A data processing method, apparatus, computer, storage medium and program product

Abstract

Embodiments of the present application disclose a data processing method, apparatus, computer, storage medium, and program product, relating to the field of computer technologies. The method includes: obtaining a service processing request text sent by a first service object in an intelligent conversation page, parsing the service processing request text to obtain service semantic information and intent requirement information; determining a service database associated with the first service object; obtaining a first target type indicated by the service semantic information, obtaining a first service data set belonging to the first target type from classification data respectively corresponding to N service types in the service database, performing information retrieval in the first service data set, obtaining retrieval content associated with a first target component, generating a service result for the service processing request text based on the intent requirement information and the retrieval content, and displaying the service result in the intelligent conversation page. By adopting the present application, the data retrieval efficiency can be improved and the labor cost can be saved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular, to a data processing method, apparatus, computer, storage medium, and program product. Background Art

[0002] With the progress of science and technology, information management technology has developed highly. The transformation of file storage from paper archiving to electronic file storage has realized the evolution from traditional to digital. Electronic file storage has introduced computer technology, scanning, and OCR (Optical Character Recognition) technology, making the storage, management, and retrieval of files more efficient and flexible. The traditional storage method is to store through key-value pairs. When retrieving, it is necessary to ensure that the key-values are exactly the same to retrieve the corresponding content in the database, which is inefficient and has a high query threshold. Summary of the Invention

[0003] Embodiments of the present application provide a data processing method, apparatus, computer, storage medium, and program product, which can reduce the retrieval volume and improve the retrieval efficiency through classified storage, and can perform targeted data retrieval by identifying the semantics of the service processing request text, thereby reducing the retrieval threshold.

[0004] On the one hand, an embodiment of the present application provides a data processing method, including:

[0005] Obtain a service processing request text sent by a first service object in an intelligent conversation page, and parse the service processing request text to obtain service semantic information and intention requirement information; the intelligent conversation page includes an object avatar corresponding to the first service object and an object message text box associated with the object avatar, and the object message text box contains the service processing request text;

[0006] Obtain a first object identifier of the first service object, and based on the first object identifier, determine a service database associated with the first object identifier; the service database includes storage data associated with virtual asset association data corresponding to N service types respectively; the virtual asset association data under each service type includes component asset data for at least one component of an aircraft; N is a positive integer;

[0007] Obtain a first target type indicated by the service semantic information, obtain a first service data set belonging to the first target type from the classification data corresponding to the N service types in the service database, and determine a first target component associated with the service semantic information based on the service semantic information; the N service types include the first target type;

[0008] Retrieve information in the first business data set to obtain retrieval content associated with the first target component, generate a business result for the business processing request text based on the intent requirement information and the retrieval content, display an intelligent avatar in the intelligent conversation page, and an intelligent message text box containing the business result associated with the intelligent avatar; the intelligent avatar is used for conversation interaction with the object avatar in the intelligent conversation page.

[0009] On the one hand, an embodiment of the present application provides a data processing device, which includes:

[0010] A data parsing module, configured to obtain a business processing request text sent by a first business object in the intelligent conversation page, parse the business processing request text to obtain business semantic information and intent requirement information; the intelligent conversation page includes an object avatar corresponding to the first business object and an object message text box associated with the object avatar, and the object message text box contains the business processing request text;

[0011] A data determination module, configured to obtain a first object identifier of the first business object, and based on the first object identifier, determine a business database associated with the first object identifier; the business database includes stored data associated with virtual asset association data corresponding to N business types respectively; the virtual asset association data under each business type includes component asset data for at least one component of an aircraft; N is a positive integer;

[0012] A data acquisition module, configured to obtain a first target type indicated by the business semantic information, obtain a first business data set belonging to the first target type from the classification data corresponding to the N business types in the business database, and determine a first target component associated with the business semantic information based on the business semantic information; the N business types include the first target type;

[0013] A data retrieval module, configured to retrieve information in the first business data set to obtain retrieval content associated with the first target component, generate a business result for the business processing request text based on the intent requirement information and the retrieval content, display an intelligent avatar in the intelligent conversation page, and an intelligent message text box containing the business result associated with the intelligent avatar; the intelligent avatar is used for conversation interaction with the object avatar in the intelligent conversation page.

[0014] In a possible implementation manner, the data processing device further includes a data storage module, and the data storage module is specifically configured to perform the following operations:

[0015] Obtain virtual asset association data for an aircraft sent by the first business object; the aircraft is composed of at least two components;

[0016] Obtain the asset data classification rules, and classify the virtual asset association data of the aircraft through N business types in the asset data classification rules to obtain N classification data; each classification data corresponds to one business type;

[0017] Generate storage data according to the N classification data, and store the storage data in the business database associated with the first business object.

[0018] In a possible implementation manner, the N business types include asset business data types, basic business data types, and transaction business data types; the N classification data include first classification data, second classification data, and third classification data; when the data storage module is used to classify the virtual asset association data of the aircraft through the N business types in the asset data classification rules to obtain N classification data, the data storage module is specifically used to perform the following operations:

[0019] Obtain the data key fields corresponding to the asset business data type; the data key fields include usage record key fields, maintenance record key fields, and depreciation record key fields;

[0020] In the virtual asset association data of the aircraft, obtain the historical usage record corresponding to the usage record key field, the historical maintenance record corresponding to the maintenance record key field, and the historical depreciation record corresponding to the depreciation record key field, and determine the historical usage record, historical maintenance record, and historical depreciation record as the first classification data corresponding to the asset business data type;

[0021] Obtain the factory key fields corresponding to the basic business data type, and in the virtual asset association data of the aircraft, obtain the factory association data corresponding to the factory key fields, and determine the factory association data as the second classification data corresponding to the basic business data type;

[0022] Obtain the transaction key fields corresponding to the transaction business data type, and in the virtual asset association data of the aircraft, obtain the transaction association data corresponding to the transaction key fields, and determine the transaction association data as the third classification data corresponding to the transaction business data type.

[0023] In a possible implementation manner, the N classification data include classification data S i , where i is a positive integer less than or equal to N; the components of the aircraft include component M j , where j is a positive integer; classification data S i includes the component asset data corresponding to A components respectively; the A components belong to at least two components in the aircraft; the A components include component M j, A is a positive integer; the data storage module is used to generate storage data according to the N types of classification data, and when storing the storage data into the business database associated with the first business object, the data storage module is specifically used to perform the following operations:

[0024] Perform vectorization processing on the business type corresponding to the classification data S i to obtain a type vectorization result;

[0025] Perform vectorization processing on A components respectively to obtain A component vectorization results;

[0026] Obtain component M j Under the classification data S i The target component asset data, based on the file type corresponding to the target component asset data, perform text parsing on the target component asset data to obtain a data parsing result corresponding to the target component asset data;

[0027] Perform vectorization processing on the data parsing result to obtain a data vectorization result corresponding to the target component asset data;

[0028] Determine the type vectorization result, the component vectorization result, and the data vectorization result as storage data, and store the storage data into the business database associated with the first business object.

[0029] In a possible implementation manner, the N types of classification data include classification data S i , where i is a positive integer less than or equal to N; the components of the aircraft include component M j , where j is a positive integer; the classification data S i Includes the component asset data corresponding to A components respectively; the A components belong to at least two components in the aircraft; the A components include component M j , A is a positive integer; the data storage module is used to generate storage data according to the N types of classification data, and when storing the storage data into the business database associated with the first business object, the data storage module is specifically used to perform the following operations:

[0030] Determine the business type corresponding to the classification data S i as the initial root node, and determine the A components as A parent nodes under the initial root node;

[0031] Obtain component M j Under the classification data S i The target component asset data, perform format conversion on P component asset sub-data in the target component asset data respectively to obtain P structured component asset sub-data; P is a positive integer;

[0032] Allocate the P structured component asset sub-data to component M respectivelyj among the child nodes under the corresponding parent node;

[0033] When generating the child nodes under the parent nodes corresponding to A components respectively, the initial root node, the A parent nodes, and the child nodes corresponding to the A parent nodes respectively are determined as the graph sub-tree for the classification data S i ; one child node is used to allocate a structured component asset sub-data;

[0034] When obtaining the graph sub-trees corresponding to N types of classification data respectively, the aircraft is determined as the graph root node, and an item knowledge graph corresponding to the aircraft is generated based on the N graph sub-trees and the graph root node. The item knowledge graph is determined as the stored data, and the stored data is stored in the business database associated with the first business object.

[0035] In a possible implementation manner, the data determination module is used to obtain the first target type indicated by the business semantic information, obtain the first business data set belonging to the first target type from the classification data corresponding to N business types in the business database, and when determining the first target component associated with the business semantic information based on the business semantic information, the data determination module is specifically used to perform the following operations:

[0036] Extract keywords from the business semantic information to obtain B business keywords, perform vectorization processing on each business keyword to obtain B keyword vectors, perform vector similarity matching based on the B keyword vectors and the type vectorization results corresponding to the N business types respectively, and determine the business type corresponding to the matched type vectorization result as the first target type; B is a positive integer;

[0037] Obtain the first business data set belonging to the first target type from the classification data corresponding to N business types in the business database;

[0038] Based on the keyword vectors corresponding to the remaining business keywords, perform vector similarity matching with the component vectorization results corresponding to each component in the first business data set, and determine the component corresponding to the matched component vectorization result as the first target component associated with the business semantic information; the remaining business keywords are the business keywords among the B business keywords except for the business keywords that match the type vectorization results corresponding to the first target type.

[0039] In a possible implementation manner, the data processing device further includes a first data update module, and the first data update module is specifically used to perform the following operations:

[0040] When a data update request for a second target component sent by a first business object is obtained on the intelligent conversation page, based on the data update request, retrieve the to-be-matched asset data belonging to the second target component from N types of classification data respectively; at least two components include the second target component; the data update request includes component update data of the second target component; the to-be-matched asset data includes the component asset data corresponding to the second target component in the N types of classification data respectively;

[0041] Match the to-be-matched asset data with the component update data. If the component update data does not exist in the to-be-matched asset data, perform vectorization processing on the component update data to obtain vectorized update data, store the vectorized update data in the business database associated with the first business object, generate a data update success notification, and display the data update success notification in the intelligent message text box on the intelligent conversation page;

[0042] If the component update data exists in the to-be-matched asset data, generate a data duplication notification and display the data duplication notification in the intelligent message text box on the intelligent conversation page.

[0043] In a possible implementation, the data determination module is used to obtain the first target type indicated by the business semantic information, retrieve the first business data set belonging to the first target type from the classification data corresponding to N business types in the business database, and when determining the first target component associated with the business semantic information based on the business semantic information, the data determination module specifically performs the following operations:

[0044] Extract keywords from the business semantic information to obtain B business keywords; B is a positive integer;

[0045] Traverse in the item knowledge graph based on the B business keywords. When a type node matching the B business keywords is traversed, determine the business type corresponding to the matching type node as the first target type;

[0046] Retrieve the first graph subtree with the first target type as the initial root node from the item knowledge graph, and determine the first graph subtree as the first business data set; the N graph subtrees in the item knowledge graph include the first graph subtree;

[0047] Traverse in the first business data set based on the first keyword. When a component node matching the first keyword is traversed, determine the matching component node as the first target component associated with the business semantic information; the first keyword is the business keyword used to indicate the component among the B business keywords;

[0048] When the data retrieval module is used to perform information retrieval in the first business data set and obtain retrieval content associated with the first target component, the data retrieval module is specifically used to perform the following operations:

[0049] Determine the business keywords other than the business keyword indicating the business type and the business keyword indicating the component among the B business keywords as the keywords to be matched;

[0050] In the first business data set, match the keywords to be matched with each sub-node under the first target component, and determine the corresponding structured component asset sub-data in the matched data sub-nodes as the retrieval content associated with the first target component.

[0051] In a possible implementation, the data processing device further includes a second data update module, and the second data update module is specifically used to perform the following operations:

[0052] When a data update request for the second target component sent by the first business object is obtained on the intelligent session page, based on the component update data of the second target component in the data update request, determine the second target type corresponding to the component update data; at least two components include the second target component; N business types include the second target type;

[0053] Based on the second target type and the item knowledge graph, determine a second graph subtree with the second target type as the initial root node, and in the second graph subtree, obtain the set of data to be matched for all sub-nodes with the second target component as the parent node;

[0054] If the component update data does not exist in the set of data to be matched, add an update sub-node to the parent node corresponding to the second target component in the second graph subtree, allocate the component update data to the update sub-node, obtain a graph update subtree, update the item knowledge graph based on the graph update subtree, obtain an item updated knowledge graph, store the item updated knowledge graph in the business database associated with the first business object, generate a data update success notification, and display the data update success notification in the intelligent message text box on the intelligent session page;

[0055] If the component update data exists in the set of data to be matched, generate a data duplication notification and display the data duplication notification in the intelligent message text box on the intelligent session page.

[0056] In a possible implementation, when the data retrieval module is used to generate a business result for the business processing request text based on the intent requirement information and the retrieval content, the data retrieval module is specifically used to perform the following operations:

[0057] Identify key information in the retrieval result to obtain key data;

[0058] If the intent requirement information indicates a data analysis requirement, perform data analysis on the key data based on the intent requirement information to obtain a data analysis result, and determine the data analysis result as the business result for the business processing request text;

[0059] If the intent requirement information indicates a drawing processing requirement, perform image drawing based on the intent requirement information and the key data to obtain a drawn data graph, and determine the drawn data graph as the business result for the business processing request text.

[0060] In a possible implementation manner, when the data retrieval module is used to perform data analysis on the key data based on the intent requirement information to obtain a data analysis result, the data retrieval module is specifically used to perform the following operations:

[0061] Obtain Q search data for the key data and the intent requirement information, and respectively perform feature extraction on the Q search data through a large language model to obtain Q first extraction features; Q is a positive integer;

[0062] Perform feature extraction on the key data through a large language model to obtain a second extraction feature, and perform cross-attention processing on the Q first extraction features and the second extraction feature to obtain attention scores corresponding to the Q first extraction features respectively;

[0063] Determine the search data associated with the attention scores greater than or equal to the attention threshold among the Q attention scores as the data analysis result.

[0064] In a possible implementation manner, when the data retrieval module is used to perform image drawing based on the intent requirement information and the key data to obtain a drawn data graph, the data retrieval module is specifically used to perform the following operations:

[0065] Identify the drawing image type according to the intent requirement information, obtain an image template with the drawing image type in the image template library, generate a Gaussian noise image according to the image template and the initial noise data, and input the Gaussian noise image, the key data, and the intent requirement information into the text-to-image model;

[0066] Perform feature extraction on the Gaussian noise image through the text-to-image model to obtain Gaussian noise features, and perform forward diffusion processing on the Gaussian noise features to obtain forward noise vectors;

[0067] Perform feature encoding on the key data through the text-to-image model to obtain data encoding features, perform feature encoding on the intent requirement information to obtain intent encoding features, and perform feature splicing on the data encoding features and the intent encoding features to obtain spliced encoding features;

[0068] Perform denoising processing on the Gaussian noise image according to the forward noise vector and the spliced encoding features to obtain a drawn data graph.

[0069] In one aspect, an embodiment of the present application provides a computer device, including a processor, a memory, and an input / output interface;

[0070] The processor is respectively connected to the memory and the input / output interface. Among them, the input / output interface is used to receive and output data, the memory is used to store a computer program, and the processor is used to call the computer program so that the computer device including the processor executes the method in one aspect of the embodiment of the present application.

[0071] In one aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program, which is adapted to be loaded and executed by a processor so that a computer device having the processor executes the method in one aspect of the embodiment of the present application.

[0072] In one aspect, an embodiment of the present application provides a computer program product or a computer program. The computer program product or the computer program includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions so that the computer device executes the methods provided in various alternative manners in one aspect of the embodiment of the present application. In other words, when the computer instructions are executed by the processor, the methods provided in various alternative manners in one aspect of the embodiment of the present application are implemented.

[0073] Implementing the embodiment of the present application will have the following beneficial effects:

[0074] In an embodiment of the present application, a service processing request text sent by a first service object is obtained on an intelligent conversation page, and the service processing request text is parsed to obtain service semantic information and intention requirement information; the intelligent conversation page includes an object avatar corresponding to the first service object and an object message text box associated with the object avatar, and the object message text box contains the service processing request text; the first object identifier of the first service object is obtained, and based on the first object identifier, a service database associated with the first object identifier is determined; the service database includes virtual asset association data corresponding to N service types respectively; the virtual asset association data under each service type includes component asset data for at least one component of an aircraft; N is a positive integer; the first target type indicated by the service semantic information is obtained, and from the classification data corresponding to the N service types in the service database, a first service data set belonging to the first target type is obtained, and based on the service semantic information, a first target component associated with the service semantic information is determined; the N service types include the first target type; information retrieval is performed in the first service data set, and retrieval content associated with the first target component is obtained, and a service result for the service processing request text is generated based on the intention requirement information and the retrieval content, and an intelligent avatar and an intelligent message text box containing the service result associated with the intelligent avatar are displayed on the intelligent conversation page; the intelligent avatar is used for conversation interaction with the object avatar on the intelligent conversation page. Through the above process, the classification storage of files based on service types is realized, the retrieval volume during retrieval is reduced, and the retrieval efficiency is improved. And through the intelligent object, the intelligent management of the file system is realized. The business requirements (service processing request text) of the first service object are obtained through the intelligent object, and the service semantic information of the service processing request text is parsed. Based on the service semantic information, targeted intelligent retrieval is performed in the file system, the service result is quickly determined, and is displayed on the intelligent conversation page. The data is classified and stored based on N service types. In the subsequent process of reading data for a certain target component (for example, the first target component) of the aircraft under a certain specified service type (for example, the first target type), only the data corresponding to the first target component under the first target type needs to be queried, which reduces the retrieval threshold, saves the data query time, and improves the data reading and retrieval efficiency. And the entire retrieval process is completed by the intelligent object, which can save labor costs. Brief Description of the Drawings

[0075] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0076] Figure 1 It is a network interaction architecture diagram provided by an embodiment of the present application;

[0077] Figure 2 It is a scenario schematic diagram of a data processing method provided by an embodiment of the present application;

[0078] Figure 3 It is a method flow of data processing provided by an embodiment of the present application Figure 1 ;

[0079] Figure 4 It is a method flow of data processing provided by an embodiment of the present application Figure 2 ;

[0080] Figure 5 It is a data classification schematic diagram provided by an embodiment of the present application;

[0081] Figure 6 It is a schematic diagram of a data processing device provided by an embodiment of the present application;

[0082] Figure 7 It is a schematic diagram of the structure of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0083] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present application.

[0084] Among them, if it is necessary to collect data of an object (such as a user, etc.) in the present application, a prompt interface or a pop-up window is displayed before and during the collection. The prompt interface or the pop-up window is used to prompt the user that some data is being collected currently. Only after obtaining the confirmation operation of the user on the prompt interface or the pop-up window, the relevant steps of data acquisition are started, otherwise it ends. Moreover, for the obtained user data, it will be used in reasonable, legal scenarios or uses, etc. Optionally, in some scenarios where user data needs to be used but the user's authorization has not been obtained, authorization can also be requested from the user, and the user data will be used when the authorization is passed.

[0085] It can be understood that in the specific implementation manners of the present application, for the user data involved, when the following embodiments of the present application are applied to specific products or technologies, the user's permission or consent needs to be obtained, and the collection, use, and processing of the relevant data need to comply with the relevant laws, regulations, and standards in the relevant regions.

[0086] In the embodiments of the present application, please refer toFigure 1 , Figure 1 is a network interaction architecture diagram provided by an embodiment of the present application. As Figure 1 shown, the network interaction architecture diagram may include a service server 101 and a cluster of terminal devices. The cluster of terminal devices may include terminal devices 102a, 102b, 102c, …, 102n. Among them, there may be communication connections between the terminal devices in the cluster of terminal devices. For example, there is a communication connection between terminal device 102a and terminal device 102b, and there is a communication connection between terminal device 102a and terminal device 102c. At the same time, there may be a communication connection between any terminal device in the cluster of terminal devices and the service server 101. For example, there is a communication connection between terminal device 102a and the service server 101. Among them, the above communication connection does not limit the connection method and can be directly or indirectly connected through a wired communication method, or can be directly or indirectly connected through a wireless communication method, or can also be connected through other methods, which are not limited in this application.

[0087] It should be understood that each terminal device in the cluster of terminal devices as Figure 1 shown may be installed with an application client having a data processing function. When the application client runs on each terminal device, an intelligent conversation page can be displayed and data interaction can be carried out with the service server 101 as Figure 1 shown respectively, so that the service server 101 can receive service data from each terminal device. Among them, the application client can be an application client with functions such as displaying data information such as text, images, audio, and video, such as a social application, an instant messaging application, a live broadcast application, a short video application, a video application, a music application, a shopping application, a novel application, a browser, etc. Among them, the application client can be an independent client or an embedded sub-client (such as an application applet, browser web page access, etc.) integrated in a certain client (such as an instant messaging client, a social client, a video client, etc.), which is not limited here.

[0088] As Figure 1As shown in the figure, any terminal device in the terminal device cluster can send a service processing request text to the service server 101 through the intelligent session page. Here, taking the terminal device 102a corresponding to the first service object as an example, the service server 101 can obtain the service processing request text sent by the first service object in the intelligent session page, that is, the first service object can send the service processing request text in the intelligent session page of the terminal device 102a. The terminal device 102a displays the object avatar corresponding to the first service object and the object message text box associated with the object avatar in the intelligent session page, where the object message text box contains the service processing request text. The terminal device 102a can send the service processing request text to the service server 101 through the intelligent session page, so that the service server 101 receives the service processing request text and performs subsequent operations on the service processing request text. The service server 101 can parse the service processing request text through natural language processing technology (Natural Language Processing, NLP) to obtain service semantic information and intention requirement information. For example, the service processing request text can be "Help me query the lease contract of XX aircraft and generate a rental change curve graph for XX aircraft according to the query content", then the service semantic information obtained by the service server after parsing can include "Query the lease contract of XX aircraft under the transaction service data type and generate a rental change curve graph for XX aircraft according to the lease contract", and the intention requirement information can be indicated as query requirement and drawing processing requirement.

[0089] The service server 101 can obtain the first object identifier of the first service object and determine the service database associated with the first object identifier based on the first object identifier, that is, the database used to store various data uploaded by the first service object. The service database includes virtual asset association data corresponding to N service types of aircraft, and the virtual asset association data under each service type includes component asset data for at least one component of the heading item. N is a positive integer. Among them, the aircraft can be a transportation device for navigation (for example, an airplane, etc.), or a related component of the transportation device. For example, when the aircraft is an airplane, the related components of the airplane can include engines, fuselages, landing gears, etc., and these components such as engines, fuselages, and landing gears can all be used as the aircraft in this application. The service server 101 can obtain the first target type indicated by the service semantic information and obtain the first service data set belonging to the first target type from the classification data corresponding to the N service types of the service database. Further, the service server 101 can determine the first target component associated with the service semantic information based on the service semantic information. Among them, the N service types include the first target type.

[0090] The service server 101 can retrieve information from the first service data set, obtain the retrieved content associated with the first target component, and generate a service result for the service processing request text based on the intent requirement information and the retrieved content. The service server 101 can send the service result to the terminal device 102a, so that the terminal device 102a can display an intelligent avatar and an intelligent message text box containing the service result associated with the intelligent avatar on its intelligent conversation page, where the intelligent avatar is used to conduct conversation interaction with the object avatar on the intelligent conversation page. Among them, the processes executed by the above service server 101 can all be completed by the intelligent object associated with the intelligent avatar (for example, Artificial Intelligence object, AI object).

[0091] Through the above process, the classification storage of files (virtual asset association data) based on N service types is realized, the retrieval volume during retrieval is reduced, and the retrieval efficiency is improved. And the intelligent object is used to realize the intelligent management of the file system. The intelligent object obtains the service requirements (service processing request text) of the first service object, parses the service semantic information of the service processing request text, and conducts targeted intelligent retrieval in the file system based on the service semantic information, quickly determines the service result, and displays it on the intelligent conversation page. The data is classified and stored based on N service types. During the subsequent data reading process for a certain target component (for example, the first target component) of the aircraft under a certain specified service type (for example, the first target type), only the data corresponding to the first target component under the first target type needs to be queried, which reduces the retrieval threshold, saves data query time, and improves data reading and retrieval efficiency. And the entire retrieval process is completed by the intelligent object, which can save labor costs.

[0092] Specifically, please refer to Figure 2 , Figure 2 is a schematic diagram of the scenario of a data processing method provided by an embodiment of the present application. As Figure 2As shown, in the intelligent conversation page 201 corresponding to the terminal device 102a, an object avatar corresponding to the first service object and an object message text box 2011 associated with the object avatar are displayed. The object message text box 2011 includes the service processing request text sent by the first service object. The intelligent object 202 corresponding to the terminal device 102a can obtain the service processing request text and send the service processing request text to the service server 101. It can be understood that the intelligent object 202 can be provided by the service provider corresponding to the service server 101. The intelligent object 202 processes the service processing request text through the server 101 and obtains the corresponding result. The service server 101 can parse the service processing request text to obtain service semantic information and intention requirement information. The service server 101 can obtain the first object identifier of the first service object. Among them, the first object identifier can be a uniquely indicative identifier created and assigned by the service server 101 when the first service object creates an object account using the application client provided by the service server 101. The service server 101 can determine the service database associated with the first object identifier based on the first object identifier. Among them, the service database includes virtual asset association data corresponding to N service types respectively; the virtual asset association data under each service type includes component asset data for at least one component of the aircraft; N is a positive integer. That is, when the first service object creates an account and uploads data in the application client, the service server 101 can classify the uploaded data (into data corresponding to N service types respectively) and store it in the service database corresponding to the first service object. The service server 101 can obtain the first target type indicated by the service semantic information, obtain the first service data set belonging to the first target type from the classification data corresponding to the N service types in the service database, and determine the first target component associated with the service semantic information based on the service semantic information. Among them, the N service types include the first target type. The service server 101 can perform information retrieval in the first service data set, obtain the retrieval content associated with the first target component, and generate a service result for the service processing request text based on the intention requirement information and the retrieval content. Further, the service server 101 sends the generated service result to the terminal device 102a, and the intelligent object corresponding to the terminal device 102a displays the service result in the intelligent conversation page. As shown in the intelligent conversation page 203, it includes an object avatar, an object message text box 2011 (displaying the service processing request text), an intelligent avatar corresponding to the intelligent object 202, and an intelligent message text box 2031 (displaying the service result).The intelligent avatar is used for conversation interaction with the object avatar on the intelligent conversation page. That is, the first business object can consult a certain content (business processing request text) on the intelligent conversation page. The intelligent object obtains the text of the consultation content for intelligent processing and obtains the final business result to reply to the consultation of the first business object.

[0093] Through the above process, the classification storage of files (virtual asset association data) based on N business types is realized, reducing the retrieval volume during retrieval and improving the retrieval efficiency. And through the intelligent object, the intelligent management of the file system is realized. The intelligent object obtains the business requirements (business processing request text) of the first business object, parses the business semantic information of the business processing request text, and performs targeted intelligent retrieval in the file system based on the business semantic information, quickly determines the business result, and displays it on the intelligent conversation page. Classified storage of data based on N business types. During the subsequent data reading process of a certain target component (for example, the first target component) of the aircraft under a certain specified business type (for example, the first target type), only the data corresponding to the first target component under the first target type needs to be queried, which reduces the retrieval threshold, saves data query time, and improves data reading and retrieval efficiency. And the entire retrieval process is completed by the intelligent object, which can save labor costs.

[0094] It can be understood that the terminal device mentioned in the embodiments of the present application can also be a computer device. The computer device in the embodiments of the present application includes, but is not limited to, a terminal device or a server. In other words, the computer device can be a server or a terminal device, or a system composed of a server and a terminal device. Among them, the above-mentioned terminal device can be an electronic device, including but not limited to mobile phones, tablet computers, desktop computers, laptop computers, palm computers, in-vehicle devices, augmented reality / virtual reality (AR / VR) devices, head-mounted displays, smart TVs, wearable devices, smart speakers, digital cameras, cameras, and other mobile internet devices (MID) with network access capabilities, or terminal devices in scenarios such as trains, ships, and flights. As Figure 1 shown, the terminal device can be a mobile phone (as shown by terminal device 102a), a desktop computer (as shown by terminal device 102b), a tablet computer (as shown by terminal device 102c), or a laptop computer (as shown by terminal device 102n), etc. Figure 1Only some of the devices are listed. Among them, the server mentioned above can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, vehicle-road collaboration, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0095] Optionally, the data involved in the embodiments of the present application can be stored in a computer device, or the data can be stored based on cloud storage technology or a blockchain network, which is not limited herein.

[0096] Further, please refer to Figure 3 , Figure 3 which is a method flow for data processing provided by the embodiments of the present application Figure 1 . This data processing method can be executed by a computer device, and the computer device can be any one of the service server 101 or the terminal device cluster as shown in Figure 1 . The following will take the execution of this data processing method by a computer device as an example for description. Among them, this data processing method can at least include the following steps S301 to step S304:

[0097] Step S301, obtain the service processing request text sent by the first service object in the intelligent session page, and parse the service processing request text to obtain service semantic information and intent requirement information; the intelligent session page includes the object avatar corresponding to the first service object and the object message text box associated with the object avatar, and the object message text box contains the service processing request text.

[0098] In the embodiments of the present application, a computer device may obtain a service processing request text sent by a first service object on an intelligent conversation page, parse the service processing request text through natural language processing technology, and obtain the service semantic information and intention requirement information indicated by the service processing request text. Optionally, the computer device may also parse the service processing request text through a pre-trained language model (PLMs) to obtain the service semantic information and intention requirement information indicated by the service processing request text. Among them, the pre-trained language model has strong context understanding and semantic capture capabilities, and can improve the accuracy and adaptability of retrieval through dynamic embedding and context awareness mechanisms. The pre-trained language model may be a Bidirectional Encoder Representations from Transformers (BERT), a Generative Pre-trained Transformer (GPT), a Text-to-Text Transfer Transformer (T5), etc., which are not limited herein. And PLMs can use self-supervised learning methods to implement predicting masked words or predicting the next sentence, so that higher retrieval performance and wider application scenario coverage can be achieved with less labeled data. Among them, the intelligent conversation page includes an object avatar corresponding to the first service object and an object message text box associated with the object avatar, and the object message text box contains the service processing request text. The service semantic information refers to the specific execution content indicated by the service processing request text, that is, the information extracted from the service processing request text that can reflect the service content and logic. The intention requirement information refers to the specific intention or requirement expressed by the service processing request. For example, if the service processing request text is "Help me query the lease contract of XX aircraft and generate a rental change curve graph for XX aircraft based on the query content", the service semantic information obtained by the computer device after parsing it may include "Query the lease contract of XX aircraft under the service data type of transaction business, and generate a rental change curve graph for XX aircraft based on the lease contract", and the intention requirement information may be indicated as query requirement and drawing processing requirement.

[0099] Step S302, obtain a first object identifier of the first service object, and based on the first object identifier, determine a service database associated with the first object identifier; the service database includes stored data associated with virtual asset association data corresponding to N service types respectively; the virtual asset association data under the service type includes component asset data for at least one component of an aircraft; N is a positive integer.

[0100] In the embodiment of the present application, the computer device may obtain the first object identifier of the first service object. When the computer device obtains the service processing request text sent by the first service object, it may also obtain the account information of the first service object and obtain the first object identifier of the first service object from the account information. Among them, the first object identifier is a uniquely indicative identifier obtained when the first service object creates an object account in the application client corresponding to the intelligent conversation page. The computer device may determine a service database associated with the first object identifier based on the first object identifier; among them, the service database includes virtual asset association data corresponding to N service types respectively; the virtual asset association data under each service type includes component asset data for at least one component of the aircraft; N is a positive integer. The service database is a database allocated by the application client for the first service object to store data in the data repository or the cloud when the first service object uploads data in the application client. Taking the aircraft engine as an example, the virtual asset association data of the aircraft may include the usage data of the components of the engine (for example, the component is installed in the engine, and the navigation-related data of the engine installed in the aircraft, such as navigation duration, navigation area, etc.), the accident proof data of the components of the engine (for example, the proof data of accidents such as fire, water ingress, and falling of the component), the asset depreciation data of the components of the engine (for example, when the component is installed in the engine and the engine is installed in the aircraft, the different navigation discount rates of the navigation areas, because the worse the environment, the more it affects the performance of the engine, so there will be different navigation discount rates in different navigation areas), the historical maintenance data of the components of the engine (such as the maintenance time, maintenance location, and the restored performance ratio after maintenance of the component), the property right certificate data of the components of the engine (for example, factory license certificate, production license certificate, manufacturer, component flight permit, etc.), the transaction data of the components of the engine (for example, the lease association data of the component: such as the lease time, lease amount, and the return status of the component after lease; for example, the purchase and sale association data of the component: such as the purchase and sale amount, purchase and sale time, etc.; for example, the modification association data of the component: such as the modification cost, modification time, etc.). The virtual asset association data of the components of the engine (including usage data, accident proof data, asset depreciation data, historical maintenance data, property right certificate data, transaction data, etc.) jointly constitute the virtual asset association data of the engine.

[0101] Step S303, obtain the first target type indicated by the service semantic information, obtain the first service data set belonging to the first target type from the classification data corresponding to the N service types in the service database, and determine the first target component associated with the service semantic information based on the service semantic information; the N service types include the first target type.

[0102] In an embodiment of the present application, a computer device may obtain a first target type indicated by business semantic information. For example, the N business types existing in a business database are respectively an asset business data type, a basic business data type, and a transaction business data type. If the keyword in the business semantic information indicates a match with the basic business data type, then the first target type is the basic business data type. The computer device may obtain a first set of business data belonging to the first target type from the classification data corresponding to the N business types in the business database. When the first target type is the basic business data type, the first set of business data obtained by the computer device includes all virtual asset association data with the business type of the basic business data type. The computer device may determine a first target component associated with the business semantic information based on the business semantic information, that is, continue to match the specific components of the aircraft under the basic business data type based on the keyword in the business semantic information, and determine the matched component as the first target component. Among them, the N business types include the first target type.

[0103] Step S304, perform information retrieval in the first set of business data to obtain retrieval content associated with the first target component, generate a business result for the business processing request text based on the intent requirement information and the retrieval content, display an intelligent avatar in the intelligent conversation page, and an intelligent message text box containing the business result associated with the intelligent avatar; the intelligent avatar is used for session interaction with the object avatar in the intelligent conversation page.

[0104] In an embodiment of the present application, a computer device may perform information retrieval in the set of business data to obtain retrieval content associated with the first target component, that is, after the computer device determines that the business type is the first target type and the specific component is the first target component, it needs to retrieve the specific virtual asset association data indicated by the business processing request text and associated with the first target component in the first set of business data. For example, the first set of business data includes virtual asset association data 1, virtual asset association data 2, virtual asset association data 3 of component 1, virtual asset association data 4 (lease contract), virtual asset association data 5 (invoice), and virtual asset association data 6 of component 2. If the business semantic information indicates that it is necessary to obtain the lease contract for component 2, the final retrieval content is virtual asset association data 4; if the business semantic information indicates that it is necessary to obtain the invoice information for component 2, the final retrieval content is virtual asset association data 5.

[0105] Further, the computer device may generate a service result for the service processing request text based on the intent requirement information and the retrieved content. That is, when the intent requirement information is a drawing requirement (such as drawing a rent change curve graph), the computer device may draw a rent change curve graph for the retrieved content (lease contract, including the rent field), and determine the rent change curve graph and the retrieved content together as the service result for the service processing request text. Further, the computer device may display an intelligent avatar and an intelligent message text box containing the service result associated with the intelligent avatar on the intelligent conversation page. The intelligent avatar is used to perform conversation interaction with the object avatar on the intelligent conversation page, that is, the intelligent object performs conversation interaction with the first service object.

[0106] It can be understood that all the operations performed by the computer device in the above process can be completed by the intelligent object calling the computer device after receiving the service processing request text sent by the first service object.

[0107] Through the above process, the classification storage of files (virtual asset association data) based on N service types is realized, the retrieval volume during retrieval is reduced, and the retrieval efficiency is improved. And through the intelligent object, the file system is intelligently managed. The intelligent object obtains the service requirements (service processing request text) of the first service object, parses the service semantic information of the service processing request text, and performs targeted intelligent retrieval in the file system based on the service semantic information, quickly determines the service result, and displays it on the intelligent conversation page. The data is classified and stored based on N service types. During the subsequent data reading process of a certain target component (such as the first target component) of the aircraft under a certain specified service type (such as the first target type), only the data corresponding to the first target component under the first target type needs to be queried, which reduces the retrieval threshold, saves data query time, and improves data reading and retrieval efficiency. And the entire retrieval process is completed by the intelligent object, which can save labor costs.

[0108] Further, please refer to Figure 4 , Figure 4 which is a method flow of data processing provided by an embodiment of the present application Figure 2 , and this data processing method can be executed by a computer device, and the computer device can be any one of the service server 101 or the terminal device cluster as shown in Figure 1 . The following will take this data processing method being executed by the computer device as an example for description. Among them, this data processing method may at least include the following steps S401 - step S406:

[0109] Step S401: Obtain the virtual asset association data for the aircraft sent by the first business object; obtain the asset data classification rules, and classify the virtual asset association data of the aircraft through N business types in the asset data classification rules to obtain N classification data; generate storage data based on the N classification data, and store the storage data in the business database associated with the first business object; the aircraft is composed of at least two components; each classification data corresponds to one business type.

[0110] In the embodiment of the present application, the computer device can obtain the virtual asset association data for the aircraft sent by the first business object. Among them, the virtual asset association data of the aircraft may include the aircraft usage data of the aircraft, aircraft accident certification data, asset depreciation data, historical maintenance data, property right certificate data, transaction data (such as lease association data), and so on. Among them, the aircraft is composed of at least two components, so the virtual asset association data of the aircraft may include the component usage data of each component, component accident certification data, component asset depreciation data, historical maintenance data of the component, property right certificate data of the component, transaction data of the component, and so on. The computer device can obtain the asset data classification rules. The N business types included in the asset data classification rules for data classification can be artificially defined. For example, the N business types may include asset business data types, basic business data types, and transaction business data types. Among them, the data corresponding to the asset business data type can be data used to record the loss of the aircraft, such as the component usage data, component accident certification data, component asset depreciation data, historical maintenance data of the component, etc. mentioned above; the data corresponding to the basic business data type can be the property right certificate data of the aircraft, such as the property right certificate data of the aircraft mentioned above (such as, component factory license certificate, component flight permit, etc.); the data corresponding to the transaction business data type can be the transaction association data of the aircraft (such as, lease association data of the aircraft, purchase and sale association data of the aircraft, modification association data of the aircraft, lease association data of aircraft components, purchase and sale association data of aircraft components, modification association data of aircraft components, etc.).

[0111] The computer device can classify the virtual asset association data of the aircraft through N business types in the asset data classification rules to obtain N classification data. For example, the virtual asset association data belonging to the same business type in the virtual asset association data of the aircraft can be classified into one category. If there are N business types, the virtual asset association data of the aircraft can be classified into N categories. Among them, the N classification data can include first classification data, second classification data, and third classification data. The specific implementation process for the computer device to classify the virtual asset association data of the aircraft to obtain N classification data can be as follows: Obtain the data keyword fields corresponding to the asset business data types; among them, the data keyword fields include usage record keyword fields, maintenance record keyword fields, and depreciation record keyword fields; subsequently, in the virtual asset association data of the aircraft, obtain the historical usage records corresponding to the usage record keyword fields, the historical maintenance records corresponding to the maintenance record keyword fields, and the historical depreciation records corresponding to the depreciation record keyword fields, and determine the historical usage records, historical maintenance records, and historical depreciation records as the first classification data corresponding to the asset business data types; it is possible to obtain the factory keyword fields corresponding to the basic business data types, obtain the factory association data corresponding to the factory keyword fields in the virtual asset association data of the aircraft, and determine the factory association data as the second classification data corresponding to the basic business data types; it is possible to obtain the transaction keyword fields corresponding to the transaction business data types, obtain the transaction association data corresponding to the transaction keyword fields in the virtual asset association data of the aircraft, and determine the transaction association data as the third classification data corresponding to the transaction business data types.

[0112] It should be understood that each business type may correspond to different key fields respectively, and each key field may correspond to a kind of data. Then, according to different key fields, data belonging to different business types can be obtained from the virtual asset associated data of the aircraft. For example, the asset business data type may correspond to the usage record key field, the maintenance record key field, and the depreciation record key field. Then, according to the usage record key field, historical usage records (such as the historical usage records of each component of the aircraft) can be obtained from the virtual asset associated data of the aircraft. According to the maintenance record key field, historical maintenance records (such as the historical maintenance records of each component of the aircraft) can be obtained from the virtual asset associated data of the aircraft. According to the depreciation record key field, historical depreciation records (such as data such as the historical accident records of each component of the aircraft) can be obtained from the virtual asset associated data of the aircraft. And the historical usage records, historical maintenance records, and historical depreciation records can all be used as the data corresponding to the asset business data type. For example, the basic business data type may correspond to the factory key field. Then, according to the factory key field, factory associated data (such as factory license certificates, production license certificates, etc.) can be obtained from the virtual asset associated data of the aircraft. For example, the transaction business data type may correspond to the transaction key field. Then, according to the transaction key field, transaction associated data (such as data such as rent, lease term, return status after lease, etc.) can be obtained from the virtual asset associated data of the aircraft.

[0113] Further, reference can be made to Figure 5 , Figure 5 which is a schematic diagram of data classification provided by an embodiment of the present application. If Figure 5 shown, the terminal device 501 may be the above-mentioned terminal device 102a, the object a may be the above-mentioned first business object, and the object a may send the virtual asset associated data 502 of the aircraft (constituted by component 1 and component 2) to the computer device 503. Among them, the virtual asset associated data 502 of the aircraft may include virtual asset associated data 1, virtual asset associated data 2, virtual asset associated data 3, virtual asset associated data 4, virtual asset associated data 5, and virtual asset associated data 6. After receiving the virtual asset associated data 502 of the aircraft, the computer device 503 may obtain an asset data classification rule, and the asset data classification rule includes N business types for data classification (including business type 1, business type 2, and business type 3); further, the virtual asset associated data 502 of the aircraft may be classified according to the business types in the asset data classification rule. For example Figure 5As shown, in the virtual asset association data 502 of the aircraft (including virtual asset association data 1, virtual asset association data 2, virtual asset association data 3, virtual asset association data 4, virtual asset association data 5, and virtual asset association data 6), virtual asset association data belonging to business type 1 (including virtual asset association data 1 and virtual asset association data 3) can be obtained, virtual asset association data belonging to business type 2 (including virtual asset association data 2 and virtual asset association data 4) can be obtained, and virtual asset association data belonging to business type 3 (including virtual asset association data 5 and virtual asset association data 6) can be obtained.

[0114] As Figure 5 shown, the virtual asset association data under one business type can be used as a type of classification data. Then, according to business type 1, business type 2, and business type 3, the virtual asset association data of the item can be divided into 3 types of classification data. These 3 types of classification data can include the classification data under business type 1 (including virtual asset association data 1 and virtual asset association data 3), the classification data under business type 2 (including virtual asset association data 2 and virtual asset association data 4), and the classification data under business type 3 (including virtual asset association data 5 and virtual asset association data 6). Optionally, the computer device can also further classify the virtual asset association data under each type of classification data according to the components of the aircraft (including component 1 and component 2). For example, for the classification data under business type 1, since virtual asset association data 1 is the virtual asset association data for component 1 and virtual asset association data 3 is the virtual asset association data for component 2, the classification data under business type 1 can be further divided into two types of data (one is virtual asset association data 1 belonging to component 1 and the other is virtual asset association data 3 belonging to component 2); similarly, for the classification data under business type 2, since both virtual asset association data 2 and virtual asset association data 4 are the virtual asset association data for component 2, the classification data under business type 2 can be further divided into 1 type of data (virtual asset association data 2 and virtual asset association data 4 belonging to component 2); similarly, for the classification data under business type 3, since both virtual asset association data 5 and virtual asset association data 6 are the virtual asset association data for component 1, the classification data under business type 3 can be further divided into 1 type of data (virtual asset association data 5 and virtual asset association data 6 belonging to component 1).

[0115] Furthermore, the computer device can generate storage data according to N types of classification data and store the storage data in the business database associated with the first business object. Specifically, the N types of classification data include classification data S i , where i is a positive integer less than or equal to N; the components of the aircraft include component M j, where j is a positive integer; classification data S i includes component asset data corresponding to A components respectively; the A components belong to at least two components in the aircraft; the A components include component M j , where A is a positive integer. The following will take the storage process S i of the classification data as an example to illustrate the process of classifying and storing the virtual asset association data for the aircraft uploaded by the first business object. The computer device can perform vectorization processing on the business type corresponding to the classification data S i to obtain a type vectorization result; that is, the computer device performs vector conversion on the text of the business type corresponding to the classification data S i to obtain a numerical form that can be processed by machine learning algorithms. The computer device can perform word segmentation processing on the text of the business type to obtain a word segmentation result, and further perform data cleaning on the word segmentation result to obtain a data cleaning result. That is, the text is split into words, phrases, or other meaningful units (referred to as word segmentation, "tokens"), and words that do not carry much meaning, such as punctuation marks or stop words in the text, are removed. For example, there is a business type of "The AssetBusiness Data Type". After word segmentation processing, the word segmentation result can be "[“The”, “Asset”, “Business”, “Data”, “Type”]", and after data cleaning, the data cleaning result can be "[“Asset”, “Business”, “Data”, “Type”]". Further, through the method of word embeddings, the data cleaning result can be mapped to a vector space to obtain a type vectorization result. Further, the computer device can perform vectorization processing on each of the A components respectively to obtain A component vectorization results; the computer device can obtain component M j in the classification data S iTarget component asset data. Based on the file type corresponding to the target component asset data, perform text parsing on the target component asset data to obtain the data parsing result corresponding to the target component asset data. The text parsing can be OCR parsing. For example, if the file type corresponding to the target component asset data is an image type (such as a PDF file), the computer device can identify text information from the image type file through OCR parsing and convert it into editable text, that is, obtain the data parsing result. Optionally, the computer device can also directly call a multi-modal large model (which can process text, pictures, or videos) to perform text parsing on the target component asset data. For example, a large language model (LLM, Large Language Model). The computer device can use the OCR ability in the LLM to perform text parsing on the target component asset data to obtain the data parsing result corresponding to the target component asset data. Since the LLM can not only process text but also support processing multi-modal data such as images and videos, and can also combine the technology of text-to-image, it can process files containing text and pictures more naturally. Thus, it enables the computer device to directly obtain the data parsing result corresponding to the target component asset data without first determining the file type corresponding to the target component asset data and then further performing OCR parsing through different file types, making file parsing more efficient. And through technologies such as Federated Learning or Transfer Learning, the LLM model can self-update as the data grows and business requirements change, maintaining high processing capabilities. Optionally, the computer device (or only the object) can combine Causal Inference technology to achieve judging causal relationships through context data, improving the intelligence of business processing and the reasoning ability for complex problems. Further, the computer device performs vectorization processing on the data parsing result to obtain the data vectorization result corresponding to the target component asset data; determines the type vectorization result, the component vectorization result, and the data vectorization result as stored data, and stores the stored data in the business database associated with the first business object. It can be understood that in this vectorized data storage method, the computer device can distinguish different business types corresponding to different data through the coordinate positions of the vectorized information of the data (such as different data vectorization results) in the vector space, which indicate different regions.

[0116] Optionally, the computer device can classify the data S i The corresponding business type is determined as the initial root node, and A components are determined as A parent nodes under the initial root node, and component M is obtained j In the classification data S iFor the target component asset data below, format conversion is performed on the P component asset sub-data in the target component asset data respectively to obtain P structured component asset sub-data; P is a positive integer. Among them, a component asset sub-data is a virtual asset association data corresponding to a component under a certain business type. The specific implementation process of a computer device for performing format conversion on a component asset sub-data to obtain a structured component asset sub-data can be: The computer device uses a library in a programming language (such as the Python-docx library in the Python language) to read the component asset sub-data, traverses the content in the component asset sub-data, that is, traverses elements such as paragraphs, tables, and pictures in the component asset sub-data, extracts the required important information (i.e., the key content in the component asset sub-data), constructs a structured object based on this key content, and writes the constructed structured object into a structured file to obtain the structured component asset sub-data. Among them, the structured component asset sub-data can be a structured file with a data format of Json. The computer device can allocate the P structured component asset sub-data to the sub-nodes under the parent node corresponding to component M j Among the sub-nodes; when generating the sub-nodes under the parent nodes corresponding to A components respectively, the computer device can determine the initial root node, the A parent nodes, and the sub-nodes corresponding to the A parent nodes respectively as the graph sub-tree for the classification data S i ; where one sub-node is used to allocate a structured component asset sub-data. When obtaining the graph sub-trees corresponding to N types of classification data respectively, the computer device can determine the aircraft as the graph root node, generate the item knowledge graph corresponding to the aircraft based on the N graph sub-trees and the graph root node, determine the item knowledge graph as the stored data, and store the stored data in the business database associated with the first business object. Among them, the graph root node is an entity node in the item knowledge graph, and the computer device can associate the N graph sub-trees to this entity node to obtain the item knowledge graph corresponding to the aircraft. Optionally, the computer device can implement the construction of the item knowledge graph through multi-scale graph embedding technology, that is, the entities in the item knowledge graph can have different representations in different contexts and scales. For example, the entity corresponding to a certain business type in the item knowledge graph is the asset business data type, and its representation at another scale can be historical usage records, historical maintenance records, and historical depreciation records, etc. That is, the concept of the entity is represented by the asset business data type, and the specific details of the entity are represented by historical usage records, historical maintenance records, and historical depreciation records, etc. It can be understood that in this structured data storage method, the computer device can distinguish different components or different business types corresponding to different data through the branch relationships of different nodes.

[0117] It can be understood that the stored data in the business database can be the vectorized results or item knowledge graphs corresponding to the various data mentioned in step S302; it can also be the various token sequences (i.e., token sequences) obtained after tokenizing the above-mentioned various data, and the corresponding sequence vectorized results or sequence item knowledge graphs. Among them, the process of obtaining the sequence vectorized results and sequence item knowledge graphs is the same as that of obtaining the vectorized results or item knowledge graphs corresponding to the various data.

[0118] It should be noted that in other data storage methods, that is, in the method where the computer device classifies the virtual asset association data by business type and then directly stores it, when performing information retrieval based on business semantic information subsequently, the computer device can adopt a multi-modal model based on contrastive learning (Contrastive Language-Image Pre-training, CLIP), which can not only process text, but also process images, videos, and other formats of content, so as to achieve multi-modal retrieval and enhance the diversity and depth of retrieval.

[0119] Step S402: Obtain the business processing request text sent by the first business object on the intelligent conversation page, and parse the business processing request text to obtain business semantic information and intent requirement information; the intelligent conversation page includes the object avatar corresponding to the first business object and the object message text box associated with the object avatar, and the object message text box contains the business processing request text.

[0120] In the embodiment of the present application, the specific implementation process of step S402 can refer to the specific description in step S301 as shown in Figure 3 and will not be elaborated here.

[0121] Step S403: Obtain the first object identifier of the first business object, and based on the first object identifier, determine the business database associated with the first object identifier; the business database includes virtual asset association data corresponding to N business types respectively; the virtual asset association data under each business type includes component asset data for at least one component of the aircraft; N is a positive integer.

[0122] In the embodiment of the present application, the specific implementation process of step S403 can refer to the specific description in step S302 as shown in Figure 3 and will not be elaborated here.

[0123] Step S404: Obtain the first target type indicated by the service semantic information. From the classification data corresponding to each of the N service types in the service database, obtain the first service data set belonging to the first target type, and based on the service semantic information, determine the first target component associated with the service semantic information; the N service types include the first target type.

[0124] In the embodiment of the present application, for the vectorized classification storage method of data during data storage, that is, the first data storage method mentioned in the above step S401, when the computer device obtains the first target type indicated by the service semantic information, it can extract keywords from the service semantic information to obtain B service keywords, perform vectorization processing on each service keyword to obtain B keyword vectors, and perform vector similarity matching based on the B keyword vectors and the type vectorization results corresponding to each of the N service types, and determine the service type corresponding to the matched type vectorization result as the first target type; B is a positive integer. A possible implementation manner for the computer device to perform vector similarity matching can refer to Formula ①:

[0125]

[0126] As shown in Formula ①, cosine_similarity(A,B) is used to represent the cosine similarity between vector A and vector B, that is, the computer device can use a keyword vector as vector A and a type vectorization result as vector B to calculate the cosine similarity between a keyword vector and a type vectorization result, that is, the vector similarity. Where “·” represents the dot product, A·B represents the dot product result between vector A and vector B, ‖A‖ represents the norm of vector A, and ‖B‖ represents the norm of vector B. When the computer device performs vector similarity matching based on the B keyword vectors and the type vectorization results corresponding to each of the N service types and calculates the vector similarity between the keyword vector and the type vectorization result, it can also be determined by calculating the Euclidean Distance, Jaccard Similarity, Manhattan Distance, Pearson Correlation Coefficient, etc. between the keyword vector and the type vectorization result, which is not limited here. The value range of the vector similarity is [-1,1], and the closer the value of the vector similarity is to 1, the more similar the two vectors are considered.

[0127] The computer device can obtain a first set of business data belonging to the first target type from the classification data corresponding to N business types in the business database. Based on the keyword vectors corresponding to the remaining business keywords, vector similarity matching is performed with the component vectorization results corresponding to each component in the first set of business data, and the component corresponding to the matched component vectorization result is determined as the first target component associated with the business semantic information. Among them, the remaining business keywords are the business keywords among B business keywords except for the business keywords that match the type vectorization results corresponding to the first target type. For example, the B business keywords include business keyword b, business keyword c, business keyword d, and business keyword f. There may be one business keyword (for example, business keyword b) among the B keywords that matches the type vectorization result corresponding to one of the N business types, then the remaining business keywords include business keyword c, business keyword d, and business keyword f. The computer device can perform vector similarity matching based on the keyword vectors corresponding to business keyword c, business keyword d, and business keyword f respectively, and the component vectorization results corresponding to each component in the first set of business data.

[0128] Optionally, for the data structured classification storage method during data storage, that is, the second data storage method mentioned in step S401 above, when the computer device obtains the first target type indicated by the business semantic information, it can extract keywords from the business semantic information to obtain B business keywords; B is a positive integer. The computer device can traverse the item knowledge graph based on the B business keywords. When a type node matching the B business keywords is traversed, the business type corresponding to the matching type node is determined as the first target type. That is, if a certain business keyword among the B business keywords is highly similar to or the same as a certain initial root node representing a business type under the graph root node in the item knowledge graph, then it is considered that the initial root node representing the business type is the type node matching the B business keywords, and the computer device can determine the business type corresponding to this type node as the first target type. For example, if the B business keywords include business keywords indicating the basic business data type, then the first target type matched by the computer device is the basic business data type. Optionally, the computer device can also implement the retrieval in the item knowledge graph through Graph Neural Networks (GNNs) technology. Specifically, the computer device can map the B business keywords to the nodes in the item knowledge graph to generate business keyword embeddings. And call a suitable GNN model to perform feature learning on the nodes in the item knowledge graph to generate node embeddings. Among them, different GNN models have different characteristics. For example, the Graph Sample and Aggregate (GraphSAGE) model and the Graph Attention Networks (GAT). There is no limitation here. Among them, GraphSAGE is suitable for large-scale graph data, and GAT can capture different importance between nodes. It can be understood that the node embeddings can capture the semantic information of the nodes and the structural information of the nodes in the item knowledge graph. Further, the computer device can calculate the similarity between the business keyword embeddings and the node embeddings in the knowledge graph, and determine the node associated with the maximum similarity as the type node, and determine the business type corresponding to this type node as the first target type. Further, the computer device can obtain the first graph subtree with the first target type as the initial root node from the item knowledge graph, and determine the first graph subtree as the first business data set. That is, the computer device retrieves the graph subtree with the first target type as the root node from the item knowledge graph, and analyzes the relevant entities and relationships corresponding to the nodes in this graph subtree, so as to construct and determine the content of the first business data set. Among the N graph subtrees in the item knowledge graph, there is the first graph subtree.That is, the first graph subtree includes an initial root node (which is the first target type at this time) and all branch nodes under the initial root node. The computer device can traverse the first business data set based on the first keyword. When a component node that matches the first keyword is traversed, the matching component node is determined as the first target component associated with the business semantic information. Among them, the first keyword is the business keyword used to indicate the component among the B business keywords; that is, in the first graph subtree of the computer device, among the child nodes under the initial root node representing the first target type, the graph query and reasoning algorithm is used to match the child nodes (including component nodes) that are the same as or highly similar to the first keyword, and the matching component nodes are determined as the first target components associated with the business semantic information. For example, under the first target type, there are child nodes corresponding to component 1, component 2, and component 3 respectively. If the first keyword matches component 2, then component 2 is determined as the first target component. Optionally, the computer device can query in the item knowledge graph based on the B business keywords using the semantic search and matching algorithm. When querying for type nodes that match the semantics of the B business keywords, the business types corresponding to all the matching type nodes are determined as potential target types, and further, the business types included in the potential target types are matched with the B business keywords, and the business types that match the B business keywords are determined as the first target types. Among them, the first target type can be the business type corresponding to the virtual asset association data that needs to be retrieved indicated by the business semantic information.

[0129] It should be noted that when the computer device performs node matching in the item knowledge graph or the first graph subtree (the first business data set) based on the business keyword, it can construct a query statement (for example, a SPARQL query statement) based on the business keyword, and perform query matching in the item knowledge graph or the first graph subtree based on this query statement. When performing query matching, the matching degree between the node and the business keyword can be determined. The ways to determine the matching degree can be exact matching (that is, checking whether the label or attribute of the node is exactly the same as the business keyword), fuzzy matching (determining the similarity between the node and the business keyword), and semantic matching (using natural language processing technology to understand the semantics of the business keyword and comparing it with the semantic features of the node), etc., which are not limited here.

[0130] Step S405: Perform information retrieval in the first business data set, obtain the retrieval content associated with the first target component, generate a business result for the business processing request text based on the intent requirement information and the retrieval content, and display a smart avatar and a smart message text box containing the business result associated with the smart avatar on the intelligent conversation page; the smart avatar is used to perform conversation interaction with the object avatar on the intelligent conversation page.

[0131] In the embodiments of the present application, for the data vectorized classification storage method during data storage, after the computer device determines the first target component, if the number of data vectorization results under the first target component in the classification data corresponding to the first target type is 1, that is, the target component asset data corresponding to the data vectorization result only includes one virtual asset association data, the computer device can directly determine the virtual asset association data corresponding to this data vectorization result as the retrieval content associated with the first target component. If the number of data vectorization results under the first target component is multiple, that is, the target component asset data corresponding to the data vectorization result includes multiple virtual asset association data, the computer device can determine the business keywords in the remaining keywords except for the business keywords that match the component vectorization result corresponding to the first target component as the keywords to be retrieved. Based on the keyword vectors corresponding to the keywords to be retrieved, match them with multiple data vectorization results under the first target component in the first business data set, and determine the virtual asset association data corresponding to the matched data vectorization result as the retrieval content associated with the first target component. Optionally, the computer device can also determine the data parsing result corresponding to the virtual asset association data as the retrieval content, where the data parsing result is an editable text obtained by performing text parsing on the virtual asset association data based on the file type corresponding to the virtual asset association data. It can be understood that there is a unique mapping relationship between the data vectorization result and the virtual asset association data, and the computer device can directly obtain the virtual asset association data corresponding to the data vectorization result through the data vectorization result.

[0132] Optionally, for the data structured classification storage method during data storage, after the computer device determines the first target component, it can determine the business keywords in the B business keywords except for the business keywords used to indicate the business type and the business keywords used to indicate the component as the keywords to be matched. In the first business data set, match the keywords to be matched with each sub-node under the first target component, and determine the corresponding structured component asset sub-data in the matched data sub-nodes as the retrieval content associated with the first target component.

[0133] Furthermore, the computer device can generate a business result for the business processing request text based on the intent requirement information and the retrieved content. Specifically, the computer device can identify key information from the retrieval result to obtain key data. If the intent requirement information indicates a data analysis requirement, the computer device can perform data analysis on the key data based on the intent requirement information to obtain a data analysis result, and determine the data analysis result as the business result for the business processing request text. Specifically, the computer device can obtain Q search data for the key data and the intent requirement information, and respectively extract features from the Q search data through a large language model to obtain Q first extraction features; Q is a positive integer. Among them, the Q search data can be sourced from the search results obtained by the computer device based on the key data and the intent requirement information through one or more different search engines. The computer device can extract features from the key data through a large language model to obtain a second extraction feature, and perform cross-attention processing on the Q first extraction features and the second extraction feature to obtain attention scores corresponding to the Q first extraction features respectively. Among them, a possible cross-attention processing method can be seen in Formula ②:

[0134]

[0135] As shown in Formula ②, Attention(Q, K, V) represents the cross-attention function, Q is used to represent the query vector query, K is used to represent the key vector key, and V is used to represent the value vector value; T represents the transpose, and K T represents the transpose matrix of K. The computer device can determine the first extraction feature as the query vector, determine the second extraction feature as the key vector and the value vector, and d k is used to represent the number of dimensions corresponding to the second extraction feature, and the calculation result of is used to represent the cross-attention score.

[0136] The computer device can determine the search data associated with the attention scores greater than or equal to the attention threshold among the Q attention scores as the data analysis result. Among them, the attention threshold can be specified manually. Optionally, the computer device can determine the search data corresponding to the maximum score among the Q attention scores as the data analysis result.

[0137] Optionally, if the intent requirement information indicates a drawing processing requirement, the computer device can perform image drawing based on the intent requirement information and key data to obtain a drawn data graph, and determine the drawn data graph as the business result for the business processing request text. Specifically, the computer device can identify the type of the drawn image according to the intent requirement information, obtain an image template with the type of the drawn image in the image template library, generate a Gaussian noise image according to the image template and the initial noise data, and input the Gaussian noise image, the key data, and the intent requirement information into the text-to-image model. The image template library includes various types of image modules, such as image templates for mathematical statistical analysis, such as statistical charts, histograms, line charts, etc. The computer device can extract features from the Gaussian noise image through the text-to-image model to obtain the mean vector and variance vector of the Gaussian noise image. The mean vector can be the average of all pixel values of the Gaussian noise image on each channel, that is, the first-order statistic of the Gaussian noise image, and the variance vector can be the average of the variances of the pixel values of each channel. That is, the second-order statistic of the Gaussian noise image. The computer device can randomly sample from the mean vector and variance vector of the Gaussian noise image to obtain a latent mean vector and a latent variance vector, and generate Gaussian noise features based on the latent mean vector and the latent variance vector. Further, in the forward diffusion network layer of the text-to-image model, the computer device obtains the latent variable distribution, where the latent variable distribution is a conceptual distribution with added noise, such as a Gaussian distribution. The computer device can continuously add random noise vectors to the Gaussian noise features in T time steps to obtain the forward noise vector. Where T is a positive integer, and the time step refers to the amplitude of adding noise to the Gaussian noise features based on the latent variable distribution. The noise can refer to unnecessary or redundant interference information existing in the image data. When T is large enough, the forward noise vector can be used to represent an image completely containing noise.

[0138] The computer device can encode key data in the text encoding layer of the text-to-image generation model to obtain data encoding features; and encode the intent requirement information through the text encoding layer to obtain intent encoding features. The data encoding features and the intent encoding features can be feature spliced to obtain spliced encoding features. The computer device can denoise the Gaussian noise image according to the forward noise vector and the spliced encoding features to obtain a drawn data graph. That is, the computer device continuously performs noise prediction on the forward noise vector through the spliced encoding features in T time steps to obtain a predicted noise vector, and denoises the forward noise vector through the predicted noise vector to obtain a target latent vector. The computer device can reconstruct an image from the target latent vector through the decoder in the text-to-image generation model to obtain a target predicted image, and determine the target predicted image as the drawn data graph for the intent requirement information and the key data. Among them, the target latent vector is used to characterize the features of the target predicted image, that is, the latent representation of the target predicted image after noise removal in the latent space. Further, the computer device can display an intelligent avatar and an intelligent message text box containing business results associated with the intelligent avatar on the intelligent conversation page. The intelligent avatar is used for conversation interaction with the object avatar on the intelligent conversation page.

[0139] Step S406, when a data update request for a second target component sent by a first business object is obtained on the intelligent conversation page, update the data in the business database based on the data update request.

[0140] In the embodiment of the present application, when the computer device obtains a data update request for a second target component sent by a first business object on the intelligent conversation page, if the data storage method in the business database is a vectorized classification storage method, the computer device can obtain the asset data to be matched belonging to the second target component from N types of classification data based on the data update request. At least two components constituting the aircraft include the second target component; the data update request includes the component update data of the second target component; the asset data to be matched includes the component asset data corresponding to the second target component in N types of classification data respectively. For example, if there is component asset data 1a (including virtual asset association data a and virtual asset association data b) of the second target component in classification data 1, component asset data 1b (including virtual asset association data c) of the second target component in classification data 2, and component asset data 1c (including virtual asset association data d and virtual asset association data f) of the second target component in classification data 3, the asset data to be matched includes asset association data a, virtual asset association data b, virtual asset association data c, virtual asset association data d, and virtual asset association data f.

[0141] The computer device can match the asset data to be matched with the component update data. If the component update data does not exist in the asset data to be matched, the component update data is vectorized to obtain vectorized update data. The vectorized update data is stored in the business database associated with the first business object, and a data update success notification is generated and displayed in the intelligent message text box on the intelligent session page. For example, the component asset data is virtual asset association data g, that is, the virtual asset association data g does not exist in the asset data to be matched (including asset association data a, virtual asset association data b, virtual asset association data c, virtual asset association data d, virtual asset association data f). The computer device can vectorize the virtual asset association data g to obtain vectorized update data and store the vectorized update data in the business database associated with the first business object. It should be noted that after the virtual asset association data under different business types is vectorized, the obtained data vectorization results have very different coordinate positions in the vector space. For example, the data vectorization result corresponding to the virtual asset association data under classification data 1 is located in the vector position in the spatial region A1 in the vector space, the data vectorization result corresponding to the virtual asset association data under classification data 2 is located in the vector position in the spatial region A2 in the vector space, and the data vectorization result corresponding to the virtual asset association data under classification data 3 is located in the vector position in the spatial region A3 in the vector space. The spatial regions A1, A2, and A3 in the spatial vector belong to three different regions. Therefore, the computer device can directly divide the virtual asset association data under different business types through different data vectorization results to achieve classified storage of data in the business database.

[0142] Optionally, if the component update data exists in the asset data to be matched, a data duplication notification is generated and displayed in the intelligent message text box on the intelligent session page. For example, the component update data is the same as the virtual asset association data c, that is, the component update data exists in the asset data to be matched (including asset association data a, virtual asset association data b, virtual asset association data c, virtual asset association data d, virtual asset association data f). The computer device does not need to update the data in the business database.

[0143] If the data storage method in the business database is a structured classification storage method, the computer device can determine the second target type corresponding to the component update data based on the component update data of the second target component in the data update request. Among them, at least two components that make up the aircraft include the second target component; N business types include the second target type. The second graph subtree with the second target type as the initial root node can be determined based on the second target type and the item knowledge graph. In the second graph subtree, the set of data to be matched for all child nodes with the second target component as the parent node is obtained.

[0144] If there is no component update data in the set of data to be matched, that is, the structured component asset sub-data in the child nodes of the set of data to be matched does not match the component update data, the computer device can add an update child node to the parent node corresponding to the second target component in the second graph subtree, convert the format of the component update data to obtain structured component update data, and allocate the structured component update data to the update child node, that is, determine the structured component update data as the content of the update child node to obtain a graph update subtree. Further, the item knowledge graph can be updated based on the graph update subtree to obtain an item updated knowledge graph, store the item updated knowledge graph in the business database associated with the first business object, generate a data update success notification, and display the data update success notification in the intelligent message text box on the intelligent session page. Specifically, the computer device can detect the inflow of new data (structured component update data) in the second graph subtree through Incremental Learning technology (that is, the process of adding an update child node to the parent node corresponding to the second target component in the second graph subtree and determining the structured component update data as the content of the update child node), and further identify the changes in the data in the second graph subtree, such as the addition of new entities (entities corresponding to the structured component update data), the update of existing entity attributes, or the formation of new relationships. Based on the changes in the data in the second graph subtree, entity alignment, etc. are realized to add the structured component update data to the item knowledge graph to obtain a graph update subtree, so as to ensure that the final item knowledge graph can maintain consistency without contradictory or redundant information after being updated to obtain the item updated knowledge graph. Optionally, if there is component update data in the set of data to be matched, a data duplication notification is generated and the data duplication notification is displayed in the intelligent message text box on the intelligent session page.

[0145] It can be understood that all the operations performed by the computer device in the above process can be completed by the intelligent object calling the computer device after receiving the business processing request text or data update request sent by the first business object.

[0146] Through the above process, two different ways of classified storage of files (virtual asset association data) based on N business types (vectorized classified storage and structured classified storage) are realized, reducing the retrieval volume during retrieval and improving the retrieval efficiency. And through the intelligent object, the file system (including the business database) is intelligently managed. The intelligent object obtains the business requirements (business processing request text) of the first business object, and parses the business semantic information of the business processing request text, so as to perform targeted intelligent retrieval in the file system. And based on the vectorized classified storage or structured classified storage method, the business database can further improve the matching efficiency between the business semantic information and the classified data in the business database, determine the business result more quickly, and display it on the intelligent conversation page. Classified storage of data based on N business types. During the subsequent data reading process of a certain target component (for example, the first target component) of the aircraft under a certain specified business type (for example, the first target type), only the data corresponding to the first target component under the first target type needs to be queried, which reduces the retrieval threshold, saves data query time, and improves data reading and retrieval efficiency. And the entire retrieval process is completed by the intelligent object, which can improve the retrieval accuracy and save labor costs. When receiving a data update request sent by the first business object, data can also be updated through the intelligent object, without manual operation and retrieving the business database to determine whether to update the data, further saving labor costs and improving business processing efficiency.

[0147] Further, please refer to Figure 6 , Figure 6 which is a schematic diagram of a data processing device provided by an embodiment of the present application. The data processing device 600 may be a computer program (including program code, etc.) running in a computer device. For example, the data processing device 600 may be an application software; the data processing device 600 may be used to execute the corresponding steps in the method provided by the embodiment of the present application. As Figure 6 shown, the data processing device 600 may be used for Figure 3 and Figure 4 the computer devices in the corresponding embodiments. Specifically, the device may include: a data storage module 11, a data parsing module 12, a data determination module 13, a data acquisition module 14, a data retrieval module 15, a first data update module 16, and a second data update module 17.

[0148] The data parsing module 12 is used to obtain the business processing request text sent by the first business object in the intelligent conversation page, and parse the business processing request text to obtain business semantic information and intention requirement information; the intelligent conversation page includes an object avatar corresponding to the first business object and an object message text box associated with the object avatar, and the object message text box contains the business processing request text;

[0149] A data determination module 13, configured to obtain a first object identifier of a first service object, and based on the first object identifier, determine a service database associated with the first object identifier; the service database includes stored data associated with virtual asset association data corresponding to N service types respectively; the virtual asset association data under each service type includes component asset data for at least one component of an aircraft; N is a positive integer;

[0150] A data acquisition module 14, configured to obtain a first target type indicated by service semantic information, obtain a first set of service data belonging to the first target type from classification data corresponding to N service types in the service database, and based on the service semantic information, determine a first target component associated with the service semantic information; the N service types include the first target type;

[0151] A data retrieval module 15, configured to perform information retrieval in the first set of service data, obtain retrieval content associated with the first target component, generate a service result for a service processing request text based on intent requirement information and the retrieval content, display a smart avatar on a smart conversation page, and a smart message text box including the service result associated with the smart avatar; the smart avatar is used for session interaction with an object avatar on the smart conversation page.

[0152] In a possible implementation manner, the data processing device 600 further includes a data storage module 11, and the data storage module 11 is specifically configured to perform the following operations:

[0153] Obtain virtual asset association data for an aircraft sent by a first service object; the aircraft is composed of at least two components;

[0154] Obtain an asset data classification rule, and classify the virtual asset association data of the aircraft through N service types in the asset data classification rule to obtain N types of classification data; each type of classification data corresponds to one service type;

[0155] Generate stored data according to the N types of classification data, and store the stored data in a service database associated with the first service object.

[0156] In a possible implementation manner, the N service types include an asset service data type, a basic service data type, and a transaction service data type; the N types of classification data include a first classification data, a second classification data, and a third classification data; when the data storage module 11 is configured to classify the virtual asset association data of the aircraft through N service types in the asset data classification rule to obtain N types of classification data, the data storage module 11 is specifically configured to perform the following operations:

[0157] Obtain the data keyword fields corresponding to the asset business data types; the data keyword fields include usage record keyword fields, maintenance record keyword fields, and depreciation record keyword fields;

[0158] In the virtual asset association data of the aircraft, obtain the historical usage records corresponding to the usage record keyword fields, the historical maintenance records corresponding to the maintenance record keyword fields, and the historical depreciation records corresponding to the depreciation record keyword fields, and determine the historical usage records, historical maintenance records, and historical depreciation records as the first classification data corresponding to the asset business data types;

[0159] Obtain the factory keyword fields corresponding to the basic business data types. In the virtual asset association data of the aircraft, obtain the factory association data corresponding to the factory keyword fields, and determine the factory association data as the second classification data corresponding to the basic business data types;

[0160] Obtain the transaction keyword fields corresponding to the transaction business data types. In the virtual asset association data of the aircraft, obtain the transaction association data corresponding to the transaction keyword fields, and determine the transaction association data as the third classification data corresponding to the transaction business data types.

[0161] In a possible implementation, the N classification data includes classification data S i , where i is a positive integer less than or equal to N; the components of the aircraft include component M j , where j is a positive integer; the classification data S i includes the component asset data corresponding to A components respectively; the A components belong to at least two components in the aircraft; the A components include component M j , where A is a positive integer; when the data storage module 11 is used to generate storage data according to the N classification data and store the storage data into the business database associated with the first business object, the data storage module 11 is specifically used to perform the following operations:

[0162] Perform vectorization processing on the business type corresponding to the classification data S i to obtain a type vectorization result;

[0163] Perform vectorization processing on the A components respectively to obtain A component vectorization results;

[0164] Obtain the target component asset data of component M j under the classification data S i , and perform text parsing on the target component asset data based on the file type corresponding to the target component asset data to obtain a data parsing result corresponding to the target component asset data;

[0165] Perform vectorization processing on the data parsing result to obtain a data vectorization result corresponding to the target component asset data;

[0166] Determine the type vectorization result, component vectorization result, and data vectorization result as stored data, and store the stored data in the business database associated with the first business object.

[0167] In a possible implementation, the N types of classification data include classification data S i , where i is a positive integer less than or equal to N; the components of the aircraft include component M j , where j is a positive integer; the classification data S i includes component asset data corresponding to A components respectively; the A components belong to at least two components of the aircraft; the A components include component M j , where A is a positive integer; when the data storage module 11 is used to generate stored data according to the N types of classification data and store the stored data in the business database associated with the first business object, the data storage module 11 is specifically used to perform the following operations:

[0168] Determine the business type corresponding to the classification data S i as the initial root node, and determine the A components as A parent nodes under the initial root node;

[0169] Obtain the target component asset data of component M j under the classification data S i , and respectively perform format conversion on the P component asset sub-data in the target component asset data to obtain P structured component asset sub-data; P is a positive integer;

[0170] Allocate the P structured component asset sub-data to the sub-nodes under the parent node corresponding to component M j respectively;

[0171] When generating the sub-nodes under the parent nodes corresponding to the A components respectively, determine the initial root node, the A parent nodes, and the sub-nodes corresponding to the A parent nodes respectively as the graph sub-tree for the classification data S i ; one sub-node is used to allocate one structured component asset sub-data;

[0172] When obtaining the graph sub-trees corresponding to the N types of classification data respectively, determine the aircraft as the graph root node, generate the item knowledge graph corresponding to the aircraft based on the N graph sub-trees and the graph root node, determine the item knowledge graph as the stored data, and store the stored data in the business database associated with the first business object.

[0173] In a possible implementation, the data determination module 13 is configured to obtain a first target type indicated by the business semantic information, obtain a first set of business data belonging to the first target type from the classification data corresponding to N business types in the business database, and when determining a first target component associated with the business semantic information based on the business semantic information, the data determination module 13 is specifically configured to perform the following operations:

[0174] Extract keywords from the business semantic information to obtain B business keywords, perform vectorization processing on each business keyword to obtain B keyword vectors, perform vector similarity matching based on the B keyword vectors and the type vectorization results corresponding to N business types respectively, and determine the business type corresponding to the matched type vectorization result as the first target type; B is a positive integer;

[0175] Obtain a first set of business data belonging to the first target type from the classification data corresponding to N business types in the business database;

[0176] Perform vector similarity matching based on the keyword vectors corresponding to the remaining business keywords and the component vectorization results corresponding to each component in the first set of business data respectively, and determine the component corresponding to the matched component vectorization result as the first target component associated with the business semantic information; the remaining business keywords are the business keywords among the B business keywords except those that match the type vectorization result corresponding to the first target type.

[0177] In a possible implementation, the data processing device 600 further includes a first data update module 16, and the first data update module 16 is specifically configured to perform the following operations:

[0178] When a data update request for a second target component sent by a first business object is obtained on the intelligent conversation page, based on the data update request, obtain the to-be-matched asset data belonging to the second target component from the N types of classification data respectively; at least two components include the second target component; the data update request includes component update data of the second target component; the to-be-matched asset data includes the component asset data corresponding to the second target component in the N types of classification data;

[0179] Match the to-be-matched asset data with the component update data. If the component update data does not exist in the to-be-matched asset data, perform vectorization processing on the component update data to obtain vectorized update data, store the vectorized update data in the business database associated with the first business object, generate a data update success notification, and display the data update success notification in the intelligent message text box on the intelligent conversation page;

[0180] If there is component update data in the asset data to be matched, a data duplication notice is generated and displayed in the intelligent message text box on the intelligent session page.

[0181] In a possible implementation, the data determination module 13 is used to obtain the first target type indicated by the business semantic information, and obtain the first set of business data belonging to the first target type from the classification data corresponding to each of the N business types in the business database. When determining the first target component associated with the business semantic information based on the business semantic information, the data determination module 13 specifically performs the following operations:

[0182] Extract keywords from the business semantic information to obtain B business keywords; B is a positive integer;

[0183] Traverse in the item knowledge graph based on the B business keywords. When a type node matching the B business keywords is traversed, the business type corresponding to the matching type node is determined as the first target type;

[0184] Obtain the first graph subtree with the first target type as the initial root node from the item knowledge graph, and determine the first graph subtree as the first set of business data; the N graph subtrees in the item knowledge graph include the first graph subtree;

[0185] Traverse in the first set of business data based on the first keyword. When a component node matching the first keyword is traversed, the matching component node is determined as the first target component associated with the business semantic information; the first keyword is the business keyword among the B business keywords used to indicate the component;

[0186] When the data retrieval module 15 is used to perform information retrieval in the first set of business data and obtain the retrieval content associated with the first target component, the data retrieval module 15 specifically performs the following operations:

[0187] Determine the business keywords other than the business keyword used to indicate the business type and the business keyword used to indicate the component among the B business keywords as the keywords to be matched;

[0188] In the first set of business data, match the keywords to be matched with each child node under the first target component, and determine the corresponding structured component asset sub-data in the matching data child nodes as the retrieval content associated with the first target component.

[0189] In a possible implementation, the data processing device 600 further includes a second data update module 17, and the second data update module 17 specifically performs the following operations:

[0190] When a data update request for a second target component sent by a first service object is obtained on the intelligent conversation page, based on the component update data of the second target component in the data update request, determine the second target type corresponding to the component update data; at least two components include the second target component; N service types include the second target type;

[0191] Based on the second target type and the item knowledge graph, determine a second graph subtree with the second target type as the initial root node. In the second graph subtree, obtain the set of data to be matched for all child nodes with the second target component as the parent node;

[0192] If the component update data does not exist in the set of data to be matched, then add an update child node to the parent node corresponding to the second target component in the second graph subtree, allocate the component update data to the update child node, obtain a graph update subtree, update the item knowledge graph based on the graph update subtree, obtain an updated item knowledge graph, store the updated item knowledge graph in the service database associated with the first service object, generate a data update success notification, and display the data update success notification in the intelligent message text box on the intelligent conversation page;

[0193] If the component update data exists in the set of data to be matched, then generate a data duplication notification and display the data duplication notification in the intelligent message text box on the intelligent conversation page.

[0194] In a possible implementation, when the data retrieval module 15 is used to generate a service result for the service processing request text based on the intent requirement information and the retrieval content, the data retrieval module 15 specifically performs the following operations:

[0195] Identify key information from the retrieval results to obtain key data;

[0196] If the intent requirement information indicates a data analysis requirement, then perform data analysis on the key data based on the intent requirement information to obtain a data analysis result, and determine the data analysis result as the service result for the service processing request text;

[0197] If the intent requirement information indicates a drawing processing requirement, then perform image drawing based on the intent requirement information and the key data to obtain a drawn data graph, and determine the drawn data graph as the service result for the service processing request text.

[0198] In a possible implementation, when the data retrieval module 15 is used to perform data analysis on the key data based on the intent requirement information to obtain a data analysis result, the data retrieval module 15 specifically performs the following operations:

[0199] Obtain Q search data for key data and intent requirement information, and respectively extract features from the Q search data through a large language model to obtain Q first extracted features; Q is a positive integer;

[0200] Extract features from the key data through a large language model to obtain second extracted features, and perform cross-attention processing on the Q first extracted features and the second extracted features to obtain attention scores corresponding to the Q first extracted features respectively;

[0201] Determine the search data associated with the attention scores greater than or equal to the attention threshold among the Q attention scores as the data analysis result.

[0202] In a possible implementation manner, when the data retrieval module 15 is used to perform image drawing based on the intent requirement information and the key data to obtain a drawn data graph, the data retrieval module 15 is specifically used to perform the following operations:

[0203] Identify the type of the drawn image according to the intent requirement information, obtain an image template with the type of the drawn image in the image template library, generate a Gaussian noise image according to the image template and the initial noise data, and input the Gaussian noise image, the key data, and the intent requirement information into the text-to-image model;

[0204] Extract features from the Gaussian noise image through the text-to-image model to obtain Gaussian noise features, and perform forward diffusion processing on the Gaussian noise features to obtain a forward noise vector;

[0205] Perform feature encoding on the key data through the text-to-image model to obtain data encoding features, perform feature encoding on the intent requirement information to obtain intent encoding features, and perform feature splicing on the data encoding features and the intent encoding features to obtain spliced encoding features;

[0206] Denoise the Gaussian noise image according to the forward noise vector and the spliced encoding features to obtain a drawn data graph.

[0207] Please refer to Figure 7 , Figure 7 is a schematic structural diagram of a computer device provided by an embodiment of the present application. As Figure 7As shown in the figure, the computer device in the embodiment of the present application may include: a processor 701, a network interface 704, and a memory 705. In addition, the computer device 700 may further include: a user interface 703 and at least one communication bus 702. Among them, the communication bus 702 is used to realize the connection and communication between these components. Among them, the user interface 703 may include a display screen (Display) and a keyboard (Keyboard). Optionally, the user interface 703 may further include a standard wired interface and a wireless interface. The network interface 704 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 705 may be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. Optionally, the memory 705 may also be at least one storage device located far from the aforementioned processor 701. As Figure 7 shown, the memory 705, as a computer-readable storage medium, may include an operating system, a network communication module, a user interface module, and a device control application program.

[0208] The network interface 704 may provide a network communication network element; while the user interface 703 is mainly used to provide an input interface for the user; and the processor 701 may be used to call the device control application program stored in the memory 705 to perform the following operations:

[0209] Obtain the service processing request text sent by the first service object in the intelligent session page, parse the service processing request text to obtain service semantic information and intention requirement information; the intelligent session page includes the object avatar corresponding to the first service object and the object message text box associated with the object avatar, and the object message text box contains the service processing request text;

[0210] Obtain the first object identifier of the first service object, and based on the first object identifier, determine the service database associated with the first object identifier; the service database includes virtual asset association data corresponding to N service types respectively; the virtual asset association data under each service type includes component asset data for at least one component of the aircraft; N is a positive integer;

[0211] Obtain the first target type indicated by the service semantic information, obtain the first service data set belonging to the first target type from the classification data corresponding to the N service types in the service database, and based on the service semantic information, determine the first target component associated with the service semantic information; the N service types include the first target type;

[0212] Retrieve information in the first business data set to obtain the retrieved content associated with the first target component, generate a business result for the business processing request text based on the intent requirement information and the retrieved content, and display an intelligent avatar and an intelligent message text box containing the business result associated with the intelligent avatar on the intelligent conversation page; the intelligent avatar is used for conversation interaction with the object avatar on the intelligent conversation page.

[0213] In addition, it should be noted here that: The embodiments of the present application also provide a computer-readable storage medium, which stores a computer program, and the computer program is suitable for being loaded and executed by the processor Figure 3 or Figure 4 the methods provided in each step in, specifically, reference can be made to the Figure 3 or Figure 4 implementation manners provided in each step in, which will not be elaborated here. In addition, the description of the beneficial effects of adopting the same method will not be elaborated either. For the technical details not disclosed in the embodiments of the computer-readable storage medium involved in the present application, please refer to the description of the method embodiments of the present application. As an example, the computer program can be deployed to be executed on a computer device, or on multiple computer devices located at one place, or, on multiple computer devices distributed at multiple places and interconnected through a communication network.

[0214] The computer-readable storage medium can be the device provided in any of the foregoing embodiments or the internal storage unit of the computer device, such as the hard disk or memory of the computer device. The computer-readable storage medium can also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the computer-readable storage medium can also include both the internal storage unit and the external storage device of the computer device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium can also be used to temporarily store the data that has been output or will be output.

[0215] The embodiments of the present application also provide a computer program product or a computer program, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes Figure 3 or Figure 4 the methods provided in the various alternative manners in, and therefore, it will not be elaborated here.

[0216] In the description, claims, and drawings of the embodiments of this application, terms such as "first" and "second" are used to distinguish different objects, rather than to describe a specific order. In addition, the term "comprising" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product, or equipment that includes a series of steps or units is not limited to the listed steps or modules, but may optionally further include steps or modules not listed, or may optionally further include other step units inherent to these processes, methods, devices, products, or equipment.

[0217] In the embodiments of this application, the term "module" or "unit" refers to a computer program with a predetermined function or a part of a computer program, which works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as a processing circuit or a memory), or a combination thereof. Similarly, one processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of an overall module or unit that includes the function of that module or unit.

[0218] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in this description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0219] The methods and related devices provided in the embodiments of this application are described with reference to the method flowcharts and / or structural schematic diagrams provided in the embodiments of this application. Specifically, each process and / or block of the method flowchart and / or structural schematic diagram, and the combination of the processes and / or blocks in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable devices generate for implementation in the process Figure 1 a process or multiple processes and / or structural schematic Figure 1a device for the functions specified in one or more boxes. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including the instruction device, or are transmitted via a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wirelessly (e.g., infrared, wireless, microwave, etc.). The instruction device implements the steps in the process Figure 1 one process or multiple processes and / or structural schematic Figure 1 the functions specified in one or more boxes. These computer program instructions can also be loaded onto a computer or other programmable device, such that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more boxes in one process or multiple processes and / or structural schematic Figure 1 one process or multiple processes and / or structural schematic steps for the functions specified in one or more boxes.

[0220] The steps in the method embodiments of this application can be adjusted, combined, and deleted according to actual needs.

[0221] The modules in the device embodiments of this application can be combined, divided, and deleted according to actual needs.

[0222] The foregoing disclosure is only for the preferred embodiments of this application, and of course cannot be used to limit the scope of rights of this application. Therefore, equivalent changes made according to the claims of this application still fall within the scope covered by this application.

Claims

1. A data processing method, characterized in that: include: Acquire a business processing request text sent by a first business object in an intelligent conversation page, parse the business processing request text, and obtain business semantic information and intention requirement information; the intelligent conversation page includes an object avatar corresponding to the first business object and an object message text box associated with the object avatar, and the object message text box contains the business processing request text; Obtaining a first object identifier of the first business object, and determining a business database associated with the first object identifier based on the first object identifier; the business database includes storage data associated with virtual asset associated data corresponding to N business types respectively; the virtual asset associated data under each business type includes component asset data for at least one component of the aircraft; N is a positive integer; Acquire a first target type indicated by the business semantic information, acquire a first business data set belonging to the first target type from the classification data corresponding to N business types in the business database, and determine a first target component associated with the business semantic information based on the business semantic information; the N business types include the first target type; Performing information retrieval in the first business data set, acquiring retrieval content associated with the first target component, generating a business result for the business processing request text based on the intention requirement information and the retrieval content, and displaying an intelligent avatar and an intelligent message text box associated with the intelligent avatar and containing the business result in the intelligent conversation page; The smart avatar is used to perform conversation interaction with the object avatar in the smart conversation page.

2. The method according to claim 1, characterized in that The method further comprises: Acquire virtual asset association data for an aircraft sent by a first business object; the aircraft is composed of at least two components; Acquire asset data classification rules, and classify virtual asset-related data of the aircraft according to N types of business types in the asset data classification rules to obtain N types of classification data; each type of classification data corresponds to one business type; Generate storage data according to the N types of classification data, and store the storage data in a business database associated with the first business object.

3. The method according to claim 2, characterized in that The N types of business types include asset business data types, basic business data types, and transaction business data types; the N types of classified data include first classified data, second classified data, and third classified data; The virtual asset associated data of the aircraft is classified according to the N business types in the asset data classification rule to obtain N types of classified data, including: Acquire the data key fields corresponding to the asset business data type; the data key fields include the usage record key field, the maintenance record key field and the depreciation record key field; In the virtual asset associated data of the aircraft, historical usage records corresponding to the usage record key field, historical maintenance records corresponding to the maintenance record key field, and historical damage records corresponding to the damage record key field are obtained, and the historical usage records, the historical maintenance records, and the historical damage records are determined as first classification data corresponding to the asset business data type; Acquire a factory key field corresponding to the basic business data type, acquire factory-related data corresponding to the factory key field in the virtual asset-related data of the aircraft, and determine the factory-related data as the second classification data corresponding to the basic business data type; Obtain a transaction key field corresponding to the transaction business data type, obtain transaction-related data corresponding to the transaction key field in the virtual asset-related data of the aircraft, and determine the transaction-related data as third classification data corresponding to the transaction business data type.

4. The method according to claim 2, characterized in that: The N types of classification data include classification data S i , i is a positive integer less than or equal to N; the aircraft components include components M j , j is a positive integer; the classification data S i The component asset data includes A components respectively corresponding to the A components; the A components belong to at least two components in the aircraft; the A components include the component M j , A is a positive integer; generating storage data according to the N types of classification data, and storing the storage data in a business database associated with the first business object, comprises: For the classification data S i The corresponding business type is vectorized to obtain a type vectorization result; Performing vectorization processing on the A components respectively to obtain vectorization results of the A components; Get the component M j In the classification data S i The target component asset data under the target component asset data, based on the file type corresponding to the target component asset data, performs text parsing on the target component asset data to obtain a data parsing result corresponding to the target component asset data; Performing vectorization processing on the data analysis result to obtain a data vectorization result corresponding to the target component asset data; The type vectorization result, the component vectorization result and the data vectorization result are determined as storage data, and the storage data is stored in a business database associated with the first business object.

5. The method according to claim 2, characterized in that: The N types of classification data include classification data S i , i is a positive integer less than or equal to N; the aircraft parts include parts M j , j is a positive integer; the classification data S i The component asset data includes A components respectively corresponding to the A components; the A components belong to at least two components in the aircraft; the A components include the component M j , A is a positive integer; generating storage data according to the N types of classification data, and storing the storage data in a business database associated with the first business object, comprises: The classification data S i The corresponding business type is determined as an initial root node, and the A components are determined as A parent nodes under the initial root node; Get the component M j In the classification data S i The target component asset data under the target component asset data are format-converted for P component asset sub-data in the target component asset data to obtain P structured component asset sub-data; P is a positive integer; The P structured component asset sub-data are respectively assigned to the component M j In the child node under the corresponding parent node; When generating child nodes under the parent nodes corresponding to the A components respectively, the initial root node, the A parent nodes and the child nodes corresponding to the A parent nodes respectively are determined as the child nodes corresponding to the classification data S i A graph subtree of; a child node is used to allocate a structured component asset sub-data; When graph subtrees corresponding to N types of classification data are obtained, the aircraft is determined as a graph root node, and an object knowledge graph corresponding to the aircraft is generated based on the N graph subtrees and the graph root node. The object knowledge graph is determined as storage data, and the storage data is stored in a business database associated with the first business object.

6. The method according to claim 4, characterized in that The step of obtaining the first target type indicated by the business semantic information, obtaining a first business data set belonging to the first target type from the classification data corresponding to the N business types in the business database, and determining a first target component associated with the business semantic information based on the business semantic information includes: Perform keyword extraction on the business semantic information to obtain B business keywords, perform vectorization processing on each business keyword to obtain B keyword vectors, perform vector similarity matching based on the B keyword vectors and the type vectorization results corresponding to the N types of business types, and determine the business type corresponding to the matching type vectorization results as the first target type; B is a positive integer; Acquire a first business data set belonging to the first target type from the classification data corresponding to the N business types in the business database; Based on the keyword vectors corresponding to the remaining business keywords, vector similarity matching is performed with the component vectorization results corresponding to each component under the first business data set, and the component corresponding to the matching component vectorization result is determined as the first target component associated with the business semantic information; the remaining business keywords are the business keywords among the B business keywords, except for the business keywords that match the type vectorization results corresponding to the first target type.

7. The method according to claim 4, characterized in that The method further comprises: When a data update request for a second target component sent by the first business object is obtained in the smart conversation page, based on the data update request, to-be-matched asset data belonging to the second target component are respectively obtained from the N types of classified data; the at least two components include the second target component; the data update request includes component update data of the second target component; the to-be-matched asset data includes component asset data corresponding to the second target component in the N types of classified data; Matching the to-be-matched asset data with the component update data; if the component update data does not exist in the to-be-matched asset data, vectorizing the component update data to obtain vectorized update data; storing the vectorized update data in a business database associated with the first business object; generating a data update success notification; and displaying the data update success notification in the smart message text box in the smart conversation page; If the component update data exists in the asset data to be matched, a data duplication notification is generated and displayed in the smart message text box in the smart conversation page.

8. The method according to claim 5, characterized in that The step of obtaining the first target type indicated by the business semantic information, obtaining a first business data set belonging to the first target type from the classification data corresponding to the N business types in the business database, and determining a first target component associated with the business semantic information based on the business semantic information includes: Extract keywords from the business semantic information to obtain B business keywords, where B is a positive integer; Traversing the item knowledge graph based on the B business keywords, when a type node matching the B business keywords is traversed, determining the business type corresponding to the matching type node as the first target type; From the item knowledge graph, a first graph subtree with the first target type as an initial root node is obtained, and the first graph subtree is determined as a first business data set; the N graph subtrees in the item knowledge graph include the first graph subtree; Traversing the first business data set based on a first keyword, when traversing to a component node matching the first keyword, determining the matching component node as a first target component associated with the business semantic information; the first keyword is a business keyword for indicating a component among the B business keywords; The performing information retrieval in the first business data set to obtain retrieval content associated with the first target component includes: Determining the business keywords other than the business keywords indicating the business type and the business keywords indicating the component among the B business keywords as keywords to be matched; In the first business data set, the to-be-matched keyword is matched with each sub-node under the first target component, and the corresponding structured component asset sub-data in the matched data sub-node is determined as the search content associated with the first target component.

9. The method according to claim 5, characterized in that The method further comprises: When a data update request for a second target component sent by the first business object is obtained in the smart conversation page, based on the component update data of the second target component in the data update request, a second target type corresponding to the component update data is determined; the at least two components include the second target component; and the N business types include the second target type; Based on the second target type and the item knowledge graph, determine a second graph subtree with the second target type as an initial root node, and in the second graph subtree, obtain a to-be-matched data set of all child nodes with the second target component as a parent node; If the component update data does not exist in the to-be-matched data set, then in the second graph subtree, an update child node is added to the parent node corresponding to the second target component, and the component update data is assigned to the update child node to obtain a graph update subtree, and the item knowledge graph is updated based on the graph update subtree to obtain an item update knowledge graph, and the item update knowledge graph is stored in the business database associated with the first business object, a data update success notification is generated, and the data update success notification is displayed in the smart message text box in the smart conversation page; If the component update data exists in the to-be-matched data set, a data duplication notification is generated and displayed in the smart message text box in the smart conversation page.

10. The method according to claim 1, characterized in that The generating of a business result for the business processing request text based on the intention requirement information and the search content includes: Identify key information of the search content to obtain key data; If the intention requirement information indicates a data analysis requirement, performing data analysis on the key data based on the intention requirement information to obtain a data analysis result, and determining the data analysis result as a business result for the business processing request text; If the intention requirement information indicates a drawing processing requirement, image drawing is performed based on the intention requirement information and the key data to obtain a drawing data graph, and the drawing data graph is determined as a business result for the business processing request text.

11. The method according to claim 10, characterized in that The performing data analysis on the key data based on the intention requirement information to obtain a data analysis result includes: Acquire Q search data for the key data and the intention demand information, and perform feature extraction on the Q search data respectively through a large language model to obtain Q first extracted features; Q is a positive integer; Performing feature extraction on the key data through the large language model to obtain a second extracted feature, performing cross attention processing on the Q first extracted features and the second extracted features to obtain attention scores corresponding to the Q first extracted features respectively; The search data associated with the attention scores greater than or equal to the attention threshold among the Q attention scores are determined as the data analysis results.

12. The method according to claim 10, characterized in that The image drawing is performed based on the intention requirement information and the key data to obtain a drawing data graph, including: Identify the type of drawn image according to the intention requirement information, obtain an image template having the type of drawn image in an image template library, generate a Gaussian noise image according to the image template and initial noise data, and input the Gaussian noise image, the key data and the intention requirement information into a Vincent graph model; Extracting features of the Gaussian noise image using the Vincent graph model to obtain Gaussian noise features, and performing forward diffusion processing on the Gaussian noise features to obtain a forward noise vector; Performing feature encoding on the key data through the document graph model to obtain data encoding features, performing feature encoding on the intention requirement information to obtain intention encoding features, and performing feature splicing on the data encoding features and the intention encoding features to obtain spliced ​​encoding features; The Gaussian noise image is denoised according to the forward noise vector and the concatenated coding feature to obtain a drawing data graph.

13. A data processing device, characterized in that: include: a data parsing module, configured to obtain a business processing request text sent by a first business object in an intelligent conversation page, parse the business processing request text, and obtain business semantic information and intention requirement information; the intelligent conversation page includes an object avatar corresponding to the first business object and an object message text box associated with the object avatar, and the object message text box contains the business processing request text; a data determination module, configured to obtain a first object identifier of the first business object, and determine a business database associated with the first object identifier based on the first object identifier; the business database includes storage data associated with virtual asset associated data corresponding to N business types respectively; the virtual asset associated data under each business type includes component asset data for at least one component of the aircraft; N is a positive integer; a data acquisition module, configured to acquire a first target type indicated by the business semantic information, acquire a first business data set belonging to the first target type from the classification data corresponding to the N business types in the business database, and determine a first target component associated with the business semantic information based on the business semantic information; the N business types include the first target type; a data retrieval module, configured to perform information retrieval in the first business data set, obtain retrieval content associated with the first target component, generate a business result for the business processing request text based on the intention requirement information and the retrieval content, and display an intelligent avatar and an intelligent message text box associated with the intelligent avatar containing the business result in the intelligent conversation page; The smart avatar is used to perform conversation interaction with the object avatar in the smart conversation page.

14. A computer device, characterized in that: Includes processor, memory, input and output interfaces; The processor is connected to the memory and the input / output interface respectively, wherein the input / output interface is used to receive and output data, the memory is used to store a computer program, and the processor is used to call the computer program so that the computer device executes the method described in any one of claims 1-12.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded and executed by a processor, so that a computer device having the processor executes the method according to any one of claims 1 to 12.

16. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 12 is implemented.

Images