A method for a digital twin XR low-code development platform

Abstract

The present application provides a method for a digital twin XR low-code development platform, including: using a static code analysis tool to perform a preliminary scan of the code modules of the digital twin XR low-code development platform, obtain the dependencies between modules, and generate a dependency map; based on the obtained code module dependency map, optimize the SaaS layer module loading and cache processing flow to ensure that the modules are loaded in the order of dependency; according to the SaaS layer module update record and volatility analysis, implement a cache invalidation strategy based on content changes and time benchmarks, and adjust the cache refresh interval in combination with business impact assessment; according to the frequency of use of the SaaS layer module and the business importance score, perform cache capacity assessment and space allocation optimization to achieve resource configuration and optimize cache allocation.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for a digital twin XR low-code development platform. Background Art

[0002] With the advent of the Industrial 4.0 era, digital twin technology, as a bridge connecting the physical world and the digital world, has become the core of intelligent manufacturing and industrial automation. The core of digital twin technology is to create a virtual copy of a system, process or equipment so that it can be simulated and analyzed in the digital domain. However, in the process of implementing this technology, especially in the low-code development platform integrated with extended reality XR technology, there are multiple technical challenges. First, although the low-code development environment simplifies the programming process, lowers the technical threshold, and enables non-professionals to participate in complex projects, it also brings challenges in code management and performance optimization. In this environment, the complex dependencies and dynamics between code modules require the platform to flexibly handle various situations to maintain efficient and stable operation. Secondly, with the complexity of application scenarios, the demand for modular design is growing. In the digital twin environment, different modules may need to process different types of data and logic, such as real-time data processing, equipment status monitoring, user interaction interface, etc. The intricate dependencies between these modules make system architecture and performance optimization a challenge. In addition, real-time is one of the key requirements of digital twin systems. Especially in applications that integrate XR technology, the fluency of user interaction experience directly affects the effectiveness of the entire system. However, due to resource constraints and technical limitations, how to maintain efficient real-time response in a low-code environment has become an important issue. Cache management is also a key issue in digital twin systems. In a dynamically changing system, how to effectively manage the cache to ensure the immediacy and accuracy of data while avoiding unnecessary waste of resources is a technical problem that needs to be solved. Especially in scenarios involving frequent data updates and high interactivity, traditional caching strategies may not meet the needs. Finally, user behavior prediction and analysis are also particularly important in low-code development platforms. The system needs to be able to understand and predict user behavior to provide more personalized and responsive services. However, the unpredictability and diversity of user behavior make this task complicated. Therefore, the digital twin XR low-code development platform faces multiple challenges of code management complexity, modular design challenges, real-time requirements, cache management difficulties, and user behavior analysis. Solving these problems is crucial to improving the performance of digital twin systems, expanding their scope of application, and improving user satisfaction. Summary of the invention

[0003] The present invention provides a method for a digital twin XR low-code development platform, which mainly includes:

[0004] Use static code analysis tools to perform a preliminary scan of the code modules of the digital twin XR low-code development platform, obtain the dependencies between modules, and generate a dependency graph;

[0005] Based on the obtained code module dependency graph, the SaaS layer module loading and cache processing process is optimized to ensure that the modules are loaded in the order of dependency;

[0006] Use the publish / subscribe model to establish an event-driven mechanism to monitor changes within the SaaS layer module. Once a key change is detected, the dynamic cache update mechanism is immediately triggered to refresh the cache.

[0007] Based on the user interaction behavior data of the digital twin XR low-code development platform, a user interaction behavior prediction model is built to predict the user's operation behavior. A differentiated caching strategy is implemented for SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold, and they are pre-loaded and cached.

[0008] According to the device performance monitoring and maintenance logs of the digital twin XR low-code development platform, implement environmental adjustments and software and hardware update strategies, and combine network status evaluation to determine the cache update requirements of the device control module and optimize cache efficiency;

[0009] Implement cache invalidation strategies based on content changes and time benchmarks based on SaaS layer module update records and volatility analysis, and adjust cache refresh intervals based on business impact assessments;

[0010] Based on the usage frequency and business importance score of the SaaS layer modules, cache capacity evaluation and space allocation optimization are carried out to implement resource configuration and optimize cache allocation.

[0011] Preferably, the static code analysis tool is used to perform a preliminary scan of the code modules of the digital twin XR low-code development platform, obtain the dependencies between modules, and generate a dependency graph, including:

[0012] Use the static code analysis tool SonarQube to scan the code modules of the digital twin XR low-code development platform, identify the properties of the structural elements in the code modules, generate a unique identifier for each code module, and record the location and function of each code module. The properties of the structural elements include method signatures, variable declarations, and access permissions; According to the module identifier, the modules are classified into basic environment layer modules, cloud rendering layer modules, low-code development editor modules, data asset library modules, and SaaS management layer modules. The module identifier includes the properties of the structural elements; Through the interfaces and methods exposed by the modules, the dependency types between modules are judged, and compile-time dependencies, runtime dependencies, or test dependencies are identified; According to the dependency direction, the source and end point of the dependency relationship are determined, and the code module dependency graph is constructed. The system uses a control flow chart to display the connections and dependency paths between modules; for each dependency relationship, the dependency weight is determined by the dependency frequency; by constructing a data flow diagram, the data transmission between modules is mapped, including the source of the monitoring data, the data processing of each module, and the data destination; by creating a control flow chart, the execution of the program is determined. Each node in the chart represents an operation in the program, and the edge represents the path when the program is executed, so as to obtain the way and path of data and control information flowing between different modules; extract the module change history, determine the reuse level of each module, and identify modules whose reuse frequency is greater than the preset frequency threshold; by checking the errors and warning information in the code, as well as the code complexity and maintainability, the module code quality is determined, and the modules with code quality lower than the preset quality threshold are marked with priority.

[0013] Preferably, the SaaS layer module loading and cache processing flow is optimized based on the obtained code module dependency graph to ensure that the modules are loaded in the order of dependency, including:

[0014] Based on the obtained code module dependency graph, the module name, version and file path are specified for the SaaS layer module, and the direction of each edge is determined to identify the dependency relationship between modules. The SaaS layer modules include company management module, project management module, permission management module and asset management module. Each edge in the graph is weighted according to the dependency weight. If there are edges in the graph that are not weighted, weights are assigned according to the estimated cost, time or resource consumption of module loading. The code module dependency graph is processed using a topological sorting algorithm to obtain a sequential queue for module loading, where each module is loaded after the modules it depends on. Based on the sequential queue obtained by topological sorting, it is determined whether the current module is already in a loading state or has been loaded. If it has been processed, it is skipped to avoid repeated loading. Load; Get the current module to be loaded and check whether the module exists in the cache. If the module exists in the cache, load it directly from the cache, otherwise continue the loading process; Before preparing to load the module, perform circular dependency detection. If a circular dependency is detected, take measures to break the cycle, including version downgrade and module replacement; Register a callback function for the loading of each module through the event callback mechanism, and perform custom processing after the module is loaded; During module loading, use a concurrency control mechanism to manage module loading requests, load multiple modules at the same time, and maintain the consistency of dependencies and loading status; Establish an error recognition mechanism to identify and classify different types of loading errors, including timeout errors and file missing errors, and pre-set corresponding recovery strategies for common error types.

[0015] Preferably, the publish / subscribe mode is used to establish an event-driven mechanism, and monitors changes in the SaaS layer module. Once a key change is detected, a dynamic cache update mechanism is immediately triggered to refresh the cache, including:

[0016] The event producer continuously monitors the status and activities within the SaaS layer module. If predefined key changes are detected, the status change information of the event source is obtained, and a structured event message is created and published. The event message includes user ID, activity type, data update, and timestamp. The event producer is a designated monitoring system. Key changes include data update and status change. Configure the message channel, including setting the channel's transmission protocol, message format, and network configuration, and provide event consumers with the function of filtering and subscribing to events. Filter according to event type. The event consumer is a cache management system. On the event consumer side, subscribe to the specified event type related to the cache content. If an event of the specified event type is received, the cache update mechanism is triggered. Set events during peak hours. The queue's buffering strategy includes setting the buffer size, adjusting the event processing priority, and ensuring the order of event processing. The peak period is the period when the number of events is greater than the preset threshold. Perform cache refresh operations on received events and add fault-tolerant mechanisms to deal with potential abnormal situations. Use the cache management mechanism to synchronize the data of all relevant cache instances, and adjust the cache expiration strategy and refresh strategy according to the timeliness and importance of the event. Configure the monitoring system to monitor the release, transmission and processing of events, and record event processing information through the logging mechanism. Establish service quality control, including the reliability of message delivery and the order of event processing. Implement security measures, including authorization and authentication of event producers and consumers and data encryption during message transmission.

[0017] Preferably, the user interaction behavior prediction model is constructed based on the user interaction behavior data of the digital twin XR low-code development platform to predict the user's operation behavior, and a differentiated caching strategy is implemented for the SaaS layer modules with a responsiveness requirement higher than a preset responsiveness threshold, and pre-loaded and cached, including:

[0018] Real-time monitoring of digital twin XR low-code development platform log records, recording user interaction behaviors on applications or websites, including click events, page browsing, search queries, and form submissions; obtaining device information, including device type and operating system, and using different data processing processes based on the user's device information; recording environmental characteristics when user interaction behaviors occur, including timestamps and geographic location information, and associating environmental characteristics with user interaction behavior data; identifying whether there are differences in user activity patterns on weekdays and weekends by counting user interaction frequencies in each time period; extracting user interaction frequencies and user classification information from historical data, applying the K-means clustering algorithm for model training, and dividing users into different groups pattern groups; according to the user interaction behavior data, the decision tree algorithm is used for model training, and a user interaction behavior prediction model is constructed to predict the user's future interaction behavior, including whether the user will click a specified button, visit a specified page, or how long the user will spend on the page; based on the user interaction behavior prediction model, the user's interaction frequency with different SaaS layer modules is predicted, and the SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold are determined; according to the predicted results of user interaction behavior, differentiated caching strategies are formulated to preload and adjust the cache update frequency of SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold; according to the user's behavior pattern and interaction frequency, the priority and update cycle of the cached content are automatically adjusted.

[0019] Preferably, the device performance monitoring and maintenance log of the digital twin XR low-code development platform is used to implement environment adjustment and software and hardware update strategies, and combined with network status evaluation, the cache update requirements of the device control module are judged to optimize cache efficiency, including:

[0020] The real-time performance monitoring system is used to obtain the real-time operating status and performance indicators of the devices of the digital twin XR low-code development platform, including online status, errors, warnings, processing speed, response time, CPU and memory usage status, and disk I / O and network I / O operations; if the equipment operating efficiency is lower than the preset efficiency threshold or abnormal performance indicators are monitored, it is determined whether environmental factors have an impact on the equipment performance and whether the equipment settings or environmental conditions need to be adjusted. Environmental factors include temperature and humidity, vibration frequency, and noise intensity; based on the equipment life cycle information, for each device, an equipment maintenance log is established, including past maintenance activities and parts replacement records. The life cycle information includes manufacturing date, expected service life, and historical maintenance records; based on the type of equipment, frequency of use, and maintenance guidelines provided by the manufacturer, the maintenance and replacement cycle of the equipment is determined. If the difference between the maintenance time and the current time is less than the preset difference threshold, the physical components, software system, and cache components of the equipment are inspected to identify signs of wear, functional abnormalities, or performance declines; if the cache efficiency declines due to equipment wear or performance degradation, optimize the cache, including increasing cache capacity, replacing aging storage media, or optimizing cache management strategies; obtain the software and firmware version information currently used by the device based on the device's log and event recording functions; if a new version is monitored, determine whether there is an update to optimize the cache based on the version update content and whether the cache needs to be updated; obtain the security vulnerability status of the device; if there is a security risk or a security patch is required, determine whether the cache needs to be updated; adjust the trigger logic, set preset thresholds and conditions, and if performance degradation or abnormal events are detected, obtain the cache size, usage rate, hit rate and failure rate data based on cache usage; if the cache failure rate is higher than the preset failure rate threshold, determine that the cache efficiency is insufficient and needs to be updated; obtain network connection quality data through network status monitoring, including latency, packet loss rate, and bandwidth usage rate, to determine whether network problems lead to cache update requirements and whether cache efficiency can be improved by alleviating network conditions;

[0021] It also includes: implementing cache adjustment coefficient calculation and application based on user interaction behavior and device performance monitoring data, while continuously optimizing cache size, update frequency and elimination strategy.

[0022] The cache adjustment coefficient calculation and application are implemented based on user interaction behavior and device performance monitoring data, while the cache size, update frequency and elimination strategy are continuously optimized, specifically including:

[0023] Use ApacheLog4j, a logging tool for the digital twin XR low-code development platform, to obtain real-time user interaction behaviors, including click depth, page dwell time, and bounce rate. Obtain device performance data, including CPU temperature, memory paging, and network latency. Perform stress tests regularly to evaluate the performance of the device under different load conditions. Based on user interaction behavior data, use the user interaction behavior prediction model to predict the user interaction frequency P user Based on the acquired equipment performance data, a linear regression algorithm is used to train the model and determine the equipment performance parameter P device , evaluate the current workload and performance status of the device. Define the cache adjustment coefficient formula based on user interaction frequency and device performance parameters,

[0024] C adjust =δ×P user +η×P device , where C adjust is the cache adjustment coefficient, δ and η are weight factors, balancing the impact of user behavior and device performance on cache adjustment. According to the defined formula C adjust Calculate the cache adjustment coefficient in real time. adjust Dynamically adjust cache size, cache update frequency, and cache elimination strategy. Continuously monitor user behavior and device performance, adjust δ and η based on feedback, and optimize cache adjustment strategy.

[0025] Preferably, the cache invalidation strategy based on content change and time reference is implemented according to the SaaS layer module update record and volatility analysis, and the cache refresh interval is adjusted in combination with the business impact assessment, including:

[0026] The update records of the SaaS layer modules are obtained through the monitoring system, including the time point and frequency of each data change; according to the data update frequency, the ARIMA algorithm is used to train the model and predict the update frequency and trend of the SaaS layer modules; the standard deviation or coefficient of variation of the data is calculated to determine the volatility of the SaaS layer modules, and identify the SaaS layer modules with volatility higher than the preset volatility threshold; for SaaS layer modules with volatility higher than the preset volatility threshold, a cache invalidation strategy based on content changes is adopted, and the cache invalidation strategy based on content changes is to update the cache immediately when the data changes; for SaaS layer modules with update frequency stability higher than the preset stability threshold, a time-based cache invalidation strategy is adopted, and the time-based cache The invalidation strategy is to update the cache regularly; according to the update trend prediction results of the SaaS layer module, set the cache refresh interval for each SaaS layer module to match the update mode and volatility of the SaaS layer module; adjust the cache refresh interval according to business needs and service level agreements; evaluate the resource limitations of the digital twin XR low-code development platform, including CPU, memory, and I / O capacity, determine the impact of resource limitations on the cache refresh frequency, and adjust the refresh strategy accordingly; according to the business decisions, operational efficiency, and customer experience of the SaaS layer module, use the support vector machine algorithm to train the model, determine the impact of the SaaS layer module on the business process, and shorten the cache refresh interval for SaaS layer modules whose impact is higher than the preset impact threshold;

[0027] It also includes: implementing real-time data dependency monitoring and threshold-based dynamic adjustment of cache refresh cycles based on the update correlation and business impact scores between SaaS layer modules, while optimizing the resource allocation and load balancing strategy of the digital twin XR low-code development platform.

[0028] According to the update correlation and business impact score between SaaS layer modules, real-time data dependency monitoring and threshold-based cache refresh cycle dynamic adjustment are implemented, while optimizing the resource allocation and load balancing strategy of the digital twin XR low-code development platform, specifically including:

[0029] According to the code module dependency graph, identify the dependencies between SaaS layer modules. Define the data dependency scoring formula based on the business impact of SaaS layer modules and the update correlation between modules. Among them, DDS i represents the data dependency score of the i-th SaaS layer module, Cov(M i ,M j ) represents the covariate between the i-th and j-th SaaS layer modules, measuring the update correlation between the two modules. Imp(M j) represents the business impact score of the jth SaaS layer module, which is quantitatively evaluated based on the importance of data to the business process. The DDS score is calculated for each SaaS layer module according to the data dependency scoring formula to quantify the dependencies between different SaaS layer modules. A real-time monitoring mechanism is established to update the DDS scores of SaaS layer modules in real time and set a threshold alarm mechanism. According to the DDS score, the cache refresh cycle of the SaaS layer module is dynamically adjusted. If the DDS score of a module exceeds the preset first scoring threshold, the cache refresh cycle of the module is shortened. If the DDS score of a module is lower than the preset second scoring threshold, the cache time of the module is increased. Combine the DDS score and the resource usage of the digital twin XR low-code development platform to optimize resource allocation and load balancing. Regularly evaluate the accuracy of the DDS formula and the effectiveness of the caching strategy, and adjust and optimize based on feedback data.

[0030] Preferably, the cache capacity evaluation and space allocation optimization are performed according to the usage frequency and business importance score of the SaaS layer module to achieve resource configuration and optimize cache allocation, including:

[0031] Use a monitoring system to monitor the frequency of use of SaaS layer modules in real time, including the number of times the module is accessed and the duration of the access; use the support vector machine algorithm to determine the business importance score of the SaaS layer module based on the type of data item, its purpose, and the degree of impact on the business process; evaluate the current cache usage, including the used space and the remaining space, based on the cache capacity limit, to determine whether cache adjustment is required; if the difference between the cache capacity and the cache upper limit is less than the preset second difference threshold, determine the data to be eliminated based on the cache elimination strategy, which is to eliminate the pages with the least number of uses; evaluate the size of the data item to determine the proportion of space occupied by the data item in the cache , optimize cache space allocation; set an expiration time for each cache item based on the update frequency of data items and business needs; use a weighted calculation method to determine the priority of each data item based on the frequency of use, business importance, data size and expiration time information; use a concurrency control mechanism to prevent multiple concurrent requests from competing for the same data item; in a distributed cache system, apply a synchronization mechanism to ensure data consistency between different nodes when adjusting cache priority; use the ARIMA algorithm to train the model based on historical access data and business patterns, predict future access trends, and adjust the priority of data in the cache to achieve resource configuration and optimize cache allocation.

[0032] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0033] The present invention provides a digital twin XR low-code development platform. By accurately managing the dependencies between code modules, the platform can handle module loading more efficiently and ensure the stability and efficiency of system operation. In addition, by responding to key changes within the module and updating the cache in real time, the risk of data delay and outdated information is greatly reduced, while optimizing resource utilization efficiency. The method in the present invention also includes an intelligent caching strategy based on user interaction and device performance, which not only improves the system response speed, but also optimizes the server load, thereby improving the user experience. By evaluating the frequency of module use and business importance, the platform can allocate resources more reasonably and optimize cache allocation to ensure the efficient operation of key business modules. In addition, the present invention also proposes a flexible cache management strategy, which enables the platform to better cope with business fluctuations and further improve operating efficiency and user experience. Overall, these strategies not only enhance the business adaptability of the platform, but also reduce resource consumption, bringing significant business benefits and user satisfaction to the digital twin XR low-code development platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a flowchart of a method for a digital twin XR low-code development platform of the present invention.

[0035] Figure 2 A schematic diagram of a method for a digital twin XR low-code development platform of the present invention.

[0036] Figure 3 This is another schematic diagram of a method for a digital twin XR low-code development platform according to the present invention. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0038] like Figure 1-3 In this embodiment, a method for a digital twin XR low-code development platform may specifically include:

[0039] S101. Use static code analysis tools to perform a preliminary scan of the code modules of the digital twin XR low-code development platform, obtain the dependencies between modules, and generate a dependency graph.

[0040] The static code analysis tool SonarQube is used to scan the code modules of the digital twin XR low-code development platform, identify the properties of the structural elements in the code modules, generate a unique identifier for each code module, and record the location and function of each code module. The properties of the structural elements include method signatures, variable declarations, and access rights. According to the module identifier, the modules are classified into basic environment layer modules, cloud rendering layer modules, low-code development editor modules, data asset library modules, and SaaS management layer modules. The module identifier includes the properties of the structural elements. Through the interfaces and methods exposed by the modules, the dependency types between modules are judged, and compile-time dependencies, runtime dependencies, or test dependencies are identified. According to the dependency direction, the source and end point of the dependency relationship are determined, and a code module dependency graph is constructed to display the connection and dependency path between modules. For each dependency relationship, the dependency weight is determined by the dependency frequency. By constructing a data flow graph, the data transmission between modules is mapped, including the source of monitoring data, the data processing of each module, and the data end point. By creating a control flow graph, the execution of the program is determined. Each node in the graph represents an operation in the program, and the edge represents the path when the program is executed. The way and path of data and control information flow between different modules are obtained. Extract the module change history, determine the reuse level of each module, and identify modules whose reuse frequency is greater than the preset frequency threshold. Determine the module code quality by checking the error and warning information in the code, as well as the code complexity and maintainability, and mark the priority of modules whose code quality is lower than the preset quality threshold.

[0041] For example, the XR low-code development platform adopts a layered design, including a basic environment layer, a cloud rendering layer, a low-code development editor, a data asset library, and a SaaS management layer. It supports development model scene material management, three-dimensional scene content production, data information visualization, simulation interaction, and intelligent editing, and can realize component-based low-code visualization of environmental modeling, physical simulation, and fast rendering. The cloud rendering layer module includes code modules such as XR streaming service, scene driving engine, physical driving engine, dynamic interaction engine, and data engine; the low-code development editor module includes code modules such as project templates, scene construction, light rendering, sound effect configuration, visualization charts, scene library applications, weather environment, physical characteristics, special effects configuration, model library applications, scene preview, animation configuration, and content publishing; the data asset library module includes code modules such as model library, scene library, project library, material library, special effects library, animation library, and UI library; the SaaS layer module includes code modules such as company management, project management, permission management, and asset management; the basic environment layer module includes code modules such as computing resources, cloud rendering resources, storage resources, network resources, and security protection resources; SonarQube is used to scan the above modules and identify the structural elements in each module, such as the method initEnvironment() in module W1, which has public access rights and generates a unique identifier for each module, and the identifier of module W1 is A1B2C3. According to the structural element attributes of the modules, they are classified, and module W1 is classified as a basic environment layer module due to its basic functions. Analyze the interfaces and methods between modules, such as the cloud rendering layer module W2 may depend on the base environment layer module W1 at runtime, and determine the types of these dependencies, such as compile-time, runtime, or test dependencies. Determine the source and end point of the dependency, such as module W2 depends on module W1 as the source and the end point, and build a graph to show these relationships, such as a graph with arrows pointing from W1 to W2. For the dependency between module W1 and module W2, if there are 10 references in the entire project, you can assign a corresponding weight to this dependency, such as a weight of 10. Map how data flows from the low-code development editor module W3 to the data asset library module W4, and create a control flow graph to show the program execution path, such as from an edit operation in module W3 to a data storage operation in module W4. Extract the change history of module W4. If it has been reused 30 times in the past month, and the preset frequency threshold is 20 times per month, it is considered to have a high reuse level. SonarQube checks the code of module W5 and finds 5 errors and 20 warnings, with a code complexity score of 7 / 10 and a maintainability of medium. If the preset code quality threshold is a maximum of 3 errors and 15 warnings, the code quality of module W6 is below the standard and needs to be processed first.

[0042] S102. Based on the obtained code module dependency graph, the SaaS layer module loading and cache processing flow is optimized to ensure that the modules are loaded in the order of dependency.

[0043] Based on the obtained code module dependency graph, the module name, version and file path are specified for the SaaS layer module, and the direction is determined for each edge to identify the dependency relationship between modules. The SaaS layer modules include company management module, project management module, permission management module and asset management module. Weights are assigned to each edge in the graph according to the dependency weight. If there are edges in the graph that are not assigned weights, weights are assigned according to the estimated cost, time or resource consumption of module loading. The code module dependency graph is processed using a topological sorting algorithm to obtain a sequential queue for module loading. Each module is loaded after its dependent modules. According to the sequential queue obtained by topological sorting, it is determined whether the current module is already in the loading state or has been loaded. If it has been processed, it is skipped to avoid repeated loading. The current module to be loaded is obtained, and the module is checked whether it exists in the cache. If the module exists in the cache, it is directly loaded from the cache, otherwise the loading process continues. Before preparing to load the module, a circular dependency detection is performed. If a circular dependency is detected, measures are taken to break the cycle, including version downgrade and module replacement. A callback function is registered for the loading of each module through the event callback mechanism, and custom processing is performed after the module is loaded. During module loading, a concurrent control mechanism is used to manage module loading requests, load multiple modules simultaneously, and maintain the consistency of dependencies and loading status. An error identification mechanism is established to identify and classify different types of loading errors, including timeout errors and file missing errors, and corresponding recovery strategies are pre-set for common error types.

[0044] For example, based on the obtained SaaS layer code module dependency graph, the dependency relationship between module A and module B is obtained. The SaaS layer modules include company management module, project management module, permission management module, and asset management module. The module name of module A is company management module, the version is 1.2.0, and the file path is / modules / UserAuthentication.js; the module name of module B is project management module, the version is 2.0.1, and the file path is / modules / UserProfile.js. In the code module dependency graph, there is a directed edge, indicating that module A depends on module B. According to the dependency weight, a weight of 0.8 is assigned to this edge, indicating that module A has a high dependency on module B. The code module dependency graph is processed using a topological sorting algorithm to obtain a loading order queue, module B, module A, and modules are loaded in order according to the loading order queue. First, check the cache. If module B exists in the cache, it is loaded directly from the cache. If it exists in the cache, the process of loading module B is skipped. Then, prepare to load module A. Before loading, perform a circular dependency detection. No circular dependency is detected, so continue to load module A. During loading, a concurrency control mechanism is used to allow multiple modules to be loaded simultaneously to improve efficiency. If a timeout error occurs when loading module A, it is identified as a timeout error type. According to the pre-set recovery strategy, module A is retried to load once. If it times out again, the backup strategy is triggered to downgrade to the previous version 1.1.0 and try to load it. Finally, after the module is loaded, the callback function is registered through the event callback mechanism to perform additional custom processing, such as initializing the module or updating the interface.

[0045] S103. An event-driven mechanism is established using a publish / subscribe model to monitor changes within the SaaS layer module. Once a key change is detected, a dynamic cache update mechanism is immediately triggered to refresh the cache.

[0046] The event producer continuously monitors the status and activities inside the SaaS layer module. If a predefined key change is detected, the status change information of the event source is obtained, and a structured event message is created and published. The event message includes user ID, activity type, data update, and timestamp. The event producer is the specified monitoring system, and the key changes include data update and status change. Configure the message channel, including setting the channel's transmission protocol, message format, and network configuration, and provide event consumers with the function of filtering and subscribing to events. Filter according to event type. The event consumer is the cache management system. On the event consumer side, subscribe to the specified event type related to the cache content. If an event of the specified event type is received, the cache update mechanism is triggered. Set the buffer strategy of the event queue during peak hours, including setting the buffer size, adjusting the event processing priority, and ensuring the order of event processing. The peak hour is the period when the number of events is greater than the preset number threshold. Perform cache refresh operations on the received events, and add a fault-tolerant mechanism to deal with potential abnormal situations. Use the cache management mechanism to synchronize the data of all relevant cache instances, and adjust the cache expiration strategy and refresh strategy according to the timeliness and importance of the event. Configure a monitoring system to monitor the publishing, transmission, and processing of events, and record information about event processing through a logging mechanism. Establish service quality controls, including the reliability of message delivery and the order of event processing. Implement security measures, including authorization and authentication of event producers and consumers, and data encryption during message transmission.

[0047] For example, the event producer is the internal monitoring system of the online collaboration platform. The event producer continuously monitors the status and activities within the SaaS layer module. When a user logs in or updates their profile, the internal monitoring system detects these key changes. If a user changes their profile, an event message is created containing the user ID, activity type, profile update, timestamp, and related content. The message channel is configured to use the WebSocket protocol, allowing real-time two-way communication. Messages are transmitted in JSON format. The network configuration ensures that data is securely transmitted within the digital twin XR low-code development platform. In addition, the cache management system subscribes to specific event types, such as user profile updates, as an event consumer. When users update their profiles, the cache management system receives the relevant events and triggers a script to update the cache of the user interface to ensure that the information seen by other collaborative users is the latest. During times when users are highly active, such as mornings and afternoons on weekdays, the buffer size of the event queue is adjusted to a higher value, such as 1,000 events, to handle a large number of user login and profile update events. Whenever a user profile update event is processed, the cache of all related interfaces is refreshed. If an abnormal situation is encountered, such as a cache service timeout, a retry is performed and an error message is recorded after failure. The cache management mechanism is used to synchronize the data of all relevant cache instances, and the cache expiration policy is adjusted according to the urgency of the event. For user login events, the cache life cycle may be shorter to ensure security. The internal monitoring system records the publishing, transmission, and processing details of all event messages, and records the processing time and results of each event in the log. The reliability of message delivery is ensured by setting the persistence and confirmation mechanism of the message queue. At the same time, queue management ensures that events are processed in the order in which they are generated. In order to protect data security, event producers and consumers need to pass SSL / TLS authentication, and all data in transit is encrypted to prevent data leakage.

[0048] S104. Based on the user interaction behavior data of the digital twin XR low-code development platform, a user interaction behavior prediction model is constructed to predict the user's operation behavior, and a differentiated caching strategy is implemented for SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold, which are pre-loaded and cached.

[0049] Monitor the log records of the digital twin XR low-code development platform in real time, record the user's interactive behavior on the application or website, including click events, page browsing, search queries, and form submissions; obtain device information, including device type and operating system, and adopt different data processing processes according to the user's device information; record the environmental characteristics when the user's interactive behavior occurs, including timestamp and geographic location information, and associate the environmental characteristics with the user's interactive behavior data. By counting the user interaction frequency in each time period, identify whether the user's active mode on weekdays and weekends is different. Extract the user's interaction frequency and user classification information from historical data, apply the K-means clustering algorithm for model training, and divide users into different behavior pattern groups. According to the user interaction behavior data, use the decision tree algorithm for model training, build a user interaction behavior prediction model, and predict the user's future interaction behavior, including predicting whether the user will click a specified button, visit a specified page, or spend time on the page. Based on the user interaction behavior prediction model, predict the user's interaction frequency with different SaaS layer modules, and determine the SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold. Based on the predicted results of user interaction behavior, differentiated cache strategies are formulated to adjust the preloading and cache update frequency of SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold. Based on the user's behavior pattern and interaction frequency, the priority and update cycle of cache content are automatically adjusted.

[0050] For example, real-time monitoring of the digital twin XR low-code development platform end log records, recording the user's interactive behavior on the application or website, including click events, page browsing, search queries, and form submissions, and obtaining that user A clicked the H button at 14:05 on April 15, 2023. Obtain device information, including device type and operating system. Different data processing processes are used according to the user's device information. User B uses an iPhone 12 running iOS 11, while user C uses a Samsung Galaxy S20 running Android 11. The data processing process may be different for different operating systems to ensure data accuracy and compatibility. Record the environmental characteristics when the user's interactive behavior occurs, including timestamps and geographic location information, and associate the environmental characteristics with the user's interactive behavior data. User D's interactive behavior at 15:30 on April 15, 2023 is recorded, and the longitude and latitude information of his geographic location is attached, with a latitude of 30522°N and a longitude of 112437°W. By counting the user interaction frequency in each time period, we can identify whether there is a difference in the user's active mode on weekdays and weekends. During the weekdays from Monday to Friday, user E interacts 25 times a day on average, while on weekends, Saturday and Sunday, the average number of user E interactions is reduced to 15 times a day. The user interaction frequency and user classification information are extracted from historical data, and the K-means clustering algorithm is used for model training to divide users into different behavior pattern groups. The 1,000 users who have undergone cluster analysis are divided into 3 groups according to their interaction habits, and each group represents a different interaction behavior pattern. According to the user interaction behavior data, the decision tree algorithm is used for model training to build a user interaction behavior prediction model to predict the user's future interaction behavior. The model predicts that user F has an 80% probability of clicking on page J in the next week. Based on the user interaction behavior prediction model, the user's interaction frequency with different SaaS layer modules is predicted, and the SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold are determined. The interaction data of user G shows that the company management module is checked every time the website is visited, so the responsiveness requirement of the company management module is set as a high priority. Based on the predicted results of user interaction behavior, differentiated cache strategies are formulated to preload and adjust the cache update frequency for SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold. Based on the prediction model, it can be decided to update the cache content of the company management module every two hours. Based on the user's behavior pattern and interaction frequency, the priority and update cycle of the cache content are automatically adjusted. If a user has only five access history per month, the content cache for that user may be set to update once a week to reduce the consumption of resources of the digital twin XR low-code development platform.

[0051] S105. Based on the device performance monitoring and maintenance logs of the digital twin XR low-code development platform, implement environmental adjustments and software and hardware update strategies, and combine network status assessment to determine the cache update requirements of the device control module and optimize cache efficiency.

[0052] The real-time performance monitoring system is used to obtain the real-time operating status and performance indicators of the devices of the digital twin XR low-code development platform, including online status, errors, warnings, processing speed, response time, CPU and memory usage status, and disk I / O and network I / O operations. If the operating efficiency of the device is lower than the preset efficiency threshold or the performance indicators are monitored to be abnormal, it is determined whether environmental factors have an impact on the performance of the device and whether the device settings or environmental conditions need to be adjusted. Environmental factors include temperature and humidity, vibration frequency, and noise intensity. Based on the device life cycle information, for each device, an equipment maintenance log is established, including past maintenance activities and component replacement records. The life cycle information includes the manufacturing date, expected service life, and historical maintenance records. According to the type of equipment, frequency of use, and maintenance guidelines provided by the manufacturer, the maintenance and replacement cycle of the equipment is determined. If the difference between the maintenance time and the current time is less than the preset difference threshold, the physical components, software system, and cache components of the device are checked to identify signs of wear, functional abnormalities, or performance degradation. If the cache efficiency is reduced due to equipment wear or performance degradation, the cache is optimized, including increasing the cache capacity, replacing aging storage media, or optimizing the cache management strategy. According to the device's log and event recording functions, obtain the software and firmware version information currently used by the device. If a new version is monitored, determine whether there is an update to optimize the cache based on the version update content and whether the cache needs to be updated. Obtain the security vulnerability status of the device. If there is a security risk or a security patch is required, determine whether the cache needs to be updated. Adjust the trigger logic and set preset thresholds and conditions. If performance degradation or abnormal events are detected, obtain the cache size, usage rate, hit rate, and failure rate data based on cache usage. If the cache failure rate is higher than the preset failure rate threshold, it is determined that the cache efficiency is insufficient and needs to be updated. Through network status monitoring, obtain network connection quality data, including latency, packet loss rate, and bandwidth usage, to determine whether network problems lead to cache update requirements and whether cache efficiency can be improved by alleviating network conditions.

[0053] For example, through the real-time performance monitoring system, the real-time operating status and performance indicators of the digital twin XR low-code development platform are obtained. The online status shows continuous online, but the processing speed is 150 requests per second, which is lower than the preset efficiency threshold of 200 requests. The CPU utilization rate is 85%, the memory utilization rate is 90%, and the disk I / O and network I / O operations show read and write delays of 20ms and 100ms respectively. These indicators indicate that the digital twin XR low-code development platform may be facing performance problems. In terms of environmental monitoring, the temperature was recorded as 35°C, which exceeded the normal operating temperature range of 20°C-30°C, and the vibration frequency and noise intensity were recorded as 50Hz and 75dB respectively, both within the range allowed by the equipment specifications. Therefore, it is judged that high temperature may be one of the environmental factors affecting the performance of the digital twin XR low-code development platform. According to the maintenance log of the digital twin XR low-code development platform, the manufacturing date of the digital twin XR low-code development platform is January 2018, and the expected service life is 5 years. The historical maintenance record shows that the digital twin XR low-code development platform was last maintained in June 2021, and the heat sink and memory bar were replaced. The maintenance guide provided by the manufacturer recommends a comprehensive maintenance every 2 years. Since the current year is 2023, 2 years have passed since the last maintenance, so it was decided to perform a maintenance inspection. During the physical inspection, it was found that the read and write heads of the hard disk showed signs of wear, resulting in a decrease in disk I / O performance. The cache management policy assessment found that the current cache capacity is 4GB, the hit rate is 70%, and the failure rate is 30%, which exceeds the preset failure rate threshold of 20%, suggesting that the cache capacity is insufficient and inefficient. The software version used by the digital twin XR low-code development platform is v5. The manufacturer released a new version v6 a week ago, which includes updates for cache management optimization. The security vulnerability status shows that the digital twin XR low-code development platform has known security risks and needs to apply the security patches included in the new version update. Network status monitoring shows that the current network latency is 50ms, the packet loss rate is 1%, and the bandwidth utilization rate is 75%, indicating that the network condition is acceptable and is not the main factor causing performance problems. In summary, choose to increase the cache capacity to 8GB, update the software to version v6 to optimize the cache management strategy, and apply security patches. In addition, schedule maintenance of the digital twin XR low-code development platform during cooler time periods to resolve overheating issues.

[0054] Based on user interaction behavior and device performance monitoring data, cache adjustment coefficient calculation and application are implemented, while continuously optimizing cache size, update frequency and elimination strategy.

[0055] Use ApacheLog4j, a logging tool for the digital twin XR low-code development platform, to obtain real-time user interaction behaviors, including click depth, page dwell time, and bounce rate. Obtain device performance data, including CPU temperature, memory paging, and network latency. Perform stress tests regularly to evaluate the performance of the device under different load conditions. Based on user interaction behavior data, use the user interaction behavior prediction model to predict the user interaction frequency P user Based on the acquired equipment performance data, a linear regression algorithm is used to train the model and determine the equipment performance parameter P device , evaluate the current workload and performance status of the device. Define the cache adjustment coefficient formula based on user interaction frequency and device performance parameters,

[0056] C adjust =δ×P user +η×P device , where C adjust is the cache adjustment coefficient, δ and η are weight factors, balancing the impact of user behavior and device performance on cache adjustment. According to the defined formula C adjust Calculate the cache adjustment coefficient in real time. adjust Dynamically adjust cache size, cache update frequency, and cache elimination strategy. Continuously monitor user behavior and device performance, adjust δ and η based on feedback, and optimize cache adjustment strategy.

[0057] For example, ApacheLog4j, a logging tool for the digital twin XR low-code development platform, is used to collect user interaction behaviors in real time, including click depth, page dwell time, and bounce rate. User A clicked on 5 pages when visiting the website, with an average page dwell time of 30 seconds and a bounce rate of 15%. User B clicked on 3 pages, with an average page dwell time of 45 seconds and a bounce rate of 10%. The performance data of the device is obtained. The CPU temperature of device A is 60°C, the memory paging rate is 20%, and the network latency is 10ms. The CPU temperature of device B is 70°C, the memory paging rate is 15%, and the network latency is 15ms. Regular stress tests are performed to evaluate the performance of the device under different load conditions. Device A performs stably under high load, with a response time of less than 5 seconds. Device B performs poorly under high load, with a response time of more than 10 seconds. Based on the user interaction behavior data, a user interaction behavior prediction model is used to predict the user interaction frequency P. user , get user A's P user The predicted value is 0.8, and user B’s P user The predicted value is 0.6. Based on the acquired equipment performance data, a linear regression algorithm is used to train the model and determine the equipment performance parameter P device , get the P of device A device The predicted value is 0.9, and the P of device Bdevice The predicted value is 0.7. Based on the user interaction frequency and device performance parameters, the cache adjustment coefficient formula is defined, C adjust =δ×P user +η×P device , set the weight factor δ = 0.6, η = 0.4. According to the defined formula C adjust Calculate the cache adjustment coefficient in real time for users A and C adjust =0.6×0.8+0.4×0.9=0.84, for users B and C adjust =0.6×0.6+0.4×0.7=0.64. According to the calculated C adjust Dynamically adjust cache size, cache update frequency, and cache elimination strategy. For user A, increase cache size, increase cache update frequency, and reduce the strictness of elimination strategy. For user B, maintain existing cache settings. Continuously monitor user behavior and device performance, adjust δ and η based on feedback results, and optimize cache adjustment strategy. Based on user A's feedback, it may be necessary to fine-tune the weight factor to obtain δ = 0.65 and η = 0.35 to further improve responsiveness. Based on user B's feedback, continue to monitor performance and adjust weight factors to cope with changes in device performance.

[0058] S106. Based on the SaaS layer module update records and volatility analysis, implement a cache invalidation strategy based on content changes and time benchmarks, and adjust the cache refresh interval in combination with business impact assessment.

[0059] Obtain the update records of the SaaS layer modules through the monitoring system, including the time point and frequency of each data change. According to the data update frequency, use the ARIMA algorithm to train the model and predict the update frequency and trend of the SaaS layer modules. Calculate the standard deviation or coefficient of variation of the data, determine the volatility of the SaaS layer modules, and identify the SaaS layer modules with volatility higher than the preset volatility threshold. For SaaS layer modules with volatility higher than the preset volatility threshold, adopt a cache invalidation strategy based on content changes. The cache invalidation strategy based on content changes is to update the cache immediately when the data changes. For SaaS layer modules with update frequency stability higher than the preset stability threshold, adopt a time-based cache invalidation strategy. The time-based cache invalidation strategy is to update the cache regularly. According to the update trend prediction results of the SaaS layer modules, set the cache refresh interval for each SaaS layer module to match the update mode and volatility of the SaaS layer modules. Adjust the cache refresh interval according to business requirements and service level agreements. Evaluate the resource constraints of the digital twin XR low-code development platform, including CPU, memory, and I / O capacity, determine the impact of resource constraints on the cache refresh frequency, and adjust the refresh strategy accordingly. Based on the business decisions, operational efficiency, and customer experience of the SaaS layer modules, the support vector machine algorithm is used for model training to determine the impact of the SaaS layer modules on the business process, and the cache refresh interval is shortened for SaaS layer modules whose impact is higher than the preset impact threshold.

[0060] For example, there are three SaaS layer modules, namely, company management module, project management module and permission management module. The update records, volatility and business impact of each module are as follows: the update records include that the company management module is updated once an hour on average, with small volatility; the project management module is updated once every 10 minutes on average, with large volatility; and the permission management module is updated twice a day on average, with moderate volatility. The volatility and business impact include that the company management module has a standard deviation of 5 and a business impact score of 0.4, which is low; the project management module has a coefficient of variation of 0.8 and a business impact score of 0.8, which is high; and the permission management module has a coefficient of variation of 0.3 and a business impact score of 0.6, which is medium. Since the update frequency of the company management module is stable, a time-based cache invalidation strategy is adopted, such as updating the cache once an hour. Since the project management module has high volatility, a content change-based cache invalidation strategy is adopted, and the cache is updated immediately when the data changes. According to the business impact and volatility, the permission management module adopts a mixed strategy, such as checking whether the cache needs to be updated every 4 hours. The resource limit of the digital twin XR low-code development platform is that the CPU usage does not exceed 70% and the memory usage does not exceed 80%. Based on the resource usage and the high business impact score of the project management module, it may be necessary to prioritize the project management module and allocate more resources. If the use of the project management module causes the CPU usage to approach 70%, it may be necessary to adjust the cache refresh interval of the company management module and the permission management module to reduce the resource burden.

[0061] Based on the update correlation and business impact score between SaaS layer modules, real-time data dependency monitoring and threshold-based dynamic adjustment of cache refresh cycle are implemented, while optimizing the resource allocation and load balancing strategy of the digital twin XR low-code development platform.

[0062] According to the code module dependency graph, identify the dependencies between SaaS layer modules. Define the data dependency scoring formula based on the business impact of SaaS layer modules and the update correlation between modules. Among them, DDS i represents the data dependency score of the i-th SaaS layer module, Cov(M i ,M j ) represents the covariate between the i-th and j-th SaaS layer modules, measuring the update correlation between the two modules. Imp(M j) represents the business impact score of the jth SaaS layer module, which is quantitatively evaluated based on the importance of data to the business process. The DDS score is calculated for each SaaS layer module according to the data dependency scoring formula to quantify the dependencies between different SaaS layer modules. A real-time monitoring mechanism is established to update the DDS scores of SaaS layer modules in real time and set a threshold alarm mechanism. According to the DDS score, the cache refresh cycle of the SaaS layer module is dynamically adjusted. If the DDS score of a module exceeds the preset first scoring threshold, the cache refresh cycle of the module is shortened. If the DDS score of a module is lower than the preset second scoring threshold, the cache time of the module is increased. Combine the DDS score and the resource usage of the digital twin XR low-code development platform to optimize resource allocation and load balancing. Regularly evaluate the accuracy of the DDS formula and the effectiveness of the caching strategy, and adjust and optimize based on feedback data.

[0063] For example, there are three SaaS layer modules, company management module, project management module, and authority management module. These modules have different degrees of influence on the business process, and there is a certain dependency relationship between the three SaaS layer modules. The covariate Cov and business impact score Imp between modules are Imp(M1)=0.8, high impact, Imp(M2)=0.3, low impact, Imp(M3)=0.5, medium impact, Cov(M1,M2)=0.6, Cov(M1,M3)=0.2, Cov(M2,M3)=0.4, Cov(M2,M1)=0.6, Cov(M3,M1)=0.2, Cov(M3,M2)=0.4. According to the business impact of the SaaS layer modules and the update correlation between modules, the data dependency score formula DDSi is defined; now, the DDS formula is used to calculate the data dependency score of each module. For the company management module,

[0064] DDSM1=Cov(M1,M2)×Imp(M2)+Cov(M1,M3)×Imp(M3)=0.6×0.3+0.2×0.5=0.18+0.1=0.28. Similarly, the DDS scores of the project management module and the permission management module can be calculated. Next, according to these scores, corresponding operations are performed. It is found that the dependency score of the company management module is relatively low, while the project management module and the permission management module may have higher dependency scores. Closely monitor the changes in these scores. If the score of the project management module increases significantly, it may indicate that its correlation with other modules is enhanced. Since the dependency score of the company management module is low, a longer cache cycle is set for the company management module. For the project management module and the permission management module, since the project management module and the permission management module have higher dependency scores, more frequent cache refreshes are selected. According to the DDS score, more resources are allocated to the project management module and the permission management module to ensure that the project management module and the permission management module can be processed effectively. Regularly evaluate the effectiveness of these strategies and adjust and optimize them based on feedback data.

[0065] S107. Based on the usage frequency and business importance score of the SaaS layer module, cache capacity evaluation and space allocation optimization are performed to implement resource configuration and optimize cache allocation.

[0066] The monitoring system is used to monitor the usage frequency of the SaaS layer module in real time, including the number of times the module is accessed and the duration of the access. The support vector machine algorithm is used to determine the business importance score of the SaaS layer module according to the type, purpose, and impact of the data item on the business process. According to the cache capacity limit, the current cache usage is evaluated, including the used space and the remaining space, to determine whether cache adjustment is required. If the difference between the cache capacity and the cache upper limit is less than the preset second difference threshold, the data to be eliminated is determined according to the cache elimination strategy, and the cache elimination strategy is to eliminate the page with the least number of uses. The size of the data item is evaluated to determine the proportion of space occupied by the data item in the cache, and the cache space allocation is optimized. Based on the update frequency and business needs of the data item, an expiration time is set for each cache item; combined with the frequency of use, business importance, data size and expiration time information, a weighted calculation method is used to determine the priority of each data item; a concurrency control mechanism is used to prevent multiple concurrent requests from competing for the same data item. In the distributed cache system, a synchronization mechanism is applied to ensure the consistency of data between different nodes when adjusting the cache priority. Based on historical access data and business patterns, the ARIMA algorithm is used for model training to predict future access trends and adjust the priority of data in the cache to implement resource configuration and optimize cache allocation.

[0067] For example, there are three SaaS layer modules, namely, company management module, project management module, and permission management module. The usage frequency, business importance score, and cache status are respectively, the usage frequency and access duration are, the company management module is accessed 100 times a day, with an average access duration of 2 minutes, the project management module is accessed 50 times a day, with an average access duration of 5 minutes, and the permission management module is accessed 20 times a day, with an average access duration of 10 minutes. The business importance score is calculated using the support vector machine algorithm. The company management module scores 0.7, medium importance, the project management module scores 0.9, high importance, and the permission management module scores 0.4, low importance. The total cache capacity is 100GB, 70GB is currently used, 30GB is left, and the cache upper limit difference threshold is set to 10GB. If the cache space allocation is optimized, the company management module occupies 15GB in the cache, the project management module occupies 25GB in the cache, and the permission management module occupies 30GB in the cache. Expiration time and data priority settings: the company management module has a high update frequency, and the expiration time is set to 1 day. The project management module has a medium update frequency, and the expiration time is set to 2 days. The permission management module has a low update frequency, and the expiration time is set to 3 days. Considering the frequency of use, business importance, and data size, the project management module has the highest priority, followed by the company management module, and the permission management module has the lowest priority. Implement a concurrency control mechanism to ensure efficient and consistent access to the project management module. In a distributed system, ensure the cache consistency of the project management module on all nodes. Use the ARIMA model to predict future access trends. The prediction results show that the access frequency of the project management module may increase.

[0068] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for a digital twin XR low-code development platform, characterized in that: The method comprises: Use static code analysis tools to perform a preliminary scan of the code modules of the digital twin XR low-code development platform, obtain the dependencies between modules, and generate a dependency graph, including: Use the static code analysis tool SonarQube to scan the code modules of the digital twin XR low-code development platform, identify the attributes of the structural elements in the code modules, generate a unique identifier for each code module, and record the location and function of each code module. The attributes of the structural elements include method signatures, variable declarations, and access rights; According to the module identifier, the modules are classified into basic environment layer modules, cloud rendering layer modules, low-code development editor modules, data asset library modules, and SaaS layer modules. The module identifier includes the attributes of the structural elements; Through the interfaces and methods exposed by the modules, the dependency types between modules are judged, and compile-time dependencies and runtime dependencies or test dependencies are identified; According to Dependency direction, determine the source and end point of the dependency relationship, build a code module dependency map, and display the connection and dependency path between modules; for each dependency relationship, determine the dependency weight through the dependency frequency; by building a data flow graph, map the data transfer between modules, including monitoring the source of data and the data processing and data end point of each module, and by creating a control flow graph, determine the execution of the program. Each node in the graph represents an operation in the program, and the edge represents the path when the program is executed, so as to obtain the mode and path of data and control information flow between different modules; extract the module change history, determine the reuse level of each module, and identify modules whose reuse frequency is greater than the preset frequency threshold; by checking the errors and warning information in the code, as well as the code complexity and maintainability, determine the code quality of the module, and mark the priority of the modules whose code quality is lower than the preset quality threshold; Based on the obtained code module dependency graph, the SaaS layer module loading and cache processing process is optimized to ensure that the modules are loaded in the order of dependency; The publish / subscribe model is used to establish an event-driven mechanism to monitor changes within the SaaS layer module. Once key changes are detected, the dynamic cache update mechanism is immediately triggered to refresh the cache. Key changes include data updates and status changes. Based on the user interaction behavior data of the digital twin XR low-code development platform, a user interaction behavior prediction model is built to predict the user's operation behavior. Based on the predicted results of the user interaction behavior, differentiated cache strategies are formulated to preload and adjust the cache update frequency for SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold; based on the user's behavior pattern and interaction frequency, the priority and update cycle of the cached content are automatically adjusted, where the responsiveness requirement refers to the priority of the responsiveness demand; According to the device performance monitoring and maintenance logs of the digital twin XR low-code development platform, implement environmental adjustment and software and hardware update strategies, and combine with network status evaluation to determine the cache update needs of the device and optimize cache efficiency, including: obtaining the real-time operation status and performance indicators of the digital twin XR low-code development platform through the real-time performance monitoring system, including online status, errors, warnings, processing speed, response time, CPU and memory usage status, and disk I / O and network I / O operations; if the device operation efficiency is lower than the preset efficiency threshold or the performance indicators are monitored to be abnormal, determine whether environmental factors have an impact on device performance and whether it is necessary to adjust the device settings or environmental conditions. Environmental factors include temperature and humidity, vibration frequency, and noise intensity; based on the device life cycle information, establish a device maintenance log for each device, including past maintenance activities and parts replacement records. The life cycle information includes manufacturing date, expected service life, and historical maintenance records; determine the device type, frequency of use, and maintenance guidelines provided by the manufacturer. Determine the maintenance and parts replacement cycle of the equipment. If the difference between the maintenance time and the current time is less than the preset difference threshold, check the physical components, software system and cache components of the equipment to identify signs of wear and abnormal functions or performance degradation. If the cache efficiency is reduced due to equipment wear or performance degradation, optimize the cache, including increasing the cache capacity and replacing aging storage media, or optimizing the cache management strategy. According to the log and event recording function of the equipment, obtain the software and firmware version information currently used by the equipment. If a new version is monitored, determine whether there is an update to optimize the cache and whether the cache needs to be updated based on the version update content. Obtain the security vulnerability status of the device. If there is a security risk or a security patch is required, determine whether the cache needs to be updated. Adjust the trigger logic, set preset thresholds and conditions. If performance degradation or abnormal events are detected, obtain the cache size, usage rate, hit rate and failure rate data based on the cache usage. If the cache failure rate is higher than the preset failure rate threshold, it is determined that the cache efficiency is insufficient and needs to be updated. According to the update records and volatility analysis of the SaaS layer modules, a cache invalidation strategy based on content changes and time benchmarks is implemented, and the cache refresh interval is adjusted in combination with the business impact assessment, wherein the update records and volatility analysis of the SaaS layer modules include: obtaining the update records of the SaaS layer modules through the monitoring system, including the time point and frequency of each data change; using the ARIMA algorithm to train the model according to the data update frequency, and predicting the update frequency and trend of the SaaS layer modules; calculating the standard deviation or coefficient of variation of the data, judging the volatility of the SaaS layer modules, and identifying the SaaS layer modules with volatility higher than the preset volatility threshold; implementing the cache invalidation strategy based on content changes and time benchmarks includes: for SaaS layer modules with volatility higher than the preset volatility threshold, adopting a cache invalidation strategy based on content changes, and the cache invalidation strategy based on content changes is to update the cache immediately when the data changes; for SaaS layer modules with update frequency stability higher than the preset stability threshold, adopting a time-based cache invalidation strategy, and the time-based cache invalidation strategy is to update the cache regularly; According to the usage frequency and business importance score of the SaaS layer module, cache capacity evaluation and space allocation optimization are carried out to realize resource configuration and optimize cache allocation, including: using a monitoring system to monitor the usage frequency of the SaaS layer module in real time, including obtaining the number of module accesses and access duration; using a support vector machine algorithm to determine the business importance score of the SaaS layer module according to the type, purpose and impact of the data item on the business process; evaluating the current cache usage according to the cache capacity limit, including the used space and the remaining space, to determine whether cache adjustment is required; if the difference between the cache capacity and the cache upper limit is less than the preset second difference threshold, the data to be eliminated is determined according to the cache elimination strategy, and the cache elimination strategy is to eliminate the page with the least number of uses ; Evaluate the size of the data item, determine the proportion of space occupied by the data item in the cache, and optimize the cache space allocation; set an expiration time for each cache item based on the update frequency and business needs of the data item; use a weighted calculation method to determine the priority of each data item based on the frequency of use, business importance, data size and expiration time information; use a concurrency control mechanism to prevent multiple concurrent requests from competing for the same data item; in the distributed cache system, apply a synchronization mechanism to ensure the consistency of data between different nodes when adjusting the cache priority; use the ARIMA algorithm to train the model based on historical access data and business patterns, predict future access trends, and adjust the priority of data in the cache to achieve resource configuration and optimize cache allocation.

2. The method according to claim 1, wherein: Based on the obtained code module dependency graph, the SaaS layer module loading and cache processing flow is optimized to ensure that the modules are loaded in the order of dependency, including: Based on the obtained code module dependency graph, specify the module name, version and file path for the SaaS layer module, and determine the direction for each edge to identify the dependency relationship between modules. The SaaS layer modules include company management module, project management module, permission management module and asset management module; Assign weights to each edge in the graph according to the dependency weights. If there are unweighted edges in the graph, assign weights based on the estimated cost and time or resource consumption of module loading. A topological sorting algorithm is used to process the code module dependency graph to obtain a sequential queue for module loading, where each module is loaded after the modules it depends on. According to the sequential queue obtained by topological sorting, determine whether the current module is already in the loading state or has been loaded. If it has been processed, skip it to avoid repeated loading; Get the current module to be loaded and check whether the module exists in the cache. If the module exists in the cache, load it directly from the cache, otherwise continue the loading process; Before loading a module, perform circular dependency detection. If a circular dependency is detected, take measures to break the cycle, including version downgrade and module replacement. Register a callback function for each module loading through the event callback mechanism, and perform custom processing after the module loading is completed; During module loading, a concurrency control mechanism is used to manage module loading requests, load multiple modules simultaneously, and maintain the consistency of dependencies and loading status; Establish an error recognition mechanism to identify and classify different types of loading errors, including timeout errors and file missing errors, and pre-set corresponding recovery strategies for common error types.

3. The method according to claim 1, wherein: The publish / subscribe mode is used to establish an event-driven mechanism to monitor changes in the SaaS layer module. Once a key change is detected, the dynamic cache update mechanism is immediately triggered to refresh the cache, including: The event producer continuously monitors the status and activities within the SaaS layer module. If a predefined key change is detected, the status change information of the event source is obtained, and a structured event message is created and published. The event message includes the user ID and activity type, data update and timestamp. The event producer is the designated monitoring system. Configure the message channel, including setting the channel's transmission protocol, message format, and network configuration, and provide event consumers with the functions of filtering and subscribing to events, filtering according to event types. Event consumers are cache management systems; On the event consumer side, subscribe to the specified event type related to the cache content. If an event of the specified event type is received, the cache update mechanism is triggered; Setting buffering strategies for event queues during peak hours, including setting buffer size, adjusting event processing priority, and ensuring event processing order. Peak hours are when the number of events is greater than a preset number threshold. Perform cache refresh operations on received events and add fault tolerance mechanisms to deal with potential abnormal situations; Use the cache management mechanism to synchronize the data of all relevant cache instances, and adjust the cache expiration and refresh policies based on the timeliness and importance of the event; Configure the monitoring system to monitor the publishing, transmission and processing of events, and record information about event processing through a logging mechanism; Establish service quality controls, including reliability of message delivery and sequentiality of event processing; Implement security measures, including authorization and authentication of event producers and consumers and data encryption during message transmission.

4. The method according to claim 1, wherein: The method of constructing a user interaction behavior prediction model based on the user interaction behavior data of the digital twin XR low-code development platform, predicting the user's operation behavior, and formulating differentiated cache strategies based on the predicted results of the user interaction behavior, and adjusting the preloading and cache update frequency of the SaaS layer modules with responsiveness requirements higher than the preset responsiveness threshold, includes: Real-time monitoring of digital twin XR low-code development platform end log records, recording user interaction behaviors on applications or websites, including click events, page views, search queries, and form submissions; Obtain device information, including device type and operating system, and use different data processing procedures based on the user's device information; Record the environmental characteristics when the user's interactive behavior occurs, including timestamp and geographic location information, and associate the environmental characteristics with the user's interactive behavior data; By counting the frequency of user interactions in each time period, we can identify whether there are differences in user activity patterns on weekdays and weekends. Extract user interaction frequency and user classification information from historical data, apply K-means clustering algorithm for model training, and divide users into different behavior pattern groups; Based on the user interaction behavior data, the decision tree algorithm is used for model training to build a user interaction behavior prediction model to predict the user's future interaction behavior, including predicting whether the user will click a specified button, visit a specified page, or how long to spend on the page.

5. The method according to claim 1, wherein: The device performance monitoring and maintenance logs of the digital twin XR low-code development platform are used to implement environmental adjustments and software and hardware update strategies, and combined with network status evaluation, the cache update requirements of the device are judged to optimize cache efficiency. The method also includes: based on user interaction behavior and device performance monitoring data, cache adjustment coefficient calculation and application are implemented, while continuously optimizing cache size, update frequency and elimination strategy. Specifically, the method includes: using the digital twin XR low-code development platform logging tool ApacheLog4j to obtain user interaction behavior in real time, including click depth, page dwell time, and bounce rate; obtaining device performance data, including CPU temperature, memory paging, and network latency; performing stress tests regularly to evaluate the performance of the device under different load conditions; using a user interaction behavior prediction model to predict user interaction frequency P based on user interaction behavior data. user ; Based on the acquired equipment performance data, a linear regression algorithm is used to train the model and determine the equipment performance parameter P device , evaluate the current workload and performance status of the device; define the cache adjustment coefficient formula, C, based on the user interaction frequency and device performance parameters adjust =δ×P user +η×P device , where C adjust is the cache adjustment coefficient, δ and η are weight factors, balancing the impact of user behavior and device performance on cache adjustment; according to the defined formula C adjust Calculate the cache adjustment coefficient in real time; based on the calculated C adjust Dynamically adjust cache size, cache update frequency, and cache elimination strategy; continuously monitor user behavior and device performance, adjust δ and η based on feedback results, and optimize cache adjustment strategy.

6. The method according to claim 1, wherein: The cache invalidation strategy based on content changes and time benchmarks is implemented according to the update records and volatility analysis of the SaaS layer modules, and the cache refresh interval is adjusted in combination with the business impact assessment, including: According to the update trend prediction results of the SaaS layer modules, set the cache refresh interval for each SaaS layer module to match the update mode and volatility of the SaaS layer modules; Adjust the cache refresh interval based on business needs and service level agreements; Evaluate the resource limitations of the digital twin XR low-code development platform, including CPU, memory, and I / O capacity, determine the impact of resource limitations on cache refresh frequency, and adjust the refresh strategy accordingly; Based on the business decisions, operational efficiency and customer experience of the SaaS layer modules, the support vector machine algorithm is used to train the model to determine the impact of the SaaS layer modules on the business process, and the cache refresh interval is shortened for the SaaS layer modules whose impact is higher than the preset impact threshold; It also includes: implementing real-time data dependency monitoring and threshold-based dynamic adjustment of cache refresh cycles based on the update correlation and business impact scores between SaaS layer modules, while optimizing the resource allocation and load balancing strategy of the digital twin XR low-code development platform. Specifically, it includes: identifying the dependencies between SaaS layer modules based on the code module dependency graph; defining the data dependency scoring formula based on the business impact of the SaaS layer modules and the update correlation between the modules. Among them, DDS i represents the data dependency score of the i-th SaaS layer module, Cov(M i ,M j ) represents the covariate between the i-th and j-th SaaS layer modules, measuring the update correlation between the two modules. Imp(M j ) represents the business impact score of the j-th SaaS layer module, which is quantitatively evaluated based on the importance of data to the business process; the DDS score is calculated for each SaaS layer module according to the data dependency scoring formula, and the dependency relationship between different SaaS layer modules is quantified; a real-time monitoring mechanism is established to update the DDS score of the SaaS layer module in real time, and a threshold alarm mechanism is set; according to the DDS score, the cache refresh cycle of the SaaS layer module is dynamically adjusted. If the DDS score of a module exceeds the preset first scoring threshold, the cache refresh cycle of the module is shortened; if the DDS score of a module is lower than the preset second scoring threshold, the cache time of the module is increased; resource allocation and load balancing are optimized based on the DDS score and the resource usage of the digital twin XR low-code development platform; the accuracy of the DDS formula and the effectiveness of the caching strategy are regularly evaluated, and adjustments and optimizations are made based on feedback data.

Images