Cross-system-level error code correction method and device and computer equipment

Abstract

The invention relates to a cross-system-level error code correction method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: marking system hierarchies associated with system error reporting information, code contents of the system hierarchies and a logical relationship among the system hierarchies, and projecting the system hierarchies and the code contents of the system hierarchies to a three-dimensional interactive interface; according to the system levels, the code content of the system levels and the logic relation between the system levels, obtaining corresponding reference error reasons and error codes; obtaining a plurality of correction schemes and corresponding reliability according to the reference error reasons and the error codes, and projecting the correction schemes and the corresponding reliability to a three-dimensional interaction interface; and in response to a target correction scheme selected by the user through the virtual reality equipment, projecting the target correction scheme to the three-dimensional interaction interface. By adopting the method, in the software development process, the error positioning and correcting efficiency can be improved, and the cross-system hierarchy of the development cycle can be shortened.

Description

Technical Field

[0001] This application relates to the field of software development technology, and in particular to a method, apparatus, computer device, computer-readable storage medium, and computer program product for cross-system level error code correction. Background Technology

[0002] In modern software development, system architecture typically encompasses multiple layers, including databases, front-end UI layers, back-end service layers, and API interface layers. While this complex architecture enhances system flexibility, it also presents challenges for error troubleshooting and code correction. Traditional code correction tools rely on preset rules and can only analyze single system layers, failing to dynamically analyze cross-layer logical coupling issues in real time. For example, a front-end UI layer error might be caused by a back-end service layer or database layer, but traditional tools struggle to effectively locate such cross-layer problems. When troubleshooting such errors, developers need to manually examine fragmented information returned from multiple system layers, such as database logs, front-end errors, and back-end exception stacks. This significantly increases the average troubleshooting time, reduces development efficiency, and extends the development cycle.

[0003] Furthermore, when troubleshooting and correcting errors, developers are still limited by traditional two-dimensional IDE interfaces, making it difficult to intuitively understand the logical relationships of the code and the interactions between different system levels. This can easily lead to secondary errors and reduce correction efficiency. Therefore, traditional technologies have significant shortcomings in locating and correcting cross-system level error codes.

[0004] Therefore, there is an urgent need for a cross-system level error code correction method, device, computer equipment, computer-readable storage medium, and computer program product that can improve the efficiency of error location and correction and shorten the development cycle. Summary of the Invention

[0005] Therefore, it is necessary to provide a cross-system level error code correction method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve the efficiency of error location and correction and shorten the development cycle, in order to address the above-mentioned technical problems.

[0006] Firstly, this application provides a method for correcting error codes across system levels, including:

[0007] In the event of system error messages during software development, obtain the logical relationships between each system level, the logical dependency weights of each system level, and the overall code.

[0008] Based on the system error messages during the software development process, and based on the logical relationships between system levels, the logical dependency weights of each system level, and the overall code, the system levels associated with the system error messages, the code content of the system levels, and the logical relationships between the system levels are marked, and the system levels associated with the system error messages and the code content of the system levels are projected onto the 3D interactive interface.

[0009] Based on the system level associated with the system error message, the code content of the system level, and the logical relationship between the system levels, obtain the corresponding reference error reason and error code;

[0010] Based on the reference error cause and error code, multiple correction schemes and corresponding reliability are obtained, and the multiple correction schemes and corresponding reliability are projected onto a three-dimensional interactive interface. The three-dimensional interactive interface is used by the user to simulate the code correction process in a virtual reality interface in the form of three-dimensional animation.

[0011] In response to the target correction scheme selected by the user through the virtual reality device on the three-dimensional interactive interface, the patch file of the target correction scheme is subjected to sandbox testing to obtain the test results, and the test results are projected onto the three-dimensional interactive interface.

[0012] In one embodiment, the process of obtaining the system error information includes:

[0013] During software development, monitor the system's operational information and trace the call chain at each system level;

[0014] When error messages are projected onto a 3D interactive interface, the eye-tracking module of the virtual reality device is used to obtain the user's pupil movement information.

[0015] Based on the user's pupil movement information, the coordinates of the interface elements of the 3D interactive interface are matched and dynamically calibrated.

[0016] Based on the coordinates of the interface elements in the 3D interactive interface, determine the system error message that the user's current gaze is focused on.

[0017] In one embodiment, obtaining the logical relationships between system levels, the logical dependency weights of each system level, and the overall code based on the system error information includes:

[0018] The system error messages are input into a pre-trained text classification model, and the system error messages are pre-classified according to the error type and urgency.

[0019] Construct an abstract syntax tree based on the current code;

[0020] The system error messages after classification and the abstract syntax tree of the current code are used as input to the graph neural network. Multi-hop reasoning is performed along the dependency relationship of the current code to output the logical relationship between each system level, the logical dependency weight of each system level and the overall code.

[0021] In one embodiment, obtaining the corresponding reference error cause and error code based on the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels includes:

[0022] Obtain front-end error messages, related system error messages, back-end service logs, and database content during the software development process;

[0023] The system layers associated with front-end error messages, related system error messages, back-end service logs, database content, system error messages, code content at the system layers, and logical relationships between system layers are input into a multimodal fusion neural network model to calculate the reference error causes and error codes for the main errors.

[0024] In one embodiment, the step of responding to a user's selection of a target correction scheme via a virtual reality device on a 3D interactive interface, performing sandbox testing on the patch file of the target correction scheme, obtaining test results, and projecting the test results onto the 3D interactive interface includes:

[0025] In response to the target correction scheme selected by the user through a virtual reality device and using gesture or voice commands, obtain the patch file for the target correction scheme;

[0026] An isolated environment is built using Docker containers to conduct sandbox testing on the patch files and obtain test results. If the test results show that the test is passed, the target modification scheme is marked as the recommended scheme and will be given priority in subsequent recommendations.

[0027] The test results are projected onto a 3D interactive interface to obtain feedback from the user based on the test results. The multimodal fusion neural network model is then updated and trained based on the feedback results.

[0028] In one embodiment, prior to sandbox testing of the patch file for the target fix, the method further includes:

[0029] When a user selects the target correction scheme through a virtual reality device, the correction steps of the target correction scheme are rendered into a three-dimensional dynamic map, and the three-dimensional dynamic map animation is projected onto the three-dimensional interactive interface using a ray tracing algorithm.

[0030] In response to the user's modification operation based on the target modification scheme, the target modification scheme is modified.

[0031] Secondly, this application also provides a cross-system level error code correction device, comprising:

[0032] The data acquisition module is used to obtain the logical relationships between system levels, the logical dependency weights of each system level, and the overall code when system error messages occur during software development.

[0033] The data processing module is used to annotate the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels based on the system error information during the software development process, the logical dependency weight of each system level, and the overall code.

[0034] The projection module is used to project the system-level information and code content associated with system error messages onto the 3D interactive interface.

[0035] The data acquisition module is also used to obtain the corresponding reference error cause and error code based on the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels;

[0036] The data processing module is also used to obtain multiple correction schemes and corresponding reliability based on the reference error cause and error code;

[0037] The projection module is also used to project multiple correction schemes and corresponding reliability onto a three-dimensional interactive interface, which is used by users to simulate the code correction process in the form of three-dimensional animation in a virtual reality interface.

[0038] The testing module is used to respond to the target correction scheme selected by the user through the virtual reality device in the 3D interactive interface, to perform sandbox testing on the patch file of the target correction scheme, and obtain the test results;

[0039] The projection module is also used to project the test results onto the three-dimensional interactive interface.

[0040] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0041] In the event of system error messages during software development, obtain the logical relationships between each system level, the logical dependency weights of each system level, and the overall code.

[0042] Based on the system error messages during the software development process, and based on the logical relationships between system levels, the logical dependency weights of each system level, and the overall code, the system levels associated with the system error messages, the code content of the system levels, and the logical relationships between the system levels are marked, and the system levels associated with the system error messages and the code content of the system levels are projected onto the 3D interactive interface.

[0043] Based on the system level associated with the system error message, the code content of the system level, and the logical relationship between the system levels, obtain the corresponding reference error reason and error code;

[0044] Based on the reference error cause and error code, multiple correction schemes and corresponding reliability are obtained, and the multiple correction schemes and corresponding reliability are projected onto a three-dimensional interactive interface. The three-dimensional interactive interface is used by the user to simulate the code correction process in a virtual reality interface in the form of three-dimensional animation.

[0045] In response to the target correction scheme selected by the user through the virtual reality device on the three-dimensional interactive interface, the patch file of the target correction scheme is subjected to sandbox testing to obtain the test results, and the test results are projected onto the three-dimensional interactive interface.

[0046] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0047] In the event of system error messages during software development, obtain the logical relationships between each system level, the logical dependency weights of each system level, and the overall code.

[0048] Based on the system error messages during the software development process, and based on the logical relationships between system levels, the logical dependency weights of each system level, and the overall code, the system levels associated with the system error messages, the code content of the system levels, and the logical relationships between the system levels are marked, and the system levels associated with the system error messages and the code content of the system levels are projected onto the 3D interactive interface.

[0049] Based on the system level associated with the system error message, the code content of the system level, and the logical relationship between the system levels, obtain the corresponding reference error reason and error code;

[0050] Based on the reference error cause and error code, multiple correction schemes and corresponding reliability are obtained, and the multiple correction schemes and corresponding reliability are projected onto a three-dimensional interactive interface. The three-dimensional interactive interface is used by the user to simulate the code correction process in a virtual reality interface in the form of three-dimensional animation.

[0051] In response to the target correction scheme selected by the user through the virtual reality device on the three-dimensional interactive interface, the patch file of the target correction scheme is subjected to sandbox testing to obtain the test results, and the test results are projected onto the three-dimensional interactive interface.

[0052] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0053] In the event of system error messages during software development, obtain the logical relationships between each system level, the logical dependency weights of each system level, and the overall code.

[0054] Based on the system error messages during the software development process, and based on the logical relationships between system levels, the logical dependency weights of each system level, and the overall code, the system levels associated with the system error messages, the code content of the system levels, and the logical relationships between the system levels are marked, and the system levels associated with the system error messages and the code content of the system levels are projected onto the 3D interactive interface.

[0055] Based on the system level associated with the system error message, the code content of the system level, and the logical relationship between the system levels, obtain the corresponding reference error reason and error code;

[0056] Based on the reference error cause and error code, multiple correction schemes and corresponding reliability are obtained, and the multiple correction schemes and corresponding reliability are projected onto a three-dimensional interactive interface. The three-dimensional interactive interface is used by the user to simulate the code correction process in a virtual reality interface in the form of three-dimensional animation.

[0057] In response to the target correction scheme selected by the user through the virtual reality device on the three-dimensional interactive interface, the patch file of the target correction scheme is subjected to sandbox testing to obtain the test results, and the test results are projected onto the three-dimensional interactive interface.

[0058] The aforementioned cross-system-level error code correction methods, devices, computer equipment, computer-readable storage media, and computer program products, when system error information exists during software development, obtain the logical relationships between each system level, the logical dependency weights of each system level, and the overall code. Combining these logical relationships, dependency weights, and the overall code, they achieve precise error location and analysis, projecting relevant information onto a 3D interactive interface, providing developers with intuitive visualization support. Through multi-hop dependency reasoning and the generation of reference error causes, developers can quickly understand the root cause of errors, avoiding the tedious process of layer-by-layer troubleshooting in traditional methods, and significantly improving error location efficiency. Simultaneously, the system provides multiple correction schemes and their reliability assessments, and verifies the effectiveness of the target correction scheme through sandbox testing, feeding the test results back to the 3D interactive interface in real time, helping developers quickly select the optimal solution and reducing the risk of secondary errors caused by error correction. Furthermore, the application of virtual reality devices and 3D interactive interfaces provides developers with an immersive development experience, breaking the limitations of traditional 2D IDE interfaces, improving developers' cognitive efficiency of code structure, optimizing the correction process, and reducing the development cycle. Overall, this technical solution significantly improves the efficiency and quality of software development, reduces development costs, and enhances the working experience of developers. Attached Figure Description

[0059] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a diagram illustrating the application environment of a cross-system-level error code correction method in one embodiment.

[0061] Figure 2 This is a flowchart illustrating a cross-system-level error code correction method in one embodiment;

[0062] Figure 3 This is a flowchart illustrating a cross-system-level error code correction method in another embodiment;

[0063] Figure 4 This is a diagram showing the main module architecture of the cross-system-level error code correction device in the most detailed embodiment;

[0064] Figure 5 This is a structural block diagram of a cross-system-level error code correction device in one embodiment;

[0065] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0067] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0068] The cross-system level error code correction method provided in this application can be applied to systems such as... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104, or it can be located in the cloud or on another network server.

[0069] When system error messages occur during software development, server 104 obtains the logical relationships between system levels, the logical dependency weights of each system level, and the overall code. Based on the system error messages during software development, and according to the logical relationships between system levels, the logical dependency weights of each system level, and the overall code, it marks the system levels, code content, and logical relationships associated with the system error messages, and projects the system levels and code content associated with the system error messages onto the 3D interactive interface of terminal 102. Based on the system levels, code content, and logical relationships associated with the system error messages, it obtains the corresponding reference error causes and error codes. Based on the reference error causes and error codes, it obtains multiple correction schemes and their corresponding reliability, and projects these multiple correction schemes and their corresponding reliability onto the 3D interactive interface. The 3D interactive interface is used by the user to simulate the code correction process in a virtual reality interface in the form of 3D animation. In response to the target correction scheme selected by the user through the virtual reality device on the 3D interactive interface, it performs sandbox testing on the patch file of the target correction scheme, obtains the test results, and projects the test results onto the 3D interactive interface.

[0070] The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection equipment. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. The server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0071] In one exemplary embodiment, such as Figure 2 As shown, a method for correcting error codes across system levels is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 202 to 210. Wherein:

[0072] Step S202: In the event of system error messages during software development, obtain the logical relationships between each system level, the logical dependency weights of each system level, and the overall code.

[0073] Specifically, system error messages refer to error messages or exception records generated during software operation when a program fails to execute normally due to various reasons. These messages typically include the following key components:

[0074] Error Description: Briefly describe the nature of the error, such as "database connection failed", "array out of bounds", "network request timed out", etc.

[0075] Error location: Indicates the specific location where the error occurred, such as the line number, function name, or file name in the code.

[0076] Stack trace: Displays the call stack at the time the error occurred, helping developers trace the error's propagation path.

[0077] Log recording: Log files generated during system runtime record the running status and operation records before and after the error occurred.

[0078] Contextual information: This may include user actions, system status, input and output data, etc. This information helps to reproduce and analyze errors.

[0079] Log recording, exception handling mechanisms, monitoring tools, user feedback, and automated testing tools can be used to obtain the logical relationships between different system levels, the logical dependency weights of each system level, and the overall code when system error messages occur during software development.

[0080] Step S204: Based on the system error information during the software development process, and based on the logical relationships between system levels, the logical dependency weights of each system level, and the overall code, mark the system level associated with the system error information, the code content of the system level, and the logical relationships between system levels, and project the system level associated with the system error information and the code content of the system level onto the 3D interactive interface.

[0081] Specifically, in software systems, complex logical relationships exist between different layers (such as the front-end UI layer, back-end service layer, database layer, API interface layer, etc.). For example, the front-end UI layer may call the back-end service layer's interface, and the back-end service layer may depend on the database query results. Through a dynamic cross-system dependency graph construction and interaction module, the system analyzes the runtime behavior and call chain of the code to generate a dynamic dependency graph. This graph clearly shows the logical relationships between each system layer, helping developers understand how errors propagate between different layers.

[0082] Logical dependency weights refer to the strength of dependencies between different system layers. For example, a backend service might highly depend on the results of a database query, resulting in a higher dependency weight; while another backend service might only occasionally call an API, leading to a lower dependency weight. Through an AI analytics engine, the system can analyze runtime data, historical fixes, and code structure information to calculate the logical dependency weights between different system layers. These weights can be calculated based on factors such as call frequency, data transfer volume, and error propagation paths.

[0083] The overall code refers to the collection of code across all levels of the system, including front-end code, back-end code, database query code, etc. The system obtains the complete codebase of the current project through an Integrated Development Environment (IDE) and uses it as the basis for analysis.

[0084] Users wear wearable virtual reality devices and use a 3D interactive integrated development environment (IDE) for software development. The 3D interactive interface serves as a visual platform for user interaction with the system, providing an immersive interactive experience through virtual reality devices (such as VR headsets and controllers). The projected content includes:

[0085] Logical relationships between system levels: The interaction paths and dependencies between system levels are displayed graphically, for example, by connecting nodes at different levels with lines to clearly show the call chain.

[0086] Logical dependency weights at each system level: The strength of dependencies is intuitively represented by colors, line thickness, etc., helping developers quickly identify key dependency points.

[0087] Full code: This displays code snippets in a visual way in a three-dimensional space. Developers can view and manipulate the code using virtual reality devices, such as zooming in, zooming out, and moving code blocks.

[0088] Step S206: Based on the system level associated with the system error message, the code content of the system level, and the logical relationship between the system levels, obtain the corresponding reference error cause and error code.

[0089] Specifically, the process of obtaining reference error causes involves: First, determining the initial location and type of the error based on the system error message. Then, tracing the error propagation path using the logical relationships between different system layers. For example, if the front-end reports an error, logical relationships can be used to trace it back to the back-end service layer and database layer. Next, combining logical dependency weights, identifying the code snippets and system layers with the greatest impact on the error. For instance, if a back-end service has a high dependency weight and that service reports an error, then that service is likely the critical source of the error. Based on the above analysis, generating possible error causes. These causes can be specific code issues (such as uninitialized variables, incorrect data format, network request timeouts, etc.) or system design issues (such as high coupling, low cohesion, etc.).

[0090] The process of obtaining error codes involves: based on the error reason and the overall code, pinpointing the specific code location. For example, if the error reason is "the database query returned an incorrect data format," then the specific database query code can be located. By examining the code's implementation, it's confirmed whether errors exist. For example, checking if the database query statement is correct and whether the backend service correctly processes the query results. The located error code snippet is then extracted and provided to developers for further analysis and correction.

[0091] Step S208: Based on the reference error cause and error code, obtain multiple correction schemes and corresponding reliability, and project the multiple correction schemes and corresponding reliability onto the three-dimensional interactive interface. The three-dimensional interactive interface is used by the user to simulate the code correction process in the form of three-dimensional animation in the virtual reality interface.

[0092] Specifically, based on the identified error causes, potential solutions are analyzed using large-scale models such as AI. For example, if the error is "the database query returned an incorrect data format," possible solutions include: modifying the database query statement to return the correct data format; adding data format validation logic in the backend service; and handling abnormal data formats in the frontend code. By referencing a database of historical code correction solutions, solutions to similar problems are identified and used as references for future solutions, generating multiple solutions, each targeting a different approach or implementation method.

[0093] Next, by combining the historical code correction scheme database, we analyze the success rate of similar schemes in the past, evaluate the feasibility of each correction scheme in the current code structure and system environment, and its advantages and disadvantages.

[0094] Considering code complexity and dependencies, assess the difficulty of implementing the proposed solution. The system generates a reliability score for each revised solution, typically expressed as a percentage.

[0095] The 3D interactive interface is an immersive visualization environment that developers can interact with using virtual reality devices (such as VR headsets). Multiple fixes and their reliability scores are graphically displayed within the interface. For example, each fix can be displayed as a virtual "card" showing a brief description and reliability score. Developers can select fixes of interest and view detailed information using gestures, voice commands, or virtual controllers. Developers can then choose a fix for further testing or application, with the system providing real-time feedback on the test results or application effects.

[0096] Step S210: In response to the target correction scheme selected by the user through the virtual reality device on the three-dimensional interactive interface, a sandbox test is performed on the patch file of the target correction scheme to obtain the test results, and the test results are projected onto the three-dimensional interactive interface.

[0097] Specifically, the 3D interactive interface visually displays multiple correction schemes, and users can select the scheme of interest through interaction methods such as gestures, voice, or virtual controllers. The correction scheme selected by the user is called the "target correction scheme." This scheme is usually the one that the user believes is most likely to solve the problem.

[0098] Sandbox testing is performed on the patch files of the target fix. The fix typically exists in the form of patch files, which contain the modifications to the code. Sandbox testing is a testing method that runs code in an isolated environment. It ensures that the testing process does not affect the actual system by creating an environment similar to but isolated from the real-world operating environment (such as a Docker container).

[0099] Testing process: Apply the patch file of the target fix to the sandbox environment, input simulated data into the sandbox environment to simulate the actual running scenario, run the code in the sandbox environment, observe whether the patch solves the problem, and check whether any new problems are introduced.

[0100] Record the running status, output results, log information, etc. during the test process. The test results include information such as whether the patch successfully solved the problem, whether it introduced new problems, and whether performance was affected.

[0101] Success: If the patch resolves the issue and does not introduce new issues, the test result is "success".

[0102] Failure: If the patch does not resolve the issue or introduces a new problem, the test result is "failure".

[0103] Test results are displayed visually in a 3D interactive interface. For example, test results can be displayed as a virtual "card" containing detailed information about the test. If the test is successful, the card can display "Test Passed" and provide details about the patch, such as the number of lines of code modified and performance metrics. If the test fails, the card can display "Test Failed" and provide the reason for the failure, such as the introduction of a new bug or a performance issue. Users can view the detailed test results through gestures, voice, or a virtual controller and select the next step based on the results, such as choosing other fixes or modifying the current solution.

[0104] The aforementioned cross-system-level error code correction method, when system error messages occur during software development, obtains the logical relationships between each system level, the logical dependency weights of each system level, and the overall code. Combining these logical relationships, dependency weights, and the overall code, it achieves precise error location and analysis, and projects this information onto a 3D interactive interface, providing developers with intuitive visualization support. Through multi-hop dependency reasoning and the generation of reference error causes, developers can quickly understand the root cause of errors, avoiding the tedious process of layer-by-layer troubleshooting in traditional methods, and significantly improving error location efficiency. Simultaneously, the system provides multiple correction schemes and their reliability assessments, and verifies the effectiveness of the target correction scheme through sandbox testing, feeding the test results back to the 3D interactive interface in real time, helping developers quickly select the optimal solution and reducing the risk of secondary errors caused by error correction. Furthermore, the application of virtual reality devices and the 3D interactive interface provides developers with an immersive development experience, breaking the limitations of traditional 2D IDE interfaces, improving developers' cognitive efficiency of code structure, optimizing the correction process, and reducing the development cycle. Overall, this technical solution significantly improves the efficiency and quality of software development, reduces development costs, and enhances the developer's work experience.

[0105] In one exemplary embodiment, such as Figure 3 As shown, the process of obtaining system error messages includes:

[0106] Step S302: During the software development process, monitor the system's operation information and trace the call chain at each system level;

[0107] Step S304: When the error message is projected onto the 3D interactive interface, the eye-tracking module of the virtual reality device is used to obtain the motion information of the user's pupils.

[0108] Step S306: Based on the user's pupil movement information, match the coordinates of the interface elements of the three-dimensional interactive interface and perform dynamic calibration.

[0109] Step S308: Determine the system error information currently focused on by the user's gaze based on the coordinates of the interface elements in the 3D interactive interface.

[0110] Specifically, the operational information of a monitoring system refers to the real-time collection of system operational status information during the software development process using monitoring tools (such as logging systems and performance monitoring tools). This information includes, but is not limited to, log files, stack traces, memory usage, network requests and responses, etc.

[0111] Call chain tracing refers to tracing the call chain across different layers of a system (such as the front-end UI layer, back-end service layer, database layer, API interface layer, etc.). This is typically achieved using distributed tracing technologies (such as Spring Cloud Sleuth) to understand how errors propagate between different layers.

[0112] The captured error information is displayed visually in a 3D interactive interface. This interface is presented to the user through a virtual reality device (such as a VR headset), allowing the user to view the error information within a virtual environment. The relevant content of the error information (such as error description, location, and call path) is rendered as 3D graphics or text and projected onto the virtual interface in the user's field of vision. The eye-tracking module (including sensors and cameras) built into the virtual reality device (such as a VR headset) can capture the user's pupil movement information in real time. This information includes the pupil's position, movement trajectory, and gaze point. When the error information is projected onto the 3D interactive interface, the eye-tracking module begins to work, recording the user's pupil movement. By analyzing the pupil movement information, the system can determine the specific position of the user's gaze within the 3D interactive interface. This involves matching the pupil's gaze point with the coordinates of elements on the interface.

[0113] To ensure matching accuracy, the system dynamically adjusts calibration parameters based on real-time pupil movement data. This is because a user's gaze can be affected by various factors (such as minute device displacements and the user's head movements), and dynamic calibration improves matching precision. Through this matching and calibration process, the system can determine the specific system error message the user is currently focusing on. For example, the user might be viewing a specific error description, call chain, or code snippet. This technology helps developers quickly locate error messages of interest without manually searching the interface. Through eye tracking, the system can automatically identify the user's focus, thus providing more accurate information support.

[0114] In this embodiment, eye tracking and a 3D interactive interface provide a more natural and intuitive interaction method. Dynamic calibration ensures accurate recognition of the user's gaze, improves the reliability of the system, and combined with virtual reality technology, provides developers with an immersive development environment and enhances the overall development experience.

[0115] In an exemplary embodiment, based on system error information, the logical relationships between system layers, the logical dependency weights of each system layer, and the overall code are obtained, including:

[0116] The system error messages are input into a pre-trained text classification model, and the system error messages are pre-classified according to the error type and urgency.

[0117] Construct an abstract syntax tree based on the current code;

[0118] The system error messages after classification and the abstract syntax tree of the current code are used as input to the graph neural network. Multi-hop reasoning is performed along the dependency relationship of the current code to output the logical relationship between each system level, the logical dependency weight of each system level and the overall code.

[0119] Specifically, system error messages typically include error descriptions, location of occurrence, and stack traces. However, this information is often scattered and difficult to directly reflect the root cause of the error or its propagation path across system levels. Therefore, a series of technical methods are needed to extract useful content from the error messages.

[0120] The pre-trained text classification model is a machine learning or deep learning-based model that has been trained using a large amount of labeled error messages. It can identify key features in the error messages, such as error type (syntax error, network anomaly, database query failure, etc.) and urgency level (crash level, warning level, etc.).

[0121] The captured system error messages are input into the model, which then categorizes them into different error types and urgency levels based on their semantic content. For example, a "database connection failed" error message is categorized as "network anomaly," with a urgency level of "high." This pre-classification quickly identifies key features of the error messages, providing a foundation for subsequent detailed analysis.

[0122] An Abstract Syntax Tree (AST) is a tree-like structure used to represent the syntactic structure of program code. Each node represents an element in the code (such as a variable, function, operator, etc.), and edges represent the syntactic relationships between them (such as function calls, variable assignments, etc.). Parsing the code of the current system generates an AST. This step provides the foundation for subsequent code structure analysis. Through the AST, the structure and syntactic relationships of the code can be clearly represented, supporting subsequent dependency analysis.

[0123] The categorized system error messages and the abstract syntax tree of the current code are used as input to a graph neural network (GNN). The GNN performs multi-hop inference along the dependencies of the current code. A GNN is a deep learning model for processing graph-structured data, capable of learning complex dependencies between nodes. The pre-categorized system error messages and the abstract syntax tree of the current code are input into the GNN. The GNN performs multi-hop inference along the dependencies of the code. This means the model considers not only directly related code snippets but also tracks deep dependencies across multiple system levels (e.g., from front-end to back-end, and then to the database). The output then shows the logical relationships between system levels, the logical dependency weights of each system level, and the overall code.

[0124] In this embodiment, multi-hop inference enables precise identification of the root cause of errors, even when the error spans multiple system levels. Combining error messages and code structure, comprehensive error analysis results are provided, helping developers quickly understand the full picture of the problem. Utilizing pre-trained models and graph neural networks, automated error analysis is achieved, reducing the time and workload of manual troubleshooting. By identifying key dependencies and weights, developers are helped to select the optimal fix, improving fix efficiency.

[0125] In an exemplary embodiment, based on the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels, the corresponding reference error cause and error code are obtained, including:

[0126] Obtain front-end error messages, related system error messages, back-end service logs, and database content during the software development process;

[0127] The system layers associated with front-end error messages, related system error messages, back-end service logs, database content, system error messages, code content at the system layers, and logical relationships between system layers are input into a multimodal fusion neural network model to calculate the reference error causes and error codes for the main errors.

[0128] Front-end error messages refer to errors or exceptions related to front-end code (such as HTML, CSS, JavaScript, etc.) in the user's browser or client application. Related systems refer to other systems that interact with the current system, such as third-party services or other modules in a microservice architecture. Error messages from related systems refer to errors or exceptions generated during their interaction with the current system. Back-end service logs record the running status, operation records, and exception information of back-end services. These logs are crucial for analyzing back-end service issues. Database content includes database query statements, execution results, table structures, and other information. Database content is key to analyzing data-related issues (such as query errors, data format problems, etc.).

[0129] A multimodal fusion neural network model is a deep learning model that combines multiple data types (such as text, code structure, logs, database content, etc.). It can process different types of data and fuse them for comprehensive analysis.

[0130] Specifically, multimodal fusion neural network models infer possible causes of errors by comprehensively analyzing various input data. These causes can be specific code problems (such as uninitialized variables or incorrect data formats) or system design problems (such as high coupling or low cohesion). By analyzing the code structure and logical relationships, multimodal fusion neural network models can pinpoint the specific location of the erroneous code. For example, if the error cause is "the database query returned incorrect data format," the model can locate the specific database query code.

[0131] In this embodiment, by combining backend service logs, database content, logical relationships, dependency weights, and overall code, the model can comprehensively analyze the causes of errors, avoiding the limitations of analyzing a single data source. The model can not only infer the cause of the error, but also accurately locate the specific error code, reducing the time developers spend troubleshooting. By using a multimodal fusion neural network model, different types of data can be processed, improving the accuracy and reliability of the analysis.

[0132] In an exemplary embodiment, in response to a user selecting a target correction scheme through a virtual reality device on a 3D interactive interface, a sandbox test is performed on the patch file of the target correction scheme to obtain test results, and the test results are projected onto the 3D interactive interface, including:

[0133] In response to the target correction scheme selected by the user through a virtual reality device and using gesture or voice commands, obtain the patch file for the target correction scheme;

[0134] An isolated environment is built using Docker containers to conduct sandbox testing on the patch files and obtain the test results. When the test results show that the test is passed, the target fix is ​​marked as the recommended solution and will be given priority in subsequent recommendations.

[0135] The test results are projected onto a 3D interactive interface to obtain user feedback based on the test results. The multimodal fusion neural network model is then updated and trained based on the feedback.

[0136] Specifically, users wear virtual reality devices (such as VR headsets) and view multiple fixes through a 3D interactive interface. Users can select a target fix using gestures (such as selection or swiping) or voice commands (such as the voice command "Select Fix A"). The fix selected by the user is called the "target fix." This fix is ​​typically the one the user believes is most likely to solve the problem. The system responds to the user's command, retrieves the patch file corresponding to the selected target fix, and prepares for subsequent testing.

[0137] Use Docker containers to build a test environment that is similar to but isolated from the actual runtime environment. Docker containers can be quickly started and destroyed, making them suitable for sandbox testing. Apply the patch file for the target fix to the Docker container, input simulated data into the container to simulate the actual running scenario, run the code in the container, observe whether the patch solves the problem, and check whether it introduces any new problems. Record the running status, output results, log information, etc. during the test process to generate test results. The test results include information such as whether the patch successfully solved the problem, whether it introduced any new problems, and whether performance was affected. Display the test results in a visual way in a 3D interactive interface. For example, the test results can be displayed as a virtual "card" containing detailed information about the test results.

[0138] Users can provide feedback on test results via gestures or voice commands, such as confirming whether the results meet expectations. User feedback, including confirming the accuracy of the results or offering other suggestions, allows the system to update and train the multimodal fusion neural network model based on user feedback. This step optimizes model performance and improves the accuracy and reliability of subsequent correction schemes.

[0139] In this embodiment, a closed-loop correction, verification, and optimization process is achieved through a combination of virtual reality interaction, sandbox testing, and model updates. Users can select correction schemes through VR devices, the system verifies the effectiveness of the schemes through sandbox testing using Docker containers, and updates the model based on user feedback. This approach not only improves the reliability of correction schemes but also optimizes the developer's workflow and enhances development efficiency.

[0140] In one exemplary embodiment, before sandbox testing the patch file of the target fix, the method further includes:

[0141] When a user selects a target correction scheme through a virtual reality device, the correction steps of the target correction scheme are rendered into a three-dimensional dynamic map, and the three-dimensional dynamic map animation is projected onto the three-dimensional interactive interface using a ray tracing algorithm.

[0142] In response to user modification operations based on the target correction scheme, the target correction scheme is modified.

[0143] Specifically, the system analyzes the patch file of the target correction plan and extracts the specific correction steps. These steps might include modifying the value of a variable, adding a conditional statement, or adjusting the order of function calls. These correction steps are arranged logically and constructed as a graph using nodes and edges. Each node represents a correction step, and edges represent dependencies between steps. Animations are used to demonstrate the execution process of each correction step, helping users intuitively understand the role and impact of each step.

[0144] Ray tracing is an advanced graphics rendering technique that simulates the propagation of light in a virtual environment, generating high-quality images and animations. Ray tracing provides more realistic visuals and enhances user immersion. It's used to render 3D dynamic maps, generating smooth animations. These rendered animations are then projected onto the user's 3D interactive interface. Users can view these animations through virtual reality devices (such as VR headsets) and intuitively understand the execution process of corrective steps. Users can pause, play, and rewind animations, or view detailed information about specific steps, using gestures, voice commands, or virtual controllers.

[0145] While viewing the 3D dynamic atlas, users may find that certain correction steps need adjustment or optimization. Users can directly modify the correction scheme through the interactive interface. The system responds to the user's modification operation, updates the 3D dynamic atlas in real time, and re-renders the animation effects, allowing users to immediately see the modified results. Based on the user's modification operation, the system regenerates the patch file of the correction scheme, preparing for subsequent sandbox testing.

[0146] In this embodiment, a three-dimensional dynamic map and interactive functions help users intuitively understand the correction steps after selecting a target correction scheme, and allow users to modify the correction scheme. This visual interactive method not only improves users' understanding of the correction scheme, but also enhances their sense of participation and control, ultimately optimizing the correction process and improving development efficiency.

[0147] The most detailed embodiment of this application is as follows:

[0148] When users wear wearable virtual reality devices and use a 3D interactive integrated development environment (IDE) for software development, existing technologies typically acquire and analyze error information from each system level separately when errors occur during system operation. This results in insufficient cross-system level correlation analysis capabilities, low anomaly localization efficiency, and difficulty in locating cross-level error codes. To address these issues, improve code correction efficiency, and reduce development cycles, this patent proposes a method for locating and correcting cross-system level error codes using virtual reality and artificial intelligence technologies. Figure 4 As shown, the architecture of this technical solution is as follows:

[0149] (1) The core of this technical solution is a cross-system error code location and correction intelligent agent, which includes three sub-functional modules: dynamic cross-system dependency graph construction and interaction module, AI analysis engine and virtual reality interaction code correction module based on federated learning.

[0150] (2) The 3D interactive integrated development environment supports cross-system error code location and correction of intelligent agents, mainly assisting in the implementation of virtual reality interactive functions. In addition, it enables rapid acquisition of environmental information to assist in error code location and correction.

[0151] (3) The historical code correction scheme database is used to store the historical code correction schemes selected by users.

[0152] The specific implementation method is as follows:

[0153] Step 0: Embed a lightweight probe in the 3D interactive integrated development environment to capture runtime stack information, memory snapshots, network requests, and historical operations; use distributed tracing technology to achieve cross-system layer (such as database, front-end UI layer, back-end service layer, API interface layer, etc.) call chain tracing (such as link tagging by integrating Sleuth); deploy a historical code correction scheme database.

[0154] Step 1: The user wears a wearable virtual reality device and uses a 3D interactive integrated development environment (IDE) for software development. When the user interface returns an error message, the virtual reality device uses eye-tracking technology to capture the user's pupil movements in real time through built-in cameras, sensors, and other modules. It matches the gaze point with the coordinates of interface elements, dynamically calibrates, and then senses and captures the system error message that the user is currently focusing on.

[0155] Step 2: The dynamic cross-system dependency graph construction and interaction module uses machine learning algorithms to classify the perceived system error information. That is, the system error information perceived in the previous step is input into a pre-trained text classification model. Semantic parsing is performed on the unparsed error information, and the error type (such as network anomaly, resource conflict, etc.) and urgency level (such as crash level, warning level, etc.) are pre-classified. The classified system error information is input into a graph neural network built based on the abstract syntax tree of the current code for multi-hop dependency inference. The output is the current code dependency relationship, the cross-system level call chain, and the code content related to the error information in other system levels (for example, when the error information is an error returned by the front end, the model can not only locate the specific front end method, but also associate it with other system levels such as back end related module methods through the graph neural network). The module then renders the relationship graph, code content, and associated code call chains into 3D nodes, projecting them onto a 3D interactive interface in front of the user using near-eye display technology and virtual devices. Users wearing virtual reality devices can view these elements and use gestures (such as pinching and unfolding fingers to zoom in and out, or swiping) to observe the flow of data across systems and code call paths. Compared to traditional debugging methods, the proposed solution combines AI and virtual reality technologies, breaking through the limitations of 2D IDE interfaces and transcending the boundaries between different system levels. It addresses the problem that current debugging tools can only locate error messages at the current single system level and lack sufficient visualization of complex code logic, significantly improving troubleshooting efficiency, optimizing software development cycles, and reducing the cost of troubleshooting complex systems.

[0156] Step 3: Based on the information output by the dynamic cross-system dependency graph construction and interaction module in the previous steps, the AI ​​code analysis engine defines the dependency weights of each system level associated with the error message in the current state (e.g., database dependency weight = 0.7, front-end dependency weight = 0.1, back-end dependency weight = 0.2, etc.).

[0157] Step 4: The AI ​​code analysis engine automatically associates backend service logs and database information in the current development environment. This information, along with the information obtained in previous steps (i.e., error codes / information, dependency weights of each system in the current state, current stack information, memory snapshots, and network requests), is input into a multimodal fusion neural network model. The model calculates the code content containing the core error and identifies the core cause of the error. For example, if a system anomaly occurs in the front-end user interface and the required data cannot be displayed, the neural network analyzes the front-end logs based on the information obtained in previous steps, including dynamic cross-system dependency graphs and call chains. It then associates the logs with relevant backend methods, traces the SQL statements that implement the query in the database, and outputs the specific error code content and its cause.

[0158] Step 5.1: The system projects the error code block, dynamic cross-system dependency graph, and the cause of the error onto the 3D interactive interface in front of the user.

[0159] Step 5.2: The virtual reality interaction code correction module based on federated learning builds a federated learning model based on a federated learning framework (such as TensorFlow Federated, Federated AI Technology Enabler, or PySyft). The model is input into the current code problem and the contents of the historical code correction solution database, and outputs several possible correction methods and patches, and gives the advantages and disadvantages of each possible solution (such as giving the expected correction success rate or the number of lines of code to be modified) for the user to choose from.

[0160] Step 6: The user wears a virtual reality device, and all candidate solutions and their advantages and disadvantages are projected onto the virtual reality interface in front of the user using near-eye display technology. The user can compare multiple solutions through gestures or voice commands and select the desired solution.

[0161] Step 7: The user confirms the fix via gesture or voice command. The system then performs sandbox testing on the corresponding patch, i.e., building an isolated environment using Docker containers, automatically deploying the patch, inputting simulated data, and testing to ensure it doesn't introduce new problems. If the test passes, the system marks the solution as a "recommended solution" and prioritizes it for similar issues in the future. If the test fails, the system reminds the user "This solution is not feasible; please try another solution," projects the test results onto the interactive interface, and updates and trains the federated learning model for the output code fix based on the feedback. This addresses issues such as the lack of systematic deduction of fixes in complex error scenarios and the unintuitive assessment of the impact of code changes, leading to secondary errors. It allows developers to better fix code and improves programming efficiency.

[0162] Step 8: After determining a suitable correction scheme, the system renders the correction steps into a 3D dynamic map and uses a ray tracing algorithm to project the animation onto the user's 3D interactive interface. The user can choose to have the system automatically correct or manually correct according to the projected correction steps, and the corresponding correction patch will be automatically submitted to the historical code correction scheme database for easy access by other team members or for future reference.

[0163] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0164] Based on the same inventive concept, this application also provides a cross-system level error code correction device for implementing the cross-system level error code correction method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more cross-system level error code correction device embodiments provided below can be found in the limitations of the cross-system level error code correction method described above, and will not be repeated here.

[0165] In one exemplary embodiment, such as Figure 5 As shown, a cross-system level error code correction device is provided, including: a data acquisition module 502, used to acquire the logical relationships between system levels, the logical dependency weights of each system level, and the overall code when system error information exists during software development;

[0166] The data processing module 504 is used to annotate the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels based on the system error information during the software development process, the logical dependency weight of each system level, and the overall code.

[0167] Projection module 506 is used to project the system level and system level code content associated with the system error information onto the three-dimensional interactive interface;

[0168] The data acquisition module 502 is also used to obtain the corresponding reference error cause and error code based on the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels;

[0169] The data processing module 504 is also used to obtain multiple correction schemes and corresponding reliability based on the reference error cause and error code;

[0170] The projection module 506 is also used to project multiple correction schemes and corresponding reliability onto a three-dimensional interactive interface, which is used by the user to simulate the code correction process in a virtual reality interface in the form of three-dimensional animation.

[0171] Test module 508 is used to respond to the target correction scheme selected by the user through the virtual reality device in the three-dimensional interactive interface, to perform sandbox testing on the patch file of the target correction scheme, and to obtain the test results.

[0172] The projection module 506 is also used to project the test results onto the three-dimensional interactive interface.

[0173] In one embodiment, the data acquisition module 502 is further configured to monitor the system's operational information and perform call chain tracing at each system level during the software development process; when error information is projected onto the 3D interactive interface, the eye-tracking module of the virtual reality device is used to acquire the user's pupil movement information; the data processing module 504 is further configured to match the interface element coordinates of the 3D interactive interface based on the user's pupil movement information and perform dynamic calibration; and determine the system error information currently focused on by the user's gaze based on the interface element coordinates of the 3D interactive interface.

[0174] In one embodiment, the data processing module 504 is further configured to input system error information into a pre-trained text classification model, pre-classify the system error information according to the error type and urgency; construct an abstract syntax tree based on the current code; use the classified system error information and the abstract syntax tree of the current code as input to a graph neural network, perform multi-hop inference along the dependency relationship of the current code, and output the logical relationship between each system level, the logical dependency weight of each system level, and the overall code.

[0175] In one embodiment, the data acquisition module 502 is further configured to acquire front-end error information, related system error information, back-end service logs, and database content during the software development process; the data processing module 504 is further configured to input the system hierarchy associated with the front-end error information, related system error information, back-end service logs, database content, and system error information, the code content of the system hierarchy, and the logical relationship between the system hierarchy into the multimodal fusion neural network model to calculate the reference error cause and error code for the main error.

[0176] In one embodiment, in response to a user selecting a target correction scheme via a virtual reality device and using gesture or voice commands, the data acquisition module 502 is further configured to acquire a patch file for the target correction scheme.

[0177] Test module 508 is specifically used to build an isolated environment using Docker containers, perform sandbox testing on patch files, and obtain test results. If the test results show that the test is passed, the target fix is ​​marked as the recommended solution and will be given priority in subsequent recommendations.

[0178] The projection module 506 is also used to project the test results onto the three-dimensional interactive interface; the data acquisition module 502 is also used to acquire the feedback results returned by the user based on the test results; and the data processing module 504 is also used to update and train the multimodal fusion neural network model according to the feedback results.

[0179] In one embodiment, the projection module 506 is further configured to render the correction steps of the target correction scheme into a three-dimensional dynamic map when the user selects the target correction scheme through the virtual reality device, and to project the three-dimensional dynamic map animation onto the three-dimensional interactive interface using a ray tracing algorithm; the data processing module 504 is further configured to modify the target correction scheme in response to the user's modification operation based on the target correction scheme.

[0180] The modules in the aforementioned cross-system level error code correction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0181] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores system error information, reference error causes, and error codes. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a cross-system level error code correction method.

[0182] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0183] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0184] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0185] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described above.

[0186] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0187] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0188] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0189] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for correcting error codes across system levels, characterized in that, The method includes: In the event of system error messages during software development, obtain the logical relationships between each system level, the logical dependency weights of each system level, and the overall code. Based on the system error messages during the software development process, and based on the logical relationships between system levels, the logical dependency weights of each system level, and the overall code, the system levels associated with the system error messages, the code content of the system levels, and the logical relationships between the system levels are marked, and the system levels associated with the system error messages and the code content of the system levels are projected onto the 3D interactive interface. Based on the system level associated with the system error message, the code content of the system level, and the logical relationship between the system levels, obtain the corresponding reference error reason and error code; Based on the reference error cause and error code, multiple correction schemes and corresponding reliability are obtained, and the multiple correction schemes and corresponding reliability are projected onto a three-dimensional interactive interface. The three-dimensional interactive interface is used by the user to simulate the code correction process in a virtual reality interface in the form of three-dimensional animation. In response to the target correction scheme selected by the user through the virtual reality device on the three-dimensional interactive interface, the patch file of the target correction scheme is subjected to sandbox testing to obtain the test results, and the test results are projected onto the three-dimensional interactive interface.

2. The method according to claim 1, characterized in that, The process of obtaining the system error information includes: During software development, monitor the system's operational information and trace the call chain at each system level; When error messages are projected onto a 3D interactive interface, the eye-tracking module of the virtual reality device is used to obtain the user's pupil movement information. Based on the user's pupil movement information, the coordinates of the interface elements of the 3D interactive interface are matched and dynamically calibrated. Based on the coordinates of the interface elements in the 3D interactive interface, determine the system error message that the user's current gaze is focused on.

3. The method according to claim 1, characterized in that, The step of obtaining the logical relationships between system levels, the logical dependency weights of each system level, and the overall code based on the system error information includes: The system error messages are input into a pre-trained text classification model, and the system error messages are pre-classified according to the error type and urgency. Construct an abstract syntax tree based on the current code; The system error messages after classification and the abstract syntax tree of the current code are used as input to the graph neural network. Multi-hop reasoning is performed along the dependency relationship of the current code to output the logical relationship between each system level, the logical dependency weight of each system level and the overall code.

4. The method according to claim 1, characterized in that, The process of obtaining the corresponding reference error cause and error code based on the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels includes: Obtain front-end error messages, related system error messages, back-end service logs, and database content during the software development process; The system layers associated with front-end error messages, related system error messages, back-end service logs, database content, system error messages, code content at the system layers, and logical relationships between system layers are input into a multimodal fusion neural network model to calculate the reference error causes and error codes for the main errors.

5. The method according to claim 4, characterized in that, The step of responding to the target correction scheme selected by the user through the virtual reality device on the 3D interactive interface, performing sandbox testing on the patch file of the target correction scheme, obtaining test results, and projecting the test results onto the 3D interactive interface includes: In response to the target correction scheme selected by the user through a virtual reality device and using gesture or voice commands, obtain the patch file for the target correction scheme; An isolated environment is built using Docker containers to conduct sandbox testing on the patch files and obtain test results. If the test results show that the test is passed, the target modification scheme is marked as the recommended scheme and will be given priority in subsequent recommendations. The test results are projected onto a 3D interactive interface to obtain feedback from the user based on the test results. The multimodal fusion neural network model is then updated and trained based on the feedback results.

6. The method according to claim 1, characterized in that, Before sandbox testing the patch files for the target fix, the following is also included: When a user selects the target correction scheme through a virtual reality device, the correction steps of the target correction scheme are rendered into a three-dimensional dynamic map, and the three-dimensional dynamic map animation is projected onto the three-dimensional interactive interface using a ray tracing algorithm. In response to the user's modification operation based on the target modification scheme, the target modification scheme is modified.

7. A cross-system level error code correction device, characterized in that, The device includes: The data acquisition module is used to obtain the logical relationships between system levels, the logical dependency weights of each system level, and the overall code when system error messages occur during software development. The data processing module is used to annotate the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels based on the system error information during the software development process, the logical dependency weight of each system level, and the overall code. The projection module is used to project the system-level information and code content associated with system error messages onto the 3D interactive interface. The data acquisition module is also used to obtain the corresponding reference error cause and error code based on the system level associated with the system error information, the code content of the system level, and the logical relationship between the system levels; The data processing module is also used to obtain multiple correction schemes and corresponding reliability based on the reference error cause and error code; The projection module is also used to project multiple correction schemes and corresponding reliability onto a three-dimensional interactive interface, which is used by users to simulate the code correction process in the form of three-dimensional animation in a virtual reality interface. The testing module is used to respond to the target correction scheme selected by the user through the virtual reality device in the 3D interactive interface, to perform sandbox testing on the patch file of the target correction scheme, and obtain the test results; The projection module is also used to project the test results onto the three-dimensional interactive interface.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images