Digital model product management method and device and computer program product

Abstract

The invention discloses a digital model product management method and device and a computer program product, and the method comprises the steps: firstly, obtaining product group information according to a received digital model product creation instruction, and then constructing an initial digital model product by using the information; secondly, performing version definition on the initial digital model product to obtain a digital model product of a target version; and finally, in response to a digital model product issuing instruction, issuing the digital model product of the target version to a target model calling platform. Through the design, the problem of version management is solved, and the publishing process is more flexible and controllable. In addition, the method can better meet data model operation requirements in various scenes, and the online application efficiency of data model products is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital model management, and in particular, to a method, device and computer program product for managing digital model products. Background Art

[0002] Currently, the management and operation and maintenance of data model products are key links. This involves how to efficiently create, define versions, and release digital model products. Traditional methods may lack effective version management or sufficient flexibility and control during the release process. Summary of the Invention

[0003] An object of the present invention is to provide a method, device and computer program product for managing digital model products.

[0004] In a first aspect, an embodiment of the present invention provides a method for managing digital model products, including: Responding to a digital model product creation instruction, and obtaining product group information included in the digital model product creation instruction; Constructing an initial digital model product according to the product group information; Defining a version of the initial digital model product to obtain a digital model product of a target version; Responding to a digital model product release instruction, and releasing the digital model product of the target version to a target model call platform.

[0005] In a possible implementation manner, the defining a version of the initial digital model product to obtain a digital model product of a target version includes: Defining parameter configuration, rule verification, data query, data processing, and model file management for the initial digital model product to obtain the digital model product of the target version.

[0006] In a possible implementation manner, the defining parameter configuration, rule verification, data query, data processing, and model file management for the initial digital model product to obtain the digital model product of the target version includes: Performing interface configuration on the initial digital model product to complete parameter configuration of the initial digital model product; Performing legality verification on the initial digital model product to complete rule verification of the initial digital model product; Performing associated data supplementation on the initial digital model product to complete data query of the initial digital model product; Configure pre - SQL for query, combination, and aggregation operations before model usage for the initial digital model product, and configure post - SQL for secondary processing of the results, original data, and parameters after model operations to complete data processing for the initial digital model product; Determine at least one execution model file for the initial digital model product, and perform request traffic ratio for the at least one execution model file to complete model file management for the initial digital model product.

[0007] In a possible implementation, after performing version definition on the initial digital model product to obtain a digital model product of the target version, the method further includes: Obtain test parameters; Input the test parameters into the digital model product of the target version to obtain test results, where the test results include task execution time and interface return results; When the test results indicate that the test passes, perform the step of responding to the digital model product release instruction to release the digital model product of the target version to the target model invocation platform.

[0008] In a possible implementation, after performing version definition on the initial digital model product to obtain a digital model product of the target version, the method further includes: Define the online time for the digital model product of the target version and generate an instruction call document for the digital model product of the target version; Configure the digital model product of the target version to an audit status; When the audit status indicates that the audit passes and the online time is triggered, perform the step of responding to the digital model product release instruction to release the digital model product of the target version to the target model invocation platform, and send the instruction call document to the target model invocation platform.

[0009] In a possible implementation, the method further includes: In response to a call operation that conforms to the instruction call document, call the digital model product of the target version from the target model invocation platform through an HTTP request.

[0010] In a possible implementation, before obtaining the product group information included in the digital model product creation instruction in response to the digital model product creation instruction, the method further includes: Obtain the user permission inheritance information, product group permissions, and digital model product permissions included in the digital model product creation instruction; On the basis that the user permission inheritance information, product group permissions, and digital model product permissions are all verified to be valid, execute the step of obtaining the product group information included in the digital model product creation instruction in response to the digital model product creation instruction.

[0011] In a second aspect, an embodiment of the present invention provides a digital model product management device, including: An acquisition module, configured to obtain the product group information included in the digital model product creation instruction in response to the digital model product creation instruction; and construct an initial digital model product according to the product group information; A management module, configured to perform version definition on the initial digital model product to obtain a digital model product of a target version; and publish the digital model product of the target version to a target model call platform in response to a digital model product release instruction.

[0012] In a third aspect, an embodiment of the present invention provides a computer device, where the computer device includes a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the computer device executes the digital model product management method in at least one possible implementation manner of the first aspect.

[0013] In a fourth aspect, an embodiment of the present invention provides a computer program product, where the computer program product includes a non-transitory computer-readable storage medium storing a computer program. When the computer program is read and executed by a computer, it implements the digital model product management method in at least one possible implementation manner of the first aspect.

[0014] Compared with the prior art, the beneficial effects provided by the present invention include: By using a digital model product management method, device, and computer program product disclosed in the present invention, the product group information can be obtained according to the received digital model product creation instruction, and then the initial digital model product is constructed using this information. Secondly, version definition is performed on the initial digital model product to obtain a digital model product of a target version. Finally, in response to the digital model product release instruction, the digital model product of the target version is published to the target model call platform. With such a design, not only the problem of version management is solved, but also the release process becomes more flexible and controllable. In addition, this method can better meet the data model operation requirements in various scenarios and improve the online application efficiency of data model products. Description of the Drawings

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0016] Figure 1 It is a schematic flowchart of the steps of the digital model product management method provided by the embodiment of the present invention; Figure 2 It is a schematic block diagram of the structure of the digital model product management device provided by the embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of the computer device provided by the embodiment of the present invention. Detailed implementation manners

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0018] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0019] The following will specifically describe the detailed implementation manners of the present invention with reference to the drawings.

[0020] To solve the technical problems in the foregoing background art, Figure 1 It is a schematic flowchart of the digital model product management method provided by the embodiment of the present disclosure. The following will introduce this digital model product management method in detail.

[0021] Step S201: In response to a digital model product creation instruction, obtain the product group information included in the digital model product creation instruction; Step S202: Construct an initial digital model product according to the product group information; Step S203: Define the version for the initial digital model product to obtain a digital model product of the target version; Step S204: In response to a digital model product release instruction, release the digital model product of the target version to the target model call platform.

[0022] In an embodiment of the present invention, by way of example, assume that an enterprise hopes to develop a digital model product for sales forecasting. The data mining modeler receives an instruction from the superior leader to create this product and provides product group information, including sales data, market trend analysis, etc. Based on the given sales data and market trend analysis, the data mining modeler begins to build the initial digital model product. They may use machine learning algorithms to train on historical sales data and adjust the model parameters according to market trends in order to accurately predict future sales. After the data mining modeler completes the initial digital model product, they need to define the version of this product. For example, they can adjust the model parameters or add new data sets for training to improve the accuracy of the model. Through continuous iteration and adjustment, they finally obtain the digital model product of the target version. After the data mining modeler completes the digital model product of the target version, the administrator receives the release instruction and approves it. Subsequently, the digital model product of the target version is released to the target model call platform. The business department can call this digital model in real time through the httpAPI interface provided by this platform for sales forecasting and obtain the results. For example, at the end of each quarter, the business department can use this digital model product to predict the sales situation of the next quarter in order to adjust the market strategy and formulate a new sales plan.

[0023] In a possible implementation manner, the foregoing step S203 can be executed by the following example.

[0024] (1) Define parameter configuration, rule verification, data query, data processing, and model file management for the initial digital model product to obtain the digital model product of the target version.

[0025] In an embodiment of the present invention, by way of example, assume that an enterprise is developing a digital model product for customer credit scoring. Data mining modelers need to configure parameters for the initial digital model product, such as selecting appropriate variables and weights, and setting parameters such as the threshold of the model, to ensure that the model can accurately predict customer credit scores. When developing a digital model product, data mining modelers may formulate some rules for verifying the output results of the model. For example, in the case of a customer credit scoring model, they can set some rules, such as the minimum credit score requirement, the conditions for rejecting applications, etc. These rules will be applied to the digital model product of the target version to ensure the compliance and accuracy of the output results. To build a digital model product, data mining modelers need to query relevant data from the company's database or other data sources. For example, in a customer purchase behavior prediction model, they can query historical order data, customer personal information, etc. to obtain the data required for training. When developing a digital model product, data mining modelers may need to perform some preprocessing operations on the original data to improve the quality and effect of the model. For example, they can perform operations such as data cleaning, missing value processing, and feature engineering on the data to obtain a more accurate and reliable data set. During the development process of the digital model product, data mining modelers will generate model files, including model parameters, weight matrices, etc. These model files need to be managed and stored for subsequent version definition and release. Administrators can use a specified system or tool to manage model files and ensure their security and integrity. Designed in this way, through the definition of the above steps, data mining modelers and administrators can configure, verify, query, process, and manage the initial digital model product according to requirements, and finally obtain the digital model product of the target version for real-time invocation and use by the business department.

[0026] In a possible implementation manner, the steps of defining the parameter configuration, rule verification, data query, data processing, and model file management for the initial digital model product to obtain the digital model product of the target version can be executed through the following examples.

[0027] (1) Perform interface configuration for the initial digital model product to complete the parameter configuration of the initial digital model product; (2) Perform legality verification on the initial digital model product to complete the rule verification of the initial digital model product; (3) Perform associated data supplementation for the initial digital model product to complete the data query of the initial digital model product; (4)Configure the pre-SQL for query, combination, and aggregation operations before model usage for the initial digital model product, and configure the post-SQL for secondary processing of the results, original data, and parameters after model operation to complete the data processing of the initial digital model product; (5)Determine at least one execution model file for the initial digital model product, and perform request traffic proportion for the at least one execution model file to complete the model file management of the initial digital model product.

[0028] In an embodiment of the present invention, by way of example, assume that an enterprise has developed a digital model product for customer recommendation. Data mining modelers need to configure interfaces on the release platform to complete the parameter configuration of the initial digital model product. They can set customer characteristics, recommendation algorithm parameters, etc., in order to obtain more accurate customer recommendation results in actual applications. The legality of the request parameter content is ensured through parameter verification rules. This process consists of three key components: parameter name, rule verification type, and rule verification expression. Parameter name: This is the identifier of the parameter to be verified. Rule verification type: This defines how the verification is performed, including two methods: regular expression and calculation. Regular expressions are usually used to check whether the parameter meets a specific format, such as a date, email address, etc.; calculation is used for more complex verification of parameter values, such as size comparison, range limitation, etc. Rule verification expression: This is the specific verification rule. When the rule verification type is regular, the rule expression is a regular expression; when the rule verification type is calculation, the rule expression is an expression based on the Lua language, which can perform complex mathematical and logical operations. When developing a digital model product, data mining modelers may need to obtain relevant data from multiple data sources to supplement the initial digital model product. For example, in a model for credit scoring, they can query the customer's transaction records, social media data, etc. for data supplementation to improve the prediction accuracy of the model. During the configuration process of the digital model product, data mining modelers can set the pre-SQL for queries, combinations, and aggregations that need to be executed before using the model. For example, in a sales prediction model, they can write SQL statements to filter, combine, and aggregate relevant sales data to generate the data input into the model. At the same time, they can also configure the post-SQL for secondary processing of the results, original data, and parameters after the model operation is completed to further optimize the output results of the model. During the management process of the digital model product, the administrator needs to determine at least one execution model file and configure the request traffic ratio. For example, in an advertising click-through rate prediction model, the administrator can manage multiple versions of the model file and adjust the request traffic ratio of each model file according to different experimental results and business requirements, so as to better evaluate and compare the effects of different models and select the optimal model for actual application.

[0029] In a possible implementation manner, after step S202, the embodiments of the present invention further provide the following examples.

[0030] (1) Obtain test parameters; (2) Input the test parameters into the digital model product of the target version to obtain test results, where the test results include task execution time and interface return results; (3)When the test result indicates that the test passes, execute the step of, in response to the digital model product release instruction, releasing the digital model product of the target version to the target model invocation platform.

[0031] In an embodiment of the present invention, exemplarily, before releasing a digital model product, an administrator needs to obtain appropriate test parameters. For example, in a model for product recommendation, the test parameters may include the number of recommended products, configuration parameters of the recommendation algorithm, etc. The administrator can select appropriate test parameters for testing and evaluation according to business requirements and experimental results. The administrator inputs the previously obtained test parameters into the digital model product of the target version for testing. For example, in a user portrait generation model, the administrator can use the test parameters as input to simulate user behavior or attribute data. Then, they can record the task execution time (the time required for model processing) and the interface return result (the generated user portrait data) as the test result. After testing the digital model product, the administrator determines whether the test passes based on the test result. For example, if the model completes the task within a predetermined time and the returned result meets the expectation, then the test can be considered passed. In this case, the administrator will continue to execute the instructions related to the release of the digital model product. After the test passes, the administrator can execute the release instruction to release the digital model product of the target version to the target model invocation platform. For example, in an online advertising click-through rate prediction model, the administrator can release the latest version of the model that has been tested and verified for the advertising system to call and use in real time, thereby improving the prediction accuracy and effect of the advertising click-through rate. Designed in this way, the administrator can test the digital model product of the target version based on the obtained test parameters and determine whether the test passes according to the test result. Once the test passes, they can execute the release instruction to release the digital model product of the target version to the target model invocation platform for the actual use of the business department or other users.

[0032] In a possible implementation manner, after step S202, the embodiment of the present invention further provides the following implementation manners.

[0033] (1)Define the online time for the digital model product of the target version and generate an instruction call document for the digital model product of the target version; (2)Configure the digital model product of the target version to the audit status; (3)When the audit status indicates that the audit passes and the online time is triggered, execute the step of, in response to the digital model product release instruction, releasing the digital model product of the target version to the target model invocation platform, and send the instruction call document to the target model invocation platform.

[0034] In an embodiment of the present invention, by way of example, an administrator needs to define the online time of a digital model product before releasing it. For example, in a model for anomaly detection, the administrator may decide to update the model at 3:00 AM every day and set the online time to 3:30 AM. At the same time, the administrator also needs to generate an instruction call document so that users can understand how to correctly call and use the model. Before releasing the digital model product, the administrator needs to configure the digital model product of the target version to the audit status. This can ensure that the product has undergone strict review and verification to meet the quality and reliability requirements of the enterprise. For example, in a model for credit scoring, the administrator can set the model to the audit status and conduct internal or external expert reviews. When the digital model product of the target version passes the audit and reaches the predetermined online time, the administrator can execute the release instruction to release the product to the target model call platform. For example, during the release process of a speech recognition model, the administrator can execute the release instruction after the audit is passed and the predetermined online time is reached, and deploy the model to the speech recognition platform for users to use. At the same time, they also send the instruction call document to the target model call platform so that users can understand how to correctly call the model. Designed in this way, after defining the online time for the digital model product of the target version, the administrator can configure it to the audit status. Once the audit is passed and the predetermined online time is reached, the administrator can execute the release instruction to release the product to the target model call platform and send the instruction call document for users to refer to. This can ensure the quality and reliability of the digital model product and provide users with a convenient way to call and use it.

[0035] In a possible implementation manner, the embodiment of the present invention further provides the following process.

[0036] (1) In response to a call operation that conforms to the instruction call document, call the digital model product of the target version from the target model call platform through an HTTP request.

[0037] In an embodiment of the present invention, exemplarily, if a user wants to use the target version of the digital model product, before using it, they will refer to the corresponding instruction call document. For example, in an image classification model, the instruction call document may contain information such as how to prepare the input image, select appropriate parameters, and how to process and interpret the output results. The user performs the call operation according to the instructions in the instruction call document and abides by the access rules and constraints. The user initiates a call from the target model call platform to the target version of the digital model product through an HTTP request. For example, in a natural language processing model, the user can use an HTTP POST request to send the text data to be processed to the target model call platform. The parameters of the request include the input text and other necessary information. Then, the target model call platform will pass the request to the target version of the digital model product for processing and return the processing results to the user. Designed in this way, the user can call the target version of the digital model product from the target model call platform through an HTTP request according to the instructions in the instruction call document. In this way, the user can easily use the model and obtain its processing results to meet their specific business needs.

[0038] In the embodiment of the present invention, illustratively, before the aforementioned step S201, the embodiment of the present invention further provides the following implementation manner.

[0039] (1) obtaining the digital model product creation instruction including user permission inheritance information, product group permissions, and digital model product permissions; (2) On the basis that the user permission inheritance information, product group permission and digital model product permission are all verified to be valid, executing the step of responding to the digital model product creation instruction and obtaining the product group information included in the digital model product creation instruction.

[0040] In an embodiment of the present invention, by way of example, assume that an administrator receives a digital model product creation instruction, which contains information about user permissions, product group permissions, and digital model product permissions. The administrator needs to parse this instruction and extract the data related to these permissions. For example, the user permission inheritance information may include specifying the permissions for a specific user to access the model, the product group permissions may stipulate the permissions that all members within the product group have, and the digital model product permissions define the access levels for specific models. After the administrator obtains the permission information in the digital model product creation instruction, they need to verify these permissions. The administrator will check whether the user permission inheritance information, product group permissions, and digital model product permissions are valid and legal. For example, when verifying user permissions, the administrator may check whether the user exists and whether they have sufficient access permissions. Similarly, the product group permissions and digital model product permissions also need to go through corresponding verifications. Once all permissions are verified as valid, the administrator can perform the operations in response to the digital model product creation instruction and obtain the product group information contained in the instruction. This may involve connecting to a database or other information storage system to retrieve the product group information related to the instruction. Designed in this way, after the administrator obtains the user permission inheritance information, product group permissions, and digital model product permissions in the digital model product creation instruction, they can verify these permissions and, after successful verification, perform the corresponding operations to obtain the product group information contained in the instruction. This can ensure that digital model products are created with the correct permissions and valid instructions and that the necessary product group information is obtained to support subsequent operations.

[0041] The following provides an overall implementation process for an embodiment of the present invention.

[0042] Before creating a model product, the product group should be created in accordance with the product group category, demand direction, demand department, and application common to the specific group. The model product creation is based on the specific sub-category to which the data product application belongs.

[0043] After creating the product, specific versions of the model product should be created within the product, and duplicate versions can be generated for the developed versions, as well as usage instruction documents, and online operations can be performed.

[0044] In an embodiment of the present invention, the product version is defined in terms of parameter configuration, parameter rule verification, data query, data processing, and model file management.

[0045] Parameter configuration: mainly used to configure the parameters and parameter types required by the interface when the product version is called through httpapi, as well as the default values that need to be automatically filled by the system when they are empty or not passed in the request, and whether the parameters need to enter the model file for calculation.

[0046] Parameter verification rules: The legality of the parameter content in the request is mainly verified by the parameter name rule, verification type rule, and verification expression. The verification types include regular and calculation. When the verification type is selected as calculation, the content filled in the rule expression is a Lua language expression.

[0047] Data query: To operate the model, it may not be enough just with the parameter content sent by the requester. Data supplement from the data warehouse or big data platform may also be required. Therefore, in data query, you can select what query conditions to use to query which data from which system, which database, and which table to supplement the operation model with data.

[0048] Data processing: Pre - SQL: If data combination or aggregation operations need to be performed on the data query results and parameters before the data enters the operation model, SQL is used to complete the data processing of the parameters and data query results. Post - SQL: Post - SQL is the function of finally re - processing the results of the operation model operation in combination with the data query and parameter content. The result of the post - SQL is the final interface return result.

[0049] Model file management: It is the model file required for the current data product version, and the request traffic allocation ratio of each model file can be configured. The sum of the total traffic ratios of multiple model files is 1.

[0050] After creation, the model version can be tested: The test will automatically substitute information such as parameter default values into the test page; the test results will include the execution time of each step and the return result of the API. When there are multiple operation model files, the operation results of each model will be returned. After the test passes, it can be published.

[0051] Publication: Only after completing the test can the model product version be published. When publishing, the online time needs to be defined or it can be put online immediately. After publishing, the product version will enter the waiting review state, and the instruction call document for this version will be generated synchronously.

[0052] Product management: Administrators can review, test, reject, modify the online time, etc. for the product versions waiting for review, and can also take offline and online actions on the reviewed products.

[0053] Permission management: Users can be added or deleted, and recharge passwords and permission assignments can be made for users.

[0054] Permission Assignment: User permission inheritance can be based on users, that is, modify the owner of the product to the successor. It is also possible to batch assign ownership for product groups or data products.

[0055] After the review and release, in the online state, other business systems can call the released data model product through the http method according to the instruction document.

[0056] In summary, the platform centered on the data model product in the embodiment of the present invention ensures the independence of new applications, continuous and friendly version management, and high-quality model management through refined data product definition and perfect permission management. In addition, it also provides a traffic allocation setting function, supports data model operations in various scenarios, and effectively isolates the new data model from the business system, so as to quickly go online with the application and avoid damaging the original business system architecture environment.

[0057] Please refer to Figure 2 , Figure 2 which is a schematic structural diagram of a digital model product management device 110 further provided by the embodiment of the present invention. The digital model product management device 110 includes: An acquisition module 1101, configured to, in response to a digital model product creation instruction, acquire product group information included in the digital model product creation instruction; and construct an initial digital model product according to the product group information; A management module 1102, configured to perform version definition on the initial digital model product to obtain a digital model product of a target version; and in response to a digital model product release instruction, release the digital model product of the target version to a target model call platform.

[0058] It should be noted that the implementation principle of the foregoing digital model product management device 110 can refer to the implementation principle of the foregoing digital model product management method, which will not be elaborated here. It should be understood that the division of each module of the above device is only a division of logical functions. In actual implementation, it can be fully or partially integrated into a physical entity, or physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also all be implemented in the form of hardware; or some modules can be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the digital model product management device 110 can be a separately established processing element, or can be integrated in a certain chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and the functions of the above digital model product management device 110 can be called and executed by a certain processing element of the above device. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together or can be independently implemented. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the integrated logic circuit in the processor element or the instructions in the form of software.

[0059] For example, the above modules can be one or more integrated circuits configured to implement the above method, such as: one or more application specific integrated circuits (ASICs), or, one or more digital signal processors (DSPs), or, one or more field programmable gate arrays (FPGAs), etc. Again, when a certain module above is implemented in the form of a processing element scheduling program code, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call program code. Again, these modules can be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0060] An embodiment of the present invention provides a computer device 100. The computer device 100 includes a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the computer device 100 executes the foregoing digital model product management device 110. As Figure 3 shown, Figure 3The computer device 100 provided in the embodiment of the present invention is a structural block diagram. The computer device 100 includes a digital model product management device 110, a memory 111, a processor 112 and a communication unit 113.

[0061] In order to realize data transmission or interaction, the memory 111, the processor 112 and the communication unit 113 are electrically connected to each other directly or indirectly. For example, the electrical connection between these elements can be realized through one or more communication buses or signal lines. The digital model product management device 110 includes at least one software function module that can be stored in the memory 111 in the form of software or firmware or solidified in the operating system (OS) of the computer device 100. The processor 112 is used to execute the digital model product management device 110 stored in the memory 111, such as the software function modules and computer programs included in the digital model product management device 110.

[0062] An embodiment of the present invention provides a readable storage medium, which includes a computer program. When the computer program is executed, the computer device where the readable storage medium is located is controlled to execute the aforementioned digital model product management method.

[0063] For illustrative purposes, the foregoing description is made with reference to specific embodiments. However, the above illustrative discussion is not intended to be exhaustive or to limit the present disclosure to the precise form disclosed. Numerous modifications and variations are possible in accordance with the above teachings. These embodiments are selected and described in order to best illustrate the principles of the present disclosure and its practical application, so that those skilled in the art can best utilize the present disclosure and utilize various embodiments with different modifications to suit the intended specific application.

Claims

1. A digital model product management method, characterized in that, including: In response to a digital model product creation instruction, obtaining product group information included in the digital model product creation instruction; Constructing an initial digital model product according to the product group information; Performing version definition on the initial digital model product to obtain a digital model product of a target version; In response to a digital model product release instruction, releasing the digital model product of the target version to a target model invocation platform.

2. The method according to claim 1, characterized in that The performing version definition on the initial digital model product to obtain a digital model product of a target version includes: Performing definition on parameter configuration, rule verification, data query, data processing, and model file management for the initial digital model product to obtain the digital model product of the target version.

3. The method according to claim 1, wherein The performing definition on parameter configuration, rule verification, data query, data processing, and model file management for the initial digital model product to obtain the digital model product of the target version includes: Performing interface configuration on the initial digital model product to complete parameter configuration of the initial digital model product; Performing legality verification on the initial digital model product to complete rule verification of the initial digital model product; Performing associated data supplementation on the initial digital model product to complete data query of the initial digital model product; Configuring a pre-SQL for query, combination, and aggregation operations before model use for the initial digital model product, and configuring a post-SQL for secondary processing of the result, original data, and parameters after model operation to complete data processing of the initial digital model product; Determining at least one execution model file for the initial digital model product, and performing request traffic ratio for the at least one execution model file to complete model file management of the initial digital model product.

4. The method according to claim 1, characterized in that, After performing version definition on the initial digital model product to obtain a digital model product of a target version, the method further includes: Obtaining test parameters; Inputting the test parameters into the digital model product of the target version to obtain a test result, where the test result includes task execution time and interface return result; When the test result indicates that the test is passed, performing the step of, in response to a digital model product release instruction, releasing the digital model product of the target version to a target model invocation platform.

5. The method according to claim 1, wherein After performing version definition on the initial digital model product to obtain a digital model product of a target version, the method further includes: Performing online time definition on the digital model product of the target version and generating an instruction call document for the digital model product of the target version; Configuring the digital model product of the target version to an audit status; When the audit status indicates that the audit is passed and the online time is triggered, performing the step of, in response to a digital model product release instruction, releasing the digital model product of the target version to a target model invocation platform, and sending the instruction call document to the target model invocation platform.

6. The method according to claim 5, wherein The method further includes: In response to a call operation that conforms to the description call document, the digital model product of the target version is called from the target model call platform through an HTTP request.

7. The method according to claim 1, characterized in that, Before obtaining the product group information included in the digital model product creation instruction in response to the digital model product creation instruction, the method further includes: Obtaining the user permission inheritance information, product group permissions, and digital model product permissions included in the digital model product creation instruction; On the basis that the user permission inheritance information, product group permissions, and digital model product permissions are all verified to be valid, execute the step of obtaining the product group information included in the digital model product creation instruction in response to the digital model product creation instruction.

8. A digital model product management device, characterized in that Includes: An acquisition module, configured to obtain the product group information included in the digital model product creation instruction in response to the digital model product creation instruction; Construct an initial digital model product according to the product group information; A management module, configured to perform version definition on the initial digital model product to obtain a digital model product of the target version; in response to a digital model product release instruction, release the digital model product of the target version to the target model call platform.

9. A computer device, characterized in that, The computer device includes a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the computer device executes the digital model product management method according to any one of claims 1-7.

10. A computer program product, characterized in that, The computer program product includes a non-transitory computer-readable storage medium storing a computer program. When the computer program is read and executed by a computer, the digital model product management method according to any one of claims 1-7 is implemented.

Images