Operation and maintenance report generation method, device and equipment

Abstract

The application provides an operation and maintenance report generation method, device and equipment. The method comprises the following steps: in response to a report generation request, accessing a knowledge base to obtain a corresponding target report template, wherein the target report template comprises template framework content and a plurality of placeholder tags; extracting each placeholder tag in the target report template, and determining model auxiliary information and a large language model processing procedure associated with each placeholder tag; obtaining project information corresponding to each placeholder tag; for each placeholder tag, executing the large language model processing procedure based on the project information and the model auxiliary information to obtain report content corresponding to the placeholder tag; and integrating the report content corresponding to each placeholder tag into the template framework content to obtain a target operation and maintenance report. The application can generate an operation and maintenance report in a segmented manner based on a report template and a large language model, thereby avoiding uncertain problems caused by “model hallucination” when a large model directly generates a report, and improving the accuracy of operation and maintenance report generation.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, specifically to a method, apparatus, and equipment for generating operation and maintenance reports. Background Technology

[0002] With the widespread use of computer systems, operations and maintenance (O&M) plays a crucial role in ensuring system operation. O&M reports, produced by O&M personnel, are an important carrier of O&M work and directly relate to system stability and data security. Currently, the traditional O&M report writing process heavily relies on manual labor, which is not only inefficient but also affected by the varying skill levels of O&M personnel.

[0003] To overcome the inefficiencies and inconsistent skill levels of manual operation and maintenance (O&M) report generation, a solution based on large language models has emerged. Currently, the traditional process of generating O&M reports using large language models involves O&M personnel directly inputting relevant prompts and data into the large language model, which then directly outputs the corresponding O&M report.

[0004] However, when traditional large language models generate operation and maintenance reports, they directly generate the full text of the operation and maintenance report using prompt words and data. This can easily lead to uncertainty caused by the "illusion" of the large language model, which in turn causes the generated operation and maintenance report to be biased. Summary of the Invention

[0005] In view of this, this application aims to provide a method, apparatus and equipment for generating operation and maintenance reports, so as to solve the problem that the uncertainty caused by the "illusion" of the large language model in traditional related technologies can easily lead to deviations in the generated operation and maintenance reports.

[0006] The first aspect of this application provides a method for generating operation and maintenance reports, including: In response to a report generation request, the knowledge base is accessed to obtain the corresponding target report template, which includes template framework content and multiple placeholder tags; Extract each placeholder label from the target report template and determine the model assistance information and large language model processing flow associated with each placeholder label; Retrieve the project information corresponding to each placeholder label; For each placeholder label, the large language model processing flow is executed based on the project information and the model auxiliary information to obtain the report content corresponding to the placeholder label; The report content corresponding to each placeholder label is integrated into the template framework content to obtain the target operation and maintenance report.

[0007] In one possible implementation of this application, the model assistance information includes preset prompt words; correspondingly, the step of executing the large language model processing flow based on the project information and the model assistance information to obtain the report content corresponding to the placeholder label includes: inputting the project information and the preset prompt words into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label.

[0008] In one possible implementation of this application, the placeholder label includes a text-based placeholder label; the preset prompt word includes a first prompt word related to the generated text content; the project information includes project text content; correspondingly, the step of inputting the project information and the preset prompt word into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label includes: inputting the project text content and the first prompt word into the large language model agent corresponding to the large language model processing flow to obtain the text-based report content corresponding to the placeholder label.

[0009] In one possible implementation of this application, the placeholder label includes a data-type placeholder label; the preset prompt word includes a second prompt word related to the generated data content; the project information includes project data content; correspondingly, the step of inputting the project information and the preset prompt word into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label includes: determining one or more sub-processes corresponding to the large language model processing flow based on the second prompt word; inputting the project data content and the second prompt word into the large language model agent according to the one or more sub-processes to obtain one or more data-type report contents corresponding to the placeholder label.

[0010] In one possible implementation of this application, the one or more sub-processes include one or more of a form sub-process, a chart sub-process, and a data summary sub-process; wherein the report content of the data class corresponding to the form sub-process is form content, the report content of the data class corresponding to the chart sub-process is a visual chart, and the report content of the data class corresponding to the data summary sub-process is data summary content.

[0011] In one possible implementation of this application, the project data content is determined by at least one of the following methods: obtaining it by accessing an external service interface; or extracting it from the tag data that comes with the data class placeholder tag.

[0012] In one possible implementation of this application, the step of inputting the project data content and the second prompt word into the large language model agent according to one or more sub-processes includes: detecting the data length of the project data content; if the data length does not exceed a preset threshold, then inputting the project data content and the second prompt word into the large language model agent according to one or more sub-processes; if the data length exceeds the preset threshold, then compressing the project data content into a data table, and inputting the data table and the second prompt word into the large language model agent according to one or more sub-processes.

[0013] In one possible implementation of this application, the placeholder label includes a file-type placeholder label; the preset prompt word includes a third prompt word related to the inserted file content; the project information includes the file content; correspondingly, the step of inputting the project information and the preset prompt word into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label includes: inputting the file content and the third prompt word into the large language model agent corresponding to the large language model processing flow to obtain the file-type report content corresponding to the placeholder label.

[0014] In one possible implementation of this application, the report generation request includes a report generation instruction statement; correspondingly, the step of accessing a knowledge base to obtain a corresponding target report template in response to the report generation request includes: receiving a report generation instruction statement sent by a user terminal; extracting key project parameters from the report generation instruction statement using a large language model; and accessing and querying the knowledge base to obtain the target report template corresponding to the key project parameters.

[0015] In one possible implementation of this application, after accessing the knowledge base and querying the knowledge base to obtain the target report template corresponding to the project key parameters, the method further includes: if the project key parameters have multiple corresponding report templates after querying the knowledge base, then a prompt to the user terminal to further refine the report generation instruction statement is returned through the large language model.

[0016] In one possible implementation of this application, the method further includes: obtaining multiple report templates and project classification information corresponding to each report template; determining key project parameters corresponding to each report template based on the project classification information; and storing the correspondence between the report templates and the key project parameters in the knowledge base.

[0017] In one possible implementation of this application, after obtaining the project information corresponding to each placeholder label, the method further includes: if the content in the model auxiliary information or project information of any placeholder label is empty, then the large language model processing flow is not executed for the placeholder label, and the placeholder label is retained in the target operation and maintenance report.

[0018] The second aspect of this application provides an operation and maintenance report generation device, including: The template acquisition unit is used to access the knowledge base to obtain the corresponding target report template in response to the report generation request, wherein the target report template includes template framework content and multiple placeholder tags; The tag recognition unit is used to extract each placeholder tag in the target report template and determine the model auxiliary information and large language model processing flow associated with each placeholder tag. The information acquisition unit is used to acquire the project information corresponding to each placeholder label; The model processing unit is used to execute the large language model processing flow for each placeholder label based on the project information and the model auxiliary information to obtain the report content corresponding to the placeholder label. The report integration unit is used to integrate the report content corresponding to each placeholder label into the template framework content to obtain the target operation and maintenance report.

[0019] A third aspect of this application provides an electronic device, including a memory and a processor; the memory is connected to the processor and is used to store a program; the processor is used to implement the operation and maintenance report generation method as described in the first aspect of this application or any possible implementation thereof by running the program in the memory.

[0020] The operation and maintenance report generation method, apparatus, and equipment provided in this application determine the corresponding target report template through a report generation request, extract each placeholder tag of the target report template, determine the model auxiliary information and large language model processing flow associated with each placeholder tag, and obtain the project information corresponding to each placeholder tag; for each placeholder tag, execute the large language model processing flow based on the project information and model auxiliary information to obtain the report content corresponding to the placeholder tag; integrate the report content corresponding to each tag into the template framework content to obtain the target operation and maintenance report. The embodiments of this application can generate operation and maintenance reports in segments based on the report template using a large language model, avoiding the uncertainty problem caused by the "model illusion" when directly generating reports from a large model, and improving the accuracy of operation and maintenance report generation. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0022] Figure 1 A schematic diagram of the implementation environment provided for the embodiments of this application.

[0023] Figure 2 A flowchart illustrating the operation and maintenance report generation method provided in this application embodiment. Figure 1 .

[0024] Figure 3 A flowchart illustrating the operation and maintenance report generation method provided in this application embodiment. Figure 2 .

[0025] Figure 4 A flowchart illustrating the operation and maintenance report generation method provided in this application embodiment. Figure 3 .

[0026] Figure 5 This is a schematic diagram of the structure of the operation and maintenance report generation device provided in the embodiments of this application.

[0027] Figure 6 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] Operations and maintenance (O&M) work is crucial for ensuring the operation of computer and other business systems, while O&M inspections are the foundation for ensuring the long-term stability of the system. The O&M reports generated by O&M inspections are a summary of O&M work and can provide O&M personnel with effective data and process support.

[0030] Currently, the generation of operation and maintenance reports mainly relies on personnel recording data. First, frontline operation and maintenance personnel conduct regular inspections according to plan and record the inspection content to form inspection logs. Then, second-line operation and maintenance personnel or experts periodically compile the inspection logs over a period of time to obtain the final operation and maintenance report. However, due to the inconsistent skill levels of frontline operation and maintenance personnel, second-line personnel or experts often need to assist in recording inspection logs. At the same time, second-line operation and maintenance personnel or experts need to integrate a large number of inspection logs in different formats, which requires significant effort and time, resulting in low efficiency. In other words, the current manual compilation of operation and maintenance reports suffers from low efficiency and inconsistent personnel skill levels.

[0031] Currently, to overcome the inefficiencies and inconsistent skill levels of manual operation and maintenance (O&M) report writing, technologies using large language models to intelligently generate O&M reports have been developed. However, traditional O&M report generation using large language models primarily relies on prompts and data to directly generate the full report. This requires the large model to process large amounts of data and prompts at once, which can lead to uncertainties caused by model "illusions," affecting the accuracy of the O&M reports.

[0032] To address the aforementioned technical issues, this application provides a method, apparatus, and device for generating operation and maintenance reports. The method determines a corresponding report template based on a user's report generation request. This report template includes multiple pre-set placeholders. Multiple placeholder tags are extracted from the report template. The model auxiliary information and large language model processing flow associated with each placeholder tag are determined. The large language model processing flow is then applied to the project information and model auxiliary information corresponding to the placeholder tags to obtain the report content corresponding to each tag. Finally, the report content corresponding to each tag is integrated into the final report. This method achieves segmented generation of operation and maintenance reports based on report templates using a large language model, avoiding the uncertainty caused by model illusion when directly generating large models and improving the accuracy of operation and maintenance report generation.

[0033] Exemplary Implementation Environment Please refer to Figure 1 , Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application. The implementation environment includes: a user terminal 101 and a server terminal 102.

[0034] The user terminal 101 can be a terminal device. For example, a terminal device can include a mobile phone, tablet computer, personal computer, or other devices. The user can enter user questions or instructions on the terminal device, and the user's input methods include, but are not limited to, keyboard input, handwriting input, or voice input.

[0035] The server 102 can be a single server or a cluster of multiple servers. A large language model is deployed on server 102, which processes user questions or instructions submitted by client 101 to generate corresponding answers or results.

[0036] Exemplary methods refer to Figure 2 , Figure 2 A flowchart illustrating the operation and maintenance report generation method provided in this application embodiment. Figure 1 In this exemplary embodiment, an operation and maintenance report generation method is provided, which can be applied to the above-mentioned... Figure 1 The server shown is described in detail below: S201: In response to the report generation request, access the knowledge base to obtain the corresponding target report template, which includes template framework content and multiple placeholder tags.

[0037] In the embodiments of this application, the report generation request is generated on the user's end and sent to the server, wherein the report generation request includes a report generation instruction statement entered by the user on the user's end.

[0038] For example, the report generation instruction statement entered by the user on the user terminal could be "Weekly maintenance report for network A of school A in region A needs to be generated".

[0039] Specifically, the system receives report generation instructions sent by the user; extracts key project parameters from the report generation instructions using a large language model; and accesses the knowledge base to query the knowledge base to obtain the target report template corresponding to the key project parameters.

[0040] In this embodiment, the knowledge base stores the correspondence between the key parameters of each user's projects and the report templates.

[0041] In this embodiment, the key parameters of the project include one or more of the following: project information, report type, and report data time range.

[0042] For example, the following key project parameters are extracted from "Weekly Operation and Maintenance Report of Network A of School A in Region A": Project information is "School A in Region A", Report type is "Network A Operation and Maintenance", and Report data time range is "Week".

[0043] For example, based on the project information "A region A school A", the report type "A network operation and maintenance", and the report data time range "week", the project key parameters of each project in the knowledge base and the corresponding relationship of the report template are accessed to determine a report template corresponding to "A region A school A", the report type "A network operation and maintenance", and the report data time range "week", which is then used as the target report template.

[0044] In the embodiments of this application, the target report template is pre-written by the user. Specifically, the target report template includes template framework content written by the user and multiple placeholder tags inserted by the user in the template framework content.

[0045] The template framework content is filled in by operations and maintenance personnel according to different projects. The template framework content refers to the structured skeleton of an operations and maintenance report. It specifies the components, logical order, and core modules of the report, and is the basic framework to ensure that the report content is complete and well-organized.

[0046] Each placeholder label can be an identifier. Optionally, the identifier can be {} or []. Placeholder labels are embedded in different positions within the template frame content, and the position of each identifier is set by the operations and maintenance personnel.

[0047] In one example, the following is a section of template frame content and its associated placeholder labels from a target report template: "In [Month], the operations and maintenance platform received a total of [Number] work orders, with a work order completion rate of [Percentage]. Among them, there were [Number] fault-related work orders, [Number] support-related work orders, and [Number] service-related work orders."

[0048] S202: Extract each placeholder label from the target report template and determine the model auxiliary information and large language model processing flow associated with each placeholder label.

[0049] In one embodiment of this application, for each placeholder tag in the target report template, the model auxiliary information associated with the placeholder tag and the large language model processing flow are obtained from the knowledge base.

[0050] The knowledge base pre-stores model assistance information and large language model processing flow corresponding to each placeholder label in the report template. In this embodiment, the target report template has placeholder labels, and the model assistance information and large language model processing flow associated with the placeholder labels are stored in the knowledge base. The model assistance information and large language model processing flow are retrieved from the knowledge base based on the placeholder labels.

[0051] In another embodiment of this application, for each placeholder tag in the target report template, the placeholder tag large language model processing flow is obtained from the knowledge base, and model auxiliary information is obtained from the tag data of the placeholder tag in the target report template.

[0052] In this embodiment, the target report template includes placeholder labels and label data. The label data includes model auxiliary information associated with the placeholder labels. The large language model processing flow associated with the placeholder labels is stored in the knowledge base.

[0053] In the embodiments of this application, the model assistance information includes preset prompts.

[0054] In one example, the default prompt could be: "Please generate a brief summary of the work priorities for month xx". The relationship between placeholder labels and label data could be: {Prompt: Please generate a brief summary of the work priorities for month xx}.

[0055] In embodiments of this application, the large language model processing flow includes the invocation of one or more agents of the large language model.

[0056] In one example, the placeholder label {prompt: Please generate a brief summary of the work priorities for xx month} corresponds to the agent of the large language model, which calls the large language model to generate natural language text paragraphs.

[0057] S203: Obtain the project information corresponding to each placeholder label.

[0058] In the embodiments of this application, placeholder labels can be categorized by type, specifically including text placeholder labels, data placeholder labels, and file placeholder labels; the corresponding project information types can include project text content, project data content, and file content.

[0059] In embodiments of this application, the project text content includes text content related to the project in relation to the user report generation request. In one example, the project text content may be a basic summary of the project.

[0060] In the embodiments of this application, the project data content includes the data content of the project related to the user report generation request. The project data content can be external data from an external service interface, internal data included in the tag data, or a combination of external data from an external service interface and internal data from the tag data.

[0061] In embodiments of this application, the file content includes the file content of items related to the user report generation request. In one example, the file content may be maintenance photos taken by maintenance personnel.

[0062] S204: For each placeholder label, execute the large language model processing flow based on project information and model auxiliary information to obtain the report content corresponding to the placeholder label.

[0063] In the embodiments of this application, the project information and model auxiliary information are processed according to the large language model processing flow corresponding to the placeholder label to obtain the report content corresponding to the placeholder label.

[0064] In the embodiments of this application, the placeholder label can be a text placeholder label, the corresponding project information is the project text content, the model auxiliary information is prompt words related to the generated text content, and the report content is the text report content.

[0065] In the embodiments of this application, the placeholder label can be a data-type placeholder label, the corresponding project information is the project data content, the model auxiliary information is prompts related to the generation of data content, and the report content is data-type report content.

[0066] In the embodiments of this application, the placeholder label can be a file-type placeholder label, the corresponding project information is the file content, the model auxiliary information is prompts related to inserting the file content, and the report content is the file-type report content.

[0067] S205: Integrate the report content corresponding to each placeholder label into the template framework content to obtain the target operation and maintenance report.

[0068] In the embodiments of this application, each placeholder in the target report template is replaced with the report content corresponding to that placeholder, so as to integrate the report content corresponding to each placeholder tag into the template framework content, thereby transforming the target report template into a target operation and maintenance report.

[0069] Optionally, the target maintenance report can be output to the user terminal for the user to view or for the user to forward to the relevant maintenance personnel.

[0070] As described above, in response to a report generation request, a corresponding target report template is determined. This target report template includes multiple placeholders. Placeholder tags are extracted from the target report template, and the associated model auxiliary information and large language model processing flow for each placeholder tag are determined. Project information corresponding to each placeholder tag is also obtained. For each placeholder tag, the large language model processing flow is executed based on the project information and model auxiliary information to obtain the report content corresponding to the placeholder tag. The report content corresponding to each tag is integrated into the template framework content to obtain the target operation and maintenance report. This embodiment of the application can generate operation and maintenance reports in a segmented manner based on the report template using a large language model, avoiding the uncertainty caused by the "model illusion" when directly generating reports from a large model, thus improving the accuracy of operation and maintenance report generation. Simultaneously, the segmented generation reduces the full-text throughput and excessively long context of the large language model, thereby improving the processing efficiency of the large language model.

[0071] In one embodiment of this application, the aforementioned model assistance information includes preset prompt words; correspondingly, step S204 specifically includes the following steps: Input the project information and preset prompts into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label.

[0072] In one embodiment of this application, the placeholder label is a text-based placeholder label, the corresponding project information is the project text content, and the model-aided information is prompts related to the generated text content. Specifically, the project text content and the prompts related to the generated text content are processed according to the large language model processing flow corresponding to the placeholder label to obtain text-based report content.

[0073] In one embodiment of this application, the placeholder label is a data-type placeholder label, the corresponding project information is the project data content, and the model-aided information is prompts related to generating the data content. Specifically, the project data content and the prompts related to generating the data content are processed according to the large language model processing flow corresponding to the placeholder label to obtain data-type report content.

[0074] In one embodiment of this application, the placeholder label is a file-type placeholder label, the corresponding item information is the file content, and the model auxiliary information is prompt words related to inserting the file content. Specifically, the file content and the prompt words related to inserting the file content are processed according to the large language model processing flow corresponding to the placeholder label to obtain the file-type report content.

[0075] In the embodiments of this application, the large language model agent corresponding to the large language model processing flow can be one agent or multiple agents.

[0076] As can be seen from the above description, for each placeholder label, the report content is obtained by processing the project information and preset prompt words using a large language model. This enables the generation of corresponding report content based on the prompt words, making the report content generation process more flexible.

[0077] In one embodiment of this application, based on the above embodiments, the placeholder label includes a text-based placeholder label; the preset prompt word includes a first prompt word related to the generated text content; the project information includes the project text content; correspondingly, the project information and the preset prompt word are input into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label, including: The project text content and the first prompt word are input into the large language model agent corresponding to the large language model processing flow to obtain the text-based report content corresponding to the placeholder label.

[0078] In this embodiment, the project text content and the first prompt word are input into the large language model agent corresponding to the large language model processing flow, so that the large language model agent performs model processing based on the project text content, the first prompt word, and the context content of the placeholder label to obtain the text-based report content corresponding to the placeholder label.

[0079] In the embodiments of this application, each text class placeholder label can correspond to one or more large language model agents, and different large language model agents have different functions. For example, in this embodiment, the large language model agent is an agent that calls the large language model to generate natural language text paragraphs.

[0080] In one example, for a text-based placeholder label, the project text content of the placeholder label is "Basic introduction of project xx", and the first prompt is "Please generate a brief introduction to the work priorities for xx month". The large language model agent corresponding to this text-based placeholder label calls the large language model to generate a natural language text paragraph. The "Basic introduction of project xx" and the first prompt "Please generate a brief introduction to the work priorities for xx month" are input into the agent that calls the large language model to generate a natural language text paragraph, so as to output "Brief introduction to the work priorities for xx month", which is the text-based report content corresponding to this text-based placeholder label.

[0081] As described above, by obtaining the corresponding project text content and generating the first prompt word related to the text content through text placeholders, and processing the project text content and the first prompt word through a large language model agent, the text report content corresponding to the text placeholders can be obtained, thus enabling the generation of text report content in the target report template.

[0082] refer to Figure 3 , Figure 3 A flowchart illustrating the operation and maintenance report generation method provided in this application embodiment. Figure 2 In one embodiment of this application, based on the above embodiments, the placeholder labels include data-type placeholder labels; the preset prompts include second prompts related to the generated data content; the project information includes project data content; accordingly, the project information and the preset prompts are input into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder labels, including: S301: Based on the second prompt word, determine one or more sub-processes corresponding to the large language model processing flow.

[0083] S302: Input the project data content and the second prompt word into the large language model agent according to one or more sub-processes to obtain the report content of one or more data classes corresponding to the placeholder label.

[0084] In this embodiment, the second prompt word can be obtained in two ways: Method 1, based on the data class placeholder tag, the second prompt word associated with the data class placeholder tag is obtained from the knowledge base; Method 2, the second prompt word is obtained from the tag data of the data class placeholder tag.

[0085] In this embodiment, the second prompt word is input into a large language model for semantic understanding to obtain one or more sub-processes corresponding to the second prompt word.

[0086] In one embodiment of this application, one or more sub-processes include one or more of a form sub-process, a chart sub-process, and a data summary sub-process. Specifically, the report content for the data category corresponding to the form sub-process is form content, the report content for the data category corresponding to the chart sub-process is a visual chart, and the report content for the data category corresponding to the data summary sub-process is a data summary.

[0087] Optionally, if the sub-process is determined to be a form sub-process based on the second prompt word, the project data content and the second prompt word are input into one or more large language model agents corresponding to the form sub-process to obtain the form content corresponding to the placeholder label, which is used as the report content of the data type.

[0088] Optionally, if the sub-process is determined to be a chart sub-process based on the second prompt word, the project data content and the second prompt word are input into one or more large language model agents corresponding to the chart sub-process to obtain the table visualization chart corresponding to the placeholder label, which serves as the data-based report content.

[0089] Optionally, if the sub-process is determined to be a form sub-process and a data summary sub-process based on the second prompt word, then the project data content and the second prompt word are input into one or more large language model agents corresponding to the form sub-process to obtain the form content corresponding to the placeholder label; and the project data content and the second prompt word are input into one or more large language model agents corresponding to the data summary sub-process to obtain the data summary content corresponding to the placeholder label; the form content and the data summary content are integrated as the data-type report content.

[0090] Optionally, if the sub-process is determined to be a chart sub-process and a data summary sub-process based on the second prompt word, then the project data content and the second prompt word are input into one or more large language model agents corresponding to the chart sub-process to obtain the visual chart corresponding to the placeholder label, and the project data content and the second prompt word are input into one or more large language model agents corresponding to the data summary sub-process to obtain the data summary content corresponding to the placeholder label. The visual chart and the data summary content are used as the data-type report content.

[0091] In one example, if the second prompt is "display this month's work order data in charts and provide a month-on-month summary", then the sub-processes determined by this second prompt include a chart sub-process and a data summary sub-process; the output data-related report content includes visualization charts and data summary content.

[0092] In this embodiment, each data class placeholder label corresponds to one or more large language model agents.

[0093] As described above, by using the second prompt word of the data class placeholder label, one or more sub-processes that need to be executed are determined. Based on one or more sub-processes, the corresponding large language model intelligent agent is used to process the project data content and the second prompt word to obtain report content of different types of data classes, thus realizing the generation of different types of data class report content in the operation and maintenance report.

[0094] In one embodiment of this application, the project data content format can be various, such as tabular data or JSON data. The project data content in the above embodiments is mainly determined through at least one of the following two methods: Method 1: Obtain by accessing an external service interface.

[0095] In this embodiment, the external service interface includes a third-party service interface. Data access requests are sent to the third-party service interface to obtain external data from the third-party service interface, which serves as one of the sources of project data content.

[0096] In embodiments of this application, the external service interface may be a third-party API (Application Programming Interface) service interface.

[0097] Method 2: Extract from the tag data that comes with the data placeholder tag.

[0098] In this embodiment, the label data in the report template can store not only model auxiliary information (second prompt words) associated with placeholder labels, but also project data content.

[0099] Specifically, internal data extracted from the tag data inherent in the data placeholder tags serves as one of the sources of project data content.

[0100] For example, an example of the tag data for a data placeholder tag is: "{{Tag Name: xxx; Tag Tip: The project is configured with electromechanical, network, and IT professional engineers. The personnel allocation for this project is organized according to these specialties; Data: {"Personnel Category": [{"Category": "IT","Number of People": 3},{"Category": "Mechanical and Electrical","Number of People": 4},{"Category": "Other","Number of People": 6}]}}}". Here, "Data: {"Personnel Category": [{"Category": "IT","Number of People": 3},{"Category": "Mechanical and Electrical","Number of People": 4},{"Category": "Other","Number of People": 6}]}" is the internal data extracted from the tag data.

[0101] As can be seen from the above description, by obtaining external data through external service interfaces and extracting internal data through built-in tag data, the diversity of the source methods of project data content for data placeholder tags is improved.

[0102] In one embodiment of this application, step S302 specifically includes S3021 to S3023: S3021: Detect the data length of the project data content.

[0103] S3022: If the data length does not exceed the preset threshold, the project data content and the second prompt word are input into the large language model agent according to one or more sub-processes.

[0104] S3023: If the data length exceeds the preset threshold, the project data content will be compressed into a data table, and the data table and the second prompt word will be input into the large language model agent according to one or more sub-processes.

[0105] In this embodiment, the length of data processed in a single session by a conventional large language model is limited by a maximum length. However, due to the large quantity of project data, the maximum length limit is often exceeded.

[0106] The maximum length limit is determined based on the processing power of the large language model itself, and a preset threshold for the maximum length limit is usually set. For example, the preset threshold for the maximum length limit of a typical large language model is 32k.

[0107] Specifically, if the data length of any item exceeds a preset threshold, the item data content is converted into a data table to compress the data length. The data table and the second prompt word are then input into the large language model agent according to one or more sub-processes, and the corresponding data-type report content is output.

[0108] In this embodiment, the project data content is compressed into a data table by calling the data conversion agent module.

[0109] As can be seen from the above description, by compressing the project data content that is too long into a data table, the situation where the large language model fails to process the data due to its excessive length can be avoided.

[0110] In one embodiment of this application, based on the above embodiments, the placeholder labels include file-type placeholder labels; the preset prompts include third prompts related to the inserted file content; the project information includes the file content; accordingly, the project information and the preset prompts are input into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder labels, including: The file content and the third prompt word are input into the large language model agent corresponding to the large language model processing flow to obtain the report content of the file class corresponding to the placeholder label.

[0111] In this embodiment, each file class placeholder label can correspond to one or more large language model agents.

[0112] For example, the third prompt of a file-type placeholder label is "Please insert maintenance photos according to the maintenance date", the file content is maintenance photos taken by maintenance personnel, and the report content of the file type is the maintenance photos inserted into the target report template.

[0113] As described above, by obtaining the corresponding file content and inserting third-party prompts related to the file content through file-class placeholder tags, and processing the file content and third-party prompts through a large language model agent, the report content of the file class corresponding to the file-class placeholder tags can be obtained, thus enabling the insertion of file-class report content into the target report template.

[0114] In one embodiment of this application, the report generation request includes a report generation instruction statement; correspondingly, in S201, in response to the report generation request, accessing the knowledge base to obtain the corresponding target report template includes: S2011: Receive reports sent by the user and generate instruction statements.

[0115] In this embodiment, the report generation instruction statement is entered by the user on the user terminal.

[0116] S2012: Extract key project parameters from report-generated instruction statements using a large language model.

[0117] In this embodiment, one or more key project parameters are obtained by semantically understanding the report generation instruction statements through a large language model.

[0118] The key parameters for a project include one or more of the following: project information, report type, and report data time range. The report data time range can include weekly, monthly, or daily time ranges.

[0119] S2013: Access the knowledge base and query the knowledge base to obtain the target report template corresponding to the key parameters of the project.

[0120] In this embodiment, by extracting one or more key project parameters, the knowledge base is accessed to query the target report template associated with one or more key project parameters.

[0121] In this embodiment, if a report template associated with one or more project key parameters is identified as a single template, then that report template is determined to be the target report template.

[0122] As described above, by processing report generation instruction statements through a large model, accurate key parameters of the report generation project can be obtained. Then, based on the key parameters of the project, a report template that meets the user's needs can be found, thereby improving the success rate of matching report templates with user requests.

[0123] In one embodiment of this application, based on the above embodiment, after step S2013, the method further includes: S2014: If the project's key parameters have multiple corresponding report templates after querying the knowledge base, then a prompt to the user is returned through the large language model to further refine the report generation instructions.

[0124] In this embodiment, if the report generation instruction statement entered by the user is relatively simple, the extracted key project parameters are insufficient, and multiple corresponding report templates may be matched after querying the knowledge base.

[0125] In one example, the user sends a report generation instruction statement, "Weekly maintenance report needs to be generated." Using a large language model, the key project parameters are extracted from the instruction statement as "Maintenance report (report type)" and "Week (report data time range)." After querying the knowledge base, it is found that the user has multiple weekly maintenance report templates corresponding to various projects. To further narrow the search scope, a prompt is returned to the user to refine the report generation instruction statement, such as "Please confirm which project's weekly maintenance report it is." The user then enters "We need to generate a weekly maintenance report for project A," thus confirming that the weekly maintenance report template for project A is the target report template.

[0126] As can be seen from the above description, if multiple matching templates are found in the knowledge base, the report generation instruction statements are refined through multiple large model interactions with the user until a matching target report template is found, thus avoiding the failure of the target report template query.

[0127] refer to Figure 4 , Figure 4 A flowchart illustrating the operation and maintenance report generation method provided in this application embodiment. Figure 3 In one embodiment of this application, based on the above embodiments, a process for configuring a report template in the knowledge base is further included, as detailed below: S401: Obtain multiple report templates and the corresponding project category information for each report template.

[0128] S402: Based on project classification information, determine the key project parameters corresponding to each report template.

[0129] S403: Store the correspondence between the report template and the key parameters of the project in the knowledge base.

[0130] In the embodiments of this application, multiple report templates can be written by the user for each project, and the project classification information corresponding to each report template can be input by the user; from the project classification information input by the user, the key parameters of the project corresponding to each report template are extracted.

[0131] In the embodiments of this application, the key parameters of the project include one or more of the following: project information, report type, and report data time range.

[0132] As described above, by extracting the key parameters of the project corresponding to the report template through project classification information, and storing the correspondence between the report template and the key parameters of the project in the knowledge base, the report templates of each user's project are stored in the database, making it convenient for users to query the target report template later.

[0133] In one embodiment of this application, the method further includes the following after step S203: If the content of the model auxiliary information or project information of any placeholder label is empty, the large language model processing flow will not be executed for any placeholder label, and the placeholder label will be retained in the target operation and maintenance report.

[0134] In one embodiment of this application, if the model auxiliary information of the placeholder tag in the knowledge base is empty, and the model auxiliary information in the tag data of the target report template of the placeholder tag is empty, it is determined that the placeholder tag will not be able to execute the large language model processing flow, and the placeholder tag will continue to be retained in the target operation and maintenance report.

[0135] In another embodiment of this application, if the project information obtained by the placeholder label is empty, it is determined that the placeholder label will not be able to execute the large language model processing flow, and the placeholder label will continue to be retained in the target operation and maintenance report.

[0136] As can be seen from the above description, for placeholder labels that cannot execute the large language model processing flow, the placeholder labels are retained in the output of the target operation and maintenance report to facilitate user location and manual filling.

[0137] It should be noted that each placeholder label in the above embodiments is associated with one or more large language model agents in the knowledge base. The large language model agents provided in this application embodiment include, but are not limited to, the following agents (refer to Table 1). By summarizing the large language model agents, it is convenient for users to call and associate them with different placeholder labels. At the same time, the large language model agents provided in this application are not limited to the scenario associated with placeholder labels, but can also be applied to different scenarios of operation and maintenance reports.

[0138] Table 1. Summary of Intelligent Agents for Large Language Models Exemplary device refer to Figure 5 , Figure 5 This is a schematic diagram of the structure of the operation and maintenance report generation device provided in an embodiment of this application. The operation and maintenance report generation device includes: a template acquisition unit 501, a tag recognition unit 502, an information acquisition unit 503, a model processing unit 504, and a report integration unit 505.

[0139] The template acquisition unit 501 is used to access the knowledge base to obtain the corresponding target report template in response to the report generation request. The target report template includes template framework content and multiple placeholder tags.

[0140] The tag recognition unit 502 is used to extract each placeholder tag in the target report template and determine the model auxiliary information and large language model processing flow associated with each placeholder tag.

[0141] The information acquisition unit 503 is used to acquire the project information corresponding to each placeholder label.

[0142] The model processing unit 504 is used to execute the large language model processing flow for each placeholder label based on the project information and the model auxiliary information to obtain the report content corresponding to the placeholder label.

[0143] The report integration unit 505 is used to integrate the report content corresponding to each placeholder label into the template framework content to obtain the target operation and maintenance report.

[0144] In one embodiment of this application, the model assistance information includes preset prompt words; the model processing unit 504 is specifically used to: input the project information and the preset prompt words into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label.

[0145] In one embodiment of this application, the placeholder label includes a text-based placeholder label; the preset prompt word includes a first prompt word related to the generated text content; the project information includes project text content; and the model processing unit 504 is specifically used to: input the project text content and the first prompt word into the large language model agent corresponding to the large language model processing flow to obtain the text-based report content corresponding to the placeholder label.

[0146] In one embodiment of this application, the placeholder label includes a data-type placeholder label; the preset prompt word includes a second prompt word related to the generated data content; the project information includes project data content; the model processing unit 504 is specifically used to: determine one or more sub-processes corresponding to the large language model processing flow based on the second prompt word; input the project data content and the second prompt word into the large language model agent according to the one or more sub-processes respectively, to obtain report content of one or more data types corresponding to the placeholder label.

[0147] In one embodiment of this application, the one or more sub-processes include one or more of a form sub-process, a chart sub-process, and a data summary sub-process; wherein the report content of the data class corresponding to the form sub-process is form content, the report content of the data class corresponding to the chart sub-process is a visual chart, and the report content of the data class corresponding to the data summary sub-process is data summary content.

[0148] In one embodiment of this application, the project data content is determined by at least one of the following methods: obtaining it by accessing an external service interface; or extracting it from the tag data that comes with the data class placeholder tag.

[0149] In one embodiment of this application, the model processing unit 504 is further configured to: detect the data length of the project data content; if the data length does not exceed a preset threshold, input the project data content and the second prompt word into the large language model agent according to one or more sub-processes respectively; if the data length exceeds the preset threshold, compress the project data content into a data table, and input the data table and the second prompt word into the large language model agent according to one or more sub-processes respectively.

[0150] In one embodiment of this application, the placeholder label includes a file-type placeholder label; the preset prompt word includes a third prompt word related to the inserted file content; the project information includes the file content; the model processing unit 504 is specifically used to: input the file content and the third prompt word into the large language model agent corresponding to the large language model processing flow to obtain the report content of the file-type corresponding to the placeholder label.

[0151] In one embodiment of this application, the report generation request includes a report generation instruction statement; the template acquisition unit 501 is specifically used for: receiving the report generation instruction statement sent by the user terminal; extracting key project parameters from the report generation instruction statement through a large language model; accessing a knowledge base and querying the knowledge base to obtain the target report template corresponding to the key project parameters.

[0152] In one embodiment of this application, the template acquisition unit 501 is further configured to: if the project key parameters have multiple corresponding report templates after querying the knowledge base, then return a prompt to the user terminal through the large language model to further refine the report generation instruction statement.

[0153] In one embodiment of this application, the apparatus further includes: a template configuration unit, configured to: acquire multiple report templates and project classification information corresponding to each report template; determine key parameters of the project corresponding to each report template based on the project classification information; and store the correspondence between the report templates and the key parameters of the project in the knowledge base.

[0154] In one embodiment of this application, the report integration unit 505 is further configured to: if the content of the model auxiliary information or project information of any placeholder label is empty, then not perform the large language model processing flow for the placeholder label, and retain the placeholder label in the target operation and maintenance report.

[0155] The apparatus provided in this embodiment belongs to the same concept as the method provided in the above embodiments of this application, and can execute the method provided in any of the above embodiments of this application, possessing the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the specific processing content of the method provided in the above embodiments of this application, and will not be repeated here.

[0156] The functions implemented by the template acquisition unit 501, the tag recognition unit 502, the information acquisition unit 503, the model processing unit 504, and the report integration unit 505 can be implemented by the same or different processors, and this application embodiment does not limit them.

[0157] It should be understood that the units in the above device can be implemented by a processor calling software. For example, the device includes a processor connected to a memory containing instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each unit in the device. The processor can be a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal or external to the device. Alternatively, the units in the device can be implemented as hardware circuits. By designing the hardware circuits, some or all of the unit functions can be implemented. The hardware circuits can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above units are implemented by designing the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented using a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files to implement the functions of some or all of the above units. All units in the above device can be implemented entirely by a processor calling software, entirely by hardware circuits, or partially by a processor calling software with the remaining parts implemented by hardware circuits.

[0158] In this application embodiment, a processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP. In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable. For example, the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.

[0159] As can be seen, each unit in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.

[0160] Furthermore, the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a System-on-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device. The at least one processor may be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.

[0161] Exemplary electronic devices Another embodiment of this application also provides an electronic device, see [link to relevant documentation] Figure 6 As shown, the device includes: Memory 601 and processor 602; The memory 601 is connected to the processor 602 and is used to store programs; The processor 602 is used to implement the operation and maintenance report generation method disclosed in any of the above embodiments by running the program stored in the memory 601.

[0162] Specifically, the aforementioned electronic device may also include: a bus, a communication interface 603, an input device 604, and an output device 605.

[0163] The processor 602, memory 601, communication interface 603, input device 604, and output device 605 are interconnected via a bus. Among them: A bus can include a pathway for transmitting information between various components of a computer system.

[0164] Processor 602 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0165] Processor 602 may include a main processor, as well as a baseband chip, modem, etc.

[0166] The memory 601 stores a program that executes the technical solution of the present invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 601 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.

[0167] Input device 604 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer, or gravity sensor.

[0168] Output device 605 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0169] The communication interface 603 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0170] The processor 602 executes the program stored in the memory 601 and calls other devices, which can be used to implement any of the steps of the operation and maintenance report generation method provided in the above embodiments of this application.

[0171] This application also proposes a chip, which includes a processor and a data interface. The processor reads and runs a program stored in a memory through the data interface to execute the operation and maintenance report generation method described in any of the above embodiments. For the specific processing procedure and its beneficial effects, please refer to the embodiments of the operation and maintenance report generation method described above.

[0172] Exemplary computer program products and storage media In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the operation and maintenance report generation method according to various embodiments of this application as described in any of the above embodiments of this specification.

[0173] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0174] Furthermore, embodiments of this application may also be storage media storing a computer program thereon, which, when run by a processor, causes the processor to execute the steps in the operation and maintenance report generation method according to various embodiments of this application as described in any of the foregoing embodiments of this specification.

[0175] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0176] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0177] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.

[0178] The modules and sub-modules in the various embodiments of the present application's devices and terminals can be merged, divided, and deleted according to actual needs.

[0179] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0180] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.

[0181] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.

[0182] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0183] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0184] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0185] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for generating operation and maintenance reports, characterized in that, include: In response to a report generation request, the knowledge base is accessed to obtain the corresponding target report template, which includes template framework content and multiple placeholder tags; Extract each placeholder label from the target report template and determine the model assistance information and large language model processing flow associated with each placeholder label; Retrieve the project information corresponding to each placeholder label; For each placeholder label, the large language model processing flow is executed based on the project information and the model auxiliary information to obtain the report content corresponding to the placeholder label; The report content corresponding to each placeholder label is integrated into the template framework content to obtain the target operation and maintenance report.

2. The method according to claim 1, characterized in that, The model assistance information includes preset prompts; Accordingly, the step of executing the large language model processing flow based on the project information and the model-aided information to obtain the report content corresponding to the placeholder labels includes: The project information and the preset prompt words are input into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label.

3. The method according to claim 2, characterized in that, The placeholder labels include text placeholder labels; the preset prompts include first prompts related to the generated text content; the project information includes project text content; Accordingly, the step of inputting the project information and the preset prompt words into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label includes: The project text content and the first prompt word are input into the large language model agent corresponding to the large language model processing flow to obtain the text-based report content corresponding to the placeholder label.

4. The method according to claim 2, characterized in that, The placeholder labels include data-related placeholder labels; the preset prompts include second prompts related to the generated data content; the project information includes project data content. Accordingly, the step of inputting the project information and the preset prompt words into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label includes: Based on the second prompt word, determine one or more sub-processes corresponding to the large language model processing flow; The project data content and the second prompt word are respectively input into the large language model agent according to one or more sub-processes to obtain report content of one or more data classes corresponding to the placeholder label.

5. The method according to claim 4, characterized in that, The one or more sub-processes mentioned therein include one or more of the form sub-process, chart sub-process, and data summary sub-process; wherein the report content of the data class corresponding to the form sub-process is form content, the report content of the data class corresponding to the chart sub-process is a visual chart, and the report content of the data class corresponding to the data summary sub-process is data summary content.

6. The method according to claim 4, characterized in that, The project data content is determined through at least one of the following methods: Obtained by accessing external service interfaces; Extract from the tag data that comes with the placeholder tags in the data class.

7. The method according to claim 4, characterized in that, The step of inputting the project data content and the second prompt word into the large language model agent according to one or more sub-processes includes: Detect the data length of the project data content; If the data length does not exceed a preset threshold, the project data content and the second prompt word are respectively input into the large language model agent according to one or more sub-processes; If the data length exceeds a preset threshold, the project data content is compressed into a data table, and the data table and the second prompt word are respectively input into the large language model agent according to one or more sub-processes.

8. The method according to claim 2, characterized in that, The placeholder labels include file-related placeholder labels; the preset prompts include third-party prompts related to inserting file content; the project information includes file content; Accordingly, the step of inputting the project information and the preset prompt words into the large language model agent corresponding to the large language model processing flow to obtain the report content corresponding to the placeholder label includes: The file content and the third prompt word are input into the large language model agent corresponding to the large language model processing flow to obtain the report content of the file class corresponding to the placeholder label.

9. The method according to any one of claims 1 to 8, characterized in that, The report generation request includes a report generation instruction statement; Accordingly, the step of accessing the knowledge base to obtain the corresponding target report template in response to the report generation request includes: Receive reports sent by the user and generate instruction statements; Key project parameters are extracted from the instruction statements generated in the report using a large language model. Access the knowledge base and query it to obtain the target report template corresponding to the key parameters of the project.

10. The method according to claim 9, characterized in that, After accessing the knowledge base and querying the knowledge base to obtain the target report template corresponding to the key parameters of the project, the process also includes: If the project's key parameters have multiple corresponding report templates after querying the knowledge base, then a prompt to the user is returned through the large language model, indicating that the report generation should be further refined.

11. The method according to claim 9, characterized in that, Also includes: Obtain multiple report templates and the corresponding project category information for each report template; Based on the project classification information, determine the key project parameters corresponding to each report template; The correspondence between the report template and the key parameters of the project is stored in the knowledge base.

12. The method according to any one of claims 1 to 8, characterized in that, After obtaining the project information corresponding to each placeholder label, the process also includes: If the content of the model auxiliary information or project information of any placeholder label is empty, the large language model processing flow will not be executed for the placeholder label, and the placeholder label will be retained in the target operation and maintenance report.

13. A maintenance report generation device, characterized in that, include: The template acquisition unit is used to access the knowledge base to obtain the corresponding target report template in response to the report generation request, wherein the target report template includes template framework content and multiple placeholder tags; The tag recognition unit is used to extract each placeholder tag in the target report template and determine the model auxiliary information and large language model processing flow associated with each placeholder tag. The information acquisition unit is used to acquire the project information corresponding to each placeholder label; The model processing unit is used to execute the large language model processing flow for each placeholder label based on the project information and the model auxiliary information to obtain the report content corresponding to the placeholder label. The report integration unit is used to integrate the report content corresponding to each placeholder label into the template framework content to obtain the target operation and maintenance report.

14. An electronic device, characterized in that, Including memory and processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the operation and maintenance report generation method as described in any one of claims 1 to 12 by running the program in the memory.

Images