A digital model product management method, device and computer program product

Abstract

The application discloses a kind of digital model product management method, device and computer program product, comprising: first, can obtain product group information according to received digital model product creation instruction, then utilize these information to build initial digital model product.Second, version definition is carried out to initial digital model product, and the digital model product of target version is obtained.Finally, in response to digital model product release instruction, the digital model product of target version is released to target model calling platform.Such design not only solves the problem of version management, but also makes the release process more flexible and controllable.In addition, the method can better meet the data model operation demand in various scenarios, and improve the online application efficiency of data model product.

Description

Technical Field

[0001] This invention relates to the field of digital model management technology, and more specifically, to a digital model product management method, apparatus, and computer program product. Background Technology

[0002] Currently, the management and operation of data model products are critical aspects. This involves how to efficiently create, version-define, 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] The purpose of this invention is to provide a digital model product management method, device, and computer program product.

[0004] In a first aspect, an embodiment of the present invention provides a digital model product management method, comprising:

[0005] In response to a digital model product creation instruction, obtain the product group information included in the digital model product creation instruction;

[0006] Based on the product group information, construct an initial digital model product;

[0007] Define a version of the initial digital model product to obtain the target version of the digital model product;

[0008] In response to the digital model product release instruction, the target version of the digital model product is released to the target model calling platform.

[0009] In one possible implementation, defining a version of the initial digital model product to obtain a target version of the digital model product includes:

[0010] The parameter configuration, rule verification, data query, data processing, and model file management of the initial digital model product are defined to obtain the target version of the digital model product.

[0011] In one possible implementation, the step of defining parameter configuration, rule validation, data query, data processing, and model file management for the initial digital model product to obtain the target version of the digital model product includes:

[0012] Configure the interface for the initial digital model product to complete the parameter configuration of the initial digital model product;

[0013] Perform a legality check on the initial digital model product to complete the rule check on the initial digital model product;

[0014] Supplement the associated data for the initial digital model product and complete the data query for the initial digital model product;

[0015] The initial digital model product is configured with pre-emptive SQL for querying, combining, and aggregating operations before model use, and post-emptive SQL for performing secondary processing on the results, original data, and parameters after model operation, thus completing the data processing of the initial digital model product.

[0016] For the initial digital model product, at least one execution model file is identified, and the request traffic ratio is calculated for the at least one execution model file to complete the model file management of the initial digital model product.

[0017] In one possible implementation, after defining a version of the initial digital model product to obtain a target version of the digital model product, the method further includes:

[0018] Obtain test parameters;

[0019] The test parameters are input into the target version of the digital model product to obtain the test results, which include the task execution time and the interface return results.

[0020] If the test results indicate that the test has passed, the step of responding to the digital model product release instruction and releasing the target version of the digital model product to the target model calling platform is executed.

[0021] In one possible implementation, after defining a version of the initial digital model product to obtain a target version of the digital model product, the method further includes:

[0022] Define the launch time for the target version of the digital model product and generate a description and calling document for the target version of the digital model product;

[0023] Configure the target version of the digital model product to an audit status;

[0024] When the review status indicates that the review has been approved and the launch time is triggered, the step of releasing the target version of the digital model product to the target model calling platform in response to the digital model product release instruction is executed, and the instruction calling document is sent to the target model calling platform.

[0025] In one possible implementation, the method further includes:

[0026] In response to a call operation that conforms to the aforementioned specification, the target version of the digital model product is invoked from the target model invocation platform via an HTTP request.

[0027] In one 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:

[0028] The digital model product creation instruction includes user permission inheritance information, product group permissions, and digital model product permissions;

[0029] Based on the verification that the user permission inheritance information, product group permissions, and digital model product permissions are all valid, 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 is executed.

[0030] Secondly, embodiments of the present invention provide a digital model product management device, comprising:

[0031] The acquisition module is used to respond 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 based on the product group information.

[0032] The management module is used to define the version of the initial digital model product to obtain the target version of the digital model product; in response to the digital model product release instruction, the target version of the digital model product is released to the target model calling platform.

[0033] Thirdly, embodiments of the present invention provide a computer device, the computer device including a processor and a non-volatile memory storing computer instructions, wherein when the computer instructions are executed by the processor, the computer device performs the digital model product management method described in at least one possible implementation of the first aspect.

[0034] Fourthly, embodiments of the present invention provide a computer program product, the computer program product including a non-transitory computer-readable storage medium storing a computer program, which, when read and executed by a computer, implements the digital model product management method described in at least one possible implementation of the first aspect.

[0035] Compared to existing technologies, the beneficial effects of this invention include: By employing the digital model product management method, apparatus, and computer program product disclosed in this invention, product group information is obtained based on received digital model product creation instructions, and then this information is used to construct an initial digital model product. Secondly, a version definition is performed on the initial digital model product to obtain a target version of the digital model product. Finally, in response to a digital model product release instruction, the target version of the digital model product is released to the target model calling platform. This design not only solves the version management problem but also makes the release process more flexible and controllable. Furthermore, this method can better meet the data model computation needs in various scenarios, improving the efficiency of deploying and applying data model products. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as limiting the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a flowchart illustrating the steps of the digital model product management method provided in an embodiment of the present invention.

[0038] Figure 2 A schematic block diagram of the structure of the digital model product management device provided in an embodiment of the present invention;

[0039] Figure 3 A schematic block diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0041] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0042] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0043] In order to solve the technical problems mentioned in the background art Figure 1 This is a flowchart illustrating the digital model product management method provided in this embodiment. The digital model product management method will now be described in detail.

[0044] Step S201: In response to the digital model product creation instruction, obtain the product group information included in the digital model product creation instruction;

[0045] Step S202: Construct an initial digital model product based on the product group information;

[0046] Step S203: Define the version of the initial digital model product to obtain the target version of the digital model product;

[0047] Step S204: In response to the digital model product release instruction, the target version of the digital model product is released to the target model calling platform.

[0048] In this embodiment of the invention, for example, suppose a company wants to develop a digital model product for sales forecasting. Data mining modelers receive instructions from their superiors to create the product, providing product group information, including sales data, market trend analysis, etc. Based on the given sales data and market trend analysis, the data mining modelers begin building an initial digital model product. They may use machine learning algorithms to train on historical sales data and adjust model parameters according to market trends to accurately predict future sales. After completing the initial digital model product, the data mining modelers need to define a version of the product. For example, they may adjust the model parameters or add new datasets for training to improve the model's accuracy. Through continuous iteration and adjustments, they finally obtain the target version of the digital model product. After the data mining modelers complete the target version of the digital model product, the administrator receives a release instruction and approves it. Subsequently, the target version of the digital model product is released to the target model calling platform. Business departments can use the platform's provided HTTP API interface to call the digital model in real time for sales forecasting and obtain results. For example, at the end of each quarter, business units can use this digital model product to forecast sales for the next quarter in order to adjust marketing strategies and develop new sales plans.

[0049] In one possible implementation, the aforementioned step S203 can be performed by the following example.

[0050] (1) Define the parameter configuration, rule verification, data query, data processing and model file management for the initial digital model product to obtain the target version of the digital model product.

[0051] In this embodiment of the invention, for example, suppose a company 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 like model thresholds, to ensure the model can accurately predict customer credit scores. When developing the digital model product, data mining modelers may formulate rules to validate the model's output. For example, in the case of a customer credit scoring model, they can set rules such as minimum credit score requirements and conditions for rejecting applications. These rules will be applied to the target version of the digital model product to ensure the compliance and accuracy of the output results. To build the 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 needed for training. When developing the digital model product, data mining modelers may need to perform some preprocessing operations on the raw data to improve the quality and effectiveness of the model. For example, they can perform data cleaning, missing value handling, feature engineering, etc., to obtain a more accurate and reliable dataset. During the development of digital model products, data mining modelers generate model files, including model parameters, weight matrices, etc. These model files need to be managed and stored for subsequent version definitions and releases. Administrators can use designated systems or tools to manage model files and ensure their security and integrity. Through this design and the defined steps above, data mining modelers and administrators can configure, validate, query, process, and manage the initial digital model product according to requirements, ultimately obtaining the target version of the digital model product for real-time use by business departments.

[0052] In one possible implementation, the steps of defining parameter configuration, rule verification, data query, data processing, and model file management for the initial digital model product to obtain the target version of the digital model product can be performed through the following example.

[0053] (1) Configure the interface for the initial digital model product and complete the parameter configuration for the initial digital model product;

[0054] (2) Perform a legality check on the initial digital model product and complete the rule check on the initial digital model product;

[0055] (3) Supplement the associated data for the initial digital model product and complete the data query for the initial digital model product;

[0056] (4) Configure the pre-SQL for querying, combining and aggregating operations before the model is used for the initial digital model product, and configure the post-SQL for secondary processing of the results, original data and parameters after the model operation is completed, so as to complete the data processing of the initial digital model product.

[0057] (5) For the initial digital model product, at least one execution model file is determined, and the request traffic ratio is calculated for the at least one execution model file to complete the model file management of the initial digital model product.

[0058] In this embodiment of the invention, for example, suppose a company develops a digital model product for customer recommendation. Data mining modelers need to configure the interface on the publishing platform to complete the parameter configuration of the initial digital model product. They can set customer characteristics, recommendation algorithm parameters, etc., to obtain more accurate customer recommendation results in practical applications. The validity of the request parameter content is ensured through parameter validation rules. This process includes three key components: parameter name, rule validation type, and rule validation expression. Parameter name: This is the identifier of the parameter to be validated. Rule validation type: This defines how validation is performed, including both regular expressions and calculations. Regular expressions are typically used to check whether parameters meet specific formats, such as dates, email addresses, etc.; calculations are used for more complex validation of parameter values, such as size comparisons, range restrictions, etc. Rule validation expression: This is the specific validation rule. When the rule validation type is regular expression, the rule expression is a regular expression; when the rule validation 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 credit scoring model, data miners can query customer transaction records, social media data, etc., to supplement the data and improve the model's predictive accuracy. During the configuration of digital model products, data mining modelers can set up pre-processing SQL statements for queries, combinations, and aggregations that need to be executed before the model is used. For example, in a sales forecasting model, they can write SQL statements to filter, combine, and aggregate relevant sales data to generate the input data for the model. They can also configure post-processing SQL statements to further optimize the model's output after the model's calculations are complete. In the management of digital model products, administrators need to identify at least one executable model file and configure the request traffic ratio. For example, in an ad click-through rate prediction model, administrators can manage multiple versions of the model file and adjust the request traffic ratio for each model file based on different experimental results and business needs. This allows for better evaluation and comparison of the effects of different models and the selection of the optimal model for practical application.

[0059] In one possible implementation, after step S202, the embodiments of the present invention also provide the following examples.

[0060] (1) Obtain test parameters;

[0061] (2) Input the test parameters into the target version of the digital model product to obtain the test results, which include the task execution time and the interface return results;

[0062] (3) If the test result indicates that the test has passed, execute the step of responding to the digital model product release instruction and releasing the target version of the digital model product to the target model calling platform.

[0063] In this embodiment of the invention, for example, before releasing a digital model product, the administrator needs to obtain appropriate test parameters. For instance, in a model for product recommendation, 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 based on business needs and experimental results. The administrator uses the aforementioned obtained test parameters to input them into the target version of the digital model product for testing. For example, in a user profile generation model, the administrator can use 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 profile data) as test results. After testing the digital model product, the administrator will determine whether the test has passed based on the test results. For example, if the model completes the task within the predetermined time and the returned result meets expectations, then the test can be considered passed. In this case, the administrator will continue to execute instructions related to the release of the digital model product. After the test passes, the administrator can execute the release instruction to release the target version of the digital model product to the target model calling platform. For example, in an online advertising click-through rate (CTR) prediction model, administrators can publish the latest, tested, and validated version of the model for the advertising system to use in real time, thereby improving the accuracy and effectiveness of CTR predictions. With this design, administrators can test the target version of the digital model product based on the acquired test parameters and determine whether it passes the test based on the results. Once the test is passed, they can execute a release command to publish the target version of the digital model product to the target model access platform for actual use by business departments or other users.

[0064] In one possible implementation, after step S202, the embodiments of the present invention also provide the following implementation.

[0065] (1) Define the launch time for the target version of the digital model product and generate a description and calling document for the target version of the digital model product;

[0066] (2) Configure the target version of the digital model product to an audit status;

[0067] (3) When the review status indicates that the review has been approved and the launch time is triggered, execute the step of releasing the target version of the digital model product to the target model calling platform in response to the digital model product release instruction, and send the instruction call document to the target model calling platform.

[0068] In this embodiment of the invention, for example, the administrator needs to define the launch time of a digital model product before publishing it. For instance, in a model used for anomaly detection, the administrator might decide to update the model at 3:00 AM every day and set the launch time to 3:30 AM. Simultaneously, the administrator also needs to generate instructional documentation to help users understand how to correctly call and use the model. Before publishing the digital model product, the administrator needs to configure the target version of the digital model product to a review status. This ensures that the product has undergone rigorous review and verification to meet the company's quality and reliability requirements. For example, in a model used for credit scoring, the administrator can set the model to a review status and conduct internal or external expert reviews. When the target version of the digital model product passes the review and reaches the predetermined launch time, the administrator can execute the publishing command to publish the product to the target model calling platform. For example, during the publishing process of a speech recognition model, the administrator can execute the publishing command after the review is passed and the predetermined launch time arrives, deploying the model to the speech recognition platform for user use. Simultaneously, they will also send instructional documentation to the target model calling platform so that users understand how to correctly call the model. This design allows administrators to define a launch time for a target version of a digital model product and then configure it to a review status. Once the review is approved and the scheduled launch time arrives, the administrator can execute the release command to publish the product to the target model access platform and send instructional documentation for user reference. This ensures the quality and reliability of the digital model product and provides users with a convenient way to access and use it.

[0069] In one possible implementation, embodiments of the present invention also provide the following process.

[0070] (1) In response to a call operation that conforms to the description of the call document, the target version of the digital model product is called from the target model call platform via an HTTP request.

[0071] In this embodiment of the invention, for example, when a user wants to use the target version of the digital model product, they will refer to the corresponding instruction manual before use. For example, in an image classification model, the instruction manual may contain information on how to prepare the input image, select appropriate parameters, and how to process and interpret the output results. The user performs the invocation operation according to the instructions in the instruction manual, complying with access rules and constraints. The user initiates the invocation from the target model invocation platform to the target version of the digital model product via 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 invocation platform. The parameters of the request include the input text and other necessary information. Then, the target model invocation platform will pass the request to the target version of the digital model product for processing and return the processing result to the user. With this design, users can invoke the target version of the digital model product from the target model invocation platform via an HTTP request according to the instructions in the instruction manual. In this way, users can conveniently use the model and obtain its processing results to meet their specific business needs.

[0072] In this embodiment of the invention, for example, prior to the aforementioned step S201, the following implementation method is also provided.

[0073] (1) Obtain the digital model product creation instruction, including user permission inheritance information, product group permissions, and digital model product permissions;

[0074] (2) On the basis that the user permission inheritance information, product group permissions and digital model product permissions are all verified to be valid, 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 is executed.

[0075] In this embodiment of the invention, exemplarily, suppose an administrator receives a digital model product creation instruction containing information about user permissions, product group permissions, and digital model product permissions. The administrator needs to parse the instruction and extract the data related to these permissions. For example, user permission inheritance information may specify the permissions that a particular user can use to access the model, product group permissions may specify the permissions that all members within the product group have, and digital model product permissions define the access level for a specific model. After obtaining the permission information from the digital model product creation instruction, the administrator needs 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 rights. Similarly, product group permissions and digital model product permissions also need to be verified accordingly. Once all permissions are verified to be valid, the administrator can perform 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 associated with the instruction. This design allows administrators to verify user permission inheritance information, product group permissions, and digital model product permissions obtained from the digital model product creation instruction. Upon successful verification, administrators can execute corresponding operations to retrieve the product group information contained in the instruction. This ensures that correct permissions and valid instructions are used to create the digital model product and obtains the necessary product group information to support subsequent operations.

[0076] The following provides an overall implementation process for an embodiment of the present invention.

[0077] Before creating a model product, a product group should be created according to the product group category, demand direction, demand department, and common applications corresponding to that specific group. Model product creation is based on the sub-category to which the data product application belongs.

[0078] After creating the product, you need to create a specific version of the model product within that product, and be able to copy the developed version, generate user manuals, and go live.

[0079] In this embodiment of the invention, the product version is defined by several aspects, including parameter configuration, parameter rule verification, data query, data processing, and model file management.

[0080] Parameter configuration: This is mainly used to configure the parameters required by the interface when calling the product version via HTTP API, the parameter types, the default values ​​that the system should automatically fill in when the request is empty or not passed, and whether the parameters need to be entered into the model file for calculation.

[0081] Parameter validation rules: These primarily validate the validity of request parameters through parameter name, validation type, and validation expression. The validation type includes regular expressions and calculations. When the validation type is selected as calculation, the content entered in the rule expression is based on Lua language expressions.

[0082] Data Query: To build a computational model, the parameters sent by the requester may not be sufficient. Data from a data warehouse or big data platform may also be needed to supplement the computational model. Therefore, in data query, we can choose which query conditions, which system, which database, which table, and which data to query to supplement the computational model.

[0083] Data processing:

[0084] Pre-processing SQL: If data query results and parameters need to be combined or aggregated before entering the calculation model, SQL is used to process the data. Post-processing SQL: Post-processing SQL is the function that combines the results of the calculation model with the data query and parameter content for final secondary processing. The result of post-processing SQL is the final result returned by the interface.

[0085] Model file management: This refers to the model files required for the current data product version, and allows configuration of the request traffic allocation ratio for each model file. The sum of the total traffic ratios for multiple model files is 1.

[0086] Once created, the model version can be tested:

[0087] The test will automatically load default values ​​and other information into the test page; the test results will include the execution time of each step and the API return results. If multiple computation model files exist, the results of each model will be returned. Once the test passes, the test can be published.

[0088] release:

[0089] Only after testing is completed can the model product version be released. When releasing, a launch time must be defined or it can be launched immediately. After release, the product version will enter a waiting-for-approval state, and the version's documentation will be generated synchronously.

[0090] Product Management:

[0091] Administrators can review, test, return, and modify the release time of product versions awaiting approval, and can also take approved products offline or bring them back online.

[0092] Access control:

[0093] Users can be added or deleted, and users can be assigned payment passwords and permissions.

[0094] Permission assignment: User permissions can be inherited based on the user, meaning the product owner can be changed to the inheritor. Ownership can also be assigned in batches to product groups or data products.

[0095] Once approved and released, other business systems can access the released data model product via HTTP, based on the documentation.

[0096] In summary, this invention's platform, centered on data model products, ensures the independence of new applications, continuous and user-friendly version management, and high-quality model management through refined data product definition and robust permission management. Furthermore, it provides traffic allocation settings to support data model computations in various scenarios and effectively isolates new data models from business systems, enabling rapid application deployment and avoiding disruption to the existing business architecture.

[0097] Please refer to the following: Figure 2 , Figure 2 A schematic block diagram of a digital model product management device 110 is also provided as an embodiment of the present invention. The digital model product management device 110 includes:

[0098] The acquisition module 1101 is used 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 based on the product group information.

[0099] The management module 1102 is used to define the version of the initial digital model product to obtain the target version of the digital model product; and in response to the digital model product release instruction, to release the target version of the digital model product to the target model calling platform.

[0100] It should be noted that the implementation principle of the aforementioned digital model product management device 110 can refer to the implementation principle of the aforementioned digital model product management method, and will not be repeated here. It should be understood that the division of the various modules of the above device is merely a logical functional division; in actual implementation, they can be fully or partially integrated into a single physical entity, or physically separated. Furthermore, these modules can all be implemented in software through processing element calls; they can all be implemented in hardware; or some modules can be implemented by processing element calls to software, while others are implemented in hardware. For example, the digital model product management device 110 can be a separately established processing element, or it can be integrated into a chip within the aforementioned device. Alternatively, it can be stored as program code in the memory of the aforementioned device, and called and executed by a processing element of the aforementioned device. The implementation of other modules is similar. Furthermore, these modules can be fully or partially integrated together, or implemented independently. 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 through the integrated logic circuits in the hardware of the processor element or through software instructions.

[0101] For example, these modules can be one or more integrated circuits configured to implement the above methods, such as one or more application-specific integrated circuits (ASICs), one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs). As another example, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a central processing unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together to implement a system-on-a-chip (SOC).

[0102] This invention provides a computer device 100, which 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 aforementioned digital model product management device 110. For example... Figure 3 As shown, Figure 3This is a structural block diagram of a computer device 100 provided in an embodiment of the present invention. The computer device 100 includes a digital model product management device 110, a memory 111, a processor 112, and a communication unit 113.

[0103] To enable data transmission or interaction, the memory 111, processor 112, and communication unit 113 are electrically connected to each other directly or indirectly. For example, these components can be electrically connected to each other 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 or embedded in the operating system (OS) of the computer device 100 in the form of software or firmware. The processor 112 is used to execute the digital model product management device 110 stored in the memory 111, such as the software function module and computer program included in the digital model product management device 110.

[0104] This invention provides a readable storage medium, which includes a computer program. When the computer program runs, it controls the computer device where the readable storage medium is located to execute the aforementioned digital model product management method.

[0105] For illustrative purposes, the foregoing description has been made with reference to specific embodiments. However, the foregoing illustrative discussions are not intended to be exhaustive or to limit the present disclosure to the precise forms disclosed. Numerous modifications and variations are possible in accordance with the foregoing teachings. These embodiments were chosen and described in order to best illustrate the principles of the present disclosure and its practical application, thereby enabling those skilled in the art to best utilize the disclosure and to employ various embodiments with different modifications to suit a particular intended application.

Claims

1. A digital model product management method characterized by, include: In response to a digital model product creation instruction, obtain the product group information included in the digital model product creation instruction; Based on the product group information, construct an initial digital model product; Define a version of the initial digital model product to obtain the target version of the digital model product; In response to the digital model product release instruction, the target version of the digital model product is released to the target model calling platform; The process of defining parameter configuration, rule validation, data query, data processing, and model file management for the initial digital model product to obtain the target version of the digital model product includes: Configure the interface for the initial digital model product to complete the parameter configuration of the initial digital model product; Perform a legality check on the initial digital model product to complete the rule check on the initial digital model product; Supplement the associated data for the initial digital model product and complete the data query for the initial digital model product; The initial digital model product is configured with pre-emptive SQL for querying, combining, and aggregating operations before model use, and post-emptive SQL for performing secondary processing on the results, original data, and parameters after model operation, thus completing the data processing of the initial digital model product. For the initial digital model product, at least one execution model file is identified, and the request traffic ratio is calculated for the at least one execution model file to complete the model file management of the initial digital model product.

2. The method according to claim 1, characterized in that, The step of defining a version of the initial digital model product to obtain the target version of the digital model product includes: The parameter configuration, rule verification, data query, data processing, and model file management of the initial digital model product are defined to obtain the target version of the digital model product.

3. The method according to claim 1, characterized in that, After defining the version of the initial digital model product to obtain the target version of the digital model product, the method further includes: Obtain test parameters; The test parameters are input into the target version of the digital model product to obtain the test results, which include the task execution time and the interface return results. If the test results indicate that the test has passed, the step of responding to the digital model product release instruction and releasing the target version of the digital model product to the target model calling platform is executed.

4. The method according to claim 1, characterized in that, After defining the version of the initial digital model product to obtain the target version of the digital model product, the method further includes: Define the launch time for the target version of the digital model product and generate a description and calling document for the target version of the digital model product; Configure the target version of the digital model product to an audit status; When the review status indicates that the review has been approved and the launch time is triggered, the step of releasing the target version of the digital model product to the target model calling platform in response to the digital model product release instruction is executed, and the instruction calling document is sent to the target model calling platform.

5. The method according to claim 4, characterized in that, The method further includes: In response to a call operation that conforms to the aforementioned specification, the target version of the digital model product is invoked from the target model invocation platform via an HTTP request.

6. 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: The digital model product creation instruction includes user permission inheritance information, product group permissions, and digital model product permissions; Based on the verification that the user permission inheritance information, product group permissions, and digital model product permissions are all valid, 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 is executed.

7. A digital model product management device, characterized in that, include: The acquisition module is used to acquire product group information included in the digital model product creation instruction in response to the digital model product creation instruction. Based on the product group information, construct an initial digital model product; The management module is used to define the version of the initial digital model product to obtain the target version of the digital model product; in response to the digital model product release instruction, the target version of the digital model product is released to the target model calling platform; The management module is specifically used for: The system performs interface configuration for the initial digital model product, completing parameter configuration; it performs legality verification for the initial digital model product, completing rule verification; it supplements associated data for the initial digital model product, completing data querying; it configures pre-emptive SQL for querying, combining, and aggregating operations before model use, and configures post-emptive SQL for secondary processing of results, original data, and parameters after model operation, completing data processing for the initial digital model product; it identifies at least one executable model file for the initial digital model product, and manages the model file for the initial digital model product by setting request traffic ratios for the at least one executable model file.

8. 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 performs the digital model product management method according to any one of claims 1-6.

9. A computer program product, characterized in that, The computer program product includes a non-transitory computer-readable storage medium storing a computer program, which, when read and executed by a computer, implements the digital model product management method according to any one of claims 1-6.

Images