Data processing method and system, electronic equipment and computer readable storage medium

Abstract

The invention discloses a data processing method and system, electronic equipment and a computer readable storage medium, and relates to the technical field of data processing and large models. The method comprises the following steps: acquiring multi-source heterogeneous data; analyzing the multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data; fusing the multi-source heterogeneous knowledge maps to obtain a target navigation map; generating multi-level description information based on the target navigation map; and performing knowledge production according to the multi-level description information, and generating a navigation page description document in the target application scene. The technical problems that the knowledge production efficiency is low and data from different sources cannot be fully utilized to complete data analysis in related technologies are solved.

Description

Technical Field

[0001] The present application relates to the fields of data processing technology and large model technology, and more specifically, to a data processing method, system, electronic device and computer-readable storage medium. Background Art

[0002] The metering automation system plays a vital role in the smart grid, responsible for the collection, processing and analysis of power metering data. With the continuous advancement of the construction of new power systems, the use of metering data is increasing, requiring the metering automation system to process a large amount of data from different sources to support decision-making, data analysis and real-time monitoring.

[0003] At present, the collection and application of metrological data mainly rely on traditional software architecture. The inherent limitations of this architecture become particularly obvious when facing data from different sources. This architecture usually relies on manual or semi-automatic methods to update navigation document descriptions. When there are major changes in the system architecture or functions, this update process is not only time-consuming and labor-intensive, but also prone to errors, resulting in the inability to ensure the real-time and accuracy of information, which in turn leads to low knowledge production efficiency. In addition, in terms of metrological data application, the current metrological automation system has a low level of intelligence and usually relies on the customized development of task systems to achieve specific data use and analysis. Therefore, to a certain extent, it limits the mining and utilization of the potential value of metrological data.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present application provide a data processing method, system, electronic device and computer-readable storage medium to at least solve the technical problems in the related art of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis.

[0006] According to one aspect of an embodiment of the present application, a data processing method is provided, including: acquiring multi-source heterogeneous data; parsing the multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data; fusing the multi-source heterogeneous knowledge graph to obtain a target navigation graph; generating multi-level description information based on the target navigation graph; performing knowledge production based on the multi-level description information to generate a navigation page description document under a target application scenario.

[0007] According to another aspect of an embodiment of the present application, a data processing method is also provided, including: acquiring multi-source heterogeneous metering data; parsing the multi-source heterogeneous metering data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous metering data; fusing the multi-source heterogeneous knowledge graph to obtain a metering navigation graph; generating multi-level description information based on the metering navigation graph; performing knowledge production based on the multi-level description information to generate a navigation question and answer page description document under a power metering data application scenario.

[0008] According to another aspect of an embodiment of the present application, a data processing method is also provided, including: obtaining a data processing request through a first application programming interface, wherein the request data carried in the data processing request includes: multi-source heterogeneous data; returning a data processing response through a second application programming interface, wherein the response data carried in the data processing response includes: a navigation page description document in a target application scenario, and the navigation page description document is generated according to any one of the above-mentioned data processing methods.

[0009] According to another aspect of an embodiment of the present application, a data processing method is also provided, including: responding to input instructions acting on an operation interface, inputting multi-source heterogeneous data; responding to processing instructions acting on the operation interface, displaying a navigation page description document in a target application scenario on the operation interface, the navigation page description document being generated according to any one of the above-mentioned data processing methods.

[0010] According to another aspect of an embodiment of the present application, a data processing system is also provided, including: a client, used to send multi-source heterogeneous data; a server, connected to the client, used to parse the multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, fuse the multi-source heterogeneous knowledge graph to obtain a target navigation graph, generate multi-level description information based on the target navigation graph, and perform knowledge production based on the multi-level description information to generate a navigation page description document in a target application scenario; the client is also used to output the navigation page description document.

[0011] According to another aspect of an embodiment of the present application, an electronic device is also provided, including: a memory storing an executable program; a processor connected to the memory via a bus and used to run the program, wherein any one of the above-mentioned data processing methods is executed when the program is running.

[0012] According to another aspect of an embodiment of the present application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned data processing methods.

[0013] According to another aspect of an embodiment of the present application, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, any one of the above-mentioned data processing methods is implemented.

[0014] In an embodiment of the present application, multi-source heterogeneous data is obtained, the multi-source heterogeneous data is parsed, a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data is generated, and then the multi-source heterogeneous knowledge graph is fused to obtain a target navigation graph, and then multi-level description information is generated based on the target navigation graph. Finally, knowledge production is performed based on the multi-level description information to generate a navigation page description document under the target application scenario, thereby achieving the purpose of fully utilizing data from different sources to complete data analysis and updating knowledge documents in real time. Therefore, the system can automatically extract, integrate and present accurate and detailed navigation page description documents (such as operation guides and description documents) according to user needs and scenario changes, greatly improving the efficiency of data use and the level of intelligence, reducing dependence on task system development, fully tapping the value of data, optimizing user operation experience, achieving data value maximization and system intelligence, and the technical effect of automatic generation and updating of knowledge documents, thereby solving the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0015] It is easy to notice that the above general description and the following detailed description are only for the purpose of exemplifying and explaining the present application, and do not constitute a limitation of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0017] Figure 1 is a schematic diagram of an application scenario of a data processing method according to an embodiment of the present application;

[0018] Figure 2 is a flow chart of a data processing method according to an embodiment of the present application;

[0019] Figure 3 is a flow chart of a knowledge generation system based on multi-source heterogeneous data fusion according to an embodiment of the present application;

[0020] Figure 4 is a flow chart of a data processing method according to an embodiment of the present application;

[0021] Figure 5 is a flow chart of a data processing method according to an embodiment of the present application;

[0022] Figure 6is a flow chart of a data processing method according to an embodiment of the present application;

[0023] Figure 7 is a structural schematic diagram of a data processing system according to an embodiment of the present application;

[0024] Figure 8 is a structural schematic diagram of a data processing device according to an embodiment of the present application;

[0025] Fig. 9 is a structural schematic diagram of another data processing device according to an embodiment of the present application;

[0026] Fig.10 is a structural schematic diagram of another data processing device according to an embodiment of the present application;

[0027] Fig.11 is a structural schematic diagram of another data processing device according to an embodiment of the present application;

[0028] Fig.12 is a structural block diagram of a computing device according to an embodiment of the present application;

[0029] Fig.13 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] The technical solution provided in this application is mainly implemented by large-scale model technology. The large model here refers to a deep learning model with large-scale model parameters, which can usually contain hundreds of millions, tens of billions, hundreds of billions, trillions or even more than ten trillion model parameters. The large model can also be called a cornerstone model / foundation model (Foundation Model). The large model is pre-trained through large-scale unlabeled corpus to produce a pre-trained model with more than 100 million parameters. This model can adapt to a wide range of downstream tasks, and the model has good generalization ability, such as large-scale language model (Large Language Model, LLM), multi-modal pre-training model (multi-modalpre-training model), etc.

[0033] It should be noted that when the large model is actually used, the pre-trained model can be fine-tuned through a small number of samples so that the large model can be applied to different tasks. For example, the large model can be widely used in natural language processing (Natural Language Processing, NLP), computer vision, speech processing and other fields, and can be specifically applied to computer vision tasks such as visual question answering (Visual Question Answering, VQA), image description (Image Caption, IC for short), image generation, etc., and can also be widely used in text-based sentiment classification, text summary generation, machine translation and other natural language processing tasks. Therefore, the main application scenarios of the large model include but are not limited to digital assistants, intelligent robots, search, online education, office software, e-commerce, intelligent design, etc. In the embodiment of the present application, data processing by a large model in a metering scenario is used as an example for explanation.

[0034] First, some nouns or terms that appear in the description of the embodiments of the present application are subject to the following explanations:

[0035] Multi-source heterogeneous data fusion: refers to the process of integrating data sets from multiple different sources and different formats (i.e. heterogeneous) into a unified and more useful information set through a series of technical means and methods. In this process, the data types that may be involved include but are not limited to structured data (such as tables in databases), semi-structured data (such as eXtensible Markup Language (XML) or JavaScript Object Notation (JSON) files) and unstructured data (such as text, pictures, etc.). This technology aims to solve the problem of data islands caused by a wide range of data sources and diverse formats, promote the effective exchange and use of cross-domain information, and thus support more accurate decision-making, analysis and prediction and other functions. In practical applications, it is widely used in smart cities, the Internet of Things, big data analysis and other fields.

[0036] Knowledge production: refers to the process of extracting information from structured, semi-structured and unstructured data sources through a series of knowledge graph technical means and methods, and converting it into semantically related knowledge units, thereby constructing or enriching the knowledge graph. This process covers multiple steps such as entity recognition, relationship extraction, and attribute filling, aiming to form a knowledge network that can express the complex connections between concepts. In practical applications, knowledge graphs can be used to improve search efficiency, support intelligent question-answering systems, and assist in decision-making analysis. With the development of big data and artificial intelligence technologies, automated knowledge graph construction tools have become increasingly mature, which has significantly improved the efficiency of knowledge production and promoted the integration and innovation of cross-domain knowledge.

[0037] The metering automation system is the core task system of the power grid. With the steady progress of the construction of new power systems, the application scenarios of metering data are becoming more and more abundant, which puts forward higher requirements on the robustness, scalability, and data usability of the metering system. The current master station system still uses traditional software architecture in data collection and application innovation. The scalability, stability, and real-time performance of the collection end cannot meet the needs of on-demand collection and end-to-end real-time refresh in the future. The degree of intelligence in the application of metering data is not high, and the use of data depends on the development of the task system, so the value of metering data cannot be fully utilized.

[0038] This application conducts end-to-end technical innovation under the guidance of the new generation of metering automation system construction specifications to further improve the stability and usability of the metering system. In the navigation document assistant scenario, because the web page architecture and function path of the metering master system have undergone a lot of changes, it is necessary to generate and update the operation description of the metering master system page.

[0039] At present, there is a knowledge management system that uses a large language model and knowledge graph to help build a more intelligent and user-friendly knowledge management system, helping users to quickly, efficiently and conveniently acquire, share, apply and innovate knowledge, improve the overall knowledge management level, empower business and create greater business value. By analyzing the content of knowledge, a two-way link relationship between knowledge is established based on knowledge graph technology, and related implicit knowledge is discovered while viewing knowledge. The knowledge management system builds an internal knowledge context based on the knowledge graph, connects knowledge nodes, document attachments, directories and paragraphs within the document, and establishes cross-document and cross-node knowledge links. By connecting upstream and downstream business knowledge and associating related knowledge, implicit business knowledge can be displayed in an explicit graphical manner. However, the knowledge management system cannot obtain and construct graphs for non-standard data on the navigation page of the metering system.

[0040] There is also a software that provides intelligent knowledge management based on knowledge graphs, which provides the function of automatically and accurately matching knowledge classification and labeling based on knowledge warehouse and atomic knowledge data, realizes intelligent classification, intelligent labeling, knowledge deduplication, batch storage, and intelligent recommendation of knowledge storage, and uses knowledge atomization technology to provide intelligent processing of knowledge content such as collection, classification, association, comparison, and update. However, this software cannot disambiguate conflicting knowledge and is not suitable for measurement knowledge generation systems.

[0041] There is also an electric power knowledge graph system that provides a distributed engine to store graph data. The middle layer includes data acquisition, graph construction and graph management modules, and finally inputs the data into the graph application layer, covering various application scenarios such as intelligent question and answer, voice assistant, and content understanding. The graph database forms a large-scale knowledge base and supports task applications by processing massive amounts of information. It can also introduce machine learning, artificial intelligence, natural language recognition and other technical means to improve the accuracy and fun of answers, save labor costs, and improve user experience. However, as a graph database product, its natural language understanding ability is poor and cannot adapt to natural language understanding in power metering scenarios.

[0042] The traditional metering automation system in the related art has the following defects.

[0043] Defect 1: It is impossible to achieve on-demand collection, that is, it is impossible to flexibly select the type, frequency and scope of collected data according to actual needs and application scenarios.

[0044] Defect 2: It is impossible to achieve end-to-end real-time refresh, that is, it is impossible to quickly complete data processing, analysis and storage after data collection to ensure the timeliness of the data.

[0045] Defect 3: The level of intelligence is not high, and the use of data depends on the development of task systems, resulting in the inability to fully utilize the value of the collected data.

[0046] With respect to the above-mentioned defects, no effective solution has been proposed before the present application.

[0047] According to an embodiment of the present application, a data processing method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0048] Considering the huge number of model parameters of the large model and the limited computing resources of the mobile terminal, the above method provided in the embodiment of the present application can be applied to Figure 1 The application scenarios shown are not limited to these. Figure 1 In the application scenario shown, the large model is deployed in the server 10, and the server 10 can be connected to one or more client devices 20 through a local area network connection, a wide area network connection, an Internet connection, or other types of data networks. The client device 20 here may include but is not limited to: a smart phone, a tablet computer, a laptop computer, a PDA, a personal computer, a smart home device, a vehicle-mounted device, etc. The client device 20 can interact with the user through a graphical user interface to implement the call of the large model, thereby implementing the method provided in the embodiment of the present application.

[0049] In an embodiment of the present application, a system composed of a client device and a server can perform the following steps: the client device executes steps of acquiring multi-source heterogeneous data and sending the acquired multi-source heterogeneous data to the server; the server executes steps of parsing the multi-source heterogeneous data based on the acquired multi-source heterogeneous data, generating a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, fusing the multi-source heterogeneous knowledge graph to obtain a target navigation graph, generating multi-level description information based on the target navigation graph, performing knowledge production based on the multi-level description information, generating a navigation page description document under the target application scenario, and returning the navigation page description document to the client device. It should be noted that when the operating resources of the client device can meet the deployment and operating conditions of the large model, the embodiment of the present application can be performed in the client device.

[0050] It should be noted that, with the rapid development of high-performance computing units, in other application scenarios, the above method provided in the embodiment of the present application can also be applied to the model all-in-one machine. In an optional embodiment, a plurality of models are built into the model all-in-one machine, and the user can choose to adjust a model as needed to obtain the user's own model, so that the high-performance computing unit built into the model all-in-one machine can directly call the adjusted model to execute the above method provided in the embodiment of the present application. In another optional embodiment, a trained model is built into the large model all-in-one machine, so that the high-performance computing unit built into the model all-in-one machine can directly call the model to execute the above method provided in the embodiment of the present application.

[0051] Furthermore, when users need to train their own models, they can also upload their own data sets through the client, which are sent by the client to the server, so that the server can adjust the pre-trained model with the data set to obtain the user's own model and then deploy it to the production environment. In order to facilitate users' needs for model adjustment, the server can provide complete adjustment tools, development frameworks and processes, and can support multiple adjustment strategies, so that the adjusted model can better adapt to applications in different fields and achieve a high degree of customization.

[0052] Under the above operating environment, this application provides Figure 2 The data processing method shown. Figure 2 is a flow chart of a data processing method according to an embodiment of the present application. Figure 2 As shown, the method may include the following steps:

[0053] Step S21, acquiring multi-source heterogeneous data;

[0054] Step S22, parsing the multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data;

[0055] Step S23, fusing the multi-source heterogeneous knowledge graphs to obtain a target navigation graph;

[0056] Step S24, generating multi-level description information based on the target navigation map;

[0057] Step S25, performing knowledge production based on the multi-level description information to generate a navigation page description document for the target application scenario.

[0058] In the embodiments of the present application, multi-source heterogeneous data can be understood as data originating from different data sources and having different structures and formats. Exemplarily, data sources may include internal systems, external network resources, sensors, social media, documents, databases, etc. The data sources are independent of each other, and the data structures, formats, and semantics may be different, which are not limited here. Data formats include unstructured text (such as emails, reports, social media posts), semi-structured data (such as XML or JSON files, web page hypertext markup language (HyperText Markup Language, HTML) code) and structured data (such as tabular data in a database), etc., which are not limited here.

[0059] It is understandable that in the measurement scenario, multi-source heterogeneous data can refer to data of various types, sources and formats involved in the measurement automation system. Multi-source heterogeneous data can include but is not limited to unstructured text (such as operation manuals, training documents), semi-structured HTML information (such as web page architecture and function descriptions) and structured tables (such as measurement data in databases), which are not limited here.

[0060] Multi-source heterogeneous data can be obtained through a variety of technical means, such as web crawlers, database queries, document reading, etc., which are not limited here.

[0061] After acquiring multi-source heterogeneous data, the multi-source heterogeneous data is parsed to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data. The multi-source heterogeneous knowledge graph can be understood as a collection of graphs extracted and constructed from multi-source heterogeneous data, that is, knowledge is extracted from data of multiple different sources and structures and organized into a collection of graphs. The multi-source heterogeneous knowledge graph uses nodes to represent entities and edges to represent the relationships between entities.

[0062] For example, when parsing multi-source heterogeneous data and generating a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, each data source can generate an independent knowledge graph, which has its own structure and information content and together constitutes a multi-source heterogeneous knowledge graph. Thus, by converting data from different sources and structures into a unified knowledge representation, the efficiency and effect of data processing and application can be improved.

[0063] Furthermore, for unstructured text, NLP technologies such as entity recognition, relationship extraction and semantic analysis can be used to build a text knowledge graph. For semi-structured HTML data, the knowledge graph related to the structure and content of the web page is generated through Document Object Model (DOM) parsing and data extraction technology. For structured data, it can be directly converted into graph nodes and relationships through database query or data mapping technology, which is not limited here.

[0064] After generating the multi-source heterogeneous knowledge graphs, the multi-source heterogeneous knowledge graphs are fused to obtain the target navigation graph, that is, the different knowledge graphs generated above are integrated into a unified target knowledge graph. Among them, the target navigation graph can be understood as a navigation manual graph, which is used to clearly display the structured knowledge representation of system pages and functions. The target navigation graph integrates multi-source heterogeneous data, so that the system can generate accurate navigation and operation guides, and improve the user's operation efficiency and experience in the system. Exemplarily, in the metering scenario, the target navigation graph can be understood as a detailed structured knowledge graph generated for a specific page or function of the metering master station system, and its purpose is to provide clear and accurate navigation paths and operation guides.

[0065] For example, when fusing multi-source heterogeneous knowledge graphs, it is necessary to consider the consistency of entities (ensuring that the same entity is represented consistently in different graphs), data conflicts (when the attribute values ​​of entities with the same name are inconsistent in different graphs, they need to be resolved according to predefined rules or algorithms), and information supplementation (using information from multiple graphs to enrich the description of entities) to obtain the target navigation graph. In this way, a more comprehensive and accurate navigation graph can be constructed to guide or explain the use process of a specific system, platform or application.

[0066] After obtaining the target navigation graph, multi-level description information is generated based on the target navigation graph. Exemplarily, the multi-level description information may include sentence-level, paragraph-level, and document-level description information. Furthermore, the sentence-level description information focuses on the simple expression of entities and relationships in the graph, which is suitable for quick search or brief definition. The paragraph-level description information provides more detailed information, combining the descriptions of multiple entities and relationships, which helps to understand complex business processes or system operations. The document-level description information is used to comprehensively cover the knowledge of the target system or subject, which can be a user manual, operating guide, or detailed business process documentation, which is not limited here.

[0067] Therefore, by dividing information into multiple levels (sentence level, paragraph level and document level), the system can perform retrieval and information generation at different granularities, thereby finding the best balance between retrieval speed and the level of detail of generated content, ensuring that the knowledge fragments collected from information of different granularities can complement each other and increase the relevance and completeness of the information in the description document. By generating multi-level description information, knowledge information of different granularities can be generated according to user needs and query context.

[0068] After generating multi-level description information, knowledge production is performed based on the multi-level description information to generate a navigation page description document under the target application scenario, that is, to generate a detailed description document for the navigation page. Among them, the target application scenario can be a navigation and knowledge content generation scenario for the metering master station system, and can also be a scenario for providing question and answer services for the intelligent question and answer system (that is, combining the context of the user's query, intelligently generating page descriptions and operation guides associated with the query, providing more accurate and more humane question and answer services), and can also be a system document update scenario (that is, when the system is updated or the function is adjusted, the relevant description documents are automatically updated to ensure that the document content is consistent with the system function and reduce maintenance costs), and can also be an online help scenario (that is, when the user encounters a problem during the operation, the description document related to the current page can be quickly accessed, providing instant help information and operation steps, and improving the efficiency of the user in solving problems), etc., which are not limited here.

[0069] Exemplarily, knowledge production based on multi-level description information may include document layout design, logical organization of information, style selection, and final document generation, thereby ensuring that the generated navigation page description document is both accurate and readable and can meet the needs of target users.

[0070] The navigation page description document not only includes specific operation guides, but also covers multi-level information such as operation logic, data sources, functional significance, etc., so that users can fully understand how to use and operate each page of the metering master station system.

[0071] It is understandable that in the metering scenario, the navigation page description document can be understood as a document that records in detail the structure, function, operation process and related knowledge of specific pages in the metering master system. This type of document includes multi-level information from sentence level, paragraph level to document level, aiming to provide system users and maintenance personnel with comprehensive operation guides and page interpretations to improve usage efficiency and user experience.

[0072] In the embodiment of the present application, by acquiring multi-source heterogeneous data, parsing the multi-source heterogeneous data, generating a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, and then fusing the multi-source heterogeneous knowledge graph to obtain a target navigation graph, and then generating multi-level description information based on the target navigation graph, and finally performing knowledge production based on the multi-level description information to generate a navigation page description document under the target application scenario. Thus, the present application extracts key information from data of different formats and sources by parsing data of different formats and sources to form a unified graph data, thereby improving the utilization rate of data, and eliminating the problem of data islands. In the metering scenario, the metering master station system can fully and accurately understand and process data from multiple sources. Furthermore, the present application can effectively solve data conflicts and ambiguity problems by fusing multi-source heterogeneous knowledge graphs, and ensure the accuracy and reliability of the knowledge graph.

[0073] At the same time, this application provides a real-time data processing and updating mechanism. Through the generation and fusion of knowledge graphs and the generation of multi-level descriptive information, it can quickly respond to system changes, improve knowledge production efficiency, and ensure the real-time and accuracy of information.

[0074] In addition, the multi-level feature generation strategy allows the system to quickly retrieve and generate information at different granularities, ensuring the real-time and responsiveness of the description document. This end-to-end real-time processing flow enables the system to quickly update and present information after data collection, meeting the needs of real-time refresh and improving the timeliness of data and user experience.

[0075] Finally, the navigation page description document obtained by this application through knowledge production through multi-level description information not only records the page structure and operation process in detail, but also incorporates relevant background knowledge and operation suggestions. It can provide richer and more user-friendly operation step instructions, greatly optimize the user experience and reduce the user's learning cost.

[0076] It can be seen that through the solution of this application, the system can automatically understand and generate information of different granularities, reducing the reliance on manual development of task systems. This automated and intelligent data processing method can not only generate knowledge documents more efficiently and accurately, but also dynamically adjust the generation and presentation of information according to user queries and usage behaviors, giving full play to the value of data.

[0077] The above-mentioned data processing method provided in the embodiment of the present application can be applied to, but is not limited to, application scenarios involving data processing in the fields of e-commerce services, educational services, legal services, medical services, conference services, social network services, financial product services, logistics services and navigation services, for example: data processing of e-commerce services, data processing of educational services, data processing of legal services, etc., which are not limited here.

[0078] By adopting the embodiment of the present application, multi-source heterogeneous data is obtained, the multi-source heterogeneous data is parsed, and a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data is generated, and then the multi-source heterogeneous knowledge graph is integrated to obtain a target navigation graph, and then multi-level description information is generated based on the target navigation graph. Finally, knowledge production is performed based on the multi-level description information to generate a navigation page description document under the target application scenario, thereby achieving the purpose of fully utilizing data from different sources to complete data analysis. Therefore, the system can automatically extract, integrate and present accurate and detailed navigation page description documents (such as operation guides and description documents) according to user needs and scenario changes, greatly improving the efficiency of data use and the level of intelligence, reducing dependence on task system development, fully tapping the value of data, optimizing user operation experience, achieving data value maximization and system intelligence, and the technical effect of automatic generation and updating of knowledge documents, thereby solving the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0079] In an optional embodiment, the multi-source heterogeneous data includes at least part of the following data: unstructured text data; semi-structured page data; structured table data.

[0080] Among them, unstructured text data can be understood as text information without a predefined data model or format, such as emails, social media posts, documents, reports, news articles, books, etc. This type of data is usually composed of natural language and lacks a clear organizational structure, such as titles, paragraphs, or tables. Through technologies such as part-of-speech tagging, named entity recognition, and semantic parsing, unstructured text can be converted into structured data for further analysis and use.

[0081] Semi-structured page data can be understood as web page content, especially data containing markup languages ​​such as HTML, XML, JSON, etc. This type of data contains both text information and structured elements such as titles, lists, tables, etc., but these elements are usually embedded in free text, and the structure may change with the changes in web page design and content.

[0082] Structured tabular data usually exists in a table or database format, with clear columns and rows defined, and each cell containing a specific type of data.

[0083] Multi-source heterogeneous data fusion improves the comprehensiveness, accuracy and intelligence of data by integrating information from different sources and in different formats, providing strong support for the operation, management and decision-making of measurement automation systems.

[0084] In an optional embodiment, in step S22, parsing the multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data includes the following method steps:

[0085] Step S221, extracting unstructured graphs from unstructured text data to obtain a text knowledge graph;

[0086] Step S222, extracting a semi-structured graph from the semi-structured page data to obtain a page knowledge graph;

[0087] Step S223, extracting a structured graph from the structured table data to obtain a table knowledge graph;

[0088] Step S224, determining at least part of the text knowledge graph, the page knowledge graph, and the table knowledge graph as the multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data.

[0089] In an embodiment of the present application, when parsing multi-source heterogeneous data and generating a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, graph extraction can be performed on data in different formats respectively to obtain the corresponding knowledge graph.

[0090] Furthermore, unstructured graph extraction is performed on unstructured text data to obtain a text knowledge graph, wherein the text knowledge graph may be an operation manual graph, a training document graph, etc., which is not limited here. Exemplarily, a text knowledge graph constructed with an operation manual or a business document as a data source may include entities and their relationships regarding system operation procedures, function descriptions, business rules, etc. For example, in the operation manual of a metering automation system, there may be an entity "power trading", under which there may be sub-entities "overview", "management view", etc., and "management view" may be associated with entities such as "user overview" and "operation path". Relationships may include "include", "guide", etc., and attributes may involve specific operation steps, Uniform Resource Locator (URL) links, and other information. The text knowledge graph helps the system understand the business processes and operation details described in the text, and generates accurate operation guides for users.

[0091] A semi-structured graph is extracted from the semi-structured page data to obtain a page knowledge graph, wherein the page knowledge graph can be a page information graph, a page structure graph, etc., which is not limited here. Exemplarily, the page knowledge graph constructed by extracting information from the metering master station system web page can reflect the structure, function and interactive elements of the web page. For example, a page may contain multiple drop-down menus, buttons and tables, and these elements and their relationships are mapped to the graph. Entities can be "drop-down menus", "query buttons", "table data", etc., while relationships may involve "trigger", "display", etc., and attributes include element identifiers (ID), class names, locations, interaction results, etc. The page knowledge graph can help the system identify and understand the dynamic behavior and interface layout of web pages, and provide a basis for intelligent navigation and page description generation.

[0092] A structured graph is extracted from the structured table data to obtain a table knowledge graph, wherein the table knowledge graph can be obtained by constructing triples, which is not limited here. Exemplarily, the table knowledge graph generated from structured data (such as Excel tables or database tables) focuses on the logical relationships and attributes of the data. In the metering system, the table may record data such as equipment type, user information, and transaction records. The constructed graph entities may include "substation", "line", "metering point", etc. The relationships involve "belong to", "association", etc., and the attributes cover specific numerical values ​​and descriptive information. This kind of graph helps the system to conduct in-depth analysis of the data and supports business indicator queries and data analysis tasks.

[0093] Afterwards, at least part of the obtained text knowledge graph, page knowledge graph, and table knowledge graph is determined as a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data. It can be understood that the obtained knowledge graphs are integrated to generate a multi-source heterogeneous knowledge graph. The generated multi-source heterogeneous knowledge graph contains information extracted from different data sources, forming a comprehensive and accurate knowledge system, which can be used to generate detailed operation descriptions, provide intelligent question and answer support, etc., effectively improving the efficiency of data use and the service capabilities of the system.

[0094] In an optional embodiment, in step S221, extracting an unstructured graph from unstructured text data to obtain a text knowledge graph includes the following method steps:

[0095] Step S2211, generating a language model extraction prompt word based on the unstructured text data, wherein the language model extraction prompt word is used to indicate extracting a target element from the unstructured text data, and formatting and outputting the target element according to a preset output format;

[0096] Step S2212, using the target prompt words to guide the pre-trained language model to perform unstructured graph extraction on the unstructured text data to obtain a text knowledge graph.

[0097] In an embodiment of the present application, when extracting unstructured graphs from unstructured text data to obtain a text knowledge graph, a language model extraction prompt word can be first generated based on the unstructured text data, wherein the language model extraction prompt word is used to indicate the extraction of target elements from the unstructured text data, and the target elements are formatted and output according to a preset output format. It can be understood that a series of specific prompt words are designed and generated by analyzing the unstructured text data. These prompt words contain specific instructions and context information, which are used to guide the pre-trained language model to more accurately extract key elements in the text in subsequent steps.

[0098] For example, the design of the prompt word needs to take into account the type of target elements to be extracted and how to structure these elements according to the preset output format. For example, if the target elements are operation steps, function descriptions, etc., the prompt word can be an instruction such as "extract operation steps" or "identify function descriptions", and also include guidance on the output format, such as "output the operation steps in the form of a list", etc., so as to guide the pre-trained language model to generate answers that meet expectations.

[0099] Among them, the pre-trained language model can be a large model or other deep learning model, which is not limited here.

[0100] After generating the language model to extract the prompt words, the target prompt words can be used to guide the pre-trained language model to extract the unstructured graph from the unstructured text data to obtain the text knowledge graph. It can be understood that the pre-trained language model is used to analyze and understand the unstructured text data based on the prompt words generated above. Furthermore, the pre-trained language model will search and extract the target elements in the text according to the instructions in the prompt words, and organize these elements according to the preset format, and finally form a text knowledge graph.

[0101] By leveraging the powerful natural language understanding and generation capabilities of the pre-trained language model, we can mine entities, attributes, relationships, and other information from free text and convert them into a graph structure. For example, from the operation manual, the pre-trained language model may extract elements such as operating steps and URL links associated with entities such as "power trading", "overview", and "management view", and construct the association relationship between entities to form a structured text knowledge graph.

[0102] It can be seen that this application parses unstructured texts such as operating manuals with the help of a large model. Under the guidance of prompt words, the large model understands the text content and extracts key information, converts it into structured entities, attributes and relationships, and generates a text knowledge graph. By utilizing the language understanding and generation capabilities of the large model, free text is converted into a queryable and analyzable knowledge structure, laying the foundation for subsequent data fusion and application. As a result, the system can effectively extract knowledge from unstructured texts, construct text knowledge graphs, and provide a structured and semantic content foundation for subsequent multi-source heterogeneous data fusion and knowledge generation tasks, further improving the intelligence level and service capabilities of the metrology automation system.

[0103] In an optional embodiment, in step S222, extracting a semi-structured graph from the semi-structured page data to obtain a page knowledge graph includes the following method steps:

[0104] Step S2221, using a preset testing tool to obtain target text content corresponding to multiple page elements in the semi-structured page data;

[0105] Step S2222, use the preset regular template to parse the target text content to obtain the page knowledge graph.

[0106] In an embodiment of the present application, when semi-structured page data is subjected to semi-structured graph extraction to obtain a page knowledge graph, a preset test tool may be used to obtain target text content corresponding to multiple page elements in the semi-structured page data. The preset test tool may be Selenium, which is a widely used automated testing tool that can simulate user behavior in a browser, such as clicking, inputting text, etc., to obtain dynamically generated page content. The preset test tool may also be other automated testing and data crawling tools, such as Puppeteer or Playwright, etc., which are not limited here.

[0107] The target text content can be understood as information extracted from specific elements (e.g., elements with predefined class attributes) in a web page by a preset test tool. For example, the target text content can include the attribute name of the drop-down box, such as the text information contained in the HTML element with the class ms-item-component, which usually corresponds to the title or descriptive label of the drop-down menu, such as "Device Type". These names are the titles of the drop-down menu options that users see on the interface, and are used to indicate the function or content category of the drop-down menu.

[0108] The target text content can also include the attribute values ​​of the drop-down box, such as the specific options contained in the HTML element with class ms-select-dropdown. For example, "substation", "interval", "user", "line" and "metering point" are all options in the "equipment type" drop-down box. These attribute values ​​represent the specific items that users can select in the drop-down menu and are important entity information when building a knowledge graph.

[0109] The target text content can also include table header information. For example, HTML elements with class cell are usually used to define cells in a table, where the header information (usually in the first row of the table) provides a description of the table columns. For example, if a table is about power trading, its header may contain information such as "trading date", "trading party", and "trading volume", which are key components of building a page knowledge graph and are used to describe the semantics of the table structure and content.

[0110] It can be seen that by matching predefined class attributes with automated tools (such as Selenium) to obtain the text content of specific elements on the page, all necessary entity, attribute and relationship information can be extracted from the semi-structured page data as accurately and completely as possible, laying a solid foundation for building knowledge graphs and knowledge fusion in subsequent steps. Selenium combined with specific class attribute matching can ensure that even if the page layout is complex or dynamic loading technology is used, the system can effectively locate and collect the required information.

[0111] After obtaining the target text content, the target text content is parsed using a preset regular template to obtain a page knowledge graph. The preset regular template may be a regular expression, which is a powerful text matching and processing tool that can match text according to a predefined pattern and extract the required information.

[0112] By using regular expression templates, the system can accurately identify key entities and attributes such as "equipment type", "substation", "interval", etc. from the acquired text content. This information will be organized into triples (i.e. entity 1-relationship-entity 2), and finally a page knowledge graph will be constructed to describe the structure, function and operation process of the page. The construction of the page knowledge graph helps the system understand and navigate complex pages in the metering automation system, and provide users with accurate page descriptions and operation guides.

[0113] Therefore, for the webpage of the metering master station system, this application uses Selenium combined with regular expressions and other technologies to parse the webpage source code, identify and extract key information in semi-structured data, such as the attribute name and value of the drop-down box, the header of the table, etc. This information is converted into a graph structure to build a page knowledge graph, so that the system can accurately capture the structured elements on the page and enhance the understanding and utilization of the page content.

[0114] In an optional embodiment, the multiple page elements include: a header element, a drop-down box attribute name element, and a drop-down box attribute value element. In step S2221, using a preset test tool to obtain target text content corresponding to the multiple page elements in the semi-structured page data includes the following method steps:

[0115] Step S22211, using a table element locator of a preset test tool to locate the header element, and extracting the header text content corresponding to the header element;

[0116] Step S22212, using a drop-down menu label locator of a preset test tool to locate the drop-down box attribute name element, and extracting the attribute name text content corresponding to the drop-down box attribute name element;

[0117] Step S22213, use the drop-down menu option value locator of the preset testing tool to locate the drop-down box attribute value element, and extract the attribute value text content corresponding to the drop-down box attribute value element.

[0118] In an embodiment of the present application, the multiple page elements include: a header element, a drop-down box attribute name element, and a drop-down box attribute value element, wherein the header element is usually used in the first row of a data table to describe the meaning or data type represented by each column. For example, in an electric power trading data table, the header element may contain titles such as "transaction date", "transaction type", and "transaction volume" to identify the specific meaning of each column of the table. These elements provide users with contextual information about the table data to help users understand the nature and purpose of each column of data.

[0119] The drop-down box attribute name element is a label or element that defines the drop-down box function, and usually contains the title or descriptive text of the drop-down box in the user interface, such as "Device Type", "Time Filter", etc. The drop-down box attribute name element provides users with context for selecting drop-down box options, and is also important entity information when building knowledge graphs.

[0120] The drop-down box attribute value element is the selectable option in the drop-down box, usually composed of <option>Tag definitions, or elements within a specific class (such as "ms-select-dropdown"). For example, in the "Equipment Type" drop-down box, there may be option value elements such as "Substation", "Interval", "User", "Line", and "Metering Point". These elements provide specific options that users can operate and are a key component of the data input and selection process. When building a knowledge graph, the drop-down box attribute value elements are used as attributes of the entity to describe the characteristics or status of the entity, thereby enriching the information and relationships in the graph.

[0121] When using the preset test tool to obtain the target text content corresponding to multiple page elements in the semi-structured page data, the table element locator (HTMLTable Element Locator) of the preset test tool can be used to locate the header element and extract the header text content corresponding to the header element. It can be understood that the preset test tool will use its built-in positioning ability to find HTML elements with the predefined class attribute "cell" on the page. These elements usually represent cells in the table. Through the table element locator, the tool can identify the table structure, and then lock the header row and extract the header text content from it. For example, in the page of power trading, the header may contain descriptions such as "trading date", "trading party", and "trading volume". This information will be extracted and used for subsequent graph construction.

[0122] Also, the drop-down menu label locator (Dropdown Menu LabelLocator) of the preset test tool is used to locate the drop-down box attribute name element, and the attribute name text content corresponding to the drop-down box attribute name element is extracted. Among them, it can be understood that the preset test tool will use a special drop-down menu label locator (such as an element with the "ms-item-component" class) to locate the drop-down box on the page. Once these elements are located, the system will extract the text content contained in these elements, that is, the attribute name of the drop-down box, such as "device type". These attribute names are key information that describes the function of the drop-down box and will be used for subsequent graph construction.

[0123] In addition, the drop-down menu option value locator (Dropdown Menu OptionValueLocator) of the preset test tool is used to locate the drop-down box attribute value element, and extract the attribute value text content corresponding to the drop-down box attribute value element. It can be understood that the preset test tool will use the drop-down menu option value locator (such as an element with the "ms-select-dropdown" class) to find all optional attribute values ​​in the drop-down box. For example, in the "Equipment Type" drop-down box, there may be options such as "Substation", "Interval", "User", "Line" and "Metering Point", and these specific attribute values ​​will be extracted one by one. The extracted attribute value text content will be used to build a knowledge graph to facilitate the system to understand and describe the functions and operation processes of the page.

[0124] Therefore, through the table element locator, drop-down menu label locator and drop-down menu option value locator in the automation tool (i.e., the preset test tool), key information in the web page, such as header elements, drop-down box attribute names and attribute values, can be accurately located and extracted, that is, key text information can be extracted from the semi-structured page data, that is, the present application can efficiently and accurately extract key text information from the semi-structured page data, thereby providing users with more detailed page descriptions and operation guides, and improving the quality of knowledge documents for navigation questions and answers in the metrology automation system.

[0125] In an optional embodiment, in step S23, fusing multi-source heterogeneous knowledge graphs to obtain a target navigation graph includes the following method steps:

[0126] Step S231, according to the preset fusion rules, the multi-source heterogeneous knowledge graphs are fused to obtain the target navigation graph, wherein the preset fusion rules include: consistency verification rules, priority setting rules and conflict resolution rules, the consistency verification rules are used to compare and analyze whether the same entities existing between different graphs in the multi-source heterogeneous knowledge graphs represent the same concept, the priority setting rules are used to assign weights to different graphs in the multi-source heterogeneous knowledge graphs according to the data source attributes of the multi-source heterogeneous knowledge graphs, and the conflict resolution rules are used to provide a multi-resolution mechanism for the contradictory contents existing in the multi-source heterogeneous knowledge graphs.

[0127] In an embodiment of the present application, when multi-source heterogeneous knowledge graphs are fused to obtain a target navigation graph, the multi-source heterogeneous knowledge graphs can be fused according to preset fusion rules to obtain a target navigation graph. Among them, the preset fusion rules include: consistency verification rules, priority setting rules, and conflict resolution rules. Further, the consistency verification rules are used to compare and analyze whether the same entities existing between different graphs in the multi-source heterogeneous knowledge graph represent the same concept. It can be understood that the rules are used to check whether the same entities existing between different graphs in the multi-source heterogeneous knowledge graph actually represent the same concept. For example, descriptions of "substation" and "power station" may appear in different graphs, and consistency verification needs to confirm whether the two can be regarded as equivalent entities. In the system, this usually involves the use of entity matching and synonym libraries to ensure that information related to the same entity can be merged together even if the expressions are different, which is crucial to eliminating redundancy and improving the accuracy of the knowledge graph.

[0128] The priority setting rules are used to assign weights to different graphs in the multi-source heterogeneous knowledge graph according to the data source attributes of the multi-source heterogeneous knowledge graph. It can be understood that the priority setting rules assign weights to different graphs based on the data source attributes of the multi-source heterogeneous knowledge graph. In this application, the priority order is defined as the priority of page structure graph>training document graph>page information graph>operation manual graph, which means that the information in the page structure graph is considered to be the most authoritative and accurate, followed by the training document graph, and so on. This priority setting helps to make decisions when information conflicts, that is, to give priority to information from higher-level graphs.

[0129] Conflict resolution rules are used to provide a multi-resolution mechanism for the contradictory content in multi-source heterogeneous knowledge graphs. It is understandable that even if priority settings are adopted, it is still possible to encounter situations where information from different graphs contradicts each other. Conflict resolution rules provide a processing mechanism for such contradictory content. When automatic processing cannot resolve information conflicts, the system will assist in decision-making based on expert review or preset logical processes. For example, if the description of the "reset" function in the operation manual graph and the page information graph is inconsistent, the system will decide which party's information to adopt based on the conflict resolution rules, or modify the fusion result through expert judgment.

[0130] Through the application of the above fusion rules, it is ensured that when fusing multi-source heterogeneous knowledge graphs, the consistency, authority and conflict issues of data can be effectively resolved. The target navigation graph generated in the end will be a high-quality knowledge graph that integrates all source information, with clear relationships between information and logical self-consistency. Such a graph can provide users with accurate page descriptions and operation guides, which is of great significance for improving the efficiency and accuracy of navigation questions and answers in the metrology automation system.

[0131] In an optional embodiment, in step S24, generating multi-level description information based on the target navigation map includes the following method steps:

[0132] Step S241, performing sentence-level retrieval on the target navigation graph to obtain sentence-level description information;

[0133] Step S242, rearrange the sentence-level description information into paragraph-level description information to obtain paragraph-level description information;

[0134] Step S243, generating document level description information at the paragraph level to obtain document level description information.

[0135] In an embodiment of the present application, when generating multi-level description information based on the target navigation map, the target navigation map can be searched at the sentence level to obtain sentence-level description information. It can be understood that the information in the target navigation map is decomposed into sentence fragments, each of which contains specific operation steps, descriptions or entity information. For example, a sentence like "Click on the module 'Power Trading Module' magnifying glass to view all detailed information." will be extracted. Through sentence-level retrieval, the system can quickly respond to users' basic queries, provide instant feedback, and ensure the speed and directness of the user experience.

[0136] After obtaining the sentence-level description information, the sentence-level description information is rearranged at the paragraph level to obtain the paragraph-level description information. It can be understood as integrating related sentences into paragraphs to construct a description with richer context. For example, all the operation steps related to the "power trading module" are organized together to form a paragraph description on how to use the module so that users can get a more comprehensive operation guide. Paragraph-level descriptions are particularly important in the fine arrangement stage, which can help users better understand the relevance of information and the coherence of the operation process.

[0137] For example, when rearranging sentence-level description information at the paragraph level, natural language processing technology can be used to analyze and compare the semantic content of each sentence to identify sentences with similar meanings or contextually related sentences. Then, considering the logical relationship and contextual coherence between sentences, sentences describing the same operation steps, the same functional module, or the same page are connected to form a coherent and logically clear paragraph to obtain paragraph-level description information.

[0138] After obtaining the paragraph-level description information, the paragraph-level description information is generated at the document level to obtain the document-level description information. It can be understood as further integrating the paragraph-level description information to generate a complete document, which not only contains the specific operation guide required by the user, but also provides background information, concept explanations and possible extended operation details related to the metering master station system. The document-level generation integrates all the collected information to form a structured and logically clear document, which can serve as a comprehensive guide for navigation questions and answers of the metering master station system, thereby ensuring that users can obtain a complete and detailed document when they need in-depth understanding or detailed operation guides, rather than just fragmented information.

[0139] For example, when generating a document level for paragraph-level description information, the topics discussed in the paragraphs can be analyzed, and paragraphs with similar topics can be classified into the same chapter to establish a clear hierarchical structure for the document. Paragraphs with the same topic can then be put together and sorted according to logical order or importance to ultimately generate a complete document.

[0140] Therefore, by generating hierarchical information from sentence level to paragraph level and then to document level, the information is gradually refined and enriched, meeting the different needs of users from quickly obtaining information to deeply understanding the operation process. This process makes full use of the rich entity, attribute and relationship information extracted from the target navigation map, and through multi-level construction, ensures the comprehensiveness, accuracy and ease of use of the description information, and improves the production efficiency and quality of navigation knowledge of the metrology automation system.

[0141] Figure 3 is a flow chart of a knowledge generation system based on multi-source heterogeneous data fusion according to an embodiment of the present application, such as Figure 3 As shown, it includes four processes: unstructured graph extraction, semi-structured graph extraction, multi-source graph fusion and multi-level feature generation.

[0142] Furthermore, in the process of extracting unstructured graphs, it is more complicated to write fixed-format parsing logic for unstructured documents, so the language understanding ability of the large model can be used to perform preliminary parsing of the document. For example, it can be implemented based on the large model prompt words, using the pre-trained large language model as the basis.

[0143] First, a pre-trained large language model is used to extract unstructured text. The model has been pre-trained through large-scale unsupervised learning and has rich semantic understanding and generation capabilities. During the extraction process, the prompt word engineering technology is used, that is, by adding specific prompt words to the input, to guide the model to generate expected answers. These prompt words include question types, scenario restrictions, etc., which help the model better understand and generate corresponding answers.

[0144] For example, the corresponding prompt word structure is as follows:

[0145]

Instruction part

[0146] You are a document information extractor. Your task is to extract the corresponding elements from the following [unstructured document] and format them according to the [output format] for output.

[0147] Unstructured Documents

[0148] 4.1 Electricity Trading

[0149] 4.1.1 Overview

[0150] 4.1.1.1 Management View

[0151] 4.1.1.1.1 User Overview

[0152] Operation process:

[0153] 1. Log in to the system and follow the path: Power Trading--Overview--Management View--Trading User Overview, and enter the page: Web Link;

[0154] 2. Switch to provincial, municipal, district and county level, and power supply station units respectively, and click the "Query" button;

[0155] 3. Switch to different time types respectively and click the "Query" button;

[0156] 4. Click the magnifying glass of the module "Power Trading Module" to view all detailed information;

[0157] 5. Click the magnifying glass of the module "Industry Classification Module" to view all detailed information;

[0158] 6. Click the "User Category Statistics" magnifying glass to view all detailed information;

[0159] 7. Click the "Reset" button and the data will return to the default query data, and the conditions will be the default query conditions.

[0160] Output format

[0161] First level heading:

[0162] Secondary heading:

[0163] Level 3 heading:

[0164] Level 4 heading:

[0165] Tab page name:

[0166] Operation path:

[0167] Web page URL:

[0168] Steps:

[0169] Now, start analyzing and outputting according to the above requirements!

[0170] The expected extraction results are as follows:

[0171] First level heading: Electricity trading

[0172] Secondary heading: Overview

[0173] Level 3 heading: Manage Views

[0174] Level 4 heading: User Overview

[0175] Tab page name: empty

[0176] Operation path: Power Trading--Overview--Management View--Trading User Overview

[0177] Web page URL link: Web page link

[0178] Steps:

[0179] 1. Log in to the system and follow the path: Power Trading--Overview--Management View--Trading User Overview, and enter the page: Web Link;

[0180] 2. Switch to provincial, municipal, district and county level, and power supply station units respectively, and click the "Query" button;

[0181] 3. Switch to different time types respectively and click the "Query" button;

[0182] 4. Click the magnifying glass of the module "Power Trading Module" to view all detailed information;

[0183] 5. Click the magnifying glass of the module "Industry Classification Module" to view all detailed information;

[0184] 6. Click the "User Category Statistics" magnifying glass to view all detailed information;

[0185] 7. Click the "Reset" button and the data will return to the default query data, and the conditions will be the default query conditions.

[0186] In the process of semi-structured graph extraction, the page structure, header information and drop-down box information on the web page can be obtained by parsing the page source code, and the page source code is semi-structured data based on HTML language. Through Selenium and regular templates, the attribute name of the drop-down box is extracted through the class ms-item-component, the attribute value of the drop-down box is extracted through the class ms-select-dropdown, and the table header information is extracted through the class cell.

[0187] In the multi-source graph fusion process, the four types of graphs, namely the operation manual graph, page information graph, training document graph and page structure graph, are effectively combined to supplement and improve the specific tab page information. In order to effectively combine the above four graphs, this application defines a set of reasonable fusion rules:

[0188] Consistency verification: When the same entity exists between different graphs, it is necessary to compare and analyze them to confirm whether they really represent the same concept.

[0189] Priority setting: A weight value is given according to the authority and timeliness of the data source, and the priority weight ranking is defined as follows: page structure map > training document map > page information map > operation manual map.

[0190] Conflict resolution mechanism: When encountering contradictory information, a choice should be made based on the pre-determined priority order. If it is difficult to judge, auxiliary decision-making will be carried out through expert review and other means.

[0191] After completing the fusion of multi-source graphs, it is necessary to use the fused knowledge base to generate the tab page details, so as to achieve the effective fusion of multiple types of graphs and provide users with richer and more detailed tab page descriptions on this basis. At the same time, the reasonable handling of potential conflicts between information is also one of the key links that cannot be ignored in the whole process.

[0192] In the multi-level feature generation process, since smaller document blocks can improve retrieval accuracy, and larger document blocks can make the content generated by the large model more comprehensive, the traditional method of "first retrieving small document blocks and then generating by the large model" may result in poor results in both the retrieval and generation stages. Based on this, when the application segments documents, knowledge fragments of different granularities are valuable in the final retrieval process.

[0193] This application designs the following three levels of information collection:

[0194] Sentence-level information: obtained by segmenting the page description into sentences. Fine-grained knowledge is suitable for the recall stage.

[0195] Paragraph-level information: It contains structured paragraph information; complete semantic knowledge is suitable for the refined sorting stage;

[0196] Document-level information: It contains as much semantic information as possible around the paragraph level. More contextual information is suitable for the large model generation stage.

[0197] This application generates supplementary information of three granularities through a large model to facilitate the decoupling of the block granularity of documents in the two stages of retrieval and generation, and realizes the indexing from small to large of "first sentence granularity retrieval - segment granularity rearrangement - structured block generation".

[0198] It can be seen that this application aims to extract information from unstructured and semi-structured data, integrate this information through multi-source graph fusion technology, and finally generate detailed tab page details. Specifically, this application is divided into four main parts:

[0199] Unstructured graph extraction: Use a pre-trained large-scale language model and use prompt word engineering technology to guide the model to parse and extract information from unstructured text. This method takes advantage of the powerful semantic understanding and generation capabilities of the large model.

[0200] Semi-structured graph extraction: For semi-structured data such as HTML-based web pages, Selenium tools are used in combination with regular expression pattern matching to extract key information. For example, information such as the attribute name, attribute value, and table header of a drop-down box can be identified and extracted based on a specific class attribute value.

[0201] Multi-source graph fusion: Information from different sources (including manuals, page information, training documents, and page structure) is integrated through a set of defined rules. This set of rules includes consistency verification to ensure that the same entity represents the same concept, setting priorities to determine which source of data is more reliable in case of conflict, and using expert review to assist decision-making when automatic processing cannot resolve the conflict.

[0202] Multi-level feature generation: To optimize the balance between retrieval efficiency and content generation quality, a multi-level knowledge fragment collection system is designed. This includes sentence-level, paragraph-level, and document-level information collection methods. This approach allows retrieval to start with fine-grained (sentences) and then gradually transition to content generation at a larger scale (paragraphs or even entire documents), thereby improving the effectiveness and flexibility of the overall process.

[0203] In summary, this application designs a knowledge generation system based on multi-source heterogeneous data fusion, which parses unstructured text, semi-structured HTML information and structured tables into structured graph data, fuses and disambiguates multi-source data, and then uses large model capabilities to generate sentence-level, paragraph-level and document-level description information to realize the knowledge production system of navigation question and answer of the metering master station system. That is, this application achieves the goal of efficiently extracting, integrating and presenting useful information from diverse data sources by comprehensively using advanced natural language processing technology and data mining methods.

[0204] It is easy to understand that the beneficial effects of the data processing method provided by the present application include the following points.

[0205] Beneficial effect (1): Combination of large-scale model-based unstructured document parsing and semi-structured data extraction: Innovative use of pre-trained large-scale language models to understand and parse unstructured text content, and enhance the model's ability to extract specific types of information through prompt word engineering. At the same time, for semi-structured data sources such as web pages, Selenium is combined with regular expressions to efficiently and accurately capture key elements (such as drop-down box attributes and header information). This method, which combines natural language processing technology and web crawler technology, enables the system to extract useful information from data sources in various formats, improving the flexibility and accuracy of information extraction.

[0206] Beneficial effect (2), Multi-source graph fusion strategy and its application: A new multi-source knowledge graph fusion framework is proposed, which aims to integrate knowledge resources from multiple channels such as operation manuals, page information, training materials, and web page structures. By defining a clear consistency verification mechanism, priority rules, and conflict resolution process, it ensures that a high-quality knowledge base can be built even when faced with complex or contradictory information. This method not only promotes the effective integration of cross-domain knowledge, but also provides a solid data foundation for subsequent application scenarios.

[0207] Beneficial effect (3), multi-level feature generation optimizes the retrieval and generation process: In view of the limitations of the existing "retrieve first, then generate" model, this application proposes a novel solution: namely, three granularity levels of information collection methods (sentence level, paragraph level, document level) are designed according to the needs of different stages, and corresponding levels of supplementary information are generated through a large model. This method breaks the boundary between retrieval and generation in traditional methods, allowing the system to more flexibly adjust its workflow to meet the needs of specific application scenarios, thereby significantly improving the overall performance.

[0208] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0209] In addition, it should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0210] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD), including several instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0211] According to the embodiment of the present application, there is also provided Figure 4 A data processing method shown in FIG. Figure 4 is a flow chart of a data processing method according to an embodiment of the present application, such as Figure 4 As shown, the method includes:

[0212] Step S41, acquiring multi-source heterogeneous measurement data;

[0213] Step S42, parsing the multi-source heterogeneous measurement data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous measurement data;

[0214] Step S43, fusing the multi-source heterogeneous knowledge graphs to obtain a measurement navigation graph;

[0215] Step S44, generating multi-level description information based on the measurement navigation map;

[0216] Step S45, performing knowledge production based on the multi-level description information, and generating a navigation question and answer page description document in the power metering data application scenario.

[0217] In the embodiment of the present application, multi-source heterogeneous metering data can be understood as data related to power metering that originates from different data sources and has different structures and formats. Exemplarily, data sources may include internal systems, external network resources, sensors, social media, documents, databases, etc. The data sources are independent of each other, and the data structure, format and semantics may be different, which is not limited here. Data formats include unstructured text (such as emails, reports, social media posts), semi-structured data (such as XML or JSON files, web page HTML code) and structured data (such as table data in a database), etc., which are not limited here.

[0218] It is understandable that multi-source heterogeneous metrology data may refer to data of various types, sources and formats involved in the metrology automation system. Multi-source heterogeneous metrology data may include, but is not limited to, unstructured text (such as operation manuals, training documents), semi-structured HTML information (such as web page architecture and function descriptions), and structured tables (such as metrology data in a database), which are not limited here.

[0219] Multi-source heterogeneous metering data can be obtained through various technical means, such as web crawlers, database queries, document reading, etc., which are not restricted here.

[0220] After acquiring multi-source heterogeneous measurement data, the multi-source heterogeneous measurement data is parsed to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous measurement data. The multi-source heterogeneous knowledge graph can be understood as a collection of graphs extracted and constructed from multi-source heterogeneous measurement data, that is, knowledge is extracted from data of multiple different sources and structures and organized into a collection of graphs. The multi-source heterogeneous knowledge graph uses nodes to represent entities and edges to represent the relationships between entities.

[0221] For example, when parsing multi-source heterogeneous measurement data and generating a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous measurement data, each data source can generate an independent knowledge graph, and these graphs have their own structure and information content, which together constitute a multi-source heterogeneous knowledge graph. Therefore, by converting data from different sources and structures into a unified knowledge representation, the efficiency and effect of data processing and application can be improved.

[0222] Furthermore, for unstructured text, NLP technologies such as entity recognition, relationship extraction and semantic analysis can be used to build a text knowledge graph. For semi-structured HTML data, the knowledge graph related to the structure and content of the web page is generated through document object model DOM parsing and data extraction technology. For structured data, it can be directly converted into graph nodes and relationships through database query or data mapping technology, which is not limited here.

[0223] After generating the multi-source heterogeneous knowledge graphs, the multi-source heterogeneous knowledge graphs are fused to obtain the metering navigation graph, that is, the different knowledge graphs generated above are integrated into a unified target knowledge graph. Among them, the metering navigation graph can be understood as a navigation manual graph, which is used to clearly display the structured knowledge representation of system pages and functions. The metering navigation graph integrates multi-source heterogeneous metering data, so that the system can generate accurate navigation and operation guides, and improve the user's operating efficiency and experience in the system. Exemplarily, in the metering scenario, the metering navigation graph can be understood as a detailed structured knowledge graph generated for a specific page or function of the metering master station system, and its purpose is to provide clear and accurate navigation paths and operation guides.

[0224] For example, when fusing multi-source heterogeneous knowledge graphs, it is necessary to consider the consistency of entities (ensuring that the same entity is represented consistently in different graphs), data conflicts (when the attribute values ​​of entities with the same name are inconsistent in different graphs, they need to be resolved according to predefined rules or algorithms), and information supplementation (using information from multiple graphs to enrich the description of entities), so as to obtain a metering navigation graph. In this way, a more comprehensive and accurate navigation graph can be constructed to guide or explain the use process of a specific system, platform or application.

[0225] After obtaining the metering navigation map, multi-level description information is generated based on the metering navigation map. Exemplarily, the multi-level description information may include sentence-level, paragraph-level, and document-level description information. Furthermore, the sentence-level description information focuses on the simple expression of entities and relationships in the map, which is suitable for quick search or brief definition. The paragraph-level description information provides more detailed information, combining the descriptions of multiple entities and relationships, which helps to understand complex business processes or system operations. The document-level description information is used to comprehensively cover the knowledge of the target system or subject, which can be a user manual, operating guide, or detailed business process documentation, which is not limited here.

[0226] Therefore, by dividing information into multiple levels (sentence level, paragraph level and document level), the system can perform retrieval and information generation at different granularities, thereby finding the best balance between retrieval speed and the level of detail of generated content, ensuring that the knowledge fragments collected from information of different granularities can complement each other and increase the relevance and completeness of the information in the description document. By generating multi-level description information, knowledge information of different granularities can be generated according to user needs and query context.

[0227] After generating multi-level description information, knowledge production is performed based on the multi-level description information to generate a navigation question and answer page description document under the power metering data application scenario, that is, to generate a detailed description document for the navigation page. Among them, the power metering data application scenario can be a navigation and knowledge content generation scenario for the metering master station system, and can also be a scenario for providing question and answer services for the intelligent question and answer system (that is, combining the context of the user's query, intelligently generating page descriptions and operation guides associated with the query, and providing more accurate and more humane question and answer services), and can also be a system document update scenario (that is, when the system is updated or the function is adjusted, the relevant description documents are automatically updated to ensure that the document content is consistent with the system function and reduce maintenance costs), and can also be an online help scenario (that is, when the user encounters a problem during the operation, the description document related to the current page can be quickly accessed, providing instant help information and operation steps, and improving the efficiency of the user in solving problems), etc., which are not limited here.

[0228] Exemplarily, knowledge production based on multi-level description information may include document layout design, logical organization of information, style selection, and final document generation, thereby ensuring that the generated navigation question and answer page description document is both accurate and readable and can meet the needs of target users.

[0229] The navigation question and answer page description document not only includes specific operation guidelines, but also covers multi-level information such as operation logic, data sources, functional significance, etc., so that users can fully understand how to use and operate each page of the metering master station system.

[0230] It is understandable that in the metering scenario, the navigation question and answer page description document can be understood as a document that records in detail the structure, function, operation process and related knowledge of specific pages in the metering master station system. This type of document includes multi-level information from sentence level, paragraph level to document level, aiming to provide system users and maintenance personnel with comprehensive operation guides and page interpretations to improve usage efficiency and user experience.

[0231] In the embodiment of the present application, by acquiring multi-source heterogeneous metering data, the multi-source heterogeneous metering data is parsed, and a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous metering data is generated, and then the multi-source heterogeneous knowledge graph is fused to obtain a metering navigation graph, and then multi-level description information is generated based on the metering navigation graph, and finally, knowledge production is performed based on the multi-level description information to generate a navigation question and answer page description document in the power metering data application scenario. Therefore, the present application parses data of different formats and sources, extracts key information from data of different formats and sources, and forms a unified graph data, thereby improving the utilization rate of data, and eliminating the problem of data islands. In the metering scenario, the metering master station system can fully and accurately understand and process data from multiple sources. Furthermore, by fusing multi-source heterogeneous knowledge graphs, the present application can effectively solve data conflicts and ambiguity problems, and ensure the accuracy and reliability of the knowledge graph.

[0232] At the same time, this application provides a real-time data processing and updating mechanism. Through the generation and fusion of knowledge graphs and the generation of multi-level descriptive information, it can quickly respond to system changes, improve knowledge production efficiency, and ensure the real-time and accuracy of information.

[0233] In addition, the multi-level feature generation strategy allows the system to quickly retrieve and generate information at different granularities, ensuring the real-time and responsiveness of the description document. This end-to-end real-time processing flow enables the system to quickly update and present information after data collection, meeting the needs of real-time refresh and improving the timeliness of data and user experience.

[0234] Finally, the navigation question-and-answer page description document obtained by this application through knowledge production through multi-level description information not only records the page structure and operation process in detail, but also incorporates relevant background knowledge and operation suggestions. It can provide richer and more user-friendly operation step instructions, greatly optimize the user experience and reduce the user's learning cost.

[0235] It can be seen that through the solution of this application, the system can automatically understand and generate information of different granularities, reducing the reliance on manual development of task systems. This automated and intelligent data processing method can not only generate knowledge documents more efficiently and accurately, but also dynamically adjust the generation and presentation of information according to user queries and usage behaviors, giving full play to the value of data.

[0236] The above-mentioned data processing method provided in the embodiment of the present application can be applied to, but is not limited to, application scenarios involving data processing in the fields of e-commerce services, educational services, legal services, medical services, conference services, social network services, financial product services, logistics services and navigation services, for example: data processing of e-commerce services, data processing of educational services, data processing of legal services, etc., which are not limited here.

[0237] By adopting the embodiment of the present application, multi-source heterogeneous metering data is obtained, the multi-source heterogeneous metering data is parsed, and a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous metering data is generated, and then the multi-source heterogeneous knowledge graph is integrated to obtain a metering navigation graph, and then multi-level description information is generated based on the metering navigation graph. Finally, knowledge production is performed based on the multi-level description information to generate a navigation question and answer page description document in the power metering data application scenario, thereby achieving the purpose of fully utilizing data from different sources to complete data analysis. Therefore, the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency and intelligence level of data use, reducing dependence on task system development, fully tapping data value, optimizing user operation experience, and realizing data value maximization and system intelligence. The technical effect of automatic generation and updating of knowledge documents solves the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0238] In an optional embodiment, the data processing method further includes the following method steps:

[0239] Step S46, receiving a power metering help question input by the target user on the navigation question and answer page;

[0240] Step S47, generating an operation step description corresponding to the power metering help question based on the navigation question and answer page description document.

[0241] In an embodiment of the present application, the data processing method further includes: receiving a power metering help question input by a target user on a navigation question and answer page. It can be understood that a user can input specific questions or help regarding power metering operations through the navigation question and answer page provided by the system. For example, the input question may be about how to perform a specific function (such as data query), how to interpret a certain indicator, or seeking help when encountering a specific operational obstacle. The system will receive the user's input and use it as a query condition for subsequent searches and information generation.

[0242] Afterwards, the operation steps corresponding to the power metering help question are generated based on the navigation question and answer page description document. It can be understood that after receiving the user's power metering help question, the system uses the navigation question and answer page description document to generate the operation steps associated with the question. Exemplarily, it can involve information retrieval, information organization, information generation and presentation results, which are not limited here.

[0243] It can be seen that by using the previously constructed knowledge graph and description documents as the information base, information retrieval and customized generation are performed according to the real-time needs of users, realizing the intelligent and personalized navigation question and answer in the power metering scenario. This process not only improves the efficiency of users in solving problems, but also improves the ease of use and user satisfaction of the system, demonstrating the powerful function of the knowledge generation system based on multi-source heterogeneous data fusion in practical applications.

[0244] It should be noted that the preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0245] According to the embodiment of the present application, there is also provided Figure 5 A data processing method shown in FIG. Figure 5 is a flow chart of a data processing method according to an embodiment of the present application, such as Figure 5 As shown, the method includes:

[0246] Step S51, obtaining a data processing request through a first application programming interface, wherein the request data carried in the data processing request includes: multi-source heterogeneous data;

[0247] Step S52, returning a data processing response through the second application programming interface, wherein the response data carried in the data processing response includes: a navigation page description document in the target application scenario, the navigation page description document being generated according to any one of the above-mentioned data processing methods.

[0248] The first application programming interface and the second application programming interface may be the same application programming interface or different application programming interfaces. In an optional embodiment, the interface parameters in the first application programming interface and the second application programming interface may include but are not limited to: interface global identifier, interface signature key, interface timestamp, interface request identifier, system call credential identifier, etc. The first application programming interface may use GET or POST as the interface request method to obtain the file processing request. The second application programming interface may use JSON format to feedback the file processing response.

[0249] The data processing request is used to request data processing of the multi-source heterogeneous data carried. The data processing response can be understood as a response to the data processing request. The data processing response carries a navigation page description document under the target application scenario. The navigation page description document is generated according to any of the above-mentioned data processing methods and will not be elaborated here.

[0250] For a detailed description, please refer to the description of the aforementioned embodiment, which will not be repeated here.

[0251] The above-mentioned data processing method provided in the embodiment of the present application can be applied to, but is not limited to, application scenarios involving data processing in the fields of e-commerce services, educational services, legal services, medical services, conference services, social network services, financial product services, logistics services and navigation services, for example: data processing of e-commerce services, data processing of educational services, data processing of legal services, etc., which are not limited here.

[0252] By adopting the embodiment of the present application, a data processing request is obtained through a first application programming interface, wherein the request data carried in the data processing request includes: multi-source heterogeneous data, and a data processing response is returned through a second application programming interface, wherein the response data carried in the data processing response includes: a navigation page description document under a target application scenario, wherein the navigation page description document is generated according to any one of the above-mentioned data processing methods, thereby achieving the purpose of fully utilizing data from different sources to complete data analysis, so that the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency and intelligence of data use, reducing dependence on task system development, fully tapping the value of data, optimizing user operation experience, achieving data value maximization and system intelligence, and the technical effect of automatic generation and updating of knowledge documents, thereby solving the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0253] It should be noted that the preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0254] According to the embodiment of the present application, there is also provided Figure 6 A data processing method is shown. Figure 6 is a flow chart of a data processing method according to an embodiment of the present application, such as Figure 6 As shown, the method includes:

[0255] Step S61, responding to an input instruction on the operation interface, inputting multi-source heterogeneous data;

[0256] Step S62, in response to the processing instruction acting on the operation interface, a navigation page description document under the target application scenario is displayed on the operation interface, and the navigation page description document is generated according to any one of the above-mentioned data processing methods.

[0257] In the embodiment of the present application, the input instruction can be understood as the input instruction of the user inputting multi-source heterogeneous data on the operation interface, and the processing instruction can be understood as the instruction for processing the multi-source heterogeneous data. After the multi-source heterogeneous data is processed, a navigation page description document under the target application scenario will be generated, and the navigation page description document under the target application scenario will be displayed on the operation interface, wherein the navigation page description document is generated according to any one of the above-mentioned data processing methods, which will not be described in detail here.

[0258] For a detailed description, please refer to the description of the aforementioned embodiment, which will not be repeated here.

[0259] The above-mentioned data processing method provided in the embodiment of the present application can be applied to, but is not limited to, application scenarios involving data processing in the fields of e-commerce services, educational services, legal services, medical services, conference services, social network services, financial product services, logistics services and navigation services, for example: data processing of e-commerce services, data processing of educational services, data processing of legal services, etc., which are not limited here.

[0260] By adopting the embodiment of the present application, by responding to the input instructions acting on the operation interface, inputting multi-source heterogeneous data, responding to the processing instructions acting on the operation interface, the navigation page description document under the target application scenario is displayed on the operation interface, and the navigation page description document is generated according to any of the above-mentioned data processing methods. Thus, the purpose of fully utilizing data from different sources to complete data analysis is achieved, so that the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency and intelligence level of data use, reducing dependence on task system development, fully tapping the value of data, optimizing user operation experience, realizing data value maximization and system intelligence, and the technical effect of automatic generation and updating of knowledge documents, thereby solving the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0261] It should be noted that the preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0262] According to the embodiment of the present application, there is also provided Figure 7 A data processing system is shown. Figure 7 is a structural diagram of a data processing system according to an embodiment of the present application, such as Figure 7 As shown, the system includes:

[0263] Client, used to send multi-source heterogeneous data;

[0264] A server, connected to a client, is used to parse multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, fuse the multi-source heterogeneous knowledge graph to obtain a target navigation graph, generate multi-level description information based on the target navigation graph, and perform knowledge production based on the multi-level description information to generate a navigation page description document under a target application scenario;

[0265] The client is also used to output navigation page description documents.

[0266] In an embodiment of the present application, a data processing system includes a client and a server that are connected to each other. The client is used to send multi-source heterogeneous data to the server. The server is used to parse the received multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, fuse the multi-source heterogeneous knowledge graph to obtain a target navigation graph, generate multi-level description information based on the target navigation graph, and perform knowledge production based on the multi-level description information to generate a navigation page description document in a target application scenario. After that, the server will feed back the navigation page description document to the client, so that the client will output the navigation page description document to feed back to the user.

[0267] For a detailed description, please refer to the description of the aforementioned embodiment, which will not be repeated here.

[0268] The above-mentioned data processing system provided in the embodiment of the present application can be applied to, but is not limited to, application scenarios involving data processing in the fields of e-commerce services, educational services, legal services, medical services, conference services, social network services, financial product services, logistics services and navigation services, for example: data processing of e-commerce services, data processing of educational services, data processing of legal services, etc., which are not limited here.

[0269] By adopting the embodiment of the present application, the purpose of fully utilizing data from different sources to complete data analysis is achieved through the above-mentioned data processing system, so that the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency of data use and the level of intelligence, reducing dependence on task system development, fully tapping the value of data, optimizing user operation experience, realizing the technical effect of maximizing data value and system intelligence, and automatically generating and updating knowledge documents, thereby solving the technical problems of low knowledge production efficiency in related technologies and the inability to fully utilize data from different sources to complete data analysis.

[0270] It should be noted that the preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0271] According to an embodiment of the present application, a device embodiment for implementing the above data processing method is also provided. Figure 8 is a structural diagram of a data processing device according to an embodiment of the present application, such as Figure 8 As shown, the device comprises:

[0272] The first acquisition module 801 is used to acquire multi-source heterogeneous data;

[0273] The first parsing module 802 is used to parse the multi-source heterogeneous data and generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data;

[0274] The first fusion module 803 is used to fuse multi-source heterogeneous knowledge graphs to obtain a target navigation graph;

[0275] The first generating module 804 is used to generate multi-level description information based on the target navigation map;

[0276] The second generation module 805 is used to perform knowledge production based on the multi-level description information and generate a navigation page description document in a target application scenario.

[0277] Optionally, the multi-source heterogeneous data includes at least part of the following data: unstructured text data; semi-structured page data; structured table data.

[0278] Optionally, the first parsing module 802 is also used to: perform unstructured graph extraction on unstructured text data to obtain a text knowledge graph; perform semi-structured graph extraction on semi-structured page data to obtain a page knowledge graph; perform structured graph extraction on structured table data to obtain a table knowledge graph; and determine at least part of the graphs in the text knowledge graph, the page knowledge graph, and the table knowledge graph as a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data.

[0279] Optionally, the first parsing module 802 is also used to: generate a language model extraction prompt word based on unstructured text data, wherein the language model extraction prompt word is used to indicate the extraction of target elements from the unstructured text data, and format the target elements according to a preset output format; use the target prompt word to guide the pre-trained language model to extract an unstructured graph from the unstructured text data to obtain a text knowledge graph.

[0280] Optionally, the first parsing module 802 is also used to: use a preset testing tool to obtain target text content corresponding to multiple page elements in the semi-structured page data; use a preset regular template to parse the target text content to obtain a page knowledge graph.

[0281] Optionally, the multiple page elements include: a header element, a drop-down box attribute name element, and a drop-down box attribute value element. The above-mentioned first parsing module 802 is also used to: use the table element locator of the preset testing tool to locate the header element, and extract the header text content corresponding to the header element; use the drop-down menu label locator of the preset testing tool to locate the drop-down box attribute name element, and extract the attribute name text content corresponding to the drop-down box attribute name element; use the drop-down menu option value locator of the preset testing tool to locate the drop-down box attribute value element, and extract the attribute value text content corresponding to the drop-down box attribute value element.

[0282] Optionally, the first fusion module 803 is also used to: fuse multi-source heterogeneous knowledge graphs according to preset fusion rules to obtain a target navigation graph; wherein the preset fusion rules include: consistency verification rules, priority setting rules and conflict resolution rules, the consistency verification rules are used to compare and analyze whether the same entities existing between different graphs in the multi-source heterogeneous knowledge graphs represent the same concept, the priority setting rules are used to assign weights to different graphs in the multi-source heterogeneous knowledge graphs according to the data source attributes of the multi-source heterogeneous knowledge graphs, and the conflict resolution rules are used to provide a multi-solution mechanism for the contradictory content existing in the multi-source heterogeneous knowledge graphs.

[0283] Optionally, the above-mentioned first generation module 804 is also used to: perform sentence-level retrieval on the target navigation map to obtain sentence-level description information; perform paragraph-level rearrangement on the sentence-level description information to obtain paragraph-level description information; perform document-level generation on the paragraph-level description information to obtain document-level description information.

[0284] By adopting the embodiment of the present application, multi-source heterogeneous data is obtained, the multi-source heterogeneous data is parsed, and a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data is generated, and then the multi-source heterogeneous knowledge graph is integrated to obtain a target navigation graph, and then multi-level description information is generated based on the target navigation graph. Finally, knowledge production is performed based on the multi-level description information to generate a navigation page description document under the target application scenario, thereby achieving the purpose of fully utilizing data from different sources to complete data analysis. Therefore, the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency and intelligence level of data use, reducing dependence on task system development, fully tapping data value, optimizing user operation experience, achieving data value maximization and system intelligence, and the technical effect of automatic generation and updating of knowledge documents, thereby solving the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0285] It should be noted here that the above-mentioned first acquisition module 801, first analysis module 802, first fusion module 803, first generation module 804 and second generation module 805 correspond to steps S21 to S25 in the embodiment, and the five modules are the same as the instances and application scenarios implemented by the corresponding steps, but are not limited to the contents disclosed in the above-mentioned embodiment. It should be noted that the above-mentioned modules or units can be hardware components or software components stored in a memory and processed by one or more processors, and the above-mentioned modules can also run in the server 10 provided in the embodiment.

[0286] According to an embodiment of the present application, another device embodiment for implementing the above data processing method is also provided. Fig. 9 is a structural diagram of another data processing device according to an embodiment of the present application, such as Fig. 9 As shown, the device comprises:

[0287] The second acquisition module 901 is used to acquire multi-source heterogeneous measurement data;

[0288] The second parsing module 902 is used to parse the multi-source heterogeneous metrology data and generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous metrology data;

[0289] The second fusion module 903 is used to fuse multi-source heterogeneous knowledge graphs to obtain a measurement navigation graph;

[0290] The third generating module 904 is used to generate multi-level description information based on the measurement navigation map;

[0291] The fourth generation module 905 is used to produce knowledge based on the multi-level description information and generate a navigation question and answer page description document in the power metering data application scenario.

[0292] Optionally, the above-mentioned device also includes: a receiving module, used to receive the power metering help question entered by the target user on the navigation question and answer page; and generate an operation step description corresponding to the power metering help question based on the navigation question and answer page description document.

[0293] By adopting the embodiment of the present application, multi-source heterogeneous metering data is obtained, the multi-source heterogeneous metering data is parsed, and a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous metering data is generated, and then the multi-source heterogeneous knowledge graph is integrated to obtain a metering navigation graph, and then multi-level description information is generated based on the metering navigation graph. Finally, knowledge production is performed based on the multi-level description information to generate a navigation question and answer page description document in the power metering data application scenario, thereby achieving the purpose of fully utilizing data from different sources to complete data analysis. Therefore, the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency and intelligence level of data use, reducing dependence on task system development, fully tapping data value, optimizing user operation experience, and realizing data value maximization and system intelligence. The technical effect of automatic generation and updating of knowledge documents solves the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0294] It should be noted here that the above-mentioned second acquisition module 901, second analysis module 902, second fusion module 903, third generation module 904 and fourth generation module 905 correspond to steps S41 to S45 in the embodiment, and the five modules are the same as the instances and application scenarios implemented by the corresponding steps, but are not limited to the contents disclosed in the above-mentioned embodiment. It should be noted that the above-mentioned modules or units can be hardware components or software components stored in a memory and processed by one or more processors, and the above-mentioned modules can also run in the server 10 provided in the embodiment.

[0295] According to an embodiment of the present application, another device embodiment for implementing the above data processing method is also provided. Fig.10 is a structural diagram of another data processing device according to an embodiment of the present application, such as Fig.10 As shown, the device comprises:

[0296] The third acquisition module 1001 is used to acquire a data processing request through the first application programming interface, wherein the request data carried in the data processing request includes: multi-source heterogeneous data;

[0297] The return module 1002 is used to return a data processing response through a second application programming interface, wherein the response data carried in the data processing response includes: a navigation page description document in a target application scenario, wherein the navigation page description document is generated according to any one of the above-mentioned data processing methods.

[0298] By adopting the embodiment of the present application, a data processing request is obtained through a first application programming interface, wherein the request data carried in the data processing request includes: multi-source heterogeneous data, and a data processing response is returned through a second application programming interface, wherein the response data carried in the data processing response includes: a navigation page description document under a target application scenario, wherein the navigation page description document is generated according to any one of the above-mentioned data processing methods, thereby achieving the purpose of fully utilizing data from different sources to complete data analysis, so that the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency and intelligence of data use, reducing dependence on task system development, fully tapping the value of data, optimizing user operation experience, achieving data value maximization and system intelligence, and the technical effect of automatic generation and updating of knowledge documents, thereby solving the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0299] It should be noted that the third acquisition module 1001 and the return module 1002 correspond to step S51 and step S52 in the embodiment, and the two modules and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules or units can be hardware components or software components stored in a memory and processed by one or more processors, and the above modules can also run in the server 10 provided in the embodiment.

[0300] According to an embodiment of the present application, another device embodiment for implementing the above data processing method is also provided. Fig.11 is a structural diagram of another data processing device according to an embodiment of the present application, such as Fig.11 As shown, the device comprises:

[0301] An input module 1101 is used to respond to input instructions acting on the operation interface and input multi-source heterogeneous data;

[0302] The display module 1102 is used to respond to the processing instruction acting on the operation interface and display the navigation page description document in the target application scenario on the operation interface, where the navigation page description document is generated according to any one of the above-mentioned data processing methods.

[0303] By adopting the embodiment of the present application, by responding to the input instructions acting on the operation interface, inputting multi-source heterogeneous data, responding to the processing instructions acting on the operation interface, the navigation page description document under the target application scenario is displayed on the operation interface, and the navigation page description document is generated according to any of the above-mentioned data processing methods. Thus, the purpose of fully utilizing data from different sources to complete data analysis is achieved, so that the system can automatically extract, integrate and present accurate and detailed operation guides and description documents according to user needs and scenario changes, greatly improving the efficiency and intelligence level of data use, reducing dependence on task system development, fully tapping the value of data, optimizing user operation experience, realizing data value maximization and system intelligence, and the technical effect of automatic generation and updating of knowledge documents, thereby solving the technical problems of low knowledge production efficiency and inability to fully utilize data from different sources to complete data analysis in related technologies.

[0304] It should be noted that the input module 1101 and the display module 1102 correspond to step S61 and step S62 in the embodiment, and the two modules and the corresponding steps implement the same examples and application scenarios, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules or units can be hardware components or software components stored in a memory and processed by one or more processors, and the above modules can also run in the server 10 provided in the embodiment.

[0305] It should be noted that the preferred implementation scheme involved in the above embodiments of the present application is the same as the scheme provided in the embodiments, as well as the application scenario and implementation process, but is not limited to the scheme provided in the embodiments.

[0306] An embodiment of the present application may provide a computing device. Fig.12 is a structural block diagram of a computing device according to an embodiment of the present application. Fig.12 As shown, the computing device A may include: one or more ( Fig.12 Only one is shown) processor 1202, memory 1204, storage controller, and peripheral interface, wherein the peripheral interface can be connected to a radio frequency module, an audio module, a display screen, etc., which is not limited here.

[0307] The above-mentioned computing device A can be understood as an integrated intelligent terminal, including but not limited to a server, a desktop computer, a PC (Personal Computer), a model all-in-one machine, etc., and the computing device can be pre-installed with the model described in the above-mentioned embodiment of the present application.

[0308] Specifically, the computing device A can preset multiple types of models, including but not limited to models in the fields of natural language processing, visual processing, speech processing, code processing, multimodal task processing, etc., so as to provide a variety of model choices. In different product forms, the computing device A can support one or more model usage methods, including but not limited to model training, model calling, model fine-tuning, model deployment, model reasoning and application, etc. In some product forms, the computing device A also supports model management, including but not limited to multi-type model management (supporting the management of multiple types of models such as discriminants and generators), model version control (supporting the control of different model versions), model evaluation (based on model evaluation tools, evaluating the performance and effect of the model), etc. In other product forms, the computing device A can also create applications based on models, provide API calling capabilities, and can call models to the created applications through API interfaces, while providing application management tools to achieve management and monitoring of applications.

[0309] Furthermore, the computing device A can also include data management (supporting the creation and management of model tuning data sets), a training center (providing rich training resources to help users learn and master AI technology), and basic management and control capabilities (providing enterprise-level basic management and control capabilities to ensure the security and efficient operation of the system). Through the above functions, a comprehensive, integrated AI development, training, deployment and application device is provided.

[0310] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implementing the methods in the above embodiments. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0311] The processor may call the executable program stored in the memory through the transmission device to execute the method described in any one of the above embodiments.

[0312] Those skilled in the art will understand that Fig.12 The structure shown is for illustration only. The computing device A may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (MID), a PAD, or other terminal devices. Fig.12 It does not limit the structure of the above computing device. For example, computing device A may also include Fig.12 More or fewer components (such as network interfaces, display devices, etc.) shown in the figure, or having Fig.12 Different configurations are shown.

[0313] An embodiment of the present application may provide an electronic device. Fig.13 is a structural block diagram of an electronic device according to an embodiment of the present application. Fig.13 As shown, the electronic device may include: an input-output device 132 ; a memory 134 and a processor 136 , wherein the processor 136 is connected to the input-output device 132 and the memory 134 via a bus 138 .

[0314] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the methods and devices in the embodiments of the present application, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, implementing the methods in the above embodiments. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0315] The processor may call the executable program stored in the memory through the transmission device to execute the method described in any one of the above embodiments.

[0316] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0317] The embodiment of the present application further provides a computer-readable storage medium. Optionally, in this embodiment, the computer-readable storage medium can be used to store the program code executed by the data processing method provided in the above embodiment.

[0318] Optionally, in this embodiment, the computer-readable storage medium may be located in any one of the computer terminals in a computer terminal group in a computer network, or in any one of the mobile terminals in a mobile terminal group.

[0319] An embodiment of the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements any one of the above-mentioned data processing methods.

[0320] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0321] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0322] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0323] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0324] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), mobile hard disk, disk or CD-ROM and other media that can store program codes.

[0325] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.< / option>

Claims

1. A data processing method, characterized in that: include: Acquire multi-source heterogeneous data; Parsing the multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data; Fusing the multi-source heterogeneous knowledge graphs to obtain a target navigation graph; Generate multi-level description information based on the target navigation map; Knowledge production is performed based on the multi-level description information to generate a navigation page description document for a target application scenario.

2. The data processing method according to claim 1, characterized in that: The multi-source heterogeneous data includes at least part of the following data: Unstructured text data; Semi-structured page data; Structured tabular data.

3. The data processing method according to claim 2, characterized in that: Parsing the multi-source heterogeneous data to generate the multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data includes: Extracting unstructured graphs from the unstructured text data to obtain a text knowledge graph; Performing semi-structured graph extraction on the semi-structured page data to obtain a page knowledge graph; Performing structured graph extraction on the structured table data to obtain a table knowledge graph; Determine at least part of the text knowledge graph, the page knowledge graph, and the table knowledge graph as the multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data.

4. The data processing method according to claim 3, characterized in that: Extracting unstructured graphs from the unstructured text data to obtain the text knowledge graph includes: Generating a language model extraction prompt word based on the unstructured text data, wherein the language model extraction prompt word is used to indicate extracting a target element from the unstructured text data, and formatting and outputting the target element according to a preset output format; The target prompt words are used to guide the pre-trained language model to perform unstructured graph extraction on the unstructured text data to obtain the text knowledge graph.

5. The data processing method according to claim 3, characterized in that: Extracting a semi-structured graph from the semi-structured page data to obtain the page knowledge graph includes: Using a preset testing tool to obtain target text content corresponding to multiple page elements in the semi-structured page data; The target text content is parsed using a preset regular template to obtain the page knowledge graph.

6. The data processing method according to claim 5, characterized in that: The multiple page elements include: a table header element, a drop-down box attribute name element, and a drop-down box attribute value element. Using the preset test tool to obtain the target text content corresponding to the multiple page elements in the semi-structured page data includes: Using the table element locator of the preset test tool to locate the header element, and extracting the header text content corresponding to the header element; The drop-down menu label locator of the preset test tool is used to locate the drop-down box attribute name element, and the attribute name text content corresponding to the drop-down box attribute name element is extracted; The drop-down menu option value locator of the preset test tool is used to locate the drop-down box attribute value element, and the attribute value text content corresponding to the drop-down box attribute value element is extracted.

7. The data processing method according to claim 1, characterized in that: The target navigation graph is obtained by fusing the multi-source heterogeneous knowledge graphs, including: According to a preset fusion rule, the multi-source heterogeneous knowledge graphs are fused to obtain the target navigation graph; Among them, the preset fusion rules include: consistency verification rules, priority setting rules and conflict resolution rules. The consistency verification rules are used to compare and analyze whether the same entities existing between different graphs in the multi-source heterogeneous knowledge graph represent the same concept. The priority setting rules are used to assign weights to different graphs in the multi-source heterogeneous knowledge graph according to the data source attributes of the multi-source heterogeneous knowledge graph. The conflict resolution rules are used to provide a multi-solution mechanism for the contradictory content existing in the multi-source heterogeneous knowledge graph.

8. The data processing method according to any one of claims 1 to 7, characterized in that: Generating the multi-level description information based on the target navigation map includes: Perform sentence-level retrieval on the target navigation graph to obtain sentence-level description information; Rearranging the sentence-level description information at paragraph level to obtain paragraph-level description information; The paragraph-level description information is subjected to document-level generation to obtain document-level description information.

9. A data processing method, characterized in that: include: Acquire multi-source heterogeneous measurement data; Parsing the multi-source heterogeneous metrology data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous metrology data; The multi-source heterogeneous knowledge graphs are integrated to obtain a metrology navigation graph; Generate multi-level description information based on the metering navigation map; Knowledge production is performed based on the multi-level description information to generate a navigation question and answer page description document for the power metering data application scenario.

10. The data processing method according to claim 9, characterized in that: The data processing method further includes: Receive power metering help questions input by target users on the navigation question and answer page; Based on the navigation question and answer page description document, an operation step description corresponding to the power metering help question is generated.

11. A data processing method, characterized in that: include: Obtaining a data processing request through a first application programming interface, wherein the request data carried in the data processing request includes: multi-source heterogeneous data; A data processing response is returned through a second application programming interface, wherein the response data carried in the data processing response includes: a navigation page description document in a target application scenario, wherein the navigation page description document is generated according to the data processing method described in any one of claims 1 to 8.

12. A data processing method, characterized in that: include: In response to input instructions acting on the operation interface, input multi-source heterogeneous data; In response to a processing instruction acting on the operation interface, a navigation page description document in a target application scenario is displayed on the operation interface, wherein the navigation page description document is generated according to the data processing method described in any one of claims 1 to 8.

13. A data processing system, characterized in that: include: Client, used to send multi-source heterogeneous data; A server, connected to the client, for parsing the multi-source heterogeneous data to generate a multi-source heterogeneous knowledge graph corresponding to the multi-source heterogeneous data, fusing the multi-source heterogeneous knowledge graph to obtain a target navigation graph, generating multi-level description information based on the target navigation graph, and performing knowledge production based on the multi-level description information to generate a navigation page description document under a target application scenario; The client is also used to output the navigation page description document.

14. An electronic device, characterized in that: include: A memory storing an executable program; A processor, used to run the program, wherein the program executes the data processing method according to any one of claims 1 to 12 when running.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is executed, the device where the computer-readable storage medium is located is controlled to execute the data processing method according to any one of claims 1 to 12.

16. A computer program product, characterized in that The invention comprises a computer program, which implements the data processing method according to any one of claims 1 to 12 when being executed by a processor.

Images