An intelligent agent response traceable graphic-textual visualized backtracking method

Abstract

This application provides a graphical and textual visualization backtracking method for agent response tracing, comprising: a server, executing: intercepting response feedback data and its corresponding response tracing data output by the agent, adding agreed-upon tags to the response tracing data to obtain processed tracing data, and sending the response feedback data and its corresponding processed tracing data to the client; and a client, executing: generating display labels corresponding to the processed tracing data, displaying the response feedback data and its corresponding display labels using a rendering layer, and, in response to click events on the display labels, rendering and displaying the processed tracing data using the graphical and textual display methods associated with the display labels. The display labels are labels that can be rendered using a formatted language mounted on the client. By converting the tracing data into display labels associated with different graphical and textual display methods, the limitations of existing client display format specifications are overcome, enabling a more comprehensive display of the tracing data.

Description

Technical Field

[0001] This application relates to the field of graphic and textual backtracking technology, and in particular to a graphic and textual visualization backtracking method for tracing the source of an agent's response. Background Technology

[0002] Among the existing technologies used by intelligent agent platforms to render and display output content, the standard Markdown format is mainly used for front-end rendering. However, Markdown, as a lightweight markup language, was originally designed for basic text layout. Its native specification does not have the technical capability to support the integrated rendering of complex text and graphics. In other words, the standard Markdown format has inherent limitations in complex information visualization scenarios, resulting in the inability to accurately and interactively present source information. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide at least one graphical visualization backtracking method for intelligent agent response tracing, which breaks through the limitations of existing display format specifications of the client by converting the tracing data into display labels associated with different graphical display methods, so that the tracing data can be displayed more comprehensively.

[0004] This application mainly includes the following aspects: In a first aspect, embodiments of this application provide a graphical and textual visualization backtracking method for tracing the source of an agent's response. The method includes: intercepting response feedback data output by the agent and its corresponding response tracing data; adding a pre-defined marker to the response tracing data to obtain processing tracing data; generating display labels corresponding to the processing tracing data; displaying the response feedback data and its corresponding display labels using a rendering layer; and, in response to a click event on the display label, rendering and displaying the processing tracing data using a graphical and textual display method associated with the display label. The display label is a label format that can be rendered by a formatted language mounted on the client.

[0005] In one possible implementation, the response feedback data includes multiple response data frames, the response tracing data includes multiple tracing data frames corresponding one-to-one with the multiple response data frames, and the processed tracing data includes multiple tracing data frames corresponding one-to-one with the multiple response data frames. For each response data frame, the response tracing data is parsed according to the given output protocol of the agent to obtain a tracing knowledge base identifier, a tracing file identifier, and a tracing slice identifier. From the target tracing knowledge base corresponding to the tracing knowledge base identifier, the target tracing file indicated by the tracing file identifier is extracted. From the target tracing file, the target knowledge slice corresponding to the tracing slice identifier is extracted. A convention-defined tag is added to the target knowledge slice to obtain the corresponding tracing data frame.

[0006] In one possible implementation, the step of generating a display label corresponding to the processing traceability data includes: continuously receiving data frames in the form of a data stream through a preset long-term connection technology; and passing the received data frames to a traceability data interceptor to identify the traceability data frames and generate a display label corresponding to the traceability data frames.

[0007] In one possible implementation, the source tracing data interceptor performs the following steps: verifying whether the data frame carries a predefined marker; if the data frame carries a predefined marker, then the received data frame is determined to be a source tracing data frame, and a display label corresponding to the source tracing data frame is generated; if the data frame does not carry a predefined marker, then the received data frame is determined to be a non-source tracing data frame, and the data frame is directly output.

[0008] In one possible implementation, the step of generating a display label corresponding to a traceability data frame includes: obtaining a traceability data interaction flag, which indicates whether the user is supported to perform an interactive operation on the target knowledge slice indicated by the display label; if the traceability data interaction flag indicates that the user is not supported to perform an interactive operation on the target knowledge slice, then the traceability data frame is converted into a first display label; if the traceability data interaction flag indicates that the user is supported to perform an interactive operation on the target knowledge slice, then the traceability data frame is converted into a second display label.

[0009] In one possible implementation, after generating a first display tag corresponding to the traceability data frame, the first display tag is associated with the hyperlink display page of the target knowledge slice corresponding to the traceability data frame; after generating a second display tag corresponding to the traceability data frame, the second display tag is associated with the page display interaction module, and the page display interaction module embeds the current display page of the response feedback data.

[0010] In one possible implementation, the step of rendering and displaying the processing traceability data using the graphic display method associated with the display tag in response to a click event on the display tag includes: in response to a click event on the first display tag, jumping to the hyperlink display page associated with the first display tag, and the hyperlink display page rendering the target knowledge slice corresponding to the first display tag.

[0011] In one possible implementation, the step of rendering and displaying the processing traceability data using the graphic display method associated with the display label in response to a click event on the display label further includes: in response to a click event on the second display label, calling the page display interaction module to open an interactive graphic input box on the current display page; and loading the target knowledge slice into the interactive graphic input box for rendering and display.

[0012] Secondly, embodiments of this application also provide a graphical and textual visualization backtracking method for tracing the source of an agent's response, applied to a server. The method includes: intercepting the response feedback data output by the agent and its corresponding response tracing data; adding a pre-defined marker to the response tracing data to obtain processing tracing data; and sending the response feedback data and its corresponding processing tracing data to the client.

[0013] Thirdly, embodiments of this application also provide a graphical and textual visualization backtracking method for intelligent agent response tracing, applied to a client. The method includes: receiving processing tracing data sent by a server; generating display labels corresponding to the processing tracing data; displaying the response feedback data and its corresponding display labels using a rendering layer; and, in response to a click event on a display label, rendering and displaying the processing tracing data using a graphical and textual display method associated with the display label. The display label is a label format that can be rendered by a formatted language mounted on the client.

[0014] This application provides a graphical visualization backtracking method for intelligent agent response tracing, comprising: intercepting response feedback data output by the intelligent agent and its corresponding response tracing data; adding agreed-upon tags to the response tracing data to obtain processed tracing data; sending the response feedback data and its corresponding processed tracing data to a client; generating display labels corresponding to the processed tracing data; displaying the response feedback data and its corresponding display labels using a rendering layer; and, in response to click events on the display labels, rendering and displaying the processed tracing data using the graphical display methods associated with the display labels. The display labels are labels that can be rendered using a formatted language mounted on the client. By converting the tracing data into display labels associated with different graphical display methods, the limitations of existing client display format specifications are overcome, enabling a more comprehensive display of the tracing data.

[0015] The advantages of this application are: 1. Based on the different interactive needs of customers for traceability data, this application designs different display tags to meet the rendering requirements of the formatted language (Markdown) on the client side. Different display tags are associated with graphic and text display methods to meet the different interactive needs of customers for traceability data. By converting traceability data into display tags and displaying them in the corresponding graphic and text display methods when the display tags are triggered, not only is the visual presentation of traceability data customized, but it can also meet the different interactive needs of users for traceability data, so that the presentation of traceability data has the ability to be customized according to user needs.

[0016] 2. During the streaming transmission between the server and the client, the traceability information is processed by predefined convention tags to ensure that the traceability information can be accurately and unambiguously identified and extracted in the data stream.

[0017] 3. When it is necessary to compare the response feedback data output by the intelligent agent with the response tracing data, the method of opening a new page by linking to load knowledge slices cannot meet the needs of this situation. The process is cumbersome, requiring frequent switching between pages and re-issuing the linked page. Neither method is convenient. To address this situation, this application introduces a second display label associated with the corresponding page display interaction module. Through the page interaction module, an interactive graphic input box is introduced into the current page. The target knowledge slice is loaded in the interactive graphic input box. This not only simplifies the operation in this situation but also supports various predefined interactive operations that users can perform on the target knowledge slice, greatly satisfying the interactive needs of different users.

[0018] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This document illustrates a flowchart of a graphical and textual visualization-based backtracking method for tracing the source of an agent's response, provided in an embodiment of this application. Figure 2 This invention provides a flowchart illustrating a method for a server to determine the processed response tracing data corresponding to a response data frame, as illustrated in an embodiment of this application. Figure 3 A flowchart illustrating the internal execution logic of a data tracing interceptor provided in an embodiment of this application is shown. Figure 4 A flowchart illustrating a display label conversion method provided in an embodiment of this application is shown; Figure 5 This illustration shows a rendering diagram of a display label provided in an embodiment of this application. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0022] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0023] Currently, mainstream AIAgent platforms have a common shortcoming in their native design: they lack the ability to trace and display the sources of information in images and text. The sources of information that AIAgent platforms rely on to generate answers mainly include the content of the system's built-in knowledge base and relevant results obtained through real-time network retrieval. These sources are the core support for ensuring the accuracy and credibility of the answers.

[0024] From a technical perspective, the underlying big model can accurately identify and associate relevant knowledge sources during the process of parsing user needs and generating corresponding answers. Furthermore, when processing the output results of the big model on the Python server side, the system itself already has the ability to extract and process the source information. It can even broadcast the cited knowledge base entries and key information fragments through voice and other means, indicating that the technology has the basic conditions to support citation tracing.

[0025] Among the existing technologies used by intelligent agent platforms to render and display output content, the standard Markdown format is mainly used for front-end rendering. However, as a lightweight markup language, Markdown was originally designed for basic text layout. Its native specification does not have the technical capability to support the integrated rendering of complex text and graphics. That is, the standard Markdown format has inherent limitations in complex information visualization scenarios, resulting in the inability to accurately and interactively present reference and traceability information. For example, even when the backend has detailed data on reference and traceability information (such as PDF original text fragments, image materials, metadata, etc.), due to the limitations of Markdown, it does not have the ability to accurately display specific text fragments from PDF documents and highlight key areas, nor does it have the ability to deeply and interactively customize the display of image content.

[0026] In summary, existing technical solutions, due to a lack of extensions and customizations to standard rendering formats, cannot effectively meet the core needs of enterprise-level AgentOps platforms for richly illustrated and precisely interactive display of knowledge base references.

[0027] Based on this, this application provides a graphical visualization backtracking method for intelligent agent response tracing. By converting tracing data into display labels associated with different graphical display methods, it breaks through the limitations of existing client display format specifications, enabling tracing data to be displayed more comprehensively, as detailed below: Please see Figure 1 , Figure 1 The flowchart illustrates a graphical and textual visualization-based backtracking method for tracing the source of an agent's response, provided in an embodiment of this application. For example... Figure 1 As shown, the method provided in this application embodiment includes the following steps: S100, interception and intelligent agent output response tracing data and its corresponding response tracing data.

[0028] S200. Add a convention marker to the response traceability data to obtain the processing traceability data.

[0029] S300: Generate display labels corresponding to the processing traceability data.

[0030] S400: The rendering layer is used to display the response feedback data and its corresponding display labels.

[0031] S500 responds to click events on display labels and renders and displays processing traceability data using the graphic and text display methods associated with the display labels.

[0032] Preferably, the display tags are in a format that can be rendered by the formatting language (e.g., Markdown) carried by the client.

[0033] In practical implementation, the graphic visualization backtracking method for intelligent agent response tracing provided in this application is applied to an intelligent interactive system, which includes a server and a client. The server executes steps S100 to S200. After executing step S200, it further executes steps S300 to S500, sending the response feedback data and its corresponding processing tracing data to the client. Considering the constraints of using standard Markdown format for front-end rendering of response tracing data in steps S100 to S200, this application adds pre-agreed tags between the server and client to the response tracing data, enabling the front-end (client) to effectively identify the response tracing data. Furthermore, by adding corresponding display tags to the identified response tracing data and calling the graphic display method associated with the display tags, the response tracing data is comprehensively displayed. Compared to directly using standard Markdown format for rendering, this achieves a more accurate and comprehensive display of the response tracing data.

[0034] In step S100, after the agent obtains the question text input by the user, it responds to the question text and outputs response feedback data. Since the server and the client interact with streaming data, the agent mounted on the server pre-agres that its output response feedback data follows the streaming data transmission format. That is, based on the question text input by the user, it automatically outputs multiple response data frames and outputs the response traceability data corresponding to each response data frame according to the given output protocol. The processed traceability data includes multiple traceability data frames that correspond one-to-one with the multiple response data frames.

[0035] In a preferred embodiment, please refer to Figure 2 , Figure 2 This document illustrates a flowchart illustrating how a server determines the processed response tracing data corresponding to a response data frame, as provided in an embodiment of this application. Figure 2 As shown, in step S200, for each response data frame, the server executes: S2001. Parse the response tracing data corresponding to the response data frame according to the given output protocol of the intelligent agent to obtain the tracing knowledge base identifier, the tracing file identifier, and the tracing slice identifier.

[0036] S2002. Extract the target traceability file indicated by the traceability file identifier from the target traceability instruction library corresponding to the traceability knowledge base identifier.

[0037] S2003. Extract the target knowledge slice corresponding to the traceability slice identifier from the target traceability file.

[0038] S2004. Add a convention marker to the target knowledge slice to obtain the source data frame corresponding to the response data frame.

[0039] In step S2001, the given output protocol specifies the encapsulation format of the response traceability data. The response traceability data includes, but is not limited to, any one of the following: traceability knowledge base identifier, traceability file identifier, and traceability slice identifier. The traceability knowledge base identifier is used to locate the traceability knowledge base. In this application, the agent is associated with multiple traceability knowledge bases. Each traceability knowledge base includes multiple traceability files, and each traceability file includes multiple knowledge slices.

[0040] For example, for the response data frame {"answer_content": "The accuracy of lithium battery SOC estimation is greatly affected by temperature. In low-temperature scenarios, an internal resistance correction model needs to be introduced."}, its response traceability data "trace_meta" is represented as follows: { "kb_id": "KB_BATTERY_001", "file_id": "DOC_SOH_202605", "slice_id": "SLICE_0087"} In steps S2002 to S2003, based on the above response tracing data, "kb_id" represents the tracing knowledge base identifier, "file_id" represents the tracing file identifier, and "slice_id" represents the tracing slice identifier.

[0041] In step S2002, based on the above analysis results, it is determined that the response tracing data originates from the target knowledge slice "SLICE_0087" in the target tracing file "DOC_SOH_202605" within the target tracing knowledge base "KB_BATTERY_001". Specifically, the target knowledge slice "SLICE_0087" is the reference source data corresponding to the response data frame {"answer_content": "The accuracy of lithium battery SOC estimation is greatly affected by temperature; an internal resistance correction model needs to be introduced in low-temperature scenarios."}

[0042] In step S2004, a pre-agreed identifier between the server and the client is added to the target knowledge slice, and encapsulated to form a traceability data frame that the client can recognize. For example, the identifier is [agentOpstarget='※']{main_id}[ / agentOps], where ※ represents the target knowledge slice to which the identifier is to be added. For example, after adding the identifier to the target knowledge slice "SLICE_0087" in the above embodiment, the resulting traceability data frame can be: [agentOps target="SLICE_0087"]{main_id}[ / agentOps] Here, main_id represents the identifier of the source knowledge base to which the target knowledge slice belongs. Each source knowledge base identifier is predefined according to actual needs.

[0043] Furthermore, multiple response data frames and the corresponding traceability data frames for each response data frame are sent to the client in the form of a data stream.

[0044] In a preferred embodiment, in step S300, the client executes: By using a pre-set long-term connection technology, data frames sent by the server are continuously received in the form of a data stream. The received data frames are then passed to the source tracing data interceptor, which identifies the source tracing data frames and generates corresponding display tags.

[0045] For example, the preset long-term connection technology can be either two-way real-time communication technology (WebSocket) or one-way streaming push technology (Server-Sent Events, or SSE). In order to process all incoming data frames uniformly, the client is pre-loaded with a source data interceptor. The source data interceptor scans and filters the data frames in the received data stream, identifies the source data frames, and generates the corresponding display labels for the source data frames.

[0046] In a preferred embodiment, please refer to Figure 3 , Figure 3 A flowchart illustrating the internal execution logic of a source tracing data interceptor provided in an embodiment of this application is shown. Figure 3 As shown, the source tracing data interceptor executes: S3001. Verify whether the data frame carries the agreed-upon marker.

[0047] S3002. If the data frame carries a conventional marker, the received data frame is determined to be a traceability data frame, and a display label corresponding to the traceability data frame is generated.

[0048] S3003. If the data frame does not carry the agreed-upon marker, then it is determined that the received data frame is not a traceable data frame, and the data frame is directly output.

[0049] Inside the source data interceptor, source data frames are filtered by recognizing conventional markers. Once a source data frame is recognized, it is converted into display tags that can be rendered by Markdown, the formatting language on the client. If the data frame is not a source data frame, it is directly output.

[0050] As shown in the above embodiments, the source tracing data interceptor identifies the convention marker [agentOps target='※']{main_id}[ / agentOps] and determines the data frame with the convention marker as the source tracing data frame.

[0051] In a preferred embodiment, please refer to Figure 4 , Figure 4 A flowchart illustrating a display label conversion method provided in an embodiment of this application is shown. Figure 4 As shown, step S3002 further includes: S30020, Obtain the traceability data interaction flag.

[0052] The data interaction flag indicates whether user interaction with the target knowledge slice indicated by the displayed label is supported.

[0053] S30021. If the traceability data interaction flag indicates that the user does not support interactive operations on the target knowledge slice, then the traceability data frame is converted into the first display label.

[0054] S30022. If the traceability data interaction flag indicates that the user can perform interactive operations on the target knowledge slice, then the traceability data frame is converted into a second display label.

[0055] In one specific embodiment, the traceability data interaction flag is retrieved from the user configuration library of the intelligent agent platform for the large model. The traceability data interaction flag can be preset according to actual needs. Preferably, in response to the change operation of the traceability data interaction flag, the switching of the traceability data display mode is completed.

[0056] In this application, based on the user's actual interaction needs with the traceability data, a traceability data interaction flag is pre-specified. For example, if the traceability data interaction flag is 0, it means that the user does not support interactive operations on the target knowledge slice indicated by the display label. If the traceability data interaction flag is 1, it means that the user supports interactive operations on the target knowledge slice indicated by the display label.

[0057] The display tags are those that can be rendered using a predefined formatting language provided by the client. For an example, please refer to [link to example]. Figure 5 , Figure 5 This illustration shows a rendering diagram of a display label provided in an embodiment of this application. For example... Figure 5 As shown, the first display tag a is a display tag with multiple hyperlink attributes, and after the first display tag corresponding to the source data frame is generated, the first display tag is associated with the hyperlink display page of the target knowledge slice corresponding to the source data frame.

[0058] Furthermore, regarding the first display label, during the process of displaying the response data frame and its corresponding first display label using the rendering layer, the first display label a is displayed in conjunction with its corresponding response data frame during rendering by the rendering layer. For example, as shown... Figure 4The response data frame is "Shape Change Assets". The first display label is an underline. When it is displayed, an underline is added under "Shape Change Assets" to give the response data frame "Shape Change Assets" a hyperlink attribute. The associated hyperlink is the target knowledge slice corresponding to "Shape Change Assets".

[0059] Specifically, the display method associated with the first display tag does not support hyperlink attributes with color or interactive requirements; it is only one of the rendering methods supported by the standard Markdown format, and no specific restrictions are imposed here.

[0060] In this application, the first display label and the second display label are different display labels supported by the standard Markdown format and set according to actual needs.

[0061] In another preferred embodiment, such as Figure 5 As shown, for the second display tag b, after converting the traceability data frame into the second display tag, the second display tag is associated with the page display interaction module, and the page display interaction module embeds the current display page of the response feedback data.

[0062] Regarding the second display label, during the process of displaying the response data frame and its corresponding second display label using the rendering layer, the second display label b is displayed next to its corresponding response data frame when it is rendered by the rendering layer. The second display label b is determined by the source knowledge base to which the target knowledge slice corresponding to the source data frame belongs. For example, the second display label b can be the identifier of the source knowledge base to which the target knowledge slice corresponding to the source data frame belongs.

[0063] Preferably, this application associates the second display tag with the page display interaction module, and displays the target knowledge slice interactively in the form of the page display interaction module. The page display interaction module supports at least interactive operations such as folding, previewing, clicking to view, color marking, content copying, and content export of the target knowledge slice. Here, no specific restrictions are placed on the interactive operations. This application achieves multiple display and interactive effects for multi-source data while retaining the compatibility of the standard Markdown format.

[0064] In a preferred embodiment, step S500 further includes: In response to a click event on the first display label, the user is redirected to the hyperlink display page associated with the first display label. The hyperlink display page will then render the target knowledge slice corresponding to the first display label.

[0065] In another preferred embodiment, step S500 further includes: In response to a click event on the second display label, the page display interaction module is invoked to open an interactive graphic input box on the current display page, and the target knowledge slice is loaded into the interactive graphic input box for rendering and display.

[0066] In one specific embodiment, the interactive text and image input box supports predefined interactive operations performed on the target knowledge slice displayed within it. For example, the interactive operations include highlighting the selected area of ​​the target knowledge slice, expanding and floating previewing the source file to which the knowledge slice belongs, marking key areas, copying and exporting content, etc., which will not be described in detail here.

[0067] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division; in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0068] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0069] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0070] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0071] The above are merely specific embodiments 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 graphical and textual visualization backtracking method for tracing the source of an intelligent agent's response, characterized in that, The method includes: Intercept the response feedback data output by the intelligent agent and its corresponding response tracing data, add a conventional tag to the response tracing data to obtain processing tracing data; Generate display labels corresponding to the processing traceability data, and use the rendering layer to display the response feedback data and its corresponding display labels; In response to a click event on the display label, the processing traceability data is rendered and displayed using the graphic display method associated with the display label. The display label is a label format that can be rendered by the formatted language on the client.

2. The method according to claim 1, characterized in that, The response feedback data includes multiple response data frames; the response tracing data includes multiple tracing data frames that correspond one-to-one with the multiple response data frames; and the processing tracing data includes multiple tracing data frames that correspond one-to-one with the multiple response data frames. Specifically, for each response data frame: The response tracing data is parsed according to the given output protocol of the intelligent agent to obtain the tracing knowledge base identifier, the tracing file identifier, and the tracing slice identifier. Extract the target traceability file indicated by the traceability file identifier from the target traceability knowledge base corresponding to the traceability knowledge base identifier; Extract the target knowledge slices corresponding to the traceability slice identifiers from the target traceability file; By adding a convention marker to the target knowledge slice, the corresponding source data frame is obtained.

3. The method according to claim 2, characterized in that, The step of generating display labels corresponding to the processing traceability data includes: By using a pre-set long-term connection technology, data frames sent by the server are continuously received in the form of a data stream; The received data frame is passed to the source data interceptor so that the source data interceptor can identify the source data frame and generate the corresponding display label.

4. The method according to claim 3, characterized in that, The source tracing data interceptor executes: Verify whether the data frame carries the agreed-upon marker; If the data frame carries the agreed-upon marker, then the received data frame is determined to be a source data frame, and a display label corresponding to the source data frame is generated; If the data frame does not carry the agreed-upon marker, it is determined that the received data frame is not a source data frame, and the data frame is directly output.

5. The method according to claim 4, characterized in that, The step of generating the display label corresponding to the source tracing data frame includes: Obtain the traceability data interaction flag, which indicates whether the user is supported to perform interactive operations on the target knowledge slice indicated by the display label; If the source data interaction flag indicates that the user does not support interactive operations on the target knowledge slice, then the source data frame is converted into a first display label; If the source data interaction flag indicates that the user can perform interactive operations on the target knowledge slice, then the source data frame is converted into a second display label.

6. The method according to claim 5, characterized in that, The method further includes: After generating the first display tag corresponding to the source traceability data frame, the first display tag is associated with the hyperlink display page of the target knowledge slice corresponding to the source traceability data frame; After generating the second display tag corresponding to the traceability data frame, the second display tag is associated with the page display interaction module, which embeds the current display page of the response feedback data.

7. The method according to claim 5, characterized in that, The step of rendering and displaying the processing traceability data in response to a click event on the display label using the graphic and text display method associated with the display label includes: In response to a click event on the first display label, the user is redirected to the hyperlink display page associated with the first display label. The hyperlink display page will render the target knowledge slice corresponding to the first display tag.

8. The method according to claim 5, characterized in that, The step of rendering and displaying the processing traceability data using the graphic and text display method associated with the display label in response to a click event on the display label further includes: In response to a click event on the second display label, the page display interaction module is invoked to open an interactive graphic input box on the currently displayed page; The target knowledge slices are loaded into the interactive graphic input box for rendering and display.

9. A graphical and textual visualization backtracking method for tracing the source of an intelligent agent's response, characterized in that, Applied to the server side, The method includes: Intercepting and responding to the intelligent agent's output, and the corresponding response tracing data; Add a predefined tag to the response tracing data to obtain processing tracing data; The response feedback data and its corresponding processing traceability data are sent to the client.

10. A graphical and textual visualization backtracking method for tracing the source of an intelligent agent's response, characterized in that, Applied to the client side, The method includes: Receive processing traceability data sent by the server; Generate display labels corresponding to the processing traceability data; The response feedback data and its corresponding display labels are displayed using a rendering layer; In response to a click event on the display label, the processing traceability data is rendered and displayed using the graphic display method associated with the display label. The display label is a label format that can be rendered by the formatted language on the client.

Images