A vue interface intelligent routing control method based on a large language model

Abstract

The application discloses a Vue interface intelligent routing control method based on a large language model, relates to the technical field of routing control, and realizes unified access and standardized processing of multi-modal data by creating a receiving nest, outputs standard natural language data, completes intention reasoning by using a pre-training large language model, generates a structured routing intention, maps the structured routing intention to a Vue Router routing configuration and completes analysis, performs fine access permission verification on a target routing by a routing guard, controls routing jump states according to a verification result, performs page information preloading on the target routing, arranges an interface flow according to rendering logic, and completes intelligent routing navigation and page rendering. The application effectively improves the intelligent level of front-end routing control, realizes integrated control of multi-modal interaction adaptation, efficient routing analysis, safe permission verification and smooth page switching, and significantly optimizes the Vue front-end interaction experience and system stability.

Description

Technical Field

[0001] This invention relates to the field of routing control technology, specifically to an intelligent routing control method for the Vue interface based on a large language model. Background Technology

[0002] Traditional Vue front-end routing relies on manual configuration and fixed rule triggering, which is difficult to adapt to multimodal natural interaction scenarios. It suffers from problems such as low intent parsing efficiency, high coupling of route mapping, coarse permission verification, and sluggish page loading and rendering. At the same time, multimodal data input lacks a unified receiving and exception handling mechanism, and resource scheduling is disordered, resulting in slow route control response, poor stability, and poor user experience. It cannot meet the needs of intelligent interaction and dynamic routing driven by large language models. Summary of the Invention

[0003] The purpose of this invention is to provide a smart routing control method for Vue interface based on a large language model, so as to solve the problems in the background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a Vue interface intelligent routing control method based on a large language model, comprising the following steps: Step S1: Create a receiving nest, input multimodal data based on the receiving nest to obtain the corresponding natural language data, perform intent reasoning on the natural language data through a pre-trained large language model, and output the corresponding structured routing intent. Step S2: Map the structured routing intent to the Vue Router route configuration, obtain the corresponding configuration information, and perform access permission verification on the target route adapted to the configuration information through route guards. Determine the execution status of route jump based on the verification result. Step S3: Preload the page information required for page rendering for the target route, arrange different interface flows by combining the page information, and execute the corresponding intelligent route navigation and page rendering for all interface flows.

[0005] In a preferred embodiment, the process of creating a receiver nest, performing multimodal data input based on the receiver nest, and obtaining the corresponding natural language data includes: Establish separate nesting structures at both the distal and proximal locations; The nesting entity corresponds to a number of nesting points. Based on the same homing information domain, nesting entities at remote and near locations are associated with each other to obtain the receiving space at the remote location. Nesting events in the receiving space are counted to obtain the nesting package for the corresponding receiving space. Establish a nesting line to connect different receiving spaces; the nesting line transmits nesting packets. Based on the modality type of the received data, the receiving space is grouped accordingly to obtain the corresponding space set. Any receiving space in the same space set is selected as the nesting starting point. The nesting event of the nesting starting point is executed to obtain the nesting packet. All nesting packets of the nesting starting point are sent to other receiving spaces in the same space set as a receiving basis for the receiving space to select to receive data of the corresponding modality type. By connecting all the receiving spaces where the corresponding nesting body is located at the far end by connecting the nesting body at the far end, the receiving nest on the corresponding receiving side is obtained. Based on the one-to-one correspondence between the nesting body at the near end and the receiving nest at the far end, the receiving nest on the storage side of each receiving space is obtained.

[0006] In a preferred embodiment, the receiving nest includes: Nesting organisms at near and far locations correspond to each other based on homing information domains. The category of the homing information domain corresponds to the modality type of the received data. There are several homing information domains of the same category, and the receiving spaces of each homing information domain are interconnected and share their respective nesting events. Nesting events represent the details of abnormal operations that occur in the input area when the receiving space where the corresponding nesting entity is located performs input of natural language data. The handling measures for resolving the operation details are shared as nesting experience with other receiving spaces that are connected to the data. The corresponding nesting packets are generated in other receiving spaces as the basis for the natural language data corresponding to the corresponding modality type to be executed. The receiving spaces that are interconnected with each other are registered to the same location on the storage side of the near end location as the storage address of the corresponding modality type.

[0007] In a preferred embodiment, the process of the nesting line sending nesting packets includes: The connection point between the nesting line and each receiving space is marked as the attachment point. All attachment points of the same space group are registered to the common operation pool. The corresponding operation space area is configured in the common operation pool. The corresponding operation is repeated for the corresponding receiving space of other space groups to determine the operation space area of ​​different space groups. The nesting package is attached to the nesting line. The nesting package moves on the nesting line. Whenever the nesting package traverses to an attachment point, it is determined whether the operation space area of ​​the common operation pool where the corresponding attachment point is located holds execution resources that can handle its own operation. If yes, then process the receiving space corresponding to the current attachment point and continue sending nesting packets to the next attachment point; if no, then fill in the resource gaps through other operating space regions and process the corresponding receiving space.

[0008] In a preferred embodiment, the process of performing intent reasoning on natural language data using a pre-trained large-scale language model and outputting the corresponding structured routing intent includes: Historical routing data is obtained, and data analysis and processing are used to obtain model training data. The model training data includes structured semantic information with several field lengths. The preliminary training model is constructed, the structured semantic information is analyzed, an input format that the model can understand is added, and the executable text in the routing control is matched and output. Determine whether the current training model has been successfully trained on the structured semantic information of each category; If so, the structured semantic information of the corresponding category is encapsulated in the input layer of the training model to obtain the corresponding analysis layer and output layer. The analysis layer is used to inject contextual information, user behavior history and application state into the structured semantic information, and the analysis results are output as structured routing intent through the output layer. If not, continue training the structured semantic information for the corresponding category until the output conditions are met.

[0009] In a preferred embodiment, the process of mapping structured routing intents to Vue Router routing configurations to obtain the corresponding configuration information includes: The predefined configuration items for Vue Router routes are defined based on the constituent elements of the structured routing intent. Each configuration item corresponds to a configuration table, which records all configuration information associated with the corresponding configuration item. The structured routing intent is mapped to the Vue Router routing configuration and parsed to obtain the configuration items of the structured routing intent in the Vue Router routing configuration, thus obtaining the corresponding configuration information.

[0010] In a preferred embodiment, the process of performing access permission verification on the target route adapted to the configuration information through route guards, and determining the execution status of route hopping based on the verification result, includes: When the route guard detects a redirection request, it obtains the configuration information of the target route, the user's identity information, role set, permission list, and login status information, extracts the permission configuration required by the target route from the metadata, and matches and compares it item by item with the user's actual permission information. According to the preset verification rules, if the user has all the required permissions, the login status is valid, the access level matches and the access constraints are met, then the access permission verification is deemed to have passed; otherwise, the access permission verification is deemed to have failed. If the access permission verification passes, the system will output "Allow Redirect" and allow the request to proceed, then proceed to the UI stream rendering. If the access permission verification fails, the system will output "Reject Redirect" and block the request, redirecting the user to the login page, the "No Permission" prompt page, or keeping the current page unchanged according to a preset strategy.

[0011] In a preferred embodiment, the process of preloading page information required for page rendering of the target route, arranging different interface flows by combining page information, and performing corresponding intelligent route navigation and page rendering on all interface flows includes: Establish several preload points and store all preload points in the loading container; A number of linked information points are deployed in the loading container. Based on the page information required by the target route, the preloaded points in all positions in the loading container are filled accordingly. Based on the linkage information points, preload points at different locations are connected to obtain linkage paths, which are used to characterize the rendering order of different page information when the target route preloads the page rendering. When the page information with the earlier rendering order is processed, the rendering data of the page information with the later rendering order in the loading container is directly called to execute the page rendering of the corresponding page information. For each preloaded page, a corresponding interface frame, interface content, and interface switching method are constructed, and standardized based on front-end program rules to obtain the interface flow of the corresponding page information. The interface flow includes the routing path, data loading sequence, and page rendering window. Based on the arrangement of page information on the linkage path, the corresponding interface flows of each page information are combined to complete the display order of the interface flows, and the corresponding intelligent routing navigation and page rendering are performed on all interface flows.

[0012] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention achieves stable acquisition of natural language data by receiving and uniformly processing multimodal data through a receiving nest, avoiding input anomalies, providing reliable data support for intent reasoning, relying on a large language model for intent reasoning, accurately outputting structured routing intents, realizing efficient conversion of natural language into routing instructions, and improving the level of intelligent routing parsing.

[0013] 2. This invention maps structured routing intents to Vue Router routing configurations to obtain corresponding configuration information, reducing configuration coupling, enabling rapid location and parsing of routing information, improving route matching accuracy, performing access permission verification through route guards, intercepting invalid jumps in advance to ensure access security, avoiding invalid rendering, improving the smoothness and security of route control, and achieving seamless connection between data loading, route jumping and page rendering through preloading and intelligent scheduling of the interface flow, thus optimizing overall rendering efficiency. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0015] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1, please refer to Figure 1 As shown in this embodiment, a Vue interface intelligent routing control method based on a large language model includes the following steps: Step S1: Create a receiving nest, input multimodal data based on the receiving nest to obtain the corresponding natural language data, perform intent reasoning on the natural language data through a pre-trained large language model, and output the corresponding structured routing intent. Step S2: Map the structured routing intent to the Vue Router route configuration, obtain the corresponding configuration information, and perform access permission verification on the target route adapted to the configuration information through route guards. Determine the execution status of route jump based on the verification result. Step S3: Preload the page information required for page rendering for the target route, arrange different interface flows by combining the page information, and execute the corresponding intelligent route navigation and page rendering for all interface flows.

[0018] It should be further explained that, in the specific implementation process, the process of creating a receiving nest, inputting multimodal data based on the receiving nest, and obtaining the corresponding natural language data includes: Establish a nested structure corresponding to the remote location and the near location. The remote location is deployed at the user's location, and the near location is deployed at the server's location. The nesting body corresponds to a number of nesting points. The nesting bodies at distant and near locations are mapped based on the same homing information domain to obtain several receiving spaces at the distant location. Each receiving space corresponds to a different number of nesting points. The nesting events that occur in each receiving space are statistically analyzed. The content of the nesting events is used as the nesting experience of the corresponding receiving space. The nesting experience is analyzed to obtain the nesting package of each nesting point in the corresponding receiving space. Establish a nesting line to connect different receiving spaces; the nesting line transmits nesting packets. Based on the modality type of the received data, the receiving space is grouped accordingly to obtain several corresponding space sets. Any receiving space in the same space set is selected as the corresponding nesting start point. The nesting event at the nesting start point is executed, and the corresponding nesting packet is obtained. All nesting packets at the nesting start point are sent to other receiving spaces in the same space set as a basis for receiving spaces to select to receive data of the corresponding modality type. By connecting all the receiving spaces where the corresponding nesting body is located at the far end by connecting the nesting body at the far end, the receiving nest on the corresponding receiving side is obtained. Based on the one-to-one correspondence between the nesting body at the near end and the receiving nest at the far end, the receiving nest on the storage side corresponding to each receiving space is obtained. The receiving side receives natural language data input in different modalities, while the storage side stores the natural language data input in all modalities at different locations on the server.

[0019] The receiving nest also includes: nest dwellers at near and far ends corresponding to homing information domains. The category of the homing information domain corresponds to the modality type of the received data. There are several homing information domains of the same category, and each homing information domain corresponds to a receiving space. The receiving spaces corresponding to homing information domains of the same category are interconnected to share their respective nesting events. A nesting event represents a receiving space where the corresponding nest dweller is located. When inputting natural language data, the operation details of abnormal operations corresponding to the input area are shared. The processing measures used to resolve the operation details are used as nesting experience for the nesting event and are shared. Within other data-connected receiving spaces, corresponding nesting packets are generated to serve as the basis for executing natural language data of the corresponding modality type as input. The receiving spaces that are interconnected with each other are registered to the same location on the storage side of the same near-end location as the storage address of the corresponding modality type. This avoids the adverse effects of the same abnormal operation on the acquisition of natural language data input, ensuring that natural language data under each modality type can be obtained well. This lays a solid data foundation in advance for VueRouter to understand the structured routing intent after natural language data processing.

[0020] The content sent by the nesting line to the nesting packet includes the following: The corresponding connection point between the nesting line and each receiving space is marked as the attachment point. All attachment points in the same space group are registered to a common operation pool. A corresponding operation space area is configured in the common operation pool. The corresponding operation is repeated on the receiving space of other space groups to determine the corresponding operation space area of ​​each space group. Different operational space areas within the public operational pool can perform interoperability based on their respective regional credentials; During the initialization of each independent operating space region, a unique encrypted region certificate is automatically generated. The region certificate includes the region number, resource permissions, and interoperability key, and is simultaneously recorded in the general ledger of the public operating pool. All operating space regions have cross-regional trust verification enabled by default. Before interoperating between operating space regions, they must verify each other's valid certificates. Once the verification is successful, a temporary trusted communication link is established. Each operating space region can only share corresponding execution resources such as computing power, scheduling, and data read / write according to the permission scope defined by the region certificate. Operating space regions without certificates or with expired certificates are prohibited from participating in interoperability replacement. The temporary trust link is in a dynamic time-limited mode, remaining valid within a single nesting package traversal processing cycle and automatically released after the task is completed. At the same time, the public operating pool records the interoperability log in real time to ensure that cross-regional resource replacement is controllable and traceable.

[0021] The nesting pack is attached to the nesting line and moves on the nesting line. Whenever the nesting pack traverses to an attachment point, it is determined whether the operation space area of ​​the common operation pool where the corresponding attachment point is located holds execution resources that can handle its own operation. If so, then process the receiving space corresponding to the current attachment point and continue sending nesting packets to the next attachment point; If not, resources are supplemented by other operating space regions that are trusted and interconnected with the current operating space region, the execution resources corresponding to the current attachment point are supplemented, and the processing of the corresponding receiving space is performed.

[0022] It should be noted that when the nesting package is attached to the nesting line and performs movement at attachment points corresponding to different receiving spaces, in addition to determining whether the receiving space of each attachment point can meet its own processing needs, the nesting package can also coordinate with other space groups to perform task distribution and resource reclamation. This ensures that the intelligent routing navigation and page rendering process in the subsequent global state can run continuously and stably. Specifically, at the task distribution level: when the nesting package traverses the attachment points, it first checks the matching degree between the remaining execution resources in the current operation space area and the tasks to be processed. If a single operation space area has overload or insufficient resources, the nesting package will split into lightweight sub-tasks. Through the point linkage mechanism of the nesting line, lightweight subtasks are distributed to the corresponding operation space areas of other space sets that have completed credit verification and have abundant resources, realizing cross-regional parallel processing and reducing the load pressure of single points. At the resource recovery level: after the nesting package completes the processing tasks of the corresponding receiving space area of ​​each attachment point, it actively traverses each operation space area that participated in resource filling or task diversion, counts the computing power, scheduling resources and cached data temporarily called by each operation space area, and transmits them back to the corresponding original space set through the nesting line, releasing the corresponding resources occupied by temporary credit, continuously refreshing the global resource status, and ensuring that there is no resource redundancy or lag in the future.

[0023] It should be further explained that, in the specific implementation process, the process of using a pre-trained large-scale language model to perform intent reasoning on natural language data and outputting the corresponding structured routing intent includes: Obtain historical routing data, and analyze and process the historical routing data to obtain model training data; The model training data includes structured semantic information of several field lengths. A preliminary training model is built through a convolutional neural network. The training model analyzes each structured semantic information and adds an input format that the model can understand. Then, it matches the corresponding executable text in the routing control and outputs the executable text. The executable text is used to represent the specific routing-related information corresponding to the structured semantic information at the routing control level. Determine whether the current training model has been successfully trained on the structured semantic information of each category; The criterion for judging whether the training is successful is the accuracy of the output results. When the accuracy reaches the preset lower threshold, it is determined that the structured semantic information of the corresponding category has not yet met the standard and the corresponding model training still needs to be performed. If so, the structured semantic information of the corresponding category is encapsulated in the input layer of the training model to obtain a corresponding analysis layer and an output layer. The analysis layer is used to inject contextual information, user behavior history and application state into the structured semantic information, and further outputs the analysis results as a structured routing intent through the output layer. Structured routing intents include intent type, destination route name, confidence level, and additional information; Identify structured routes as The structured routing intent is as follows: =< , , , >; in, The intent type representing a structured routing intent includes redirect, return, redirect with parameters, and redirect, etc. This indicates the target route name and the corresponding route path to the target page (such as / home, user / profile, etc.). This represents the confidence level of the current structured routing intent, with a corresponding value range of (0, 1). This indicates additional information about the structured routing intent, including navigation methods (such as push, replace), whether authentication conditions are configured, and priorities.

[0024] If not, continue training the structured semantic information for the corresponding category until the output conditions are met.

[0025] It should be further explained that, in the specific implementation process, the process of mapping structured routing intents to VueRouter routing configurations and obtaining the corresponding configuration information includes: The predefined configuration items for Vue Router routes are defined based on the components of a structured route intent, namely the intent type, target route name, confidence level, and additional information mentioned above. Each configuration item corresponds to a configuration table. The configuration table is used to record all configuration information associated with the corresponding configuration item. It maps the structured routing intent to the Vue Router routing configuration for parsing, obtains the configuration item corresponding to the structured routing intent in the Vue Router routing configuration, and then obtains the corresponding configuration information.

[0026] It should be noted that during the mapping and parsing process, the target route name in the structured routing intent is used as the core matching keyword, the intent type is used as the matching constraint, the confidence level is used as the route validity filtering criterion, and additional information is used as supplementary conditions for route attributes, forming a multi-dimensional and multi-level matching mechanism. The configuration table corresponding to each configuration item is stored independently using a key-value pair structure, ensuring that the configuration information is isolated from each other and does not interfere with each other. During route mapping, configuration items can be quickly located and configuration information can be quickly read, avoiding matching errors, parsing delays, or information corruption caused by the coupling of route configurations. Through the above-mentioned methods of pre-defining configuration items, independently storing configuration tables, multi-dimensional mapping and parsing, and integrating and encapsulating configuration information, the structured routing intent can be efficiently and stably converted into standard route configurations that Vue Router can recognize and execute, greatly improving the accuracy of route mapping and system execution efficiency.

[0027] It should be further explained that, in the specific implementation process, the process of performing access permission verification on the target route adapted to the configuration information through route guards, and determining the execution status of route hopping based on the verification results, includes: When the route guard detects a redirection request for the target route, it automatically obtains the configuration information of the target route that has been parsed. The configuration information includes the route path, route name, metadata, permission identifier, navigation mode, and additional information. The route guard obtains the current user's identity information, role set, permission list, and login status information. The route guard extracts the required permission configuration from the metadata of the configuration information corresponding to the target route, including the required role type, access level, login status requirements, function permission identifier, and access constraints, and compares them with the actual permission information held by the current user item by item. The route guard performs access permission checks according to preset permission verification rules; specifically as follows: If the user has all the permissions required by the target route, and the login status is valid, the access level matches, and the access constraints are met, then the access permission verification is considered successful. If a user lacks any of the permissions required for the target route, or if their login status is invalid, their access level is insufficient, or their access constraints are not met, then the access permission verification is deemed to have failed.

[0028] The final execution status of the route jump is determined based on the permission verification result. When the permission verification passes, the route guard outputs the execution status of allowing the jump, allowing the route jump request and entering the rendering stage of the subsequent interface flow. When the permission verification fails, the route guard outputs the execution status of rejecting the jump, blocking the route jump request, and executing a redirect to the login page, the no permission prompt page, or keeping the current route page unchanged according to the preset policy.

[0029] The preset permission verification rules include the following: Login status verification rules: Verify whether the current user is in a valid login state. If the login state is invalid, the permission verification will be directly deemed to have failed. Role matching verification rules: The user's set of roles is compared with the role types required by the target route. If the user's roles do not include the required roles, the permission verification is deemed to have failed. Access level verification rules: The user's current access level is compared with the access level required by the target route. If the user's access level is lower than the required level, the permission verification is deemed to have failed. Function permission verification rules: The user's list of function permissions is compared with the function permission identifiers required by the target route item by item. If any function permission is missing, the permission verification is deemed to have failed. Access constraint verification rules: Verify whether the current access time, access device, access environment, user status and other constraints meet the requirements of the target route. If any constraint is not met, the permission verification is deemed to have failed. Permission priority rule: When both allow and deny permissions exist, the decision-making strategy of denying permissions takes precedence over allowing permissions; The route guard performs the verification in sequence according to the above rules. If all rules are satisfied, the access permission verification is deemed to have passed. If any rule is not satisfied, the access permission verification is deemed to have failed.

[0030] It should be noted that during the access permission verification process, the route guard adopts a global pre-interception mechanism to complete the full-process verification before the route jump is executed, without causing page flickering or invalid rendering, ensuring the security and smoothness of route control. At the same time, the route guard supports multi-level permission combination verification, dynamic permission real-time update verification, and extended permission judgment based on additional information, making access permission control more granular and reliable.

[0031] It should be further explained that, in the specific implementation process, the process of preloading the page information required for page rendering of the target route, arranging different interface flows by combining the page information, and executing the corresponding intelligent route navigation and page rendering for all interface flows includes: Establish several preload points and store all preload points in the loading container; A number of linked information points are set up in the loading container. Based on all the page information required when rendering the preloaded page corresponding to the target route, the preloaded points in all positions in the loading container are filled accordingly. Linkage paths are obtained by connecting preload points at different locations based on linkage information points. The linkage paths are used to represent the rendering order of different page information when the target route preloads the page rendering. For page information that is rendered earlier in the linkage information point, the data processing related to page rendering in the corresponding preload point is directly executed to obtain the rendering data of the corresponding page information in real time. For page information that is rendered later in the linkage information point, the corresponding page rendering timestamp and rendering method are set, and the data processing is performed in advance in the corresponding preload point to obtain the corresponding rendering data and put it into the loading container. After the page information in the previous preloading point has been processed, the rendering data of the page information that is rendered later in the rendering order is directly called in the loading container, and then the relevant page rendering processing of the corresponding page information is executed. For each preloaded page, a corresponding interface frame, interface content, and interface switching method are constructed. The interface frame represents the size of the page to be displayed for the current page information, the interface content corresponds to some visual presentation forms of the page information, including but not limited to text, audio and video, jump links, and interactive components, etc., and the interface switching method corresponds to the transition form between different page information. The interface frame, content, and switching method of each page are standardized based on the front-end program rules to obtain an interface flow for the corresponding page information. The interface flow includes the routing path, data loading sequence, and page rendering window. Among them, the routing path represents the route from the starting page corresponding to the current page information to the target page corresponding to the required page information; the data loading sequence represents the order of data requests required to jump to the target page; and the page rendering window is used to define the display duration of each page information after rendering. Based on the arrangement of page information on the linkage path, the corresponding interface flows of each page information are combined to complete the display order of the interface flows. All interface flows are arranged according to the display order, and corresponding intelligent routing navigation and page rendering are executed.

[0032] It should be noted that during the formal execution phase after the global UI flow layout is completed, an intelligent routing dynamic scheduling mechanism is activated. Relying on the global scheduling capabilities of the loading container, the routing path jump, data loading sequence execution, and page rendering window display are triggered level by level according to the UI flow display order and the rendering sequence of the linkage path. During the intelligent routing navigation process, the routing status is updated in real time, the jump connection effect of upstream and downstream pages is monitored, and the start time of the next UI flow is dynamically scheduled according to the page rendering completion status, so as to achieve seamless connection of routing jump, data loading, page rendering, and UI switching.

[0033] In this process, after the rendering and display of a single interface flow are completed and the interface switching is executed, the idle resources of the current preloaded point are reclaimed in real time, and the relationship between effective linkage information points is preserved to provide resource support for the rendering and execution of subsequent interface flows. At the same time, the running data such as the route jump time, data loading efficiency, page rendering time, and interface switching smoothness of each interface flow are recorded throughout the process to form a route rendering log. After all interface flows have completed the intelligent route navigation and page rendering and the complete interface display process in sequence, a global closed-loop verification is performed to verify the execution integrity of all linkage paths, the data processing completion of all preloaded points, and the compliance of the display order of each interface flow. After confirming that there are no unexecuted, unrendered, or jump abnormal nodes, the redundant data, expiration timestamps, and temporary rendering parameters of the loading container are uniformly reset to complete the execution of the target route preload page rendering and intelligent route navigation in this round.

[0034] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A Vue interface intelligent routing control method based on a large language model, characterized in that, Includes the following steps: Step S1: Create a receiving nest, input multimodal data based on the receiving nest to obtain the corresponding natural language data, perform intent reasoning on the natural language data through a pre-trained large language model, and output the corresponding structured routing intent. Step S2: Map the structured routing intent to the Vue Router route configuration, obtain the corresponding configuration information, and perform access permission verification on the target route adapted to the configuration information through route guards. Determine the execution status of route jump based on the verification result. Step S3: Preload the page information required for page rendering for the target route, arrange different interface flows by combining the page information, and execute the corresponding intelligent route navigation and page rendering for all interface flows.

2. The intelligent routing control method for Vue interface based on a large language model according to claim 1, characterized in that, The process of creating a receiver nest, inputting multimodal data based on the receiver nest, and obtaining the corresponding natural language data includes: Establish separate nesting structures at both the distal and proximal locations; The nesting entity corresponds to a number of nesting points. Based on the same homing information domain, nesting entities at remote and near locations are associated with each other to obtain the receiving space at the remote location. Nesting events in the receiving space are counted to obtain the nesting package of the corresponding receiving space. Establish a nesting line to connect different receiving spaces; the nesting line transmits nesting packets. Based on the modality type of the received data, the receiving space is grouped accordingly to obtain the corresponding space set. Any receiving space in the same space set is selected as the nesting starting point. The nesting event of the nesting starting point is executed to obtain the nesting packet. All nesting packets of the nesting starting point are sent to other receiving spaces in the same space set as a receiving basis for the receiving space to select to receive data of the corresponding modality type. By connecting all the receiving spaces where the corresponding nesting body is located at the far end by connecting the nesting body at the far end, the receiving nest on the corresponding receiving side is obtained. Based on the one-to-one correspondence between the nesting body at the near end and the receiving nest at the far end, the receiving nest on the storage side of each receiving space is obtained.

3. The intelligent routing control method for Vue interface based on a large language model according to claim 2, characterized in that, The receiving nest includes: Nesting organisms at near and far locations correspond to each other based on homing information domains. The category of the homing information domain corresponds to the modality type of the received data. There are several homing information domains of the same category, and the receiving spaces of each homing information domain are interconnected and share their respective nesting events. Nesting events represent the details of abnormal operations that occur in the input area when the receiving space where the corresponding nesting entity is located performs input of natural language data. The handling measures for resolving the operation details are shared as nesting experience with other receiving spaces that are connected to the data. The corresponding nesting packets are generated in other receiving spaces as the basis for the natural language data corresponding to the corresponding modality type to be executed. The receiving spaces that are interconnected with each other are registered to the same location on the storage side of the near end location as the storage address of the corresponding modality type.

4. The intelligent routing control method for Vue interface based on a large language model according to claim 3, characterized in that, The process of sending nesting packets via the serial nesting line includes: The connection point between the nesting line and each receiving space is marked as the attachment point. All attachment points of the same space group are registered to the common operation pool. The corresponding operation space area is configured in the common operation pool. The corresponding operation is repeated for the corresponding receiving space of other space groups to determine the operation space area of ​​different space groups. The nesting package is attached to the nesting line. The nesting package moves on the nesting line. Whenever the nesting package traverses to an attachment point, it is determined whether the operation space area of ​​the common operation pool where the corresponding attachment point is located holds execution resources that can handle its own operation. If yes, then process the receiving space corresponding to the current attachment point and continue sending nesting packets to the next attachment point; if no, then fill in the resource gaps through other operating space regions and process the corresponding receiving space.

5. The intelligent routing control method for Vue interface based on a large language model according to claim 4, characterized in that, The process of performing intent reasoning on natural language data using a pre-trained large-scale language model and outputting the corresponding structured routing intent includes: Historical routing data is obtained, and data analysis and processing are used to obtain model training data. The model training data includes structured semantic information of several field lengths. The preliminary training model is constructed, the structured semantic information is analyzed, an input format that the model can understand is added, and the executable text in the routing control is matched and output. Determine whether the current training model has been successfully trained on the structured semantic information of each category; If so, the structured semantic information of the corresponding category is encapsulated in the input layer of the training model to obtain the corresponding analysis layer and output layer. The analysis layer is used to inject contextual information, user behavior history and application state into the structured semantic information, and the analysis results are output as structured routing intent through the output layer. If not, continue training the structured semantic information for the corresponding category until the output conditions are met.

6. The intelligent routing control method for Vue interface based on a large language model according to claim 5, characterized in that, The process of mapping structured routing intents to Vue Router route configurations and obtaining the corresponding configuration information includes: The predefined configuration items for Vue Router routes are defined based on the constituent elements of the structured routing intent. Each configuration item corresponds to a configuration table, which records all configuration information associated with the corresponding configuration item. The structured routing intent is mapped to the Vue Router routing configuration and parsed to obtain the configuration items of the structured routing intent in the VueRouter routing configuration, thus obtaining the corresponding configuration information.

7. The intelligent routing control method for Vue interface based on a large language model according to claim 6, characterized in that, The process of performing access permission checks on the target routes adapted to the configuration information through route guards, and determining the execution status of route hopping based on the check results, includes: When the route guard detects a redirection request, it obtains the configuration information of the target route, the user's identity information, role set, permission list, and login status information, extracts the permission configuration required by the target route from the metadata, and matches and compares it item by item with the user's actual permission information. According to the preset verification rules, if the user has all the required permissions, the login status is valid, the access level matches and the access constraints are met, then the access permission verification is deemed to have passed; otherwise, the access permission verification is deemed to have failed. If the access permission verification passes, the system will output "Allow Redirect" and allow the request to proceed, then proceed to the UI stream rendering. If the access permission verification fails, the system will output "Reject Redirect" and block the request, redirecting the user to the login page, the "No Permission" prompt page, or keeping the current page unchanged according to a preset strategy.

8. The intelligent routing control method for Vue interface based on a large language model according to claim 7, characterized in that, The process of preloading page information required for page rendering for the target route, arranging different interface flows by combining page information, and executing corresponding intelligent route navigation and page rendering for all interface flows includes: Establish several preload points and store all preload points in the loading container; A number of linked information points are deployed in the loading container, and the preloaded points in all positions in the loading container are filled accordingly based on the page information required by the target route. Based on the linkage information points, preload points at different locations are connected to obtain linkage paths, which are used to characterize the rendering order of different page information when the target route preloads the page rendering. When the page information with the earlier rendering order is processed, the rendering data of the page information with the later rendering order in the loading container is directly called to execute the page rendering of the corresponding page information. For each preloaded page, a corresponding interface frame, interface content, and interface switching method are constructed, and standardized based on front-end program rules to obtain the interface flow of the corresponding page information. The interface flow includes the routing path, data loading sequence, and page rendering window. Based on the arrangement of page information on the linkage path, the corresponding interface flows of each page information are combined to complete the display order of the interface flows, and the corresponding intelligent routing navigation and page rendering are performed on all interface flows.

Images