Structured data processing method and device, computer equipment and storage medium

Abstract

The invention relates to a structured data processing method and device, computer equipment, a storage medium and a computer program product. The method can be applied to artificial intelligence, such as a scene in which structured data processing is carried out on a service through dialogue interaction; the method comprises the following steps: generating first inquiry information according to knowledge data of a service, and inquiring entity data according to the first inquiry information; generating second inquiry information according to the entity data and the service logic, and replying the second inquiry information to obtain first structured data; and when the entity data is updated to the target entity data, determining third inquiry information for data derivation according to the target entity data, the entity data and the first structured data, and generating second structured data corresponding to the target entity data according to the third inquiry information. By adopting the method, the efficiency of updating the structured data of the business can be improved when entity data update is generated in business development.

Description

Technical Field

[0001] The present application relates to the field of computer technologies, and particularly to a method, apparatus, computer device, storage medium, and computer program product for processing structured data. Background Art

[0002] Structured data is also known as Structured Query Language (SQL) data. SQL data is widely used in business, such as for managing business data, facilitating operations for adding, deleting, modifying, and querying business data.

[0003] With the update and iteration of business, business data will change accordingly, and the instructions for performing operations of adding, deleting, modifying, and querying data will also change accordingly to correctly execute the updated business. In related technologies, it is usually manual to update the SQL data of the business according to the development of the business, resulting in low processing efficiency of SQL data. Summary of the Invention

[0004] Based on this, to address the above technical problems, it is necessary to provide a method, apparatus, computer device, computer-readable storage medium, and computer program product for processing structured data, which can improve the efficiency of updating the structured data of the business when entity data is updated due to business development.

[0005] In a first aspect, the present application provides a method for processing structured data. The method includes:

[0006] Generating first query information based on the knowledge data of the business, and querying the entity data of the business according to the first query information; determining the business logic based on the entity data; generating second query information based on the entity data and the business logic, and performing a reply process on the second query information to obtain first structured data; when the entity data is updated to target entity data, determining third query information for data derivation based on the target entity data, the entity data, and the first structured data, and generating second structured data corresponding to the target entity data according to the third query information.

[0007] In a second aspect, the present application further provides a device for processing structured data. The device includes:

[0008] An entity data determination module, configured to generate first query information based on the knowledge data of the business, and query the entity data of the business according to the first query information;

[0009] A business logic determination module, configured to determine the business logic based on the entity data;

[0010] The first structured data determination module is configured to generate a second query message based on entity data and business logic, and perform a response process on the second query message to obtain first structured data;

[0011] The second structured data determination module is configured to, when the entity data is updated to target entity data, determine a third query message for data derivation based on the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third query message.

[0012] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0013] Generate a first query message based on the knowledge data of the business, and query the entity data of the business according to the first query message; determine the business logic based on the entity data; generate a second query message based on the entity data and the business logic, and perform a response process on the second query message to obtain first structured data; when the entity data is updated to target entity data, determine a third query message for data derivation based on the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third query message.

[0014] In some embodiments, the first structured data determination module is further configured to obtain the entities included in the second query message, the entity attributes of the entities, and the entity relationships; determine the table construction instructions of the entities according to the other entities corresponding to the entities, the entity attributes, and the entity relationships; generate the table structure of the entities according to the table construction instructions of the entities; determine the data processing instructions corresponding to the business logic based on the table structure of the entities; the first structured data includes the table construction instructions of the entities, the table structure of the entities, and the data processing instructions.

[0015] In some embodiments, the second structured data determination module is further configured to determine the difference data between the target entity data and the entity data included in the third query message; perform data processing on the first structured data included in the third query message according to the difference data to obtain second structured data corresponding to the target entity data.

[0016] In some embodiments, the second structured data determination module is further configured to: search for a table creation instruction of a first target entity corresponding to the differential data in the first structured data included in the third query information; perform data processing on the table creation instruction corresponding to the first target entity according to the differential data to obtain an updated table creation instruction of the first target entity; generate a table structure of the updated first target entity according to the updated table creation instruction of the first target entity; search for a data processing instruction related to the first target entity in the first structured data; perform data processing on the found data processing instruction according to the table structure of the updated first target entity to obtain a first target data processing instruction; the second structured data includes the updated table creation instruction of the first target entity, the table structure of the updated first target entity, and the first target data processing instruction.

[0017] In some embodiments, the processing apparatus for structured data further includes: a third structured data determination module, configured to, when the business logic is updated to a target business logic, determine fourth query information according to the target business logic, the business logic, the entity data, and the first structured data; generate second structured data corresponding to the target business logic according to the fourth query information.

[0018] In some embodiments, the third structured data determination module is further configured to: determine a second target entity related to the target business logic included in the fourth query information; search for a table creation instruction of the second target entity in the first structured data; perform data processing on the table creation instruction of the second target entity according to the difference between the target business logic and the business logic to obtain an updated table creation instruction of the second target entity; generate a table structure of the updated second target entity according to the updated table creation instruction of the second target entity; determine a second target data processing instruction corresponding to the target business logic according to the table structure of the updated second target entity; the second structured data includes the updated table creation instruction of the second target entity, the table structure of the updated second target entity, and the second target data processing instruction.

[0019] In some embodiments, the third structured data determination module is further configured to perform reply processing on the fourth query information through a first large model to obtain second structured data corresponding to the target business logic.

[0020] In some embodiments, the entity data determination module is further configured to perform entity data processing on the first query information through a second large model to obtain entity data of the business; the first structured data determination module is further configured to perform reply processing on the second query information through a third large model to obtain first structured data; the second structured data determination module is further configured to perform reply processing on the third query information through the first large model to obtain second structured data.

[0021] In some embodiments, the knowledge data of the service is the knowledge data of the payment service; the service logic is the resource query logic; the entity data determination module is further configured to generate a first query message according to the knowledge data of the payment service, and process the first query message through a second large model to obtain the entity data of the payment service; the first structured data determination module is further configured to generate a second query message according to the entity data of the payment service and the resource query logic, and perform a reply process on the second query message through a third large model to obtain a table creation instruction for the entity data, a table structure of the entity data, and a resource query instruction corresponding to the resource query logic; the second structured data determination module is further configured to, when the target entity data in the first data derivation query message has a new regional attribute of the account entity compared with the entity data, perform a reply process on the third query message through a data derivation large model to obtain a table creation instruction for the account entity after the new regional attribute is added, a table structure of the account entity after the new regional attribute is added, and a resource query instruction after the new regional attribute is added.

[0022] In some embodiments, the structured data processing device further includes: a first training module, configured to obtain sample knowledge data of a service, generate a fifth query message according to the sample knowledge data; perform a reply process on the fifth query message through a second large model to obtain training entity data; adjust parameters of the second large model according to the training entity data and the entity data label corresponding to the sample knowledge data until the second large model converges to obtain a converged second large model.

[0023] In some embodiments, the structured data processing device further includes: a second training module, configured to generate a sixth query message according to the sample entity data and sample service logic of the service; perform a reply process on the sixth query message through a third large model to obtain training structured data; adjust parameters of the third large model according to the structured data label corresponding to the sample entity data and sample service logic and the training structured data until the third large model converges to obtain a converged third large model.

[0024] Fourthly, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the following steps are implemented:

[0025] Generate a first query message according to the knowledge data of the service, and query the entity data of the service according to the first query message; determine the service logic based on the entity data; generate a second query message according to the entity data and the service logic, and perform a reply process on the second query message to obtain first structured data; when the entity data is updated to target entity data, determine a third query message for data derivation according to the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third query message.

[0026] In a fifth aspect, the present application also provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0027] Generate first inquiry information based on the knowledge data of the business, and query the entity data of the business according to the first inquiry information; determine the business logic based on the entity data; generate second inquiry information based on the entity data and the business logic, and perform a reply process on the second inquiry information to obtain first structured data; when the entity data is updated to target entity data, determine third inquiry information for data derivation based on the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third inquiry information.

[0028] The above-mentioned method, apparatus, computer device, storage medium, and computer program product for processing structured data generate first inquiry information through the knowledge data of the business, query the entity data of the business according to the first inquiry information, generate second inquiry information according to the entity data and the business logic, convert the entity data into first structured data according to the second inquiry information, and when the entity data is updated to target entity data, generate third inquiry information according to the target entity data, the entity data, and the first structured data, and perform data derivation according to the third inquiry information to obtain second structured data; through an intelligent question-and-answer task, gradually determine the entity data and the first structured data of the business, and when the entity data is updated to target entity data, determine the second structured data corresponding to the updated target entity data through the intelligent question-and-answer task. That is to say, in the case where the entity data changes due to business development, data derivation can be performed through a dialogue interaction method based on the original entity data, the first structured data, and the updated target entity data to obtain second structured data corresponding to the target entity data, without the need for manual rewriting of the original first structured data, thereby improving the efficiency of updating the structured data of the business when the entity data changes due to business development. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is an application environment diagram of the method for processing structured data in an embodiment;

[0030] Figure 2 It is a flowchart of the method for processing structured data in an embodiment;

[0031] Figure 3 It is a flowchart of deriving second structured data from the first structured data when the entity data is updated in an embodiment;

[0032] Figure 4Schematic flowchart of generating the second structured data based on the third interrogation information in an embodiment;

[0033] Figure 5 Schematic flowchart of processing the first structured data based on the difference data to obtain the second structured data in an embodiment;

[0034] Figure 6 Schematic flowchart of determining the table creation instruction and table structure of the target entity data and determining the target data processing instruction when updating the entity data in an embodiment;

[0035] Figure 7 Schematic flowchart of determining the second structured data when updating the business logic in an embodiment;

[0036] Figure 8 Schematic flowchart of deriving the second structured data from the first structured data when implementing entity data update based on the large model in an embodiment;

[0037] Figure 9 Schematic flowchart of the processing method of structured data in the payment business scenario in an embodiment;

[0038] Figure 10 Schematic flowchart of adjusting the parameters of the second large model in an embodiment;

[0039] Figure 11 Schematic flowchart of adjusting the parameters of the third large model in an embodiment;

[0040] Figure 12 Schematic flowchart of the processing method of structured data in another embodiment;

[0041] Figure 13 Schematic flowchart of the processing method of structured data in another embodiment;

[0042] Figure 14 Block diagram of the structure of the processing device for structured data in an embodiment;

[0043] Figure 15 Internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0044] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0045] Artificial Intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a way similar to human intelligence. Artificial intelligence also studies the design principles and implementation methods of various intelligent machines, enabling the machines to have the functions of perception, reasoning, and decision-making.

[0046] Artificial intelligence technology is an interdisciplinary subject with a wide range of fields involved, including both hardware-level and software-level technologies. The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, pre-trained model technology, operation / interaction systems, mechatronics, etc. Among them, the pre-trained model, also known as the large model or the foundation model, can be widely applied to downstream tasks in various directions of artificial intelligence after fine-tuning. The software technologies of artificial intelligence mainly include several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0047] With the research and progress of artificial intelligence technology, artificial intelligence technology has been studied and applied in multiple fields. For example, common ones include smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, driverless, autonomous driving, drones, digital twins, virtual humans, robots, artificial intelligence-generated content (AIGC), dialogue interaction, intelligent healthcare, intelligent customer service, game AI, etc. It is believed that with the development of technology, artificial intelligence technology will be applied in more fields and play an increasingly important role.

[0048] The method for processing structured data provided in the embodiments of this application can be applied to an application environment as Figure 1 shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed in the cloud or other network servers; the method for processing structured data can be executed by the terminal 102 or the server 104, or jointly executed by the terminal 102 and the server 104.

[0049] Taking the execution by server 104 using a method for processing structured data as an example, server 104 generates first query information based on the knowledge data of the business, and queries the entity data of the business according to the first query information. Server 104 determines the business logic based on the entity data; server 104 generates second query information based on the entity data and the business logic, and performs a reply process on the second query information to obtain first structured data; when the entity data is updated to target entity data, server 104 determines third query information for data derivation based on the target entity data, the entity data, and the first structured data, and generates second structured data corresponding to the target entity data according to the third query information.

[0050] Among them, the terminal 102 can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, an Internet of Things device, and a portable wearable device. The Internet of Things device can be a smart speaker, a smart TV, a smart air conditioner, a smart in-vehicle device, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc.

[0051] Server 104 can be an independent physical server or a service node in a blockchain system. The service nodes in the blockchain system form a peer-to-peer network.

[0052] In addition, server 104 can also be a server cluster composed of multiple physical servers, and can be a cloud server providing 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, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0053] The connection between the terminal 102 and the server 104 can be made through communication connection methods such as Bluetooth, USB (Universal Serial Bus), or network. This application does not limit this here.

[0054] In some embodiments, as Figure 2 shown, a method for processing structured data is provided. This method is executed by the server or terminal in Figure 1 , or can also be executed collaboratively by the server and terminal in Figure 1 . Taking the execution of this method by a computer device as an example for illustration, the computer device can be the server or terminal in Figure 1 , including the following steps:

[0055] Step 202, generate first query information based on the knowledge data of the business, and query the entity data of the business according to the first query information.

[0056] Among them, the knowledge data of the service is the knowledge content related to the service. The knowledge data may include knowledge content such as the definition of the service, the usage scenarios of the service, and the key technical points of the service. In practical applications, the service may be a payment service, a search service, a resource transfer service, etc.

[0057] The first query information is used to query the entity data included in the knowledge data. For example, the first query information is used to indicate what the entity data that can be extracted according to the knowledge data is.

[0058] The entity data includes entities, entity attributes of the entities, and entity relationships. The entity attributes can represent the characteristics or properties of the entities. For example, if the entity is an organization, the entity attributes of the organization may include: the number of employees, the registered capital, etc. The scale of the organizational structure can be determined through the entity attributes.

[0059] The entity relationship is used to represent the relationship between entities. For example, entity E1 is an account entity, entity E2 is a user entity, and the entity relationship of entity E1 includes the ownership relationship between entity E1 and entity E2.

[0060] In some embodiments, the computer device generates the first query information based on the knowledge data of the service. It may be to obtain the knowledge data of the service, obtain the first query template, embed the knowledge data of the service into the first query template, and obtain the first query information. Exemplarily, the first query template may be: "Please extract entity data from < >", and the "< >" in the first query template is used to embed the knowledge data of the service.

[0061] The computer device queries the entity data of the service according to the first query information. It may be to process the first query information through a second large model to obtain the entity data. The second large model can process intelligent question-and-answer tasks. In practical applications, the second large model can be implemented through the GPT model, and the GPT model is a Generative Pre-Trained Transformer.

[0062] Exemplarily, the computer device obtains the knowledge data of the service, embeds the knowledge data of the service into the first query template, and obtains the first query information: "Please extract entity data from <knowledge data of the service>", inputs the first query information into the second large model, and the reply output by the second large model is the entity data of the service.

[0063] Step 204, determine the business logic based on the entity data.

[0064] Among them, the business logic is used to represent the operations when executing a business. Specifically, it can be the operations that may be executed when a user uses the business. The business logic can be in the form of natural language, that is, the business logic can be represented by text. For example, if the business is a payment business, the business logic can be: "Query the transaction data for a set date", and the business logic can be: "Query the balance".

[0065] Since the entity data includes entities, entity attributes of the entities, and entity relationships, executing the business logic may involve one entity and at least one entity attribute of that entity, or may involve at least two entities, at least one entity attribute of each of the two entities, and the entity relationship between the two entities.

[0066] In some embodiments, for each entity in the entity data, the computer device obtains the entity attributes of the entity and generates business logic for querying the entity attributes; the computer device determines another entity according to the entity relationship of the entity, and generates business logic for the associated attributes between the entity and the other entity based on the entity with the entity relationship and the other entity.

[0067] In some embodiments, the business logic can be generated by a staff member based on the entity data. The staff member sets the business logic based on the entities, entity attributes of the entities, and entity relationships in the entity data and inputs it into the computer device so that the computer device obtains the business logic determined based on the entity data.

[0068] Step 206: Generate second query information based on the entity data and the business logic, and perform a reply process on the second query information to obtain first structured data.

[0069] Among them, the second query information is used to query the structured data (SQL data) corresponding to the entity data and the SQL data corresponding to the business logic. For example, the second query information is used to represent what the SQL data of the entity data and the business logic are.

[0070] The first structured data includes the SQL data corresponding to the entity data and the SQL data corresponding to the business logic.

[0071] The computer device generates the second query information based on the entity data and the business logic, which can be to obtain a second query template, embed the entity data and the business logic into the second query template to obtain the second query information. Exemplarily, the second query template can be: "The entity data is <1>, the business logic is <2>, what are the SQL data of the entity data and the business logic", where "<1>" in the second query template is used to embed the knowledge data of the business, and "<2>" is used to embed the business logic.

[0072] The computer device processes the second inquiry information to obtain the first structured data, which can be obtained by processing the second inquiry information through the third large model; that is, taking the second inquiry information as a question and asking the third large model, the third large model processes the second inquiry information and outputs the SQL data corresponding to the entity data and the SQL data corresponding to the business logic; the third large model can be implemented by GPT.

[0073] In some embodiments, the second inquiry information includes: the first structured inquiry information and the second structured inquiry information; step 206 may include: generating the first structured inquiry information based on the entity data, processing the first structured inquiry information to obtain the structured data corresponding to the entity data; generating the second structured inquiry information based on the structured data of the entity data and the business logic, and processing the second structured inquiry information to obtain the structured data corresponding to the business logic.

[0074] In practical applications, the SQL data of the entity data can be obtained by processing the first structured inquiry information through the third large model, and similarly, the SQL data corresponding to the business logic can be obtained by processing the second structured inquiry information through the third large model.

[0075] In some embodiments, processing the second inquiry information to obtain the first structured data includes: obtaining the entities included in the second inquiry information, the entity attributes of the entities, and the entity relationships; determining the table creation instructions of the entities based on the entities, entity attributes, and other entities corresponding to the entity relationships; generating the table structure of the entities based on the table creation instructions of the entities; determining the data processing instructions corresponding to the business logic based on the table structure of the entities; the first structured data includes the table creation instructions of the entities, the table structure of the entities, and the data processing instructions.

[0076] Among them, the table creation instruction is an SQL instruction for creating a table structure. When the table creation instruction is run, the corresponding table structure can be generated; the table structure is the data structure of the SQL table corresponding to the entity. The entity attributes of the entity are the fields in the table structure. The entity relationships of the entity can be represented by the relationship keys in the table structure. In the table structures of two entities with entity relationships, there are the same relationship keys; the data processing instruction is an SQL instruction for implementing the business logic.

[0077] The number of entities included in the second inquiry information can be at least one. The number of entity attributes of each entity can be at least one. For a certain entity among the at least one entity, there may be no entity relationship or there may be at least one entity relationship.

[0078] Specifically, the computer device obtains the entities included in the second inquiry information, the entity attributes of the entities, and the entity relationships. For each entity, based on the entity attributes and entity relationships of the entity, it determines the table name, column name, and column constraints, and determines the table creation instruction for the entity according to the table name, column name, and column constraints; when the computer device runs the table creation instruction of the entity, it can generate a table structure.

[0079] Exemplarily, the entity attributes of the account entity include account identifier, user identifier, and balance. There is an entity relationship between the account entity and the user entity. The table creation instruction for the account entity is:

[0080] sql

[0081] CREATE TABLE Accounts (

[0082] id INT PRIMARY KEY,

[0083] user_id INT,

[0084] balance DECIMAL(10, 2),

[0085] FOREIGN KEY (user_id) REFERENCES Users(id) );

[0087] Among them, "CREATE TABLE Accounts" indicates that the name of the table structure is Accounts (account); "id INT PRIMARY KEY" defines a column named id (representing the account identifier), with the data type being an integer, and the values in this column are unique, that is, each record in the table structure of Accounts has a unique id value.

[0088] "user_id INT" defines a column named user_id (user identifier), with the data type being INT (in integer form), and "balance DECIMAL(10, 2)," defines the balance field, which is a decimal field and can store up to 10 digits, with 2 digits being the decimal part; "FOREIGN KEY (user_id) REFERENCES Users(id)" indicates that user_id is a foreign key that references the id field of TABLE Users, and this statement can represent the entity relationship between the user entity and the account entity.

[0089] When running the above table creation instruction for the account entity, the corresponding table structure for the account entity can be created, and this table structure is used to store relevant information of the user.

[0090] Exemplarily, the entity attributes of the user entity include name and email, and the table creation instruction for the user entity is:

[0091] sql

[0092] CREATE TABLE Users (

[0093] id INT PRIMARY KEY,

[0094] name VARCHAR(255),

[0095] email VARCHAR(255) )

[0097] Among them, "CREATE TABLE Users" indicates that the name of the table structure is Users; "id INT PRIMARY KEY" defines a column named id with an integer data type, and the values in this column are unique, that is, each record in the table structure of Users has a unique id value; "name VARCHAR(255)" defines a column named name (name) with a variable-length string data type and a maximum length of 255; "email VARCHAR(255)" defines a column named email (email) with a variable-length string data type and a maximum length of 255. When the above table creation instruction for the user entity is run, the corresponding table structure of the user entity can be created, and this table structure is used to store relevant information of the user, such as name and email.

[0098] The computer device determines the data processing instruction corresponding to the business logic according to the table structure of the entity, which can be to determine the entity and entity attributes related to the business logic, and determine the query conditions of the business logic, and generate a data processing instruction based on the entity and entity attributes related to the business logic, as well as the query conditions.

[0099] Exemplarily, the business logic is: query the email of user A. It can be known that the entity related to the business logic is user A, and the entity attribute is email. Based on the table structure of user A, the data processing instruction can be generated:

[0100] sql

[0101] SELECT email FROM UsersA.

[0102] It should be noted that in the case of processing the second query information through the second large model to obtain the first structured data, specifically, the entities, entity attributes, and entity relationships included in the second query information can be obtained through the second large model, and the table creation instructions for the entities can be determined based on the entities, entity attributes, and other entities corresponding to the entity relationships; the table structure of the entities can be generated based on the table creation instructions of the entities, and the data processing instructions corresponding to the business logic can be determined based on the table structure of the entities.

[0103] In the above embodiment, based on the entities, entity attributes, and entity relationships included in the second query information, the table creation instructions for the entities are generated, and then the table structure of the entities is determined. Based on the table structure, the data processing instructions corresponding to the business logic are generated. The table creation instructions, table structure, and data processing instructions all belong to SQL data, and the data processing efficiency during business execution can be improved according to the SQL data.

[0104] Step 208, when the entity data is updated to the target entity data, determine the third query information for data derivation based on the target entity data, entity data, and the first structured data, and generate the second structured data corresponding to the target entity data based on the third query information.

[0105] Among them, the target entity data is obtained after adding, deleting, or modifying the entity data; specifically, the entity attributes can be added, deleted, or modified to obtain the target entity data; for example, the entity data includes a user entity and the entity attributes of the user entity: name and email, and the entity attribute of the user entity: balance is added to obtain the target entity data, and the target entity data includes the user entity and the entity attributes of the user entity: name, email, and balance.

[0106] For ease of explanation, the entity data is referred to as the entity data in the first stage, and the updated entity data is referred to as the target entity data in the second stage.

[0107] The third query information is used to query the second structured data obtained by data derivation of the first structured data based on the target entity data and entity data. For example, after the entity data in the first stage is updated to the target entity data in the second stage, what is the structured data in the second stage obtained by data derivation of the first structured data in the first stage.

[0108] After updating the entity data to the target entity data, in order to ensure that the overall data of the business matches the update of the entity data, it is necessary to determine the table creation instructions and table structure of the target entity data, and based on the evolution of the entity data updated to the target entity data, deduce the evolution of the business logic, and then determine the data processing instructions corresponding to the updated business logic; therefore, the second structured data corresponding to the target entity data includes: the table creation instructions and table structure of the target entity in the target entity data, and the data processing instructions corresponding to the updated business logic.

[0109] In practical applications, when entity data is updated due to business development, it is necessary to synchronously update the entity data and the structured data corresponding to the business logic, that is, the structured data of the business needs to be updated accordingly. This process is also called structured data migration. By performing data derivation on the first structured data using the target entity data and the entity data, the efficiency of structured data migration can be improved.

[0110] In some embodiments, the entity data can be updated to the target entity data according to the actual needs of business development. The target entity data can be obtained by adding, deleting, or modifying the entity data by the staff.

[0111] The computer device obtains the entity data and the first structured data in the first stage, obtains the target entity data in the second stage, obtains the third inquiry template, and embeds the entity data, the first structured data, and the target entity data into the third inquiry template to obtain the third inquiry information. Exemplarily, the third inquiry template can be: "The entity data in the first stage is <1>, the structured data is <2>, and the entity data in the second stage is <3>. Please deduce the structured data in the second stage", where "<1>" in the third inquiry template is used to embed the entity data, "<2>" is used to embed the first structured data, and "<3>" is used to embed the target entity data.

[0112] The computer device generates the second structured data corresponding to the target entity data based on the third inquiry information, which can be by processing the third inquiry information through the first large model to obtain the second structured data. The first large model can be implemented through the GPT model.

[0113] Exemplarily, as Figure 3 shown, the entity data includes a user entity, the entity attributes of the user entity: name and email, the business logic is to query user information, and the first structured data in the first stage includes: the table creation instructions of the user entity, the table structure of the user entity, and the data processing instructions corresponding to querying user information.

[0114] Among them, the table creation instructions of the user entity include:

[0115] sql

[0116] CREATE TABLE Users (

[0117] id INT PRIMARY KEY,

[0118] name VARCHAR(255),

[0119] email VARCHAR(255) )

[0121] Running the table creation instruction for the user entity can obtain the table structure of the user entity.

[0122] The data processing instruction corresponding to querying user information is: SELECT Name,email FROM User.

[0123] In the second stage, the target entity data includes the updated user entity, the entity attributes of the updated user entity: name, email, and balance. That is to say, compared with the entity data, the target entity data has a new entity attribute: balance.

[0124] Determine the third interrogation information based on the entity data, the target entity data, and the first structured data. Perform a response process on the third interrogation information through the first large model to obtain the second structured data corresponding to the target entity data. The second structured data includes: the table creation instruction for the updated user entity, the table structure of the updated user entity, and the data processing instruction corresponding to querying the updated user information.

[0125] Among them, the table creation instruction for the updated user entity includes:

[0126] sql

[0127] CREATE TABLE Users (

[0128] id INT PRIMARY KEY,

[0129] name VARCHAR(255),

[0130] email VARCHAR(255),

[0131] Debt DECIMAL(10, 2) )

[0133] Running the table creation instruction for the updated user entity can obtain the table structure of the updated user entity.

[0134] The data processing instruction corresponding to querying the updated user information is: SELECT Name,email,Debt FROMUser.

[0135] In the above method for processing structured data, first inquiry information is generated based on the knowledge data of the business, entity data of the business is queried according to the first inquiry information, second inquiry information is generated based on the entity data and business logic, the entity data is converted into first structured data according to the second inquiry information, and when the entity data is updated to target entity data, third inquiry information is generated based on the target entity data, the entity data, and the first structured data, and data derivation is performed according to the third inquiry information to obtain second structured data; through an intelligent question-and-answer task, the entity data and the first structured data of the business are gradually determined. When the entity data is updated to target entity data, the second structured data corresponding to the updated target entity data is determined through the intelligent question-and-answer task. That is to say, in the case where the entity data changes due to business development, data derivation can be performed through a dialogue interaction method based on the original entity data, the first structured data, and the updated target entity data to obtain the second structured data corresponding to the target entity data, without the need for manual rewriting of the original first structured data, thus improving the efficiency of updating the structured data of the business when the entity data is updated due to business development.

[0136] In some embodiments, as Figure 4 shown, generating the second structured data corresponding to the target entity data according to the third inquiry information includes: Step 401, determining the difference data between the target entity data and the entity data included in the third inquiry information; Step 402, performing data processing on the first structured data included in the third inquiry information according to the difference data to obtain the second structured data corresponding to the target entity data.

[0137] Among them, the difference data may be an entity or an entity attribute by which the target entity data changes compared with the entity data; exemplarily, the entity data includes: entity A and entity A's entity attributes: A1, A2; entity B and entity B's entity attributes: B1, B2; the target entity data includes: entity A and entity A's entity attributes: A1, A2, A3; entity B and entity B's entity attributes: B1, B2.

[0138] In some embodiments, the first structured data includes: a table creation instruction for an entity, a table structure of the entity, and a data processing instruction corresponding to the business logic; performing data processing on the first structured data according to the difference data may be to determine the target entity corresponding to the difference data in the entity, regenerate the table creation instruction for the target entity, generate the table structure of the target entity according to the regenerated table creation instruction, determine the business logic related to the target entity, update the business logic according to the difference data, and determine the updated data processing instruction according to the table structure of the target entity and the updated business logic.

[0139] In some embodiments, data processing is performed on the first structured data according to the differential data, including determining updated target entities according to the differential data, obtaining the table creation instructions of the target entities in the first structured data, performing data derivation processing on the table creation instructions of the target entities according to the differential data to obtain the table creation instructions of the target entities, generating the table structure of the target entities according to the table creation instructions obtained by data derivation, performing data derivation processing on the business logic according to the differential data to obtain the updated business logic, and determining the updated data processing instructions according to the table structure of the target entities and the updated business logic.

[0140] In the above embodiments, data processing is performed on the first structured data through the differential data to obtain the second structured data corresponding to the target entity data. The second structured data corresponding to the target entity data can be derived based on the original entity data, the first structured data, and the updated target entity data, without the need for manual rewriting of the original first structured data, improving the efficiency of updating the structured data of the business when entity data is updated due to business development.

[0141] In some embodiments, as Figure 5 shown, data processing is performed on the first structured data included in the third inquiry information according to the differential data to obtain the second structured data corresponding to the target entity data, including: Step 501, searching for the table creation instructions of the first target entity corresponding to the differential data in the first structured data included in the third inquiry information; Step 502, performing data processing on the table creation instructions of the first target entity according to the differential data to obtain the updated table creation instructions of the first target entity; Step 503, generating the table structure of the updated first target entity according to the updated table creation instructions of the first target entity; Step 504, searching for the data processing instructions related to the first target entity in the first structured data; Step 505, performing data processing on the found data processing instructions according to the table structure of the updated first target entity to obtain the first target data processing instructions. The second structured data includes the table creation instructions of the updated first target entity, the table structure of the updated first target entity, and the first target data processing instructions.

[0142] Among them, the first target entity is the entity corresponding to the differential data; exemplarily, the entity data includes: entity A and the entity attributes of entity A: A1, A2; entity B and the entity attributes of entity B: B1, B2; the target entity data includes: entity A and the entity attributes of entity A: A1, A2, A3; entity B and the entity attributes of entity B: B1, B2, and the differential data is the entity attribute A3. Based on the differential data, the first target entity can be determined to be entity A.

[0143] The computer device obtains the table creation instruction of the first target entity in the first stage, performs data derivation processing on the table creation instruction of the first target entity in the first stage according to the differential data, obtains the updated table creation instruction of the first target entity in the second stage, and generates the table structure of the updated first target entity according to the updated table creation instruction of the first target entity; obtains the business logic related to the first target entity, determines the data processing instruction corresponding to the business logic in the first structured data, and performs data derivation processing on the data processing instruction according to the table structure of the first target entity to obtain the second structured data.

[0144] Exemplarily, as Figure 6 shown, in the above example, according to the entity attribute A3 of entity A, data derivation processing is performed on the table creation instruction of entity A in the first stage to obtain the table creation instruction of entity A in the second stage, and the table structure of entity A in the second stage is generated based on the table creation instruction of entity A in the second stage; the business logic of entity A in the first stage is to query the information of entity A, obtain the data processing instruction for querying the information of entity A, and perform data derivation on this data processing instruction according to the table structure of entity A in the second stage to obtain the target data processing instruction for querying the information of entity A in the second stage.

[0145] It should be noted that in the case of obtaining the second structured data by performing reply processing on the third inquiry information through the first large model, it may be to obtain, through the first large model, in the first structured data included in the third inquiry information, the table creation instruction of the first target entity corresponding to the differential data; perform data processing on the table creation instruction corresponding to the first target entity according to the differential data to obtain the updated table creation instruction of the first target entity; generate the table structure of the updated first target entity according to the updated table creation instruction of the first target entity; find the data processing instruction related to the first target entity in the first structured data; perform data processing on the found data processing instruction according to the table structure of the updated first target entity to obtain the first target data processing instruction.

[0146] In the above embodiment, by deriving the table creation instruction corresponding to the first target entity through the differential data, the updated table creation instruction of the first target entity is obtained, and then, according to the updated table creation instruction of the first target entity, data derivation is performed on the data processing instruction related to the first target entity to obtain the first target data processing instruction, without the need for manual rewriting of the original first structured data, which improves the efficiency of updating the structured data of the business when entity data is updated due to business development.

[0147] In some embodiments, such as Figure 7As shown in the figure, the processing of structured data further includes: Step 701, when the business logic is updated to the target business logic, determining the fourth query information based on the target business logic, the business logic, and the first structured data; Step 702, generating the second structured data corresponding to the target business logic based on the fourth query information.

[0148] Among them, the target business logic is obtained by rewriting the business logic. Specifically, it can be to rewrite the entities or entity attributes involved in the business logic; Exemplarily: The business logic is: find the balance of entity A in the first currency, and the target business logic is: find the balance of entity A in the first currency and the second currency; Thus, it can be seen that the entity attribute of entity A involved in the business logic includes the first currency, and the entity attributes of entity A involved in the target business logic include the first currency and the second currency; Therefore, the target business logic is obtained by rewriting the entity attributes of entity A in the business logic.

[0149] The fourth query information is used to query the second structured data derived from data inference based on the target business logic, the business logic, and the first structured data.

[0150] After updating the business logic to the target business logic, in order to ensure that the overall data of the business matches the update of the business logic, it is necessary to determine the entities involved in the target business logic and synchronously update the table creation instructions and table structures of the entities.

[0151] The second structured data corresponding to the target business logic includes: the target data processing instructions corresponding to the target business logic, the target table creation instructions and table structures of the entities related to the target business logic.

[0152] In some embodiments, according to the actual needs of business development, the business logic can be updated to the target business logic; the computer device obtains the entity data and the first structured data in the first stage, obtains the target business logic in the second stage, and embeds the entity data and the first structured data in the first stage and the target business logic in the second stage into the fourth query template according to the third query template to obtain the fourth query information. Exemplarily, the fourth query template can be: "The structured data in the first stage is <1>, the business logic is <2>, the target business logic in the second stage is <3>, please derive the structured data in the second stage", where "<1>" in the fourth query template is used for structured data, "<2>" is used to embed the business logic, and "<3>" is used to embed the target business logic.

[0153] In some embodiments, generating the second structured data corresponding to the target business logic based on the fourth query information includes: performing reply processing on the fourth query information through the first large model to obtain the second structured data corresponding to the target business logic.

[0154] Specifically, input the fourth interrogation information into the first large model, and the first large model outputs an answer based on the fourth interrogation information, that is, the second structured data. By using the first large model to process the response to the fourth interrogation information, the second structured data corresponding to the target business logic can be obtained, which can improve the efficiency of determining the second structured data.

[0155] In some embodiments, generating the second structured data corresponding to the target business logic based on the fourth interrogation information includes: determining a second target entity related to the target business logic included in the fourth interrogation information; searching for the table creation instruction of the second target entity in the first structured data, and performing data processing on the table creation instruction of the second target entity according to the difference between the target business logic and the business logic to obtain the updated table creation instruction of the second target entity; generating the table structure of the updated second target entity according to the updated table creation instruction of the second target entity; determining the second target data processing instruction corresponding to the target business logic according to the table structure of the updated second target entity; the second structured data includes the updated table creation instruction of the second target entity, the table structure of the updated second target entity, and the second target data processing instruction.

[0156] Among them, the second target entity is the entity involved in the target business logic; for example, the target business logic includes: searching for the balance of entity A in the first currency and the second currency, then the second target entity is entity A.

[0157] The difference between the target business logic and the business logic may be the difference in the entity attributes of the second target entity. For example, the business logic is: searching for the balance of entity A in the first currency, and the target business logic is: searching for the balance of entity A in the first currency and the second currency; thus, it can be seen that the difference between the target business logic and the business logic is the entity attribute of entity A: the second currency.

[0158] For the convenience of description, the second target entity in the first stage is denoted as the second target entity in the first stage, and the updated second target entity is denoted as the second target entity in the second stage; the table creation instruction of the second target entity in the first stage may be different from the table creation instruction of the second target entity in the second stage, and further, the table structure of the second target entity in the first stage may be different from the table structure of the second target entity in the second stage.

[0159] In one implementation, the computer device determines the second target entity in the target business logic, searches for the table creation instruction of the second target entity in the second structured data, and performs data derivation on the table creation instruction of the second target entity based on the difference between the target business logic and the business logic to obtain the updated table creation instruction of the second target entity, and obtains the updated table creation instruction of the second target entity.

[0160] In another implementation, the computer device determines a second target entity in the target business logic, and based on the differences between the target business logic and the business logic, performs data processing on the second target entity to obtain an updated second target entity, and performs a structured transformation on the updated second target entity to obtain a table creation instruction for the updated second target entity.

[0161] The computer device runs the table creation instruction of the updated second target entity to obtain the table structure of the updated second target entity; and based on the table structure of the updated second target entity, performs a structured transformation on the target business logic to obtain a second target data processing instruction.

[0162] Exemplarily, the business logic of the first stage is: to find the balance of entity A in the first currency, and the target business logic of the second stage is: to find the balance of entity A in the first currency and the second currency; find the table creation instruction of entity A in the first structured data, and based on the differences between the business logic and the target business logic, perform data derivation on the table creation instruction of entity A to obtain the table creation instruction of the updated entity A; it can be understood that the updated entity A includes entity attributes including: the first currency and the second currency.

[0163] Table creation instruction of the updated entity A:

[0164] sql

[0165] CREATE TABLE Users (

[0166] id INT PRIMARY KEY,

[0167] name VARCHAR(255),

[0168] currency1_balance DECIMAL(10, 2),

[0169] currency2_balance DECIMAL(10, 2) );

[0171] Running the table creation instruction of the updated entity A can obtain the table structure of the updated entity A.

[0172] The target business logic is: to query the total balance of the user, then the target data processing instruction includes:

[0173] SELECT

[0174] id,

[0175] name,

[0176] (currency1_balance + currency2_balance) AS total_balance

[0177] FROM

[0178] Users;

[0179] It should be noted that when the first large model processes and replies to the fourth query information to obtain the second structured data corresponding to the target business logic, it can be to determine the second target entity related to the target business logic included in the fourth query information through the first large model; search for the table creation instruction of the second target entity in the first structured data; perform data processing on the table creation instruction of the second target entity according to the difference between the target business logic and the business logic to obtain the updated table creation instruction of the second target entity; generate the table structure of the updated second target entity according to the updated table creation instruction of the second target entity; determine the second target data processing instruction corresponding to the target business logic according to the table structure of the updated second target entity.

[0180] In the above embodiment, when the business logic is updated due to business development, the first structured data can be deduced based on the original business logic and the updated target business logic to obtain the second structured data corresponding to the target business logic, without the need for manual rewriting of the original structured data for the target business logic, thus improving the efficiency of updating the structured data of the business when the business logic is updated due to business development.

[0181] In some embodiments, querying the entity data of the business according to the first query information includes: processing the entity data of the business by the second large model for the first query information to obtain the entity data of the business; processing and replying to the second query information to obtain the first structured data, including: processing and replying to the second query information by the third large model to obtain the first structured data; generating the second structured data corresponding to the target entity data according to the third query information, including: processing and replying to the third query information by the first large model to obtain the second structured data.

[0182] Specifically, such as Figure 8As shown, the computer device generates first inquiry information based on the knowledge data of the service, inputs the first inquiry information into the second large model, and through the second large model, processes the reply to the first inquiry information to obtain the entity data of the service; the computer device generates second inquiry information based on the entity data and service logic of the service, inputs the second inquiry information into the third large model, and through the third large model, processes the reply to the second inquiry information to obtain the first structured data; when the entity data is updated to the target entity data, the computer device determines the third inquiry information based on the target entity data, the entity data, and the first structured data, inputs the third inquiry information into the first large model, and through the first large model, processes the reply to the third inquiry information to obtain the second structured data.

[0183] In the above embodiment, through the second large model, the third large model, and the first large model, through multiple intelligent question-and-answer tasks, the entity data and the first structured data of the service are gradually determined. When the entity data of the service is updated, the updated structured data caused by the update of the entity data is deduced through the intelligent question-and-answer task, improving the efficiency of updating the structured data of the service.

[0184] In some embodiments, as Figure 9 shown, the knowledge data of the service is the knowledge data of the payment service; the service logic is the resource query logic; generating the first inquiry information based on the knowledge data of the service and querying the entity data of the service according to the first inquiry information includes: Step 901, generating the first inquiry information based on the knowledge data of the payment service and processing the first inquiry information through the second large model to obtain the entity data of the payment service; generating the second inquiry information based on the entity data and the service logic and processing the reply to the second inquiry information to obtain the first structured data, including: Step 902, generating the second inquiry information based on the entity data of the payment service and the resource query logic and processing the reply to the second inquiry information through the third large model to obtain the table creation instruction of the entity data, the table structure of the entity data, and the resource query instruction corresponding to the resource query logic; generating the second structured data corresponding to the target entity data based on the third inquiry information, including: Step 903, when the target entity data in the first data derivation inquiry information has a new regional attribute of the account entity compared with the entity data, processing the reply to the third inquiry information through the data derivation large model to obtain the table creation instruction of the account entity after adding the new regional attribute, the table structure of the account entity after adding the new regional attribute, and the resource query instruction after adding the new regional attribute.

[0185] Specifically, when the service is the payment service, the computer device generates the first inquiry information according to the knowledge data of the payment service, inputs the first inquiry information into the second large model, and the second large model executes the question-and-answer task for the first inquiry information, and the output answer is the entity data of the payment service; this entity data is the entity data of the payment service in the first stage.

[0186] Exemplarily, the entity data of the payment service in the first stage includes: user entity, entity attributes of the user entity, account entity, entity attributes of the account entity, and entity relationship between the user entity and the account entity. Among them, the entity attributes of the user entity include: name and email; the entity attributes of the account entity include: name and balance.

[0187] The computer device generates a second query message based on the entity data and business logic of the payment service, inputs the second query message into the third large model, and the third large model performs a question-and-answer task for the second query message, and the output answer is the first structured data of the payment service in the first stage; the first structured data includes: table creation instructions and table structures of the user entity, table creation instructions and table structures of the account entity, and data processing instructions corresponding to the business logic.

[0188] Exemplarily, the table creation instruction of the user entity in the first stage is:

[0189] sql

[0190] CREATE TABLE Users (

[0191] id INT PRIMARY KEY,

[0192] name VARCHAR(255),

[0193] email VARCHAR(255) )

[0195] The table creation instruction of the account entity in the first stage is:

[0196] sql

[0197] CREATE TABLE Accounts (

[0198] id INT PRIMARY KEY,

[0199] user_id INT,

[0200] balance DECIMAL(10, 2),

[0201] FOREIGN KEY (user_id) REFERENCES Users(id) );

[0203] Among them, FOREIGN KEY (user_id) REFERENCES Users(id) indicates that user_id is a foreign key that references the id field of TABLE Users. This statement can represent the entity relationship between the user entity and the account entity.

[0204] Running the table creation instruction for the user entity obtains the table structure of the user entity, and running the table creation instruction for the account entity obtains the table structure of the account entity.

[0205] The resource query logic in the first stage can be to query the negative resources of a certain user, which can be understood as querying the debts of a certain user; the resource query instruction generated by the third large model for the business logic is:

[0206] sql

[0207] SELECT SUM(balance) AS total_debt

[0208] FROM Accounts

[0209] WHERE user_id = {user_id}

[0210] AND balance < 0;

[0211] When the entity data in the first stage is updated to the target entity data in the second stage, the target entity data includes: the entity attributes of the user entity include: name and email, and the entity attributes of the account entity: name, balance, and region; the computer device generates the third query information based on the target entity data, the entity data, and the first structured data, and inputs the third query information into the first large model. The first large model performs a question-and-answer task for the third query information, and the output answer is the second structured data; the second structured data includes: the table creation instruction and table structure of the user entity in the second stage, the table creation instruction and table structure of the account entity in the second stage, and the resource query instruction in the second stage.

[0212] Exemplarily, the table creation instruction and table structure of the user entity in the second stage are the same as those of the user entity in the first stage.

[0213] The table creation instruction for the account entity in the second stage (the table creation instruction for the account entity after adding the region attribute) is:

[0214] CREATE TABLE Accounts (

[0215] id INT PRIMARY KEY,

[0216] user_id INT,

[0217] balance DECIMAL(10, 2),

[0218] region VARCHAR(50),

[0219] FOREIGN KEY (user_id) REFERENCES Users(id) );

[0221] Executing the table creation instruction for the account entity in the second phase can obtain the table structure of the account entity in the second phase (the table structure of the account entity after the new region attribute is described).

[0222] Based on the table structures of the user entity and the account entity in the second phase, data derivation is performed on the resource query instruction in the first phase to obtain the resource query instruction in the second phase (the resource query instruction after the new region attribute is added):

[0223] SELECT SUM(balance) AS total_debt

[0224] FROM Accounts

[0225] WHERE user_id ={user_id}

[0226] AND region ={region id}

[0227] AND balance < 0;

[0228] In the above embodiments, the large model can be used to execute the Q&A task to obtain the entity data involved in the payment service and the first structured data corresponding to the entity data and business logic. When new region attributes are added to the account entity due to the development of the payment service, the first structured data can be deduced and processed based on the changes in the entity data through the dialogue interaction method to obtain the table creation instruction for the account entity after the new region attribute is added, the table structure of the account entity after the new region attribute is added, and the resource query instruction after the new region attribute is added, without the need for manual rewriting of the original first structured data, thus improving the efficiency of updating the structured data of the business when entity data is updated during business development.

[0229] In some embodiments, such as Figure 10As shown, before generating the first query information based on the knowledge data of the business, it further includes: Step 1001, obtaining the sample knowledge data of the business and generating the fifth query information based on the sample knowledge data; Step 1002, performing a response process on the fifth query information through the second large model to obtain training entity data; Step 1003, adjusting the parameters of the second large model according to the training entity data and the entity data labels corresponding to the sample knowledge data until the second large model converges to obtain a converged second large model.

[0230] Among them, the sample knowledge data is the knowledge content related to the business; the fifth query information is used to query the training entity data included in the sample knowledge data; the entity data label is the real entity data in the sample knowledge data.

[0231] Specifically, the computer device obtains the sample knowledge data of the business, embeds the sample knowledge data into the first query template to obtain the fifth query information, and inputs the fifth query information into the second large model. The second large model performs a question-and-answer task for the fifth query information, and the output answer is the training entity data. The computer device can determine the first cross-entropy loss function according to the training entity data and the entity data labels corresponding to the sample knowledge data, adjust the parameters of the second large model according to the first cross-entropy loss function, and adjust the learning rate according to the optimization algorithm (such as the ADAM optimization algorithm) to make the parameter update more stable until the second large model converges to obtain a converged second large model.

[0232] Determining the first cross-entropy loss function according to the training entity data and the entity data labels corresponding to the sample knowledge data may be to extract the feature representation of the training entity data and the feature representation of the entity data label, and determine the second cross-entropy loss function based on the feature representation of the training entity data and the feature representation of the entity data label.

[0233] In some embodiments, when the computer device obtains the sample knowledge data, it can preprocess the sample knowledge data to obtain the preprocessed sample knowledge data, and input the preprocessed sample knowledge data into the second large model, so that the second large model learns the content of entities, entity attributes, and entity relationships related to the business. Furthermore, the second large model can answer questions about extracting entity data for specific businesses; among them, preprocessing the sample knowledge data can be to perform entity relationship modeling on the sample knowledge data to obtain an entity relationship graph, and input the entity relationship graph into the second large model, so that the second large model learns the content of entities, entity attributes, and entity relationships related to the business based on the entity relationship graph.

[0234] In the above embodiments, according to the sample knowledge data and the entity data labels corresponding to the sample knowledge data, the parameters of the second largest model are adjusted so that the converged second largest model can process questions including entity data in the query knowledge data, improving the efficiency of determining the entity data of the service.

[0235] In some embodiments, as Figure 11 shown, before generating the entity data query information based on the knowledge data of the service, it further includes: Step 1101, generating a sixth query information according to the sample entity data of the service and the sample service logic; Step 1102, performing a reply process on the sixth query information through the third largest model to obtain training structured data; Step 1103, adjusting the parameters of the third largest model according to the structured data labels corresponding to the sample entity data and the sample service logic and the training structured data until the third largest model converges to obtain a converged third largest model.

[0236] Among them, the sample service logic is used to represent the operations when the service is executed, which may involve the entities and entity attributes included in the sample entity data; the sixth query information is used to query the structured data corresponding to the training entity data and the structured data corresponding to the sample service logic; the training structured data includes the training structured data corresponding to the sample entity data and the training structured data corresponding to the sample service logic, specifically including the training table creation instructions and training table structures of the sample entities in the sample entity data, and the training data processing instructions corresponding to the sample service logic; the structured data labels include the structured data labels corresponding to the sample entity data and the data processing instruction labels corresponding to the sample service logic.

[0237] Specifically, the computer device obtains the sample entity data and the corresponding sample service logic, embeds the sample entity data and the sample service logic into the second query template to obtain the sixth query information, inputs the sixth query information into the third largest model, and the third largest model performs a question-and-answer task for the sixth query information, and the output answer is the training structured data; the computer device can determine the second cross-entropy loss function according to the training structured data and the structured data labels, adjust the parameters of the third largest model according to the second cross-entropy loss function, and adjust the learning rate according to the optimization algorithm to make the parameter update more stable until the third largest model converges to obtain a converged third largest model.

[0238] Determining the second cross-entropy loss function according to the training structured data and the structured data labels may be to extract the feature representation of the training structured data and the feature representation of the structured data labels, and determine the second cross-entropy loss function based on the feature representation of the training structured data and the feature representation of the structured data labels.

[0239] In some embodiments, the computer device obtains sample entity data and corresponding structured data labels, and sample business logic and corresponding structured data labels, and inputs the sample entity data and structured data labels, and the sample business logic and corresponding structured data labels into the second large model, so that the second large model can learn the ability to represent entity information using structured language, and the ability to represent business logic using structured language. Furthermore, the second large model can answer questions related to business that convert entity data and business logic into structured data.

[0240] In the above embodiments, according to the structured data labels corresponding to the sample entity data and sample business logic of the business, the parameters of the third large model are adjusted, so that the converged third large model can process questions about structured data that inquire about entity data and business logic, improving the efficiency of converting entity data and business logic into structured data.

[0241] In some embodiments, before generating entity data query information based on the knowledge data of the business, it further includes: generating seventh query information based on the sample target entity data, sample entity data, and training structured data of the business; performing a reply process on the seventh query information through the first large model to obtain first training target structured data, and adjusting the parameters of the first large model according to the first target structured data label corresponding to the sample target entity data and the first training target structured data until the first large model converges to obtain a converged first large model.

[0242] Among them, the sample target entity data is obtained by updating the sample entity data. In practical applications, the sample target entity data can be obtained by modifying a certain entity in the sample entity data, or by modifying the entity attributes of a certain entity in the sample entity data.

[0243] The seventh query information is used to query the first training target structured data derived from the training structured data based on the sample target entity data and sample entity data.

[0244] In some embodiments, the computer device embeds the sample target entity data, sample entity data, and training structured data into the third query template to obtain seventh query information, inputs the seventh query information into the first large model, and performs a question and answer task on the seventh query information through the first large model to obtain first training target structured data; the computer device determines the third cross-entropy loss function according to the first training target structured data and the first target structured data label, adjusts the parameters of the first large model according to the third cross-entropy loss function, and adjusts the learning rate according to the optimization algorithm to make the parameter update more stable until the first large model converges.

[0245] In the process of performing a question-and-answer task on the seventh query information through the first large model, the first large model can learn the differences between the sample target entity data and the sample entity data, and deduce the training data processing instructions corresponding to the sample business logic in the training structured data based on the differences, to obtain the first training target structured data; adjust the parameters of the first large model according to the third cross-entropy loss function, so that the first training target structured data output by the first large model is close to the first target structured data label, improving the transfer ability of the first large model for structured data, and enabling the converged first large model to deduce the second structured data corresponding to the target entity data from the first structured data based on the differences between the entity data and the updated target entity data.

[0246] In some embodiments, the computer device obtains the sample entity data, sample business logic, and training structured data of the service in the first stage, and obtains the target sample entity data, sample target business logic, and the first training target structured data of the service in the second stage; wherein, the target sample entity data is obtained by updating the sample entity data, the target sample business logic is updated based on the sample entity data and corresponds to the updated sample business logic, and the first training target structured data is determined based on the target sample entity data and the target business logic.

[0247] Input the sample entity data, sample business logic, and training structured data of the service in the first stage, the target sample entity data, sample target business logic, and the first training target structured data of the service in the second stage into the first large model, so that the first large model can learn the entity data differences between the sample entity data in the first stage and the sample entity data in the second stage, and learn the business logic differences between the target sample business logic and the sample business logic, and can learn the relationship between the entity data differences and the business logic differences, so as to have the ability to deduce the business logic in the second stage from the business logic in the first stage according to the entity data changes in the first stage and the second stage.

[0248] In the above embodiments, adjust the parameters of the first large model according to the sample entity data, sample business logic, and training structured data in the first stage, and the target sample entity data and the corresponding training target structured data label in the second stage, so that the converged first large model can deduce the second structured data corresponding to the target entity data from the first structured data based on the differences between the entity data and the updated target entity data, improving the efficiency of updating the structured data of the service when the entity data is updated during the business development.

[0249] In some embodiments, before generating entity data query information based on the knowledge data of the service, it further includes: generating eighth query information based on the sample entity data of the service, the sample service logic, the training structured data, and the target sample service logic; processing the eighth query information through the first large model to obtain second training target structured data, and adjusting the parameters of the first large model according to the second target structured data label corresponding to the target sample service logic and the second training target structured data until the first large model converges to obtain a converged first large model.

[0250] The first large model can learn the difference between the sample service logic and the target sample service logic, and deduce the sample entity data according to this difference to obtain second training target structured data. Adjust the parameters of the first large model according to the second target structured data label corresponding to the target sample service logic and the second training target structured data, so that the converged first large model can deduce the structured data corresponding to the entity data based on the update of the service logic, improving the migration ability of the first large model to structured data. When the service logic is updated due to business development, the efficiency of updating the structured data of the entity data is improved.

[0251] In some embodiments, as Figure 12 shown, the processing method of structured data includes:

[0252] Step 1201, obtain the sample knowledge data of the service, and generate fifth query information based on the sample knowledge data; process the fifth query information through the second large model to obtain training entity data; adjust the parameters of the second large model according to the training entity data and the entity data label corresponding to the sample knowledge data until the second large model converges to obtain a converged second large model;

[0253] Step 1202, generate sixth query information based on the sample entity data and the sample service logic of the service; process the sixth query information through the third large model to obtain training structured data; adjust the parameters of the third large model according to the structured data label corresponding to the sample entity data and the sample service logic and the training structured data until the third large model converges to obtain a converged third large model;

[0254] Step 1203, process the first query information through the second large model to obtain the entity data of the service;

[0255] Step 1204, determine the service logic based on the entity data;

[0256] Step 1205, generate second query information based on the entity data and the service logic;

[0257] Step 1206, perform reply processing on the second inquiry information through the third large model to obtain the first structured data;

[0258] Step 1207A, when the entity data is updated to the target entity data, determine the third inquiry information for data derivation based on the target entity data, entity data, and the first structured data;

[0259] Step 1208A, perform reply processing on the third inquiry information through the first large model to obtain the second structured data;

[0260] Step 1207B, when the business logic is updated to the target business logic, determine the fourth inquiry information based on the target business logic, business logic, entity data, and the first structured data;

[0261] Step 1208B, perform reply processing on the fourth inquiry information through the first large model to obtain the second structured data corresponding to the target business logic.

[0262] In some embodiments, as Figure 13 shown, the processing method of structured data includes:

[0263] Step 1301, generate the first inquiry information based on the knowledge data of the business, and query the entity data of the business according to the first inquiry information;

[0264] Step 1302, determine the business logic based on the entity data;

[0265] Step 1303, generate the second inquiry information based on the entity data and the business logic;

[0266] Step 1304, obtain the entities included in the second inquiry information, the entity attributes of the entities, and the entity relationships; determine the table creation instructions for the entities according to the entities, entity attributes, and other entities corresponding to the entity relationships; generate the table structure of the entities based on the table creation instructions of the entities; determine the data processing instructions corresponding to the business logic based on the table structure of the entities; the first structured data includes the table creation instructions of the entities, the table structure of the entities, and the data processing instructions;

[0267] Step 1305A, when the entity data is updated to the target entity data, determine the third inquiry information for data derivation based on the target entity data, entity data, and the first structured data;

[0268] Step 1306A, determine the difference data between the target entity data and the entity data included in the third inquiry information; in the first structured data included in the third inquiry information, search for the table creation instructions of the first target entity corresponding to the difference data; perform data processing on the table creation instructions corresponding to the first target entity according to the difference data to obtain the updated table creation instructions of the first target entity;

[0269] Step 1307A, generate the table structure of the updated first target entity according to the table creation instruction of the updated first target entity; search for data processing instructions related to the first target entity in the first structured data; perform data processing on the found data processing instructions according to the table structure of the updated first target entity to obtain the first target data processing instruction; the second structured data includes the table creation instruction of the updated first target entity, the table structure of the updated first target entity, and the first target data processing instruction;

[0270] Step 1305B, when the business logic is updated to the target business logic, determine the fourth query information according to the target business logic, the business logic, the entity data, and the first structured data;

[0271] Step 1306B, determine the second target entity related to the target business logic included in the fourth query information; search for the table creation instruction of the second target entity in the first structured data; perform data processing on the table creation instruction of the second target entity according to the difference between the target business logic and the business logic to obtain the updated table creation instruction of the second target entity; generate the table structure of the updated second target entity according to the updated table creation instruction of the second target entity; determine the second target data processing instruction corresponding to the target business logic according to the table structure of the updated second target entity; the second structured data includes the updated table creation instruction of the second target entity, the table structure of the updated second target entity, and the second target data processing instruction.

[0272] In the above method for processing structured data, the first query information is generated through the knowledge data of the business, the entity data of the business is queried according to the first query information, the second query information is generated according to the entity data and the business logic, the entity data is converted into the first structured data according to the second query information, when the entity data is updated to the target entity data, the third query information is generated according to the target entity data, the entity data, and the first structured data, and the second structured data is obtained through data derivation according to the third query information; through the intelligent question-and-answer task, the entity data and the first structured data of the business are gradually determined, when the entity data is updated to the target entity data, the second structured data corresponding to the updated target entity data is determined through the intelligent question-and-answer task, that is to say, in the case of entity data update caused by business development, the second structured data corresponding to the target entity data can be obtained through data derivation in a dialogue interaction manner based on the original entity data, the first structured data, and the updated target entity data, without manual rewriting of the original first structured data, improving the efficiency of updating the structured data of the business when entity data is updated due to business development.

[0273] Based on the same inventive concept, an embodiment of the present application further provides a structured data processing apparatus for implementing the above-mentioned structured data processing method. The solution provided by this apparatus for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more of the following structured data processing apparatus embodiments can refer to the limitations on the structured data processing method in the foregoing, and will not be elaborated here.

[0274] In some embodiments, as Figure 14 shown, a structured data processing apparatus is provided, including: an entity data determination module 1401, a service logic determination module 1402, a first structured data determination module 1403, and a second structured data determination module 1404; where

[0275] The entity data determination module 1401 is configured to generate first query information based on the knowledge data of the service, and query the entity data of the service according to the first query information;

[0276] The service logic determination module 1402 is configured to determine the service logic based on the entity data;

[0277] The first structured data determination module 1403 is configured to generate second query information based on the entity data and the service logic, and perform a reply process on the second query information to obtain first structured data;

[0278] The second structured data determination module 1404 is configured to, when the entity data is updated to target entity data, determine third query information for data derivation based on the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third query information.

[0279] In some embodiments, the first structured data determination module 1403 is further configured to obtain the entities included in the second query information, the entity attributes of the entities, and the entity relationships; determine the table creation instructions of the entities according to the entities, the entity attributes, and other entities corresponding to the entity relationships; generate the table structures of the entities according to the table creation instructions of the entities; determine the data processing instructions corresponding to the service logic based on the table structures of the entities; the first structured data includes the table creation instructions of the entities, the table structures of the entities, and the data processing instructions.

[0280] In some embodiments, the second structured data determination module 1404 is further configured to determine the difference data between the target entity data and the entity data included in the third query information; perform data processing on the first structured data included in the third query information according to the difference data to obtain second structured data corresponding to the target entity data.

[0281] In some embodiments, the second structured data determination module 1404 is further configured to search for a table creation instruction of a first target entity corresponding to the difference data in the first structured data included in the third query information; perform data processing on the table creation instruction corresponding to the first target entity according to the difference data to obtain an updated table creation instruction of the first target entity; generate a table structure of the updated first target entity according to the updated table creation instruction of the first target entity; search for a data processing instruction related to the first target entity in the first structured data; perform data processing on the found data processing instruction according to the table structure of the updated first target entity to obtain a first target data processing instruction; the second structured data includes the updated table creation instruction of the first target entity, the table structure of the updated first target entity, and the first target data processing instruction.

[0282] In some embodiments, the processing apparatus for structured data further includes: a third structured data determination module, configured to determine fourth query information according to the target business logic, the business logic, the entity data, and the first structured data when the business logic is updated to the target business logic; generate second structured data corresponding to the target business logic according to the fourth query information.

[0283] In some embodiments, the third structured data determination module is further configured to determine a second target entity related to the target business logic included in the fourth query information; search for a table creation instruction of the second target entity in the first structured data; perform data processing on the table creation instruction of the second target entity according to the difference between the target business logic and the business logic to obtain an updated table creation instruction of the second target entity; generate a table structure of the updated second target entity according to the updated table creation instruction of the second target entity; determine a second target data processing instruction corresponding to the target business logic according to the table structure of the updated second target entity; the second structured data includes the updated table creation instruction of the second target entity, the table structure of the updated second target entity, and the second target data processing instruction.

[0284] In some embodiments, the third structured data determination module is further configured to perform reply processing on the fourth query information through a first large model to obtain second structured data corresponding to the target business logic.

[0285] In some embodiments, the entity data determination module 1401 is further configured to perform entity data processing on the first query information through a second large model to obtain entity data of the business; the first structured data determination module 1403 is further configured to perform reply processing on the second query information through a third large model to obtain first structured data; the second structured data determination module 1404 is further configured to perform reply processing on the third query information through the first large model to obtain second structured data.

[0286] In some embodiments, the knowledge data of the service is the knowledge data of the payment service; the service logic is the resource query logic; the entity data determination module 1401 is further configured to generate a first query message based on the knowledge data of the payment service, and process the first query message through a second large model to obtain the entity data of the payment service; the first structured data determination module 1403 is further configured to generate a second query message based on the entity data of the payment service and the resource query logic, and perform a reply process on the second query message through a third large model to obtain a table creation instruction for the entity data, a table structure of the entity data, and a resource query instruction corresponding to the resource query logic; the second structured data determination module 1404 is further configured to, when the target entity data in the first data derivation query message has a new regional attribute of the account entity compared with the entity data, perform a reply process on the third query message through a data derivation large model to obtain a table creation instruction for the account entity after adding the new regional attribute, a table structure of the account entity after adding the new regional attribute, and a resource query instruction after adding the new regional attribute.

[0287] In some embodiments, the processing device for structured data further includes: a first training module, configured to obtain sample knowledge data of a service, generate a fifth query message based on the sample knowledge data; perform a reply process on the fifth query message through a second large model to obtain training entity data; and adjust parameters of the second large model according to the training entity data and entity data labels corresponding to the sample knowledge data until the second large model converges to obtain a converged second large model.

[0288] In some embodiments, the processing device for structured data further includes: a second training module, configured to generate a sixth query message based on sample entity data and sample service logic of a service; perform a reply process on the sixth query message through a third large model to obtain training structured data; and adjust parameters of the third large model according to structured data labels corresponding to the sample entity data and the sample service logic and the training structured data until the third large model converges to obtain a converged third large model.

[0289] Each module in the above processing device for structured data can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in a processor in a computer device in a hardware form or be independent of it, or be stored in a memory in a computer device in a software form, so that the processor can call and execute operations corresponding to the above respective modules.

[0290] In some embodiments, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 15As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data related to the processing method of structured data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for processing structured data.

[0291] Those skilled in the art can understand that Figure 15 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0292] In some embodiments, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the following steps are implemented:

[0293] Generate first query information based on the knowledge data of the service, and query the entity data of the service according to the first query information; determine the business logic based on the entity data; generate second query information based on the entity data and the business logic, and perform a reply process on the second query information to obtain first structured data; when the entity data is updated to target entity data, determine third query information for data derivation based on the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third query information.

[0294] In some embodiments, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0295] Generate first inquiry information based on the knowledge data of the service, and query the entity data of the service according to the first inquiry information; determine the business logic based on the entity data; generate second inquiry information based on the entity data and the business logic, and perform a reply process on the second inquiry information to obtain first structured data; when the entity data is updated to target entity data, determine third inquiry information for data derivation based on the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third inquiry information.

[0296] In some embodiments, a computer program product is provided, including a computer program which, when executed by a processor, implements the following steps:

[0297] Generate first inquiry information based on the knowledge data of the service, and query the entity data of the service according to the first inquiry information; determine the business logic based on the entity data; generate second inquiry information based on the entity data and the business logic, and perform a reply process on the second inquiry information to obtain first structured data; when the entity data is updated to target entity data, determine third inquiry information for data derivation based on the target entity data, the entity data, and the first structured data, and generate second structured data corresponding to the target entity data according to the third inquiry information.

[0298] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data that have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0299] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0300] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0301] The above-described embodiments only represent several implementation manners of this application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of this application. It should be noted that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can still be made, and these all belong to the protection scope of this application. Therefore, the protection scope of this application should be subject to the appended claims.

Claims

1. A method for processing structured data, characterized in that, The method includes: generating first inquiry information based on the knowledge data of the service, and querying the entity data of the service according to the first inquiry information; determining the business logic based on the entity data; generating second inquiry information based on the entity data and the business logic, and performing a reply process on the second inquiry information to obtain first structured data; when the entity data is updated to target entity data, determining third inquiry information for data derivation based on the target entity data, the entity data, and the first structured data, and generating second structured data corresponding to the target entity data according to the third inquiry information.

2. The method according to claim 1, wherein The performing a reply process on the second inquiry information to obtain first structured data includes: obtaining the entities included in the second inquiry information, the entity attributes of the entities, and the entity relationships; determining the table creation instructions of the entity according to the entities, the entity attributes, and other entities corresponding to the entity relationships; generating the table structure of the entity according to the table creation instructions of the entity; determining the data processing instructions corresponding to the business logic based on the table structure of the entity; the first structured data includes the table creation instructions of the entity, the table structure of the entity, and the data processing instructions.

3. The method according to claim 1, wherein The generating the second structured data corresponding to the target entity data according to the third inquiry information includes: determining the difference data between the target entity data and the entity data included in the third inquiry information; performing data processing on the first structured data included in the third inquiry information according to the difference data to obtain second structured data corresponding to the target entity data.

4. The method according to claim 3, characterized in that The performing data processing on the first structured data included in the third inquiry information according to the difference data to obtain second structured data corresponding to the target entity data includes: searching for the table creation instructions of the first target entity corresponding to the difference data in the first structured data included in the third inquiry information; performing data processing on the table creation instructions of the first target entity according to the difference data to obtain the updated table creation instructions of the first target entity; generating the table structure of the updated first target entity according to the updated table creation instructions of the first target entity; searching for the data processing instructions related to the first target entity in the first structured data; performing data processing on the found data processing instructions according to the table structure of the updated first target entity to obtain the first target data processing instructions; the second structured data includes the updated table creation instructions of the first target entity, the table structure of the updated first target entity, and the first target data processing instructions.

5. The method according to claim 1, characterized in that The method further includes: when the business logic is updated to target business logic, determining fourth inquiry information based on the target business logic, the business logic, and the first structured data; generating second structured data corresponding to the target business logic according to the fourth inquiry information.

6. The method according to claim 5, characterized in that The generating the second structured data corresponding to the target business logic according to the fourth inquiry information includes: Determine a second target entity related to the target business logic included in the fourth inquiry information; Search for the table creation instruction of the second target entity in the first structured data; Perform data processing on the table creation instruction of the second target entity according to the difference between the target business logic and the business logic to obtain the updated table creation instruction of the second target entity; Generate the table structure of the updated second target entity according to the updated table creation instruction of the second target entity; Determine the second target data processing instruction corresponding to the target business logic according to the table structure of the updated second target entity; the second structured data includes the table creation instruction of the updated second target entity, the table structure of the updated second target entity, and the second target data processing instruction.

7. The method according to claim 5, characterized in that, The generating the second structured data corresponding to the target business logic according to the fourth inquiry information includes: Performing a reply process on the fourth inquiry information through a first large model to obtain the second structured data corresponding to the target business logic.

8. The method according to claim 1, wherein The querying the entity data of the business according to the first inquiry information includes: Performing entity data processing on the first inquiry information through a second large model to obtain the entity data of the business; The performing a reply process on the second inquiry information to obtain the first structured data includes: Performing a reply process on the second inquiry information through a third large model to obtain the first structured data; The generating the second structured data corresponding to the target entity data according to the third inquiry information includes: Performing a reply process on the third inquiry information through a first large model to obtain the second structured data.

9. The method according to claim 8, wherein The knowledge data of the business is the knowledge data of the payment business; the business logic is the resource query logic; The generating the first inquiry information according to the knowledge data of the business and querying the entity data of the business according to the first inquiry information includes: Generating the first inquiry information according to the knowledge data of the payment business and performing processing on the first inquiry information through the second large model to obtain the entity data of the payment business; The generating the second inquiry information according to the entity data and the business logic and performing a reply process on the second inquiry information to obtain the first structured data includes: Generating the second inquiry information according to the entity data of the payment business and the resource query logic and performing a reply process on the second inquiry information through the third large model to obtain the table creation instruction of the entity data, the table structure of the entity data, and the resource query instruction corresponding to the resource query logic; The generating the second structured data corresponding to the target entity data according to the third inquiry information includes: When the target entity data in the first data derivation inquiry information has a new regional attribute of the account entity compared with the entity data, performing a reply process on the third inquiry information through a data derivation large model to obtain the table creation instruction of the account entity after adding the regional attribute, the table structure of the account entity after adding the regional attribute, and the resource query instruction after adding the regional attribute.

10. The method according to claim 8, characterized in that Before generating the first query information based on the business-related knowledge data, the following steps are also included: Obtain the sample knowledge data of the business, and generate the fifth query information based on the sample knowledge data; Perform a response process on the fifth query information through the second large model to obtain training entity data; Adjust the parameters of the second large model according to the training entity data and the entity data labels corresponding to the sample knowledge data until the second large model converges, and obtain the converged second large model.

11. The method according to claim 8, wherein Before generating the entity data query information based on the business-related knowledge data, the following steps are also included: Generate the sixth query information based on the sample entity data and sample business logic of the business; Perform a response process on the sixth query information through the third large model to obtain training structured data; Adjust the parameters of the third large model according to the structured data labels corresponding to the sample entity data and the sample business logic and the training structured data until the third large model converges, and obtain the converged third large model.

12. A processing device for structured data, characterized in that The device includes: An entity data determination module, configured to generate the first query information based on the business-related knowledge data, and query the entity data of the business according to the first query information; A business logic determination module, configured to determine the business logic based on the entity data; A first structured data determination module, configured to generate the second query information based on the entity data and the business logic, and perform a response process on the second query information to obtain the first structured data; A second structured data determination module, configured to, when the entity data is updated to the target entity data, determine the third query information for data derivation according to the target entity data, the entity data, and the first structured data, and generate the second structured data corresponding to the target entity data according to the third query information.

13. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 11.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 11.

15. A computer program product comprising a computer program, characterized in that, When this computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 11.

Images