Intelligent test data generation and optimization method and system based on low-code platform

Abstract

The invention discloses an intelligent test data generation and optimization method and system based on a low-code platform, and belongs to the technical field of software engineering and artificial intelligence, and the method comprises the steps: analyzing a business rule and data model definition information of a low-code application; according to the business rule and the data model definition information, utilizing a generative model to generate initial test data; performing quality evaluation and fault prediction on the generated initial test data; collecting test execution feedback information, including fault types caused by data, performance bottleneck positions and the like; formulating a data generation parameter adjustment strategy according to the test execution feedback information and the service target; and regenerating test data according to the adjustment strategy, and outputting the optimized test data for a subsequent iterative test. And the optimized test data is subjected to effect evaluation, so that the test quality and efficiency are continuously improved. According to the invention, the test data dynamically adapts to business demand changes, the test coverage rate and the fault discovery efficiency are improved, and the labor cost is reduced.

Description

Technical Field

[0001] The present invention relates to the fields of software engineering and artificial intelligence technology, and specifically to an intelligent test data generation and optimization method and system based on a low-code platform. Background Art

[0002] With the widespread adoption of low-code platforms, the demand for efficient and intelligent test data generation and optimization is becoming increasingly urgent in R&D and testing. Traditional test data generation methods are either overly manual or lack deep integration with low-code features and business scenarios. Data optimization also struggles to adapt in real time based on test feedback, limiting testing efficiency and quality and hindering the full potential of low-code platforms. Existing technologies lack mature intelligent test data solutions tailored to specific low-code environments. Summary of the Invention

[0003] The technical task of the present invention is to address the above shortcomings and provide an intelligent test data generation and optimization method and system based on a low-code platform, organically integrating the low-code platform with artificial intelligence technology, and automatically and accurately generating test data that fits the actual scenario according to business rules; through real-time feedback optimization strategy, the test data can dynamically adapt to changes in business needs, greatly improving test coverage and fault detection efficiency, and reducing labor costs.

[0004] The technical solution adopted by the present invention to solve its technical problem is:

[0005] The intelligent test data generation and optimization method based on the low-code platform includes the following steps:

[0006] Parse the business rules and data model definition information of low-code applications;

[0007] Generate initial test data using generative models based on business rules and data model definition information;

[0008] Perform quality assessment and failure prediction on the generated initial test data;

[0009] Collect test execution feedback information, including the types of failures caused by data and the location of performance bottlenecks;

[0010] Develop data generation parameter adjustment strategies based on test execution feedback and business goals;

[0011] Regenerate test data based on the adjustment strategy and output optimized test data for subsequent iterative testing;

[0012] Evaluate the effectiveness of optimized test data and continuously improve test quality and efficiency.

[0013] Furthermore, by parsing the metadata interface (such as JSON configuration files, database schema) or API calls of the low-code platform, the functional structure and business rules of the low-code application are extracted. The functional structure includes form fields and workflow nodes; the business rules include data validation logic and permission constraints; thus, the business rules and data model definition information are extracted;

[0014] The initial test data is generated by using a generative model based on business rules and data model definition information: for form-type applications, the text generation model is combined with field types and associated constraints to generate text data simulating user input; for task-based applications involving workflows, historical process data and business scenarios are analyzed to generate test data instances covering various branch processes and exception handling paths to ensure that the test data fully covers the functional points and business processes of low-code applications.

[0015] Furthermore, the generative model includes a Transformer-based text generation model and a conditional generative adversarial network, which are used to generate form-type text data and workflow-type structured data.

[0016] Furthermore, the step of performing quality assessment and fault prediction on the generated initial test data includes performing assessment and prediction using a data quality assessment model and a fault prediction model; assessing the quality of the generated test data and predicting possible fault locations.

[0017] Furthermore, quality assessment indicators include: field filling completeness, logical consistency, and boundary coverage;

[0018] Fault prediction includes: null pointer exception probability and permission conflict risk level.

[0019] Furthermore, the step of formulating a data generation parameter adjustment strategy based on test execution feedback information and business objectives includes dynamically adjusting data generation parameters, distribution characteristics or business scenario combinations according to the fault type and performance bottleneck location, so as to increase the generation weight of data related to modules with high failure rates and correct the proportion of data that causes performance problems.

[0020] Furthermore, data storage management is implemented through the test data management layer, supporting the indexing and efficient retrieval of historical test data (stored by application module classification), business metadata (such as field types, constraint rules) and model training materials (labeled fault data, performance logs).

[0021] The present invention also claims protection for an intelligent test data generation and optimization system based on a low-code platform, comprising:

[0022] The low-code platform interface layer is used to capture the business rules and data model definition information of low-code applications, interact with the artificial intelligence model layer, transmit business rules and data model information, and receive data generation results and model requirement information feedback;

[0023] The artificial intelligence model layer integrates a generative model and an analytical model. The generative model is used to generate test data that conforms to business logic based on business rules and data model definitions, and the analytical model is used to perform quality assessment and fault prediction on the generated test data.

[0024] The test data management layer is used to store historical test data, business metadata, and AI model training materials, provide training data materials for the AI ​​model layer, and transmit test data and business metadata information to the optimization decision layer;

[0025] The optimization decision layer is used to formulate data generation parameter adjustment strategies based on test execution feedback and business goals, and output optimized test data for subsequent iterative testing. The optimization decision layer includes an optimization result evaluation module, which analyzes the optimization effect through quantitative indicators (such as the improvement in test case pass rate and the reduction in defect density), and feeds the evaluation results back to the artificial intelligence model layer to adjust the generation strategy;

[0026] The system realizes intelligent test data generation and optimization through the above method.

[0027] The present invention also claims an intelligent test data generation and optimization device based on a low-code platform, comprising: at least one memory and at least one processor;

[0028] The at least one memory is configured to store a machine-readable program;

[0029] The at least one processor is configured to call the machine-readable program to implement the above method.

[0030] The present invention also claims protection for a computer-readable medium having computer instructions stored thereon, which implement the above method when executed by a processor.

[0031] Compared with the prior art, the intelligent test data generation and optimization method and system based on the low-code platform of the present invention have the following beneficial effects:

[0032] Existing low-code platforms often generate test data manually or out of touch with business scenarios. This invention organically integrates low-code platforms with artificial intelligence technology to automatically and accurately generate test data tailored to real-world scenarios based on business rules. Through real-time feedback optimization strategies, test data dynamically adapts to changing business needs, significantly improving test coverage and fault detection efficiency while reducing labor costs.

[0033] This invention is applicable to low-code R&D and testing platforms, such as the test module supporting Inspur's low-code development tools. Its implementation significantly improves the automation and intelligence of platform testing, shortens product development cycles, and enhances product stability and competitiveness. This demonstrates the company's technological innovation capabilities, helps expand market share, seize technological leadership in the low-code development field, create greater economic benefits for the company, and lay the foundation for long-term development. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a diagram illustrating the architecture of an intelligent test data generation and optimization system based on a low-code platform provided by one embodiment of the present invention;

[0035] Figure 2 This is a flowchart of an intelligent test data generation process based on a low-code platform provided by one embodiment of the present invention;

[0036] Figure 3 This is a real-time data optimization decision logic diagram provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The present invention will be further described below with reference to specific embodiments.

[0038] An embodiment of the present invention provides an intelligent test data generation and optimization method based on a low-code platform. The implementation of this method includes the following steps:

[0039] Parse the business rules and data model definition information of low-code applications;

[0040] Generate initial test data using generative models based on business rules and data model definition information;

[0041] Perform quality assessment and failure prediction on the generated initial test data;

[0042] Collect test execution feedback information, including the types of failures caused by data and the location of performance bottlenecks;

[0043] Develop data generation parameter adjustment strategies based on test execution feedback and business goals;

[0044] Regenerate test data based on the adjustment strategy and output optimized test data for subsequent iterative testing;

[0045] Evaluate the effectiveness of optimized test data and continuously improve test quality and efficiency.

[0046] The specific implementation of this method is as follows:

[0047] 1. If Figure 1As shown in the figure, a four-layer intelligent test data system is constructed, which includes a low-code platform interface layer, an artificial intelligence model layer, a test data management layer, and an optimization decision layer.

[0048] The low-code platform interface layer accurately extracts the functional structure (such as form fields, workflow nodes) and business rules (such as data validation logic, permission constraints) of the application under test by parsing the metadata interface (such as JSON configuration files, database schema) or API calls of the low-code platform; thereby extracting business rules and data model definition information.

[0049] The AI ​​model layer integrates multiple generative models (such as Transformer-based text data generation models and conditional generative adversarial networks for structured data generation) and analytical models (such as data quality assessment models and fault prediction models);

[0050] The test data management layer includes a data storage management module that supports indexing and efficient retrieval of historical test data (classified and stored by application module), business metadata (such as field types and constraint rules), and model training materials (annotated fault data and performance logs).

[0051] The optimization decision-making layer formulates data optimization strategies based on test result feedback and business goals.

[0052] 2. Intelligent generation method:

[0053] like Figure 2 The figure below illustrates the intelligent test data generation process. Based on the business rules and data model definitions extracted from the low-code platform, a generative model is used to automatically generate initial test data that conforms to business logic. For form-based applications, the text generation model is combined with field types and associated constraints to generate text data that simulates user input. For task-based applications involving workflows, historical process data and business scenarios are analyzed to generate test data instances covering various branch processes and exception handling paths, ensuring comprehensive coverage of the low-code application's functional areas and business processes.

[0054] The generative model includes a Transformer-based text generation model and a conditional generative adversarial network, which is used to generate form-type text data and workflow-type structured data.

[0055] Perform quality assessment and fault prediction on the generated initial test data, including using data quality assessment models and fault prediction models for assessment and prediction; assess the quality of the generated test data and predict possible fault locations.

[0056] Quality assessment indicators include: field filling completeness, logical consistency, and boundary coverage;

[0057] Fault prediction includes: null pointer exception probability and permission conflict risk level.

[0058] 3. Real-time optimization strategy:

[0059] like Figure 3 As shown in the figure, during the test execution process, AI analysis models are used to monitor test data performance in real time, such as the types of failures caused by the data and the location of performance bottlenecks. Based on the monitoring results, the optimization decision-making layer dynamically adjusts the test data generation parameters, distribution characteristics, or business scenario combinations. For example, the generation weight of data related to modules with high failure rates is increased, and the proportion of extreme data that causes performance issues is corrected. This makes the data for subsequent test iterations more targeted and effective, continuously improving test quality and efficiency.

[0060] This method implements data storage management through the test data management layer, supporting the indexing and efficient retrieval of historical test data (classified and stored by application module), business metadata (such as field type and constraint rules), and model training materials (annotated fault data and performance logs).

[0061] This approach implements intelligent test data generation and optimization for low-code platforms by building a four-layer architecture. The low-code platform interface layer captures business rules, the AI ​​model layer generates and evaluates data using generative and analytical models, the test data management layer stores data and assets, and the optimization decision layer formulates strategies based on feedback and outputs optimized data. These layers work closely together to improve test data quality and efficiency, meeting the testing needs of low-code application development.

[0062] An embodiment of the present invention further provides an intelligent test data generation and optimization system based on a low-code platform, comprising:

[0063] The low-code platform interface layer is used to capture the business rules and data model definition information of low-code applications, interact with the artificial intelligence model layer, transmit business rules and data model information, and receive data generation results and model requirement information feedback.

[0064] The artificial intelligence model layer integrates a generative model and an analytical model. The generative model is used to generate test data that conforms to business logic based on business rules and data model definitions, and the analytical model is used to perform quality assessment and fault prediction on the generated test data.

[0065] The test data management layer is used to store historical test data, business metadata, and artificial intelligence model training materials, provide training data materials for the artificial intelligence model layer, and transmit test data and business metadata information to the optimization decision-making layer.

[0066] The optimization decision layer is used to formulate data generation parameter adjustment strategies based on test execution feedback information and business goals, and output optimized test data for subsequent iterative testing. The optimization decision layer includes an optimization result evaluation module, which analyzes the optimization effect through quantitative indicators (such as the improvement rate of test case pass rate and the reduction rate of defect density), and feeds the evaluation results back to the artificial intelligence model layer to adjust the generation strategy.

[0067] Among them, the low-code platform interface layer extracts business rules and data model definition information by parsing and extracting the functional structure (such as form fields, workflow nodes) and business rules (such as data validation logic, permission constraints) of the low-code application.

[0068] The generative model includes a Transformer-based text generation model and a conditional generative adversarial network, which is used to generate form-type text data and workflow-type structured data.

[0069] The analysis model includes a data quality assessment model and a fault prediction model, which are used to assess the quality of generated test data and predict possible fault locations.

[0070] The test data management layer also includes a data storage management module for managing the storage and retrieval of stored historical test data, business metadata, and artificial intelligence model training materials.

[0071] The optimization decision layer also includes an optimization result evaluation module for evaluating the effect of the optimized test data, such as whether the test quality and efficiency are effectively improved, and further optimizing and adjusting the data generation parameters based on the evaluation results.

[0072] The system specifically implements intelligent test data generation and optimization through the intelligent test data generation and optimization method based on the low-code platform described in the above embodiment. The implementation steps include:

[0073] Parse the business rules and data model definition information of low-code applications;

[0074] Generate initial test data using generative models based on business rules and data model definition information;

[0075] Perform quality assessment and failure prediction on the generated initial test data;

[0076] Collect test execution feedback information, including the types of failures caused by data and the location of performance bottlenecks;

[0077] Develop data generation parameter adjustment strategies based on test execution feedback and business goals;

[0078] Regenerate test data based on the adjustment strategy and output optimized test data for subsequent iterative testing;

[0079] Evaluate the effectiveness of optimized test data and continuously improve test quality and efficiency.

[0080] The specific application cases of this system are as follows:

[0081] In Inspur Group's internal low-code R&D project, this system was integrated into a unified R&D and testing platform. For a low-code enterprise resource management application, the system parsed the procurement module's form fields (such as supplier name, material code, purchase quantity, etc.) and approval process definitions through the interface layer. A text generation model was used to generate supplier name text that conforms to actual business needs based on the industry corpus. A conditional generative adversarial network injected business rule constraints (such as the minimum / maximum purchase quantity and the formula relating unit price to total price) to ensure that the generated structured data conforms to actual business logic. Diverse test cases were also generated based on historical data distribution. During test execution, the fault prediction model used a random forest algorithm to train a classification model based on fault types (such as permission conflicts and data format errors) and location data in historical test logs. This model predicted the fault modules and probabilities that might be triggered by newly generated test data. The optimized decision layer then adjusted the generation logic of the relevant data for that node in subsequent test data, adding test cases under different permission role combinations. This effectively improved the depth and accuracy of testing for this module, facilitating rapid iterative optimization and high-quality delivery of low-code applications.

[0082] In another specific application case, for the customer service ticket system developed with low code, the system parses the ticket flow rules (such as priority allocation and automatic transfer conditions) through the interface layer. The generative model combines historical ticket data and fault scenarios (such as timeout and unprocessed, permission exceeded) to generate test data containing abnormal timestamps and illegal operation roles. The fault prediction model identifies potential conflicts in the transfer logic and optimizes the decision-making layer to increase the transfer path coverage to 95%.

[0083] An embodiment of the present invention further provides an intelligent test data generation and optimization device based on a low-code platform, comprising: at least one memory and at least one processor;

[0084] The at least one memory is configured to store a machine-readable program;

[0085] The at least one processor is used to call the machine-readable program to implement the intelligent test data generation and optimization method based on the low-code platform described in the above embodiment.

[0086] An embodiment of the present invention further provides a computer-readable medium having computer instructions stored thereon, which, when executed by a processor, causes the processor to execute the intelligent test data generation and optimization method based on a low-code platform described in the above embodiment. Specifically, a system or device equipped with a storage medium can be provided, on which software program code that implements the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or device is caused to read and execute the program code stored in the storage medium.

[0087] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute part of the present invention.

[0088] Examples of storage media for providing program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, the program code can be downloaded from a server computer via a communication network.

[0089] In addition, it should be clear that the functions of any of the above embodiments can be achieved not only by executing the program code read by the computer, but also by enabling the operating system operating on the computer to complete part or all of the actual operations based on the instructions of the program code.

[0090] In addition, it can be understood that the program code read from the storage medium is written into the memory provided in the expansion board inserted into the computer or into the memory provided in the expansion unit connected to the computer, and then based on the instructions of the program code, the CPU installed on the expansion board or expansion unit is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above embodiments.

[0091] The present invention has been shown and described in detail above through the accompanying drawings and preferred embodiments. However, the present invention is not limited to these disclosed embodiments. Based on the above multiple embodiments, those skilled in the art can know that the code review methods in the above different embodiments can be combined to obtain more embodiments of the present invention, and these embodiments are also within the scope of protection of the present invention.

Claims

1. An intelligent test data generation and optimization method based on a low-code platform, characterized in that: The implementation of this method includes the following steps: Parse the business rules and data model definition information of low-code applications; Generate initial test data using generative models based on business rules and data model definition information; Perform quality assessment and failure prediction on the generated initial test data; Collect test execution feedback information, including the types of failures caused by data and the location of performance bottlenecks; Develop data generation parameter adjustment strategies based on test execution feedback and business goals; Regenerate test data based on the adjustment strategy and output optimized test data for subsequent iterative testing; Evaluate the effectiveness of optimized test data and continuously improve test quality and efficiency.

2. The intelligent test data generation and optimization method based on the low-code platform according to claim 1 is characterized in that: By parsing the metadata interface or API call of the low-code platform, the functional structure and business rules of the low-code application are extracted. The functional structure includes form fields and workflow nodes; the business rules include data validation logic and permission constraints; thus, the business rules and data model definition information are extracted; The initial test data is generated by using a generative model based on business rules and data model definition information: for form-type applications, the text generation model is combined with field types and associated constraints to generate text data simulating user input; for task-based applications involving workflows, historical process data and business scenarios are analyzed to generate test data instances covering various branch processes and exception handling paths to ensure that the test data fully covers the functional points and business processes of low-code applications.

3. The intelligent test data generation and optimization method based on the low-code platform according to claim 1 or 2 is characterized in that: The generative model includes a Transformer-based text generation model and a conditional generative adversarial network, which is used to generate form-type text data and workflow-type structured data.

4. The intelligent test data generation and optimization method based on the low-code platform according to claim 1 is characterized in that: The steps of performing quality assessment and fault prediction on the generated initial test data include performing assessment and prediction using a data quality assessment model and a fault prediction model; assessing the quality of the generated test data and predicting possible fault locations.

5. The intelligent test data generation and optimization method based on the low-code platform according to claim 1 or 4 is characterized in that: Quality assessment indicators include: field filling completeness, logical consistency, and boundary coverage; Fault prediction includes: null pointer exception probability and permission conflict risk level.

6. The intelligent test data generation and optimization method based on the low-code platform according to claim 1 is characterized in that: The step of formulating a data generation parameter adjustment strategy based on test execution feedback information and business objectives includes dynamically adjusting data generation parameters, distribution characteristics or business scenario combinations according to the fault type and performance bottleneck location, so as to increase the generation weight of data related to modules with high failure rates and correct the proportion of data that causes performance problems.

7. The intelligent test data generation and optimization method based on the low-code platform according to claim 1 is characterized in that: Data storage management is implemented through the test data management layer, supporting indexing and efficient retrieval of historical test data, business metadata, and model training materials.

8. Intelligent test data generation and optimization system based on low-code platform, characterized by: include: The low-code platform interface layer is used to capture the business rules and data model definition information of low-code applications, interact with the artificial intelligence model layer, transmit business rules and data model information, and receive data generation results and model requirement information feedback; The artificial intelligence model layer integrates a generative model and an analytical model. The generative model is used to generate test data that conforms to business logic based on business rules and data model definitions, and the analytical model is used to perform quality assessment and fault prediction on the generated test data. The test data management layer is used to store historical test data, business metadata, and AI model training materials, provide training data materials for the AI ​​model layer, and transmit test data and business metadata information to the optimization decision layer; The optimization decision layer is used to formulate data generation parameter adjustment strategies based on test execution feedback and business goals, and output optimized test data for subsequent iterative testing. The optimization decision layer includes an optimization result evaluation module, which analyzes the optimization effect through quantitative indicators and feeds the evaluation results back to the artificial intelligence model layer to adjust the generation strategy; The system realizes intelligent test data generation and optimization through the method described in any one of claims 1 to 7.

9. An intelligent test data generation and optimization device based on a low-code platform, characterized in that: include: at least one memory and at least one processor; The at least one memory is configured to store a machine-readable program; The at least one processor is configured to call the machine-readable program to implement the method according to any one of claims 1 to 7.

10. A computer-readable medium, characterized in that The computer readable medium stores computer instructions, which, when executed by a processor, implement the method according to any one of claims 1 to 7.

Images