
[0192] 4. Identify user report needs by separating intent from large model components, and identify the issue scenario with the highest similarity to the need as the organizational percentage issue (i.e., the organizational percentage maintenance component can solve the issue) by classifying the large model components for problem categories, and then enter that branch link.
[0193] 5. Extract large model components using parameters and identify operation and maintenance parameters; among them, the equipment type is coded equipment, and the indicators are online status and online rate.
[0194] For example, if the operation and maintenance parameters are incomplete, such as when a user asks "generate an online rate report", the large model components can be extracted by the parameters, and the user can be further asked "please provide the online rate of what type of device".
[0195] 6. Convert the extracted operation and maintenance parameters into an operation and maintenance parameter instruction set. For example, convert the encoding device to encoderdevice and the online status to onlinestatus.
[0196] 7. The operation and maintenance report controller element connects the corresponding HTTP interface data element, organization proportion operation and maintenance element, etc., based on the operation and maintenance parameter instruction set.
[0197] 8. Based on the operation and maintenance parameter instruction set, call the HTTP interface data element to call the network management service's interface for "querying the number of online (i.e., the number of devices in online status) and the total number of coded devices under different organizations".
[0198] For example, one can query a "detailed list of online status of coding devices under different organizations", which contains each device and its online status. Then, the number of online coding devices under each organization can be calculated by summing the data using the maintenance elements of an organization.
[0199] 9. Utilize organizational proportion maintenance components to determine the online rate of coding equipment under different organizations.
[0200] 10. Based on the online rate of coding devices under different organizations, generate report H5 code using code large model components.
[0201] 11. Use the report generation element to render the H5 code generated by the large model element and output a visual report page.
[0202] 12. Using the report export component, export the report page to PDF, and finally obtain an operation and maintenance service report showing the distribution of the online rate of coding devices in each organization's region.
[0203] For example, the generated operation and maintenance service report can be combined with a large inference model to provide some analytical suggestions.
[0204] 13. End.
[0205] The method provided in this application has been described above. The apparatus provided in this application is described below:
[0206] Please see Figure 4 The image shows a schematic diagram of a maintenance service report generation device based on a large model scheduling and orchestration, as provided in an embodiment of this application. Figure 4 As shown, the operation and maintenance service report generation device based on large model scheduling and orchestration may include:
[0207] The acquisition unit is configured to acquire a natural language requirement description used to generate an operation and maintenance service report;
[0208] The controller unit is configured to perform intent analysis on the natural language demand description using intent analysis big model elements to obtain intent analysis results;
[0209] The controller unit is also configured to determine the corresponding target problem scenario based on the intent analysis results and using the problem classification big model components;
[0210] The controller unit is also configured to extract operation and maintenance parameters from the natural language requirement description based on the target problem scenario and the parameter extraction large model component to obtain the target operation and maintenance parameters.
[0211] The controller unit is also configured to generate a corresponding target operation and maintenance parameter instruction set based on the target operation and maintenance parameters, and to acquire target operation and maintenance data using the corresponding target data element based on the target operation and maintenance parameter instruction set.
[0212] The controller unit is also configured to generate target code based on the target problem scenario, using code big model components to generate target code for the target operation and maintenance data, and generate a visual operation and maintenance service report page based on the target code.
[0213] The specific implementation process of generating operation and maintenance service reports based on large model scheduling and orchestration in each unit of the operation and maintenance service report generation device can be found in the relevant descriptions in the above method embodiments.
[0214] Please see Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. The electronic device may include a processor 501, a communication interface 502, a memory 503, and a communication bus 504. The processor 501, communication interface 502, and memory 503 communicate with each other via the communication bus 504. The memory 503 stores a computer program; the processor 501 can execute the program stored in the memory 503 to perform the operation and maintenance service report generation method based on large-model scheduling and orchestration described above.
[0215] The memory 503 mentioned in this document can be any electronic, magnetic, optical, or other physical storage device that can contain or store information, such as executable instructions, data, etc. For example, memory 503 can be: RAM (Random Access Memory), volatile memory, non-volatile memory, flash memory, storage drive (such as hard disk drive), solid-state drive, any type of storage disk (such as optical disc, DVD, etc.), or similar storage media, or combinations thereof.
[0216] This application also provides a non-transitory machine-readable storage medium storing a computer program, such as... Figure 5 The memory 503 in the memory, the computer program can be generated by Figure 5 The processor 501 in the electronic device shown executes to implement the operation and maintenance service report generation method based on large model scheduling and orchestration described above.
[0217] This application also provides a computer program stored on a non-transitory machine-readable storage medium, such as... Figure 5 The memory 503 in the processor, and when the processor executes the computer program, it causes the processor 501 to execute the operation and maintenance service report generation method based on large model scheduling and orchestration described above.
Claims
1. A method for generating operation and maintenance service reports based on large-scale model scheduling and orchestration, characterized in that, include: Obtain a natural language requirement description for generating operation and maintenance service reports; Using large-scale intent analysis model components, intent analysis is performed on the natural language demand description to obtain intent analysis results; Based on the intent analysis results, the corresponding target problem scenario is determined using the components of the problem classification model. Based on the target problem scenario, the operation and maintenance parameters are extracted from the natural language requirement description using the parameter extraction large model components to obtain the target operation and maintenance parameters; Based on the target operation and maintenance parameters, a corresponding target operation and maintenance parameter instruction set is generated, and based on the target operation and maintenance parameter instruction set, the target operation and maintenance data is obtained using the corresponding target data element; Based on the target problem scenario, target code is generated using code large model components for the target operation and maintenance data, and a visual operation and maintenance service report page is generated based on the target code.
2. The method according to claim 1, characterized in that, After extracting operation and maintenance parameters from the natural language requirement description based on the target problem scenario using large model components to obtain the target operation and maintenance parameters, the process further includes: Based on the parameter types corresponding to the target problem scenario, a completeness analysis is performed on the target operation and maintenance parameters; If the target maintenance parameters are determined to be incomplete, a prompt message is output, which is used to prompt relevant personnel to supplement the requirements. Supplement the input requirements with natural language descriptions, extract large model components using parameters to determine the supplemented operation and maintenance parameters, and update the target operation and maintenance parameters based on the supplemented operation and maintenance parameters; The step of generating a corresponding target operation and maintenance parameter instruction set based on the target operation and maintenance parameters includes: If the current target operation and maintenance parameters are complete, generate the corresponding target operation and maintenance parameter instruction set based on the current target operation and maintenance parameters.
3. The method according to claim 1, characterized in that, Based on the target problem scenario, the step of generating target code for the target operation and maintenance data using large code model components includes: Based on the target problem scenario, the target operation and maintenance data is statistically analyzed using the corresponding target operation and maintenance components to obtain the operation and maintenance statistical results; Based on the target problem scenario, target code is generated using code large model components based on the operation and maintenance statistics results.
4. The method according to claim 3, characterized in that, The step of performing maintenance statistics on the target maintenance data using the corresponding target maintenance components includes: The target operation and maintenance data obtained from multiple different target data elements are used to perform operation and maintenance statistics.
5. The method according to claim 1, characterized in that, Based on the target problem scenario, the step of generating target code for the target operation and maintenance data using large code model components includes: Based on the target operation and maintenance data cached in the cache element, operation and maintenance statistics are performed using the statistical large model element to obtain the operation and maintenance statistics results; Based on the target problem scenario, target code is generated using code large model components based on the operation and maintenance statistics results.
6. The method according to claim 1, characterized in that, After generating the visual operation and maintenance service report page based on the target code, the method further includes: Export the operation and maintenance component using the report, and export a snapshot PDF of the visualized operation and maintenance service report page; And / or, Export maintenance components and detailed maintenance data using the details.
7. The method according to claim 1, characterized in that, For any large model component, it encapsulates component identification information, component description information, program package, and parameter configuration information. Among them, the component description information is used to describe the function of the large model component, the program package is used to run and call the associated large model to implement the function of the large model component, and the parameter configuration information includes the large model information associated with the large model component and the prompt word project corresponding to the problem scenario. Among them, for any type of large model component, different problem scenarios correspond to different large model components, or different problem scenarios correspond to the same large model component, and the large model component encapsulates prompt word projects corresponding to different problem scenarios; And / or, The data element includes data elements corresponding to different data sources, used to acquire and transmit data from the corresponding data sources respectively; wherein, the data sources include interfaces, databases, caches or message queues.
8. A device for generating operation and maintenance service reports based on large-scale model scheduling and orchestration, characterized in that, include: The acquisition unit is configured to acquire a natural language requirement description used to generate an operation and maintenance service report; The controller unit is configured to perform intent analysis on the natural language demand description using intent analysis large model elements to obtain intent analysis results; The controller unit is further configured to determine the corresponding target problem scenario based on the intent analysis results and using the problem classification big model elements; The controller unit is further configured to extract operation and maintenance parameters from the natural language requirement description based on the target problem scenario using a parameter extraction large model component, thereby obtaining the target operation and maintenance parameters. The controller unit is also configured to generate a corresponding target operation and maintenance parameter instruction set based on the target operation and maintenance parameters, and to acquire target operation and maintenance data using the corresponding target data element based on the target operation and maintenance parameter instruction set. The controller unit is also configured to generate target code for the target operation and maintenance data using code big model components based on the target problem scenario, and generate a visual operation and maintenance service report page based on the target code.
9. An electronic device, characterized in that, include: The processor, communication interface, memory, and communication bus are connected, with the processor, communication interface, and memory communicating with each other via the communication bus. Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the method as described in any one of claims 1 to 7.