Intelligent questioning and answering method, device and equipment for network operation and maintenance, storage medium and product

Abstract

The invention discloses a network operation and maintenance intelligent question answering method and device, equipment, a storage medium and a product, and relates to the technical field of data query, and the network operation and maintenance intelligent question answering method comprises the following steps: when a query question which is input by a user and is related to network operation and maintenance is received, performing word segmentation processing on the query question, obtaining a query keyword and a problem variable word; querying to obtain network resource data and reference knowledge based on the query keyword and the problem variable word; the query question, the network resource data and the reference knowledge are input into a preset question and answer model to optimize a query result, an optimized question and answer result is obtained, and the question and answer result comprises reference document information associated with the question and answer result. The problem that the working efficiency is difficult to improve due to the fact that operation and maintenance personnel spend a lot of time to screen and integrate information can be avoided.

Description

Technical Field

[0001] This application relates to the field of network operation and maintenance technology, and in particular to a network operation and maintenance intelligent question-and-answer method, device, equipment, storage medium and product. Background Technology

[0002] In the basic network operation and maintenance scenarios of large and medium-sized government and enterprise units, the operation and maintenance personnel often need to spend a lot of time filtering and integrating information when carrying out daily operation and maintenance and emergency troubleshooting. This is due to the complexity of the network environment, the variety of equipment, and the vast amount of knowledge and information. As a result, it is difficult to improve work efficiency. Summary of the Invention

[0003] The main purpose of this application is to provide a network operation and maintenance intelligent question-and-answer method, device, equipment, storage medium and product, which aims to solve the technical problem that operation and maintenance personnel spend a lot of time screening and integrating information, resulting in difficulty in improving work efficiency.

[0004] To achieve the above objectives, this application proposes an intelligent question-and-answer method for network operation and maintenance, which includes:

[0005] When a user inputs a query related to network operation and maintenance, the query is segmented to obtain query keywords and question variable words.

[0006] Based on the query keywords and the question variables, network resource data and reference knowledge are obtained through querying.

[0007] The query question, the network resource data, and the reference knowledge are input into a preset question-and-answer model to optimize the query results, thereby obtaining optimized question-and-answer results. The question-and-answer results include reference document information associated with the question-and-answer results.

[0008] In one embodiment, the step of retrieving network resource data and reference knowledge based on the query keywords and the question variable words includes:

[0009] In a distributed manner, a sub-database of the network knowledge database is used to search for whether a knowledge fragment is related to the query keywords and question variable words.

[0010] If a correlation exists, a preset number of reference knowledge fragments are selected based on the correlation threshold set for each of the related sub-libraries, and the source document of the reference knowledge fragment is determined.

[0011] The aforementioned knowledge fragments and the aforementioned source documents are used as reference knowledge.

[0012] Based on the aforementioned question variables, network resource data is retrieved.

[0013] In one embodiment, the step of querying and obtaining network resource data based on the question variable includes:

[0014] Determine the target field corresponding to the user's permission authentication level, and query the field ledger model to see if there is a matching field for the question variable in the target field;

[0015] If the matching field exists, then search for the query method corresponding to the matching field in the data table ledger model;

[0016] If the query method is a knowledge base query, then the target sub-library corresponding to the permission authentication level is determined in the network resource database, and a distributed query is performed in the target sub-library to find the first knowledge fragment related to the question variable word and the source document corresponding to the first knowledge fragment;

[0017] If the query method is a real-time query, then through the real-time query interface, the second knowledge fragment related to the question variable word and the source document corresponding to the second knowledge fragment are obtained, wherein the second knowledge fragment is within the query range of the permission authentication level;

[0018] The first knowledge fragment, the second knowledge fragment, and the source documents corresponding to the first and second knowledge fragments are used as network resource data.

[0019] In one embodiment, if the query method is a real-time query, the step of retrieving the second knowledge fragment related to the question variable and the corresponding source document through the real-time query interface includes:

[0020] If the query method is real-time query, then query the query conditions corresponding to the matching field in the field ledger model;

[0021] The query expression is obtained by combining the query conditions corresponding to different matching fields;

[0022] By using the real-time query interface corresponding to the permission authentication level, the initial second knowledge fragment related to the question variable word corresponding to the query expression can be obtained.

[0023] Based on the number of times the initial second knowledge fragment matches the matching keywords, the initial second knowledge fragment is filtered to obtain a target number of second knowledge fragments, and the source document corresponding to the second knowledge fragment is determined, wherein the target number is determined based on the throughput of the real-time query interface.

[0024] In one embodiment, the step of performing word segmentation on the query question to obtain query keywords and question variable words includes:

[0025] Query the key variable words in the query question, wherein the key variable words include line number, equipment brand and equipment model;

[0026] Extract the remaining variable terms from the query questions that do not include the key variable terms;

[0027] The key variable words and the remaining variable words are used as question variable words;

[0028] The extracted query question is then subjected to keyword extraction to obtain the query keywords.

[0029] In one embodiment, the step of performing word segmentation on the query question to obtain query keywords and question variable words upon receiving a user-input query question includes:

[0030] Obtain the original document and preprocess it to obtain the target document after format conversion. The preprocessing includes format conversion, document splitting, and removal of table of contents and header.

[0031] The target document is uploaded to a network knowledge database. The target document is then sliced ​​in the network knowledge database based on preset slicing rules. The sliced ​​knowledge fragments belonging to the same document are stored in the same sub-database of the network knowledge database.

[0032] Furthermore, to achieve the above objectives, this application also proposes a data query device, which includes:

[0033] The word segmentation module is used to segment the query question into words when it receives a query question related to network operation and maintenance input from a user, so as to obtain query keywords and question variable words.

[0034] The query module is used to retrieve network resource data and reference knowledge based on the query keywords and the question variable words;

[0035] The optimization module is used to input the query question, the network resource data, and the reference knowledge into a preset question-answering model to optimize the query results and obtain optimized question-answering results, wherein the question-answering results include the reference document information associated with the question-answering results.

[0036] In addition, to achieve the above objectives, this application also proposes a data query device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the network operation and maintenance intelligent question-and-answer method as described above.

[0037] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the network operation and maintenance intelligent question-and-answer method described above.

[0038] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the network operation and maintenance intelligent question-and-answer method described above.

[0039] One or more technical solutions proposed in this application have at least the following technical effects:

[0040] Compared to the network operation and maintenance scenarios in large and medium-sized government and enterprise units, where maintenance personnel often spend a significant amount of time filtering and integrating information due to the complex network environment, diverse equipment, and vast amount of knowledge resources during routine maintenance and emergency troubleshooting, resulting in low work efficiency, this application, upon receiving a user's query related to network operation and maintenance, performs word segmentation on the query to obtain query keywords and question variable words; based on the query keywords and question variable words, it retrieves network resource data and reference knowledge; and it inputs the query, network resource data, and reference knowledge into a preset question-answering model to perform the query. The results are optimized to obtain optimized question-and-answer results, which include reference document information associated with the question-and-answer results. After receiving the user's input question, this application can automatically perform word segmentation on the question and retrieve network resource data and reference knowledge based on the query keywords and question variable words obtained from the word segmentation. This eliminates the need for operation and maintenance personnel to filter and integrate information. Furthermore, the query results are optimized through a preset question-and-answer model, and finally, a large model of question-and-answer results is output. The question-and-answer results also include reference document information associated with the question-and-answer results, which avoids the problem of operation and maintenance personnel spending a lot of time filtering and integrating information, resulting in poor work efficiency. Attached Figure Description

[0041] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0042] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1This is a flowchart illustrating an embodiment of the intelligent question-and-answer method for network operation and maintenance provided in this application.

[0044] Figure 2 This is a flowchart illustrating the question-and-answer process of the intelligent question-and-answer method for network operation and maintenance in this application.

[0045] Figure 3 This is a schematic diagram of the knowledge base classification tree for the intelligent question-answering method for network operation and maintenance in this application;

[0046] Figure 4 This is a flowchart illustrating the network knowledge database retrieval process of the intelligent question-and-answer method for network operation and maintenance in this application.

[0047] Figure 5 This is a flowchart illustrating the network resource data acquisition process of the intelligent question-and-answer method for network operation and maintenance in this application.

[0048] Figure 6 This is a flowchart illustrating Embodiment 2 of the intelligent question-and-answer method for network operation and maintenance in this application.

[0049] Figure 7 A word segmentation flowchart is provided for the intelligent question-answering method for network operation and maintenance in this application;

[0050] Figure 8 Flowchart for the knowledge base construction of the intelligent question-answering method for network operation and maintenance in this application;

[0051] Figure 9 This is a schematic diagram of the module structure of the data query device according to an embodiment of this application;

[0052] Figure 10 This is a schematic diagram of the device structure of the hardware operating environment involved in the intelligent question-and-answer method for network operation and maintenance in this embodiment of the application.

[0053] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0054] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0055] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0056] The main solution of this application embodiment is as follows: when a user inputs a query question related to network operation and maintenance, the query question is segmented to obtain query keywords and question variable words; based on the query keywords and question variable words, network resource data and reference knowledge are retrieved; the query question, the network resource data, and the reference knowledge are input into a preset question-answering model to optimize the query results and obtain optimized question-answering results, wherein the question-answering results include reference document information associated with the question-answering results.

[0057] In the basic network operation and maintenance scenarios of large and medium-sized government and enterprise units, the operation and maintenance personnel often need to spend a lot of time filtering and integrating information when carrying out daily operation and maintenance and emergency troubleshooting. This is due to the complexity of the network environment, the variety of equipment, and the vast amount of knowledge and information. As a result, it is difficult to improve work efficiency.

[0058] After receiving a user's input question, this application can automatically segment the question into words and retrieve network resource data and reference knowledge based on the query keywords and question variables obtained from the segmentation. This eliminates the need for operations and maintenance personnel to filter and integrate information. Furthermore, it optimizes the query results through a preset question-and-answer model and finally outputs the answer and reference document information of a large model. This avoids the problem of operations and maintenance personnel spending a lot of time filtering and integrating information, which leads to poor work efficiency.

[0059] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or data query device capable of performing the above functions. The following description uses a data query device as an example to illustrate this embodiment and the subsequent embodiments.

[0060] Based on this, embodiments of this application provide a network operation and maintenance intelligent question-answering method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the intelligent question-and-answer method for network operation and maintenance in this application.

[0061] In this embodiment, the intelligent question-and-answer method for network operation and maintenance includes steps S10 to S30:

[0062] Step S10: When a user inputs a query related to network operation and maintenance, the query is segmented to obtain query keywords and question variable words.

[0063] It should be noted that the execution entity in this embodiment is a data query device. This data query device has a front-end entry point, which is used to obtain the query question input by the user. The data query device then performs word segmentation on the query question to obtain query keywords and question variable words.

[0064] Furthermore, the query keywords are a list of keywords used for knowledge base queries. The knowledge base is built on RAG (Retrieval-augmented generation), a platform that combines retrieval and generation artificial intelligence technologies. Its operation typically involves retrieving text fragments related to the user's question from a text database. Question variables include IP addresses, device names, line numbers, interface names, and brand / model numbers.

[0065] Furthermore, after receiving a query, the data query device will authenticate the user to obtain the user's authorization level. Specifically, the data query device will determine whether the user is a trusted user and whether they are on a trusted network, and determine the user's authorization level based on the judgment result.

[0066] Specifically, the authorization level is divided into three levels: Level 1 (high-privilege users on trusted networks), Level 2 (high-privilege users on untrusted networks), and Level 3 (low-privilege users on untrusted networks). High-privilege users refer to relevant users within the organization who are responsible for network operation and management, low-privilege users are other users within the organization, and trusted networks are usually the organization's internal networks.

[0067] Step S20: Based on the query keywords and the question variable words, retrieve network resource data and reference knowledge;

[0068] Understandably, the data query device also includes a query model. This model is used to retrieve online resource data and reference knowledge based on query keywords and question variables.

[0069] Step S30: Input the query question, the network resource data, and the reference knowledge into a preset question-answering model to optimize the query results and obtain optimized question-answering results, wherein the question-answering results include reference document information associated with the question-answering results.

[0070] It should be noted that the reference document information includes the document name. The data query device integrates the query question, network resource data, and reference knowledge into the question-answering model, significantly improving the accuracy, interpretability, and practicality of intelligent question answering. This is particularly suitable for the complex network operation and maintenance environments of large and medium-sized government and enterprise units. Figure 2 , Figure 2 A flowchart of the question-and-answer process is provided.

[0071] Specifically, the data query device acts as a professional network operations and maintenance expert, based on a pre-set question-and-answer model that prompts the query question. It is responsible for answering the user's questions about network operations and maintenance by referring to reference knowledge, real-time network resources, and the query question entered by the user.

[0072] Furthermore, the pre-defined question-and-answer model will link the user's historical questions and answers during the process of generating question-and-answer results. It will not only be based on the user's current question, but also refer to the user's previous questions and answers, thereby generating more coherent and personalized answers.

[0073] Understandably, through intelligent word segmentation, multi-source data fusion, and large-scale model question-answering optimization, the accuracy, interpretability, security, and practicality of the network operation and maintenance question-answering system have been significantly improved, providing efficient, intelligent, and compliant knowledge service support for large and medium-sized government and enterprise units.

[0074] In one feasible implementation, step S20 may include the following steps:

[0075] In a distributed manner, a sub-database of the network knowledge database is used to search for whether a knowledge fragment is related to the query keywords and question variable words.

[0076] It should be noted that the network knowledge database is divided into several sub-databases: technical document repositories from different vendors, regulatory standards, industry standards, basic network standards, network protocol standards, case studies, and network technology books. Specifically, based on the richness of the documents and actual needs, the vendor technical document repository can be further subdivided according to device type, and other sub-databases can also be further subdivided layer by layer according to the document content, ultimately resulting in, as follows: Figure 3 The diagram shows a knowledge base classification tree. In the distributed knowledge retrieval method, the data query device performs parallel queries on all leaf nodes of the knowledge base classification tree to check for the existence of knowledge fragments related to the query keywords and question variables.

[0077] If a correlation exists, a preset number of reference knowledge fragments are selected based on the correlation threshold set for each of the related sub-libraries, and the source document of the reference knowledge fragment is determined.

[0078] Understandably, each sub-database in the network knowledge database has its own relevance threshold and a preset number of segments that can be retained. By adjusting the relevance threshold and the preset number, the weight of each sub-database in the final result can be flexibly adjusted, making the output results more focused on frequently used documents, which helps optimize retrieval efficiency and result quality. The data query device simultaneously searches all sub-databases in parallel based on the knowledge base search keyword list and some question variables related to basic operation and maintenance knowledge. For each sub-database, multiple knowledge segments are retrieved and sorted by relevance. The device then filters out the preset number of knowledge segments that exceed the threshold and determines the source document corresponding to that knowledge segment, i.e., determines the document name mentioned before the knowledge segmentation, referring to... Figure 4 , Figure 4 A flowchart for retrieving information from a network knowledge database is provided. Figure 4 The parameters T0-Tx, N0-Nx, etc., can be set according to actual needs. Specifically, for different sub-databases, different search strategies, output result quantities, and data result similarity thresholds can be set based on actual conditions such as importance and representativeness. That is, each sub-database has a corresponding number of filters and similarity thresholds.

[0079] The aforementioned knowledge fragments and the aforementioned source documents are used as reference knowledge.

[0080] It should be noted that the data query device uses reference knowledge fragments and their corresponding source documents as reference knowledge.

[0081] Based on the aforementioned question variables, network resource data is retrieved.

[0082] Understandably, the data query device retrieves network resource data based on the question variable keywords.

[0083] In one feasible implementation, the step of querying network resource data based on the question variable words includes:

[0084] Determine the target field corresponding to the user's permission authentication level, and query the field ledger model to see if there is a matching field for the question variable in the target field;

[0085] It should be noted that the field ledger model includes the following information: table ID, field, field name, field access level, field to be matched, and field matching method. Because the field ledger model stores field access levels, the data query device needs to first determine the target field range corresponding to the user's access level, i.e., filter out the target fields, and then determine whether the question variable matches these fields, i.e., whether matching fields exist.

[0086] Furthermore, when the data query device queries network resource data through the query model, it carries the permission authentication level. The query model will query network resource data with the corresponding openness level according to the permission authentication level.

[0087] Furthermore, the data query device establishes a data source ledger, which includes a data table ledger model and a data field ledger model. Each data source ledger must have an access level set, allowing users to query data with access permissions at or below the access level.

[0088] If the openness level is 0, large model queries are not currently available.

[0089] If the access level is 1, the data is accessible to queries with access level 1, but not to queries with access levels 2 and 3.

[0090] At open level 2, the data is accessible to queries with authentication levels 1 and 2, but not to queries with authentication level 3.

[0091] If the access level is 3, the data is open to all queries with access levels 1, 2, and 3.

[0092] Furthermore, when the openness level is 0, the data query device will not return data to the user.

[0093] If the matching field exists, then search for the query method corresponding to the matching field in the data table ledger model;

[0094] Understandably, the data table ledger model contains the following information: data table ID, data table name, database connection information where the table resides, large model query method, data table access level, and a list of associated RAG knowledge base document IDs. After obtaining the matching field, the data query device will search for the corresponding query method in the data table ledger model based on that matching field.

[0095] Specifically, the query methods for network resource databases include static network resource data retrieval from the knowledge base and real-time network data query interface queries. For network resource data tables with small data volumes and low update frequencies, the data query device converts each row of data into a knowledge fragment and automatically uploads it to the corresponding network resource data knowledge base of the RAG platform via a scheduled task that calls the RAG platform interface. The data table name is equivalent to the document name. During network resource data queries, these data can be directly retrieved by the retrieval capabilities provided by the RAG platform. For network resource data tables with large data volumes or high real-time requirements, the real-time network data query interface is used to search and match variable values ​​such as IP address, device name, line number, interface name, brand, and model.

[0096] If the query method is a knowledge base query, then the target sub-library corresponding to the permission authentication level is determined in the network resource database, and a distributed query is performed in the target sub-library to find the first knowledge fragment related to the question variable word and the source document corresponding to the first knowledge fragment;

[0097] It should be noted that the network resource database comprises three knowledge bases: a network resource data knowledge base (Level 1), a network resource data knowledge base (Level 2), and a network resource data knowledge base (Level 3), used to store network resource data knowledge fragments corresponding to the openness level. The target sub-base is the knowledge base corresponding to the user's access level. The data query device performs a distributed query in the knowledge base corresponding to the user's access level, based on the query method for the reference knowledge fragment, to find the first knowledge fragment related to the question variable and the source document corresponding to the first knowledge fragment.

[0098] Furthermore, the data query device contains all the data from data interfaces / knowledge bases with lower openness levels. Specifically, the data query device converts data tables into data fields of corresponding openness levels. For example, it converts data rows from a data table with an openness level of 3 into knowledge fragments. Depending on the required openness level data, the conversion is divided into data field knowledge fragment conversion for fields with an openness level of 3, data fields with openness levels of 3 and 2, and data field knowledge fragment conversion for fields with openness levels of 1-3.

[0099] Specifically, knowledge fragments converted from data fields with an openness level of 3 can be queried or stored by all data interfaces / knowledge bases at levels 1, 2, and 3; knowledge fragments converted from data fields with an openness level of 2 can be queried or stored by data interfaces / knowledge bases at levels 1 and 2; and knowledge fragments converted from data fields with an openness level of 1 can only be queried or stored by data interfaces / knowledge bases at level 1. Similarly, the conversion of data rows from data tables with openness levels of 2 and 1 into knowledge fragments can be understood.

[0100] Furthermore, when converting data rows of a data table with an openness level of 3 into knowledge fragments of data fields with an openness level of 3, key-value pairs are constructed according to the field name and the corresponding field data value to obtain the knowledge fragment in JSON format for that row of data. Similarly, the implementation method for converting knowledge fragments of other openness level data tables can be obtained.

[0101] If the query method for the large model in the data table ledger model is "RAG Knowledge Base Static Network Resource Data Knowledge Base Retrieval", then for the JSON format knowledge fragments of the obtained row data, it is further determined whether the number of characters in the knowledge fragment is greater than the fragmentation threshold of 256 (the threshold can be adjusted according to the RAG platform limitations, knowledge base retrieval settings, and large model throughput). If it is greater than the threshold, then fragmentation is performed, that is, the JSON containing key-value pairs is split into multiple JSONs. Each knowledge fragment must retain the key-value pairs of the data field whose "match field" value is "yes" (i.e., the matching field is displayed for fragment matching) for the retrieval of knowledge fragments. Finally, one or more knowledge fragments are obtained.

[0102] Furthermore, after completing the knowledge fragment conversion of data fields with openness levels of 1-3 in a data table with an openness level of 3, each knowledge fragment is written into a document. The RAG platform interface is then called to upload the knowledge fragment document to the RAG knowledge bases "Network Resource Data Knowledge Base (Level 3)," "Network Resource Data Knowledge Base (Level 2)," and "Network Resource Data Knowledge Base (Level 1)" respectively. The slicing method is set to adapt to the JSON format to ensure that one JSON is sliced ​​into one knowledge fragment. The same logic applies to other cases.

[0103] Furthermore, if the large model query method in the data table ledger model is "real-time network data query interface", the interface will query network data in real time and generate a knowledge fragment response of the corresponding level after being called.

[0104] If the query method is a real-time query, then through the real-time query interface, the second knowledge fragment related to the question variable word and the source document corresponding to the second knowledge fragment are obtained, wherein the second knowledge fragment is within the query range of the permission authentication level;

[0105] Understandably, when a field match is successful and the query method is real-time query, the data query device can obtain a second knowledge fragment that meets the conditions and mark the source document by calling the real-time interface. This can provide users with accurate, real-time, and traceable network operation and maintenance knowledge support while ensuring access security.

[0106] The first knowledge fragment, the second knowledge fragment, and the source documents corresponding to the first and second knowledge fragments are used as network resource data.

[0107] It should be noted that the data query device treats the first knowledge fragment, the second knowledge fragment, and the source documents corresponding to the aforementioned knowledge fragments as online resource data, referring to... Figure 5 , Figure 5 A flowchart for obtaining network resource data is provided.

[0108] In one feasible implementation, if the query method is a real-time query, the step of retrieving the second knowledge fragment related to the question variable and the corresponding source document through the real-time query interface includes:

[0109] If the query method is real-time query, then query the query conditions corresponding to the matching field in the field ledger model;

[0110] Understandably, query conditions include =, like, and in, which are common matching methods for database fields and values. When the query method is real-time query, the data query device will query the query conditions corresponding to the matching fields in the field ledger model. These query conditions are used to construct the final SQL or API request parameters to obtain relevant information from the real-time data source.

[0111] Specifically:

[0112]

[0113] The query expression is obtained by combining the query conditions corresponding to different matching fields;

[0114] It should be noted that when the data query device obtains a query expression by combining query conditions, it uses OR logic to associate different query conditions in order to obtain more data fragments.

[0115] By using the real-time query interface corresponding to the permission authentication level, the initial second knowledge fragment related to the question variable word corresponding to the query expression can be obtained.

[0116] It is understandable that different permission authentication levels correspond to different real-time query interfaces. The data query device retrieves the initial second knowledge fragments related to the question variable words through the real-time query interface corresponding to the permission authentication level. By binding the user's permission authentication level with different real-time query interfaces, differentiated, refined, and compliant network resource access services can be provided to different user groups while ensuring data security.

[0117] Based on the number of times the initial second knowledge fragment matches the matching keywords, the initial second knowledge fragment is filtered to obtain a target number of second knowledge fragments, and the source document corresponding to the second knowledge fragment is determined, wherein the target number is determined based on the throughput of the real-time query interface.

[0118] It should be noted that the data query device sorts the data in reverse order based on the number of matching keywords hit by each initial second knowledge fragment, and takes the first M data rows as knowledge fragments as the final result response. The default value of M is dynamically calculated based on the amount of response data, that is, the number of characters in the final response data does not exceed the threshold. The threshold is 1024 by default, which can be adjusted according to the actual throughput capacity of the large model.

[0119] In this embodiment, a highly secure, highly accurate, and highly interpretable intelligent question-answering system is constructed by combining natural language questions, variable word recognition, multi-source knowledge retrieval, and access control mechanisms, and by introducing a question-answering model to optimize the results.

[0120] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 6 Step S20, the network operation and maintenance intelligent question-and-answer method further includes steps S01 to S04:

[0121] Step S01: Query the key variable words in the query question, wherein the key variable words include line number, equipment brand and equipment model;

[0122] Understandably, in order to prevent non-standard but crucial variables such as device number, asset number, and alarm ID from being incorrectly segmented or misjudged, the data query device will call external APIs to query key variable words that cannot be matched by simple regular expressions, and extract the key variable words.

[0123] Step S02: Extract the query questions that do not include the key variable words to obtain the remaining variable words;

[0124] It should be noted that the data query device extracts variable values ​​such as IP, device name, interface name, brand, and model from the remaining content using regular expression matching.

[0125] Step S03: The key variable words and the remaining variable words are used as problem variable words.

[0126] Understandably, the data query device uses key variable words and remaining variable words as question variable words, referring to... Figure 7 , Figure 7 A flowchart for word segmentation processing is provided.

[0127] Step S04: Extract keywords from the extracted query question to obtain query keywords;

[0128] Understandably, the data query device uses a large model (knowledge base retrieval keyword extraction large model Prompt) to identify the remaining keywords of the input question and obtain the query keywords.

[0129] Specifically, the large model uses Prompt to extract key network variables from user questions by identifying keywords in the knowledge base retrieval process. It prompts the large model to further extract key concepts and variable values ​​from the user question input based on conceptual knowledge in the network operations and maintenance field and the already identified variable values. Simultaneously, through examples, the large model is guided to combine, transform, and refine some abbreviations, shortened expressions, and colloquialisms of concepts based on its understanding of the user question, ultimately obtaining a keyword list for knowledge base retrieval.

[0130] In one possible implementation, the following steps are included prior to step S10:

[0131] Obtain the original document and preprocess it to obtain the target document after format conversion. The preprocessing includes format conversion, document splitting, and removal of table of contents and header.

[0132] It should be noted that the data query device processes the original document through multiple steps, including format conversion, document segmentation, and removal of table of contents and headers, using a document processing script, before uploading it to the corresponding knowledge base of the RAG platform. Removing the table of contents and headers prevents keyword searches after segmentation from hitting table of contents and header fragments, reducing interference with the actual text content. The RAG platform knowledge base stores knowledge fragments for each document. During retrieval, the RAG platform knowledge base must be specified, and all knowledge fragments from all documents in that base are matched using a preset relevance matching algorithm. Finally, the system outputs knowledge fragments exceeding a set relevance threshold, along with the corresponding document information and relevance score.

[0133] The target document is uploaded to a network knowledge database. The target document is then sliced ​​in the network knowledge database based on preset slicing rules. The sliced ​​knowledge fragments belonging to the same document are stored in the same sub-database of the network knowledge database.

[0134] Understandably, after the target document is uploaded to the online knowledge database, the data query device will slice it based on preset slices using the RAG platform document slicing tool, and store the sliced ​​data fragments belonging to the same document in the same sub-database.

[0135] Specifically, the default slicing rules are as follows: For documents in doc / docx / ppt / pdf / txt / md / json formats, the document structure is automatically identified, and segments are generated according to the hierarchical structure, producing a tree-structured segmentation with hierarchical information; for csv / xls / xlsx documents, they are sliced ​​by line, with each line being a slice, and each slice containing a header and corresponding data rows; for jpg / png / img images, they are recognized by OCR and parsed into a complete slice. In actual processing, the slicing rules can be adjusted according to the document's characteristics, referring to... Figure 8 , Figure 8 A flowchart for knowledge base construction is provided.

[0136] In this implementation, by performing structured word segmentation on user questions and standardizing preprocessing of original documents before orderly storage into the knowledge base, the accuracy of knowledge retrieval, the completeness of variable identification, the efficiency of system response, and the interpretability and security of knowledge content are significantly improved.

[0137] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the intelligent question-and-answer method for network operation and maintenance in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0138] This application also provides a data query device, please refer to... Figure 9 The data query device includes:

[0139] The word segmentation module 10 is used to segment the query question into words when it receives a query question related to network operation and maintenance input from a user, so as to obtain query keywords and question variable words.

[0140] Query module 20 is used to retrieve network resource data and reference knowledge based on the query keywords and the question variable words;

[0141] The optimization module 30 is used to input the query question, the network resource data, and the reference knowledge into a preset question-answering model to optimize the query results and obtain optimized question-answering results, wherein the question-answering results include reference document information associated with the question-answering results.

[0142] Optionally, the word segmentation module includes:

[0143] The extraction submodule is used to query the key variable words in the query question, wherein the key variable words include line number, equipment brand and equipment model; extract the remaining variable words from the query question that does not include the key variable words; extract keywords from the extracted query question to obtain query keywords; and use the key variable words and the remaining variable words as question variable words.

[0144] The slicing submodule is used to acquire the original document and preprocess it to obtain a target document after format conversion. The preprocessing includes format conversion, document segmentation, and removal of the table of contents and headers. The target document is then uploaded to a network knowledge database. In the network knowledge database, the target document is sliced ​​based on preset slicing rules, and the sliced ​​knowledge fragments belonging to the same document are stored in the same sub-database of the network knowledge database.

[0145] Optionally, the query module includes:

[0146] The filtering submodule is used to perform distributed queries in the sub-libraries of the network knowledge database to determine whether there are knowledge fragments related to the query keywords and question variables. If there are related fragments, a preset number of reference knowledge fragments are selected based on the relevance threshold set for each sub-library, and the source document of the reference knowledge fragments is determined. The reference knowledge fragments and the source document are used as reference knowledge. Based on the question variables, network resource data is retrieved.

[0147] Optionally, the filtering submodule includes:

[0148] The query unit is used to determine the target field corresponding to the user's permission authentication level, and to query in the field ledger model whether the question variable term exists in the target field; if the matching field exists, the query method corresponding to the matching field is searched in the data table ledger model; if the query method is a knowledge base query, the target sub-library corresponding to the permission authentication level is determined in the network resource database, and a distributed query is performed in the target sub-library for the first knowledge fragment related to the question variable term and the source document corresponding to the first knowledge fragment; if the query method is a real-time query, the second knowledge fragment related to the question variable term and the source document corresponding to the second knowledge fragment are obtained through the real-time query interface, wherein the second knowledge fragment is within the query range of the permission authentication level; the first knowledge fragment, the second knowledge fragment, and the source document corresponding to the first knowledge fragment and the second knowledge fragment are used as network resource data.

[0149] Optionally, the query unit includes:

[0150] The query subunit is used to query the query conditions corresponding to the matching field in the field ledger model if the query method is real-time query; combine the query conditions corresponding to different matching fields to obtain a query expression; query the initial second knowledge fragment related to the question variable word corresponding to the query expression through the real-time query interface corresponding to the permission authentication level; filter the initial second knowledge fragment based on the number of times the initial second knowledge fragment hits the matching keyword to obtain a target number of second knowledge fragments, and determine the source document corresponding to the second knowledge fragment, wherein the target number is determined based on the throughput of the real-time query interface.

[0151] The data query device provided in this application, employing the intelligent question-and-answer method for network operation and maintenance described in the above embodiments, can solve the technical problem of data query. Compared with the prior art, the beneficial effects of the data query device provided in this application are the same as those of the intelligent question-and-answer method for network operation and maintenance described in the above embodiments, and other technical features in the data query device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0152] This application provides a data query device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the network operation and maintenance intelligent question-and-answer method in Embodiment 1 above.

[0153] The following is for reference. Figure 10 The diagram illustrates a structural schematic of a data query device suitable for implementing embodiments of this application. The data query device in these embodiments may include, but is not limited to, mobile terminals such as mobile phones, laptops, tablets, digital broadcast receivers, PDAs (Personal Digital Assistants), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 10 The data query device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0154] like Figure 10As shown, the data query device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the data query device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the data query device to communicate wirelessly or wiredly with other devices to exchange data. Although the figures show data query devices with various systems, it should be understood that it is not required to implement or possess all of the systems shown. More or fewer systems may be implemented alternatively.

[0155] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0156] The data query device provided in this application, employing the network operation and maintenance intelligent question-and-answer method described in the above embodiments, can solve the technical problem of data query. Compared with the prior art, the beneficial effects of the data query device provided in this application are the same as those of the network operation and maintenance intelligent question-and-answer method described in the above embodiments, and other technical features of this data query device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0157] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0158] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0159] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the network operation and maintenance intelligent question-and-answer method described in the above embodiments.

[0160] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0161] The aforementioned computer-readable storage medium may be included in the data query device; or it may exist independently and not assembled into the data query device.

[0162] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by a data query device, the data query device: upon receiving a user-input query related to network operation and maintenance, performs word segmentation on the query to obtain query keywords and question variable words; based on the query keywords and question variable words, retrieves network resource data and reference knowledge; and inputs the query, the network resource data, and the reference knowledge into a preset question-answering model to optimize the query results, obtaining optimized question-answering results, wherein the question-answering results include reference document information associated with the question-answering results.

[0163] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0164] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0165] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0166] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., computer programs) for executing the above-described intelligent question-and-answer method for network operation and maintenance, thereby solving the technical problem of data query. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the intelligent question-and-answer method for network operation and maintenance provided in the above embodiments, and will not be repeated here.

[0167] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the network operation and maintenance intelligent question-and-answer method described above.

[0168] The computer program product provided in this application can solve the technical problem of data query. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the intelligent question-and-answer method for network operation and maintenance provided in the above embodiments, and will not be repeated here.

[0169] The above description is only a part of the embodiments of this application and does not limit the scope of protection of this application. All equivalent structural transformations made under the technical concept of this application and using the content of this application specification and drawings, or direct / indirect applications in other related technical fields, are included in the scope of protection of this application.

Claims

1. A network operation and maintenance intelligent question-and-answer method, characterized in that, The intelligent question-and-answer method for network operation and maintenance includes: When a user inputs a query related to network operation and maintenance, the query is segmented to obtain query keywords and question variable words. Based on the query keywords and the question variables, network resource data and reference knowledge are obtained through querying. The query question, the network resource data, and the reference knowledge are input into a preset question-and-answer model to optimize the query results, thereby obtaining optimized question-and-answer results. The question-and-answer results include reference document information associated with the question-and-answer results.

2. The intelligent question-and-answer method for network operation and maintenance as described in claim 1, characterized in that, The steps for retrieving network resource data and reference knowledge based on the query keywords and the question variables include: In a distributed manner, a sub-database of the network knowledge database is used to search for whether a knowledge fragment is related to the query keywords and question variable words. If a correlation exists, a preset number of reference knowledge fragments are selected based on the correlation threshold set for each of the related sub-libraries, and the source document of the reference knowledge fragment is determined. The aforementioned knowledge fragments and the aforementioned source documents are used as reference knowledge. Based on the aforementioned question variables, network resource data is retrieved.

3. The intelligent question-and-answer method for network operation and maintenance as described in claim 2, characterized in that, The step of querying network resource data based on the question variable words includes: Determine the target field corresponding to the user's permission authentication level, and query the field ledger model to see if there is a matching field for the question variable in the target field; If the matching field exists, then search for the query method corresponding to the matching field in the data table ledger model; If the query method is a knowledge base query, then the target sub-library corresponding to the permission authentication level is determined in the network resource database, and a distributed query is performed in the target sub-library to find the first knowledge fragment related to the question variable word and the source document corresponding to the first knowledge fragment; If the query method is a real-time query, then through the real-time query interface, the second knowledge fragment related to the question variable word and the source document corresponding to the second knowledge fragment are obtained, wherein the second knowledge fragment is within the query range of the permission authentication level; The first knowledge fragment, the second knowledge fragment, and the source documents corresponding to the first and second knowledge fragments are used as network resource data.

4. The intelligent question-and-answer method for network operation and maintenance as described in claim 3, characterized in that, If the query method is a real-time query, the step of retrieving the second knowledge fragment related to the question variable and the corresponding source document through the real-time query interface includes: If the query method is real-time query, then query the query conditions corresponding to the matching field in the field ledger model; The query expression is obtained by combining the query conditions corresponding to different matching fields; By using the real-time query interface corresponding to the permission authentication level, the initial second knowledge fragment related to the question variable word corresponding to the query expression can be obtained. Based on the number of times the initial second knowledge fragment matches the matching keywords, the initial second knowledge fragment is filtered to obtain a target number of second knowledge fragments, and the source document corresponding to the second knowledge fragment is determined, wherein the target number is determined based on the throughput of the real-time query interface.

5. The intelligent question-and-answer method for network operation and maintenance as described in claim 1, characterized in that, The step of performing word segmentation on the query question to obtain query keywords and question variable words includes: Query the key variable words in the query question, wherein the key variable words include line number, equipment brand and equipment model; Extract the remaining variable terms from the query questions that do not include the key variable terms; The key variable words and the remaining variable words are used as question variable words; The extracted query question is then subjected to keyword extraction to obtain the query keywords.

6. The intelligent question-and-answer method for network operation and maintenance as described in claim 1, characterized in that, Before the step of receiving a user's input query question and performing word segmentation on the query question to obtain query keywords and question variable words, the following steps are included: Obtain the original document and preprocess it to obtain the target document after format conversion. The preprocessing includes format conversion, document splitting, and removal of table of contents and header. The target document is uploaded to a network knowledge database. The target document is then sliced ​​in the network knowledge database based on preset slicing rules. The sliced ​​knowledge fragments belonging to the same document are stored in the same sub-database of the network knowledge database.

7. A data query device, characterized in that, The device includes: The word segmentation module is used to segment the query question into words when it receives a query question related to network operation and maintenance input from a user, so as to obtain query keywords and question variable words. The query module is used to retrieve network resource data and reference knowledge based on the query keywords and the question variable words; The optimization module is used to input the query question, the network resource data, and the reference knowledge into a preset question-answering model to optimize the query results and obtain optimized question-answering results, wherein the question-answering results include the reference document information associated with the question-answering results.

8. A data query device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the network operation and maintenance intelligent question-and-answer method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the network operation and maintenance intelligent question-and-answer method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the network operation and maintenance intelligent question-and-answer method as described in any one of claims 1 to 6.

Images