A method, system, and device for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface.

Abstract

This invention relates to the field of interactive technology based on graphical user interfaces (GUIs), specifically a method, system, and device for dynamic rendering interaction of intelligent agent streaming data based on a GUI. The method includes: progressively presenting the target text stream in a first view window as the streaming progresses, forming a text progress bar; extracting local state data that is streamed along with boundary markers; initiating initial data reconstruction in parallel in a second view window, and simultaneously establishing dynamic placeholder anchors in the second view window; continuously and incrementally reconstructing the underlying data of the second view window, and rendering a functional canvas within the dynamic placeholder anchors in real time, coordinating with the text progress stepping of the first view window. This ensures users receive low-latency, word-by-word streaming text feedback while achieving smooth transitions and spatiotemporal updates of heterogeneous functional canvases at the spatial geometric level.

Description

Technical Field

[0001] This invention relates to the field of interactive technology based on graphical user interfaces, and in particular to a method, system, and device for dynamic rendering and interaction of intelligent agent streaming data based on graphical user interfaces. Background Technology

[0002] With the explosive growth of artificial intelligence and Large Language Model (LLM) technology, agents are playing an increasingly important interactive role in graphical user interfaces (GUIs). Currently, when agents output data to the front-end user interface, they commonly use streaming technologies based on server-sent events (SSE) or WebSockets to present text output in a word-by-word, progressive manner on the front end, reducing perceived latency for the user.

[0003] However, the interactive output of modern intelligent agents is no longer limited to plain text, but is frequently accompanied by the dynamic generation of heterogeneous functional components (such as automated data charts, logical orchestration streams, interactive operation canvases, etc.). During the streaming process, the structured data (such as JSON, XML, or structured configuration fields in specific formats) on which these heterogeneous components depend flows to the front end in a highly fragmented and incomplete state, mixed sequentially in the text stream.

[0004] Traditional front-end UI rendering engines suffer from severe black-box and technical gaps in such streaming large-model interaction scenarios: If the front-end rendering engine adopts a safe waiting strategy, that is, it must wait until all the streaming data corresponding to the current heterogeneous component has been received and the data closure is completed before calling the main thread to perform full rendering, it will cause the interface to have a long white screen or lag, weakening the real-time response advantage brought by streaming transmission; If the front-end rendering engine adopts a real-time dynamic parsing strategy, that is, it performs high-frequency trial parsing of unclosed and broken data, since the final spatial geometric parameters, physical viewport outline or deep data model of streaming data cannot be predicted before it is fully received, it will cause large-scale and high-frequency layout reflow and interface flickering in the front-end view as the data source continues to flow in. In severe cases, it may even cause the parser to crash due to unclosed syntax, directly blocking the main rendering thread of the graphical interface.

[0005] In short, there is a lack of a dynamic rendering interaction scheme in the existing technology that can safely and with low latency eliminate the spatiotemporal logical barriers between high-frequency progressive text synchronization display and incremental reconstruction of the underlying heterogeneous data model. Summary of the Invention

[0006] In this section, as well as in the abstract and title of this application, some simplifications or omissions may be made to avoid obscuring the purpose of this section, the abstract, and the title of this application, and such simplifications or omissions shall not be used to limit the scope of the invention.

[0007] To address the spatiotemporal logical barrier between high-frequency progressive text synchronization and incremental reconstruction of the underlying heterogeneous data model in existing technologies, one objective of this invention is to provide a dynamic rendering and interaction method for intelligent agent streaming data based on a graphical user interface. This method includes: progressively presenting a target text stream and forming a text progress bar in a first view window, following the streaming progress; capturing boundary markers embedded in the target text stream in real time during the presentation of the target text stream, and extracting local state data that is streamed along with the boundary markers; initiating initial data reconstruction in parallel in a second view window based on the local state data, calculating the viewport layout parameters of the second view window according to the current reconstruction progress, and simultaneously establishing dynamic placeholder anchor points in the second view window; receiving subsequent incremental streaming data, continuously incrementally reconstructing the underlying data of the second view window, and rendering a functional canvas that coordinates with the text progress stepping of the first view window within the dynamic placeholder anchor points.

[0008] As a preferred embodiment of the intelligent agent streaming data dynamic rendering interaction method based on a graphical user interface of the present invention, the step of capturing the boundary markers embedded in the target text stream in real time and extracting the local state data that is streamed along with the boundary markers includes: reading the data blocks of the target text stream during the streaming transmission process, wherein each data block contains a type identifier, content data, and metadata as boundary markers; during the presentation of the target text stream, sequentially scanning the character sequence in the target text stream, and in response to the start guide identifier of the captured boundary marker, driving the streaming state machine to switch from the text presentation state to the data parsing state; in response to the streaming state machine switching from the text presentation state to the data parsing state, determining the target component type according to the type identifier in the current data block, and extracting the content data or metadata in the current data block as local state data.

[0009] As a preferred embodiment of the intelligent agent streaming data dynamic rendering interaction method based on the graphical user interface of the present invention, the step of calculating the viewport layout parameters of the second view window according to the current reconstruction progress includes: performing structured field retrieval on the metadata of the data block to extract standardized spatial coordinates; obtaining the current viewport scaling factor and canvas displacement parameters of the second view window; performing a product scaling transformation on the standardized spatial coordinates and the viewport scaling factor to obtain intermediate scaling coordinates, and performing displacement vector superposition of the intermediate scaling coordinates and the canvas displacement parameters to generate target display layout coordinates, which are used as the viewport layout parameters of the second view window.

[0010] As a preferred embodiment of the intelligent agent streaming data dynamic rendering interaction method based on the graphical user interface of the present invention, the steps of establishing a dynamic placeholder anchor point in the second view window and rendering a functional canvas that coordinates with the text progress step of the first view window in real time within the dynamic placeholder anchor point include: dividing a target graphic area within the second view window according to the target display layout coordinates, and activating a transition placeholder component with a preset geometric contour within the target graphic area to establish the dynamic placeholder anchor point; during the continuous incremental reconstruction of the underlying data, monitoring the completion progress parameter of the streaming incremental data in real time, and based on the completion progress parameter, synchronously decreasing the first opacity component of the transition placeholder component and correspondingly increasing the second opacity component of the functional canvas to perform pixel-level overlapping gradient between the dynamic placeholder anchor point and the functional canvas until the transition placeholder component is converted to a hidden state and the functional canvas is converted to a presented state within the target graphic area.

[0011] As a preferred embodiment of the intelligent agent streaming data dynamic rendering interaction method based on the graphical user interface of the present invention, it further includes: when the action intention identified from the underlying data of the second view window matches the operation type in the pre-stored service interface library; retrieving the preset feature identifier contained in the streaming incremental data, reading the parameter data bound to the preset feature identifier, filling the parameter data into the parameter item corresponding to the operation type in the service interface library, so as to assemble the standardized request body; and distributing the standardized request body to the external system in a non-blocking manner in a background worker thread independent of the main rendering thread of the interface, so as to asynchronously call the target service interface corresponding to the action intention.

[0012] As a preferred embodiment of the intelligent agent streaming data dynamic rendering interaction method based on a graphical user interface of the present invention, the method further includes: when distributing a standardized request body to an external system in a non-blocking manner, the method further includes: in response to the recognition of the action intent, obtaining the current execution timestamp and introducing a preset random factor, concatenating the execution timestamp and the preset random factor to obtain a unique verification feature code; writing the unique verification feature code into a reserved header data item in the standardized request body; distributing the standardized request body carrying the unique verification feature code, so that when the external system retryes the network, it can compare with the stored historical execution log to see if there is a historical feature code that matches the unique verification feature code, and refuse to execute the same target service interface call a second time if the comparison result is that it exists.

[0013] As a preferred embodiment of the intelligent agent streaming data dynamic rendering interaction method based on the graphical user interface of the present invention, it further includes: persistently storing the text progress of the first view window and the display update status of the function canvas of the second view window in real time as a time-series status log containing a log timestamp field, a component status field, and a service call result field; configuring a timeline interaction control in the current interface containing the first view window and the second view window, and in response to the trigger command of the timeline interaction control, reading the time-series status log, and reconstructing the text presentation status and function canvas of the corresponding historical moment in the first view window and the second view window according to the linear order of the log timestamp field.

[0014] As a preferred embodiment of the intelligent agent streaming data dynamic rendering interaction method based on the graphical user interface of the present invention, it further includes: configuring a graphics interruption control component in the function canvas; in response to a user-triggered event on the graphics interruption control component, a background listening thread distributes a thread termination instruction to the first worker thread responsible for streaming rendering, so that the first worker thread terminates the current streaming data reading loop, switches the second view window to a suspended state, and simultaneously closes the network socket connection corresponding to the first worker thread and reclaims the memory address space.

[0015] Another objective of this invention is to provide a graphical user interface-based intelligent agent streaming data dynamic rendering interaction system, employing the aforementioned graphical user interface-based intelligent agent streaming data dynamic rendering interaction method, comprising: a view synchronization module, used to provide a parallel first view window and a second view window, and in the first view window, progressively presenting the target text stream and forming text progress according to the streaming progress; a boundary capture module, used to capture the boundary markers embedded in the target text stream in real time during the continuous presentation of the target text stream in the first view window, and extract the local state data that is streamed along with the boundary markers; a parallel reconstruction module, used to initiate initial data reconstruction in parallel in the second view window based on the local state data, calculate the viewport layout parameters of the second view window according to the current reconstruction progress, and synchronously establish dynamic placeholder anchor points in the second view window; and an incremental collaborative rendering module, used to receive subsequent streaming incremental data, continuously incrementally reconstruct the underlying data of the second view window, and render a functional canvas that is in real time coordinated with the text progress step of the first view window within the dynamic placeholder anchor points.

[0016] Another object of the present invention is to provide an electronic device including one or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and when the computer programs are executed by the one or more processors, the electronic device performs the above-described method for dynamic rendering interaction of intelligent agent streaming data based on a graphical user interface.

[0017] The beneficial effects of this invention are as follows: By constructing parallel first and second view windows, this invention decouples the progressive presentation of the plain text stream from the incremental reconstruction of the heterogeneous functional canvas in terms of space and thread. While the first view window performs highly real-time text presentation, boundary capture is performed on the text stream in real time, thereby enabling the early extraction of local state data representing the component architecture from the fragmented streaming data stream. Based on this local state data, in the dynamic logical gaps where the underlying data is not yet fully closed, initialization data reconstruction is initiated in parallel in advance, completing the prediction of the physical viewport parameters of the second view window and establishing dynamic placeholder anchor points.

[0018] Breaking away from the traditional black-box approach of waiting for full-scale streaming component rendering, this approach enables continuously flowing incremental streaming data to be continuously completed and collaboratively rendered within pre-defined dynamic placeholder anchor points in an aligned and data-driven incremental manner. While ensuring users receive low-latency, word-by-word streaming text feedback, it achieves smooth transitions and spatiotemporal updates of heterogeneous functional canvases at the spatial geometry level, eliminating technical defects such as high-frequency redrawing of the graphical interface, white screen deadlocks, or flickering collapses caused by data corruption. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a display interface diagram of the intelligent agent streaming data dynamic rendering and interaction method based on a graphical user interface according to the present invention.

[0021] Figure 2 This is a flowchart of the local state data of the intelligent agent streaming data dynamic rendering interaction method based on the graphical user interface of the present invention.

[0022] Figure 3 This is a flowchart illustrating the display of the functional canvas of the intelligent agent streaming data dynamic rendering and interaction method based on a graphical user interface according to the present invention.

[0023] Figure 4This is a schematic diagram of the operation of the timeline interactive control in the intelligent agent streaming data dynamic rendering interaction method based on the graphical user interface of the present invention.

[0024] Figure 5 This is a schematic diagram of the system structure of the intelligent agent streaming data dynamic rendering and interaction system based on a graphical user interface according to the present invention.

[0025] In the diagram: 101, First view window; 101a, Target text stream; 102, Second view window; 102a, Dynamic placeholder anchor point; 102b, Timeline interactive control; 103, Streaming state machine; 200, View synchronization module; 300, Boundary snapping module; 400, Parallel reconstruction module; 500, Incremental collaborative rendering module. Detailed Implementation

[0026] To make the objectives, features, and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0027] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0028] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0029] Example 1, see Figures 1-4 This is the first embodiment of the present invention, which provides a method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface.

[0030] It is worth noting that, in terms of technical and physical aspects, the intelligent agent in the embodiments of the present invention specifically refers to an artificial intelligence large language model (LLM) inference engine deployed on a cloud server or a local computing chip, or an automated instruction execution process with streaming data output capability.

[0031] First, in the first view window 101 of the graphical user interface, the system receives data continuously distributed by the network side via a streaming protocol. At this time, the graphics rendering thread progressively (e.g., word-by-word, token-by-token) presents the target text stream 101a in the first view window 101 as the streaming progresses. As the target text stream 101a extends, it visually accumulates to form a clear text progress, serving as an immediate perceptual clue for the user to understand the evolution of the intelligent agent's thinking.

[0032] During the progressive presentation of the target text stream 101a, the system does not treat the input stream as a plain text black box. Instead, it initiates a background interception stream to capture implicitly embedded boundary markers in the target text stream 101a in real time. These boundary markers are non-explicit control characters embedded by the agent on the network side according to a preset data protocol. Once a boundary marker is captured, it signifies the start of input of a structured data block for a specific interactive function canvas. The system then initiates an extraction process, instantly stripping away the local state data that is streamed along with the boundary markers. At this point, the local state data is typically an unclosed, incomplete, or broken fragment in terms of syntactic structure.

[0033] Following this, the system overcomes logical gaps where data is not fully closed, and initiates initial data reconstruction in parallel within an independent second view window 102 based on local state data. The system does not passively wait for data transmission to complete; instead, it predicts and calculates the viewport layout parameters of the second view window 102 based on the current reconstruction progress. Specifically: First, the current reconstruction progress refers to the cumulative completeness percentage of structured data blocks input and received in real-time with the streaming protocol. When the streaming state machine 103 switches from text presentation to data parsing, the overall data reconstruction progress is in the initialization phase because the specific content data has not yet been transmitted. Since the streaming protocol stipulates that metadata representing the topological shell of the target component takes precedence over the transmission of specific business payload content, the system can extract the standardized spatial coordinates of the target component before the content data arrives by performing structured parsing on the first arriving data blocks during the initialization phase. The standardized spatial coordinates include standard relative horizontal coordinates, standard relative vertical coordinates, standard relative width, and standard relative height, used to represent the relative geometric contour of the target component.

[0034] Secondly, the system obtains the physical environment parameters of the current second view window 102 in real time by calling the application programming interface of the front-end graphics system. The physical environment parameters include the current viewport scaling factor and the current canvas displacement parameters. The viewport scaling factor includes the horizontal and vertical physical scaling ratios, reflecting the current pixel density of the display screen and the user's dynamic scaling state of the window; the canvas displacement parameters include the horizontal and vertical translation distances, reflecting the real-time physical displacement of the functional canvas reference origin in the second view window 102 relative to the absolute pixel origin of the display screen.

[0035] Finally, the graphics processor invokes its internal computation pipeline to perform the following linear spatial transformation control on the standardized spatial coordinates extracted earlier and the real-time acquired physical environment parameters: Performing a product scaling transformation: The system multiplies the standard relative width with the horizontal physical scaling ratio to obtain the physical pixel width of the target component; simultaneously, it multiplies the standard relative height with the vertical physical scaling ratio to obtain the physical pixel height of the target component. Similarly, the standard relative horizontal and vertical coordinates are multiplied by the horizontal and vertical physical scaling ratios, respectively, to obtain the intermediate scaled horizontal and vertical coordinates.

[0036] Perform displacement vector superposition: The system sums and superimposes the intermediate scaled x-coordinates calculated above with the horizontal translation distance of the current functional canvas to determine the target physical absolute x-coordinate of the target component in the current display screen absolute coordinate system; at the same time, it sums and superimposes the intermediate scaled y-coordinates calculated above with the vertical translation distance of the current functional canvas to determine the target physical absolute y-coordinate of the target component in the current display screen absolute coordinate system.

[0037] After the above two-step spatial transformation process, the system finally outputs the target display layout coordinates, which consist of the target's physical absolute horizontal coordinate, the target's physical absolute vertical coordinate, the physical pixel width, and the physical pixel height, as the viewport layout parameters of the second view window 102.

[0038] Using the calculated viewport layout parameters, the system synchronously establishes a dynamic placeholder anchor point 102a in the second view window 102. The dynamic placeholder anchor point 102a matches the final interactive component to be presented in terms of spatial geometry, thereby locking the geometric space of the functional canvas in advance.

[0039] Finally, the system continuously receives subsequent streaming incremental data. This data is specifically designed to complete the functional canvas. After receiving the streaming incremental data, the system continuously and incrementally reconstructs the underlying data of the second view window 102, gradually closing the underlying structured data model. Simultaneously, within the established dynamic placeholder anchor point 102a, the functional canvas is rendered in real-time, coordinating with the text progress in the first view window 101. Through this step-by-step coordination between the two windows (first view window 101 and second view window 102), while text is streaming out in the first view window 101, the functional canvas in the second view window 102 also grows and evolves smoothly and synchronously.

[0040] By capturing boundary markers in real time during the presentation of the target text stream 101a in the first view window 101, the system can obtain local state data in advance. This allows the system to predict viewport layout parameters and establish dynamic placeholder anchor points 102a in the second view window 102. This eliminates the white screen stuttering caused by incomplete data from heterogeneous components, enabling subsequent received streaming incremental data to smoothly reconstruct a functional canvas that coordinates with the text progress steps within the dynamic placeholder anchor point 102a, thus eliminating the technical risks of frequent interface redraws and spatial layout flickering.

[0041] Example 2, see Figures 1-4 This is the second embodiment of the present invention, which is based on embodiment 1.

[0042] Specifically, in the micro-steps of capturing the boundary markers embedded in the target text stream 101a in real time and extracting the local state data that is streamed along with the boundary markers, the system first reads the data blocks of the target text stream 101a during the streaming process. Each data block is packaged into an independent frame or message at the transport layer, and its structure includes a type identifier, content data, and metadata as boundary markers. During the rendering of the target text stream 101a, the system's underlying scanner sequentially scans the character sequences in the target text stream 101a. When the scanner matches a specific character combination and responds to the capture of the start indicator of the boundary marker (such as a specific escape character or a custom component start tag), the system drives a streaming state machine 103 to undergo a state transition, that is, switch from the text rendering state to the data parsing state. Once the streaming state machine 103 switches from the text rendering state to the data parsing state, the system turns off the text echoing to the front-end interface and instead starts the underlying deserialization. The system determines the target component type based on the type identifier in the current data block (e.g., whether the current incoming data is an automated chart component or a logic tree component), and extracts the content data or metadata from the current data block as local state data. Utilizing the streaming state machine 103, the system switches from text rendering state to data parsing state when it captures the starting guide identifier of the boundary marker. This allows for the independent separation of plain text display and component metadata parsing within a single data stream, ensuring the accuracy of local state data extraction.

[0043] Preferably, when calculating the viewport layout parameters of the second view window 102 based on the current reconstruction progress, the system first performs structured field retrieval on the metadata of the data blocks to extract standardized spatial coordinates. Standardized spatial coordinates are virtual coordinates that are absolutely proportional and independent of a specific screen resolution. Next, the system obtains the current viewport scaling factor and canvas displacement parameters of the second view window 102 through the graphics system interface. The system performs a product scaling transformation on the standardized spatial coordinates and the viewport scaling factor to obtain intermediate scaling coordinates, which are used to adapt to the current screen pixel density. Subsequently, the system performs displacement vector superposition on the intermediate scaling coordinates and the canvas displacement parameters, completing the mapping from the relative coordinate system to the absolute screen coordinate system through vector superposition, thereby generating the target display layout coordinates as the viewport layout parameters of the second view window 102. By performing a product scaling transformation on the standardized spatial coordinates and the viewport scaling factor, and then performing displacement vector superposition with the canvas displacement parameters, the resolution differences between different physical hardware terminals can be smoothed out. The target display layout coordinates are calculated in the pre-reconstruction stage, avoiding viewport offset or image stretching of the second view window 102 during subsequent rendering.

[0044] Preferably, in the micro-step of establishing a dynamic placeholder anchor point 102a in the second view window 102 and performing pixel-level overlapping gradient, the system first divides a precise target graphic area within the second view window 102 based on the target display layout coordinates calculated in the previous step. The system then activates a transition placeholder component with a preset geometric contour (e.g., a dynamic skeleton component with a wavy skeleton screen texture and a gray semi-transparent contour) within the target graphic area, thereby establishing a dynamic placeholder anchor point 102a in the user's vision and providing the user with pre-emptive feedback that the component is operating in parallel. During the continuous incremental reconstruction of the underlying data, the system monitors the completion progress parameters of the streaming incremental data in real time (e.g., by calculating the proportion of received bytes to the total number of bytes in the metadata declaration). Based on the completion progress parameters, the system activates a bidirectional opacity driving engine during rendering: synchronously decreasing the first opacity component of the transition placeholder component (e.g., linearly decreasing it from 1.0 to 0.0), and correspondingly increasing the second opacity component of the functional canvas (e.g., linearly increasing it from 0.0 to 1.0). Therefore, the system performs a pixel-level cross-fade animation between the dynamic placeholder anchor point 102a and the functional canvas at the physical layer until the transition placeholder component smoothly transitions to a completely hidden state, and the functional canvas transitions to a full-color display state within the target graphic area. Within the target graphic area, the transition placeholder component is activated to establish the dynamic placeholder anchor point 102a. By synchronizing the completion progress parameters and adjusting the first and second opacity components in reverse, a pixel-level cross-fade animation between the transition placeholder component and the functional canvas is achieved, eliminating abrupt visual jumps and interface flickering during the dynamic loading of the transition placeholder component.

[0045] Furthermore, when the action intent identified from the underlying data of the second view window 102 (such as a user click or an external API call instruction generated autonomously by the agent) matches the operation type in the pre-stored service interface library within the system, the system initiates external request assembly. The system retrieves preset feature identifiers flowing in with the streaming incremental data and reads the parameter data bound to the preset feature identifiers. Subsequently, the system fills the parameter data into the parameter items in the service interface library corresponding to the operation type (for example, filling the device ID extracted from the streaming into the Slot of the standard control API), thereby automatically assembling a standardized request body. In order not to block the user's interactive interface, the system distributes the standardized request body to the external system in a non-blocking manner in a background worker thread independent of the main interface rendering thread, so as to asynchronously call the target service interface corresponding to the action intent. By assembling the standardized request body by filling the extracted parameter data into the corresponding parameter items in the service interface library and distributing it in a non-blocking manner in an independent background worker thread, the system isolates the time-consuming network I / O communication from the main interface rendering thread, ensuring that even if the external system responds slowly, the streaming text presentation and interaction of the interface will not freeze.

[0046] Furthermore, when distributing standardized request bodies to external systems in a non-blocking manner, the system responds to the recognition of the action intent by obtaining the current high-precision execution timestamp and introducing a preset random factor (such as a set of encrypted random noise). The system concatenates the execution timestamp with the preset random factor (or performs a hash operation) to generate a unique verification signature. The system writes the unique verification signature into the reserved header field in the standardized request body. When network jitter triggers the distribution of a standardized request body carrying the unique verification signature, the external system, upon receiving the network retry request, will compare its stored historical execution logs to see if a historical signature matching the unique verification signature exists. If the external system finds a match, it determines that the request is a duplicate request caused by a network timeout retry, thus rejecting the second execution of the same target service interface call (implementing server-side idempotency control). By using the execution timestamp and a preset random factor to generate a unique verification feature code and writing it into the header data item of the standardized request body, the external system can perform deduplication comparison based on the historical feature code during network retries. This avoids the risk of the external system repeatedly executing the same target service interface call due to network jitter retries at the protocol level.

[0047] Furthermore, during runtime, the system captures a full snapshot of the text progress in the first view window 101 and the display update status of the function canvas in the second view window 102 in real time, and persists this snapshot as a time-series status log containing log timestamp fields, component status fields, and service call result fields. A timeline interactive control 102b is configured in the current interface containing the first view window 101 and the second view window 102. In response to user drag-and-click trigger commands on the timeline interactive control 102b, the system reads the time-series status log and performs a reverse or forward retrieval of the log according to the linear order of the log timestamp fields. The system sequentially reconstructs the text presentation status and function canvas at corresponding historical moments in the first view window 101 and the second view window 102. By persistently storing the text progress and function canvas display update status as a time-series status log containing multi-dimensional fields, and using the timeline interactive control 102b for linear sequential reading, the system gains content auditing and state restoration capabilities, improving the usability of the intelligent agent interaction.

[0048] Furthermore, the system configures an explicit graphics interruption control component in the function canvas (such as a red interactive button to stop generation or interrupt rendering). In response to a user-triggered event on the graphics interruption control component (e.g., the user notices the large model output deviates from expectations and actively calls for a halt), a persistent background listener thread dispatches a thread termination command to the first worker thread responsible for streaming rendering. Upon receiving the command, the first worker thread forcibly terminates the current streaming data read loop, disconnects the data receiving stream, and switches the second view window 102 to a suspended state. Simultaneously, the system closes the underlying network socket connection corresponding to the first worker thread and forcibly reclaims the memory address space it occupies, restoring the system to a clean initial waiting state. Configuring the graphics interruption control component in the function canvas and dispatching a thread termination command to the first worker thread via a background listener thread to terminate the streaming data read loop and close the network socket connection establishes an underlying circuit breaker mechanism. This mechanism releases network and memory resources, preventing local resource leaks and UI freezes caused by uncontrolled large model output.

[0049] Example 3, see Figures 1-5 This embodiment is the third embodiment of the present invention. This embodiment is based on Embodiment 1 and Embodiment 2. This embodiment provides an intelligent agent streaming data dynamic rendering interaction system based on a graphical user interface.

[0050] The view synchronization module 200 serves as the core of the interface control, providing parallel first view window 101 and second view window 102. Its internal rendering driving logic controls the first view window 101 to progressively present the target text stream 101a in accordance with the streaming progress, thereby forming text progress.

[0051] The boundary capture module 300 acts as a data stream interceptor, establishing a data-level cascade with the view synchronization module 200. It is used to capture implicitly embedded boundary markers in the physical network stream in real time during the continuous presentation of the target text stream 101a in the first view window 101, and to extract local state data that is streamed along with the boundary markers.

[0052] The parallel reconstruction module 400 serves as a spatial prediction and initialization engine, with its input connected to the output of the boundary capture module 300. It is used to initiate parallel initialization data reconstruction based on local state data within a logically isolated second view window 102, calculate the viewport layout parameters of the second view window 102 according to the current reconstruction progress, and simultaneously establish dynamic placeholder anchor points 102a within the second view window 102.

[0053] The incremental collaborative rendering module 500, as the ultimate reconstruction and hybrid rendering engine, maintains a synchronization lock with the view synchronization module 200 and the parallel reconstruction module 400, respectively. It is used to receive subsequent streaming incremental data, continuously incrementally reconstruct the underlying data of the second view window 102, and render a functional canvas that coordinates with the text progress step of the first view window 101 in real time within the dynamic placeholder anchor point 102a.

[0054] This embodiment also provides an electronic device. The electronic device, at the hardware physical level, includes: one or more processors (such as a central processing unit (CPU), a graphics processing unit (GPU), or a specific artificial intelligence acceleration chip core); memory (such as high-speed dynamic random access memory (DRAM), flash memory, etc.); and one or more computer programs. The one or more computer programs are statically or dynamically stored in the memory. When the computer program is invoked and executed by one or more processors, the processor's underlying microcode control logic drives the hardware pipeline inside the electronic device, strictly executing each step of the above-described intelligent agent streaming data dynamic rendering interaction method based on a graphical user interface.

[0055] To facilitate understanding of the technical solution of this invention, its working principle is explained in detail below: When the agent starts continuously distributing streaming data to the front end, the view synchronization module 200 starts the first working thread and types out the target text stream 101a word by word in the first view window 101 in the form of streaming text.

[0056] During this period, the boundary capture module 300 maintains high-frequency monitoring. Once an implicit start guide marker is detected in the text, the streaming state machine 103 transitions, cuts off text display, and starts structured chunked deserialization to extract local state data containing type, content, and metadata.

[0057] After obtaining the local state data, the parallel reconstruction module 400 takes over the calculation, extracts the standardized spatial coordinates from the metadata in the second view window 102, and predicts the target display layout coordinates through matrix-vector transformation of the viewport scaling factor and canvas displacement parameters. Based on these target display layout coordinates, the system delineates the target graphic area in the second view window 102 and illuminates a transition placeholder component with a skeleton screen texture, forming a dynamic placeholder anchor point 102a.

[0058] Subsequently, as more and more streaming incremental data is continuously added, the incremental collaborative rendering module 500 continuously completes the underlying broken data model. During this process, based on the completion percentage of the streaming data, the system synchronously lowers the opacity of the transition placeholder component and raises the opacity of the final functional canvas at opposite slopes, so that the functional canvas gradually unfolds within the dynamic placeholder anchor point 102a in a pixel-level seamless gradient.

[0059] If external calls are involved, the system assembles a standardized request body with a unique verification signature in a separate background worker thread, enabling non-blocking asynchronous network calls with idempotent anti-duplicate functionality. Simultaneously, the progress of the entire interaction process is persistently written to a time-series status log in real time. Users can drag the progress bar at any time using the timeline interactive control 102b to reproduce the text presentation and function canvas snapshots at any historical moment. They can also click the graphical interruption control component in the function canvas at any time to circuit-break the network socket connection of the first worker thread and reclaim memory address space through the background listening thread, thus achieving a closed loop in the interaction cycle.

[0060] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface, characterized in that, include: In the first view window (101), the target text stream (101a) is presented progressively with the streaming progress and a text progress is formed; During the presentation of the target text stream (101a), the boundary markers embedded in the target text stream (101a) are captured in real time, and the local state data that is streamed along with the boundary markers is extracted. Based on local state data, the initialization data reconstruction is started in parallel in the second view window (102). The viewport layout parameters of the second view window (102) are calculated according to the current reconstruction progress, and dynamic placeholder anchor points (102a) are established in the second view window (102) simultaneously. Receive subsequent streaming incremental data, continuously incrementally reconstruct the underlying data of the second view window (102), and render a functional canvas that coordinates with the text progress step of the first view window (101) in real time within the dynamic placeholder anchor point (102a).

2. The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface according to claim 1, characterized in that, The steps of capturing the embedded boundary markers in the target text stream (101a) in real time and extracting the local state data that is streamed along with the boundary markers include: Read the data blocks of the target text stream (101a) during the streaming process, wherein each data block contains a type identifier as a boundary marker, content data and metadata; During the presentation of the target text stream (101a), the character sequence in the target text stream (101a) is scanned sequentially, and in response to the capture of the start guide flag of the boundary marker, the streaming state machine (103) is driven to switch from the text presentation state to the data parsing state. In response to the streaming state machine (103) switching from text presentation state to data parsing state, the target component type is determined based on the type identifier in the current data block, and the content data or metadata in the current data block is extracted as local state data.

3. The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface according to claim 2, characterized in that, The steps for calculating the viewport layout parameters of the second view window (102) based on the current reconstruction progress include: The metadata of the data blocks is retrieved using structured fields to extract standardized spatial coordinates; Obtain the current viewport scaling factor and canvas displacement parameters of the second view window (102); The standardized spatial coordinates are multiplied and scaled by the viewport scaling factor to obtain intermediate scaling coordinates. The intermediate scaling coordinates are then superimposed with the canvas displacement parameters to generate target display layout coordinates, which are used as the viewport layout parameters of the second view window (102).

4. The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface according to claim 3, characterized in that, The steps of establishing a dynamic placeholder anchor point (102a) in the second view window (102) and rendering a functional canvas that coordinates with the text progress stepping of the first view window (101) in real time within the dynamic placeholder anchor point (102a) include: Based on the target display layout coordinates, the target graphic area is divided in the second view window (102), and a transition placeholder component with a preset geometric contour is activated in the target graphic area to establish the dynamic placeholder anchor point (102a). During the continuous incremental reconstruction of the underlying data, the completion progress parameter of the streaming incremental data is monitored in real time. Based on the completion progress parameter, the first opacity component of the transition placeholder component is decreased synchronously, and the second opacity component of the functional canvas is increased accordingly to perform pixel-level overlapping gradient between the dynamic placeholder anchor point (102a) and the functional canvas until the transition placeholder component is converted to a hidden state and the functional canvas is converted to a presented state within the target graphic area.

5. The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface according to any one of claims 1 to 4, characterized in that, Also includes: When the action intent identified from the underlying data of the second view window (102) matches the operation type in the pre-stored service interface library, Retrieve the preset feature identifier contained in the streaming incremental data, read the parameter data bound to the preset feature identifier, and fill the parameter data into the parameter item corresponding to the operation type in the service interface library to assemble a standardized request body; In a background worker thread independent of the main UI rendering thread, standardized request bodies are distributed to external systems in a non-blocking manner to asynchronously call the target service interface corresponding to the action intent.

6. The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface according to claim 5, characterized in that, When distributing standardized request bodies to external systems in a non-blocking manner, it also includes: In response to the recognition of the action intent, the current execution timestamp is obtained and a preset random factor is introduced. The execution timestamp and the preset random factor are concatenated to obtain a unique verification feature code. Write the unique verification signature into the reserved header data field in the standardized request body; Distribute a standardized request body carrying a unique verification signature, so that when the external system retryes the network, it can compare the stored historical execution logs to see if there is a historical signature that matches the unique verification signature, and refuse to execute the same target service interface call a second time if the comparison result is that it exists.

7. The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface according to any one of claims 1 to 4, characterized in that, Also includes: The text progress of the first view window (101) and the display update status of the function canvas of the second view window (102) are continuously stored as a time-series status log containing log timestamp fields, component status fields and service call result fields. A timeline interactive control (102b) is configured in the current interface containing the first view window (101) and the second view window (102). In response to a trigger command for the timeline interactive control (102b), the timeline status log is read, and the text presentation status and function canvas of the corresponding historical moment are reconstructed in the first view window (101) and the second view window (102) according to the linear order of the log timestamp field.

8. The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface according to claim 7, characterized in that, Also includes: Configure the graphics interrupt control component in the function canvas. In response to a user-triggered event for the graphics interrupt control component, the background listening thread distributes a thread termination instruction to the first worker thread responsible for streaming rendering, so that the first worker thread terminates the current streaming data reading loop, switches the second view window (102) to the suspended state, closes the network socket connection corresponding to the first worker thread, and reclaims the memory address space.

9. A dynamic rendering and interactive system for intelligent agent streaming data based on a graphical user interface, characterized in that, The method for dynamic rendering and interaction of intelligent agent streaming data based on a graphical user interface as described in claim 8 includes: The view synchronization module (200) is used to provide a first view window (101) and a second view window (102) in parallel, and in the first view window (101), the target text stream (101a) is presented progressively with the progress of the streaming transmission and a text progress is formed; The boundary capture module (300) is used to capture the boundary markers embedded in the target text stream (101a) in real time during the continuous presentation of the target text stream (101a) in the first view window (101), and extract the local state data that is output in a streaming manner along with the boundary markers. The parallel reconstruction module (400) is used to initiate initial data reconstruction in parallel in the second view window (102) based on local state data, calculate the viewport layout parameters of the second view window (102) according to the current reconstruction progress, and simultaneously establish dynamic placeholder anchor points (102a) in the second view window (102). The incremental collaborative rendering module (500) is used to receive subsequent streaming incremental data, continuously incrementally reconstruct the underlying data of the second view window (102), and render the functional canvas that is in real time with the text progress stepping of the first view window (101) within the dynamic placeholder anchor point (102a).

10. An electronic device, characterized in that, include: One or more processors; Memory; And one or more computer programs, wherein the one or more computer programs are stored on the memory, and when the computer programs are executed by the one or more processors, cause the electronic device to perform the intelligent agent streaming data dynamic rendering interaction method based on a graphical user interface as described in claim 8.

Images