Interface interaction method and device, electronic equipment and storage medium

Abstract

The invention provides an interface interaction method and device, electronic equipment and a storage medium, and relates to the field of operation and maintaining.The method comprises the steps that when an operation and maintenance dialogue text input by a user is received, the operation and maintenance dialogue text is analyzed through a pre-training language model to determine to-be-retrieved information and user viewing intention information; searching a database for target operation and maintenance data according to the to-be-retrieved information, determining a target rendering template in a plurality of candidate rendering templates according to the data type of the operation and maintenance data and the user viewing intention information, and rendering the target operation and maintenance data into a to-be-displayed operation and maintenance window based on the target rendering template; and destroying the operation and maintenance window currently displayed in the display interface, and outputting the operation and maintenance window to be displayed to the display interface. According to the method, a dialogue interaction mode can be provided for the user, the to-be-displayed operation and maintenance window can be automatically generated according to the dialogue content with the user, and the to-be-displayed operation and maintenance window can be dynamically displayed in the display interface, so that the use convenience of operation and maintenance personnel can be improved.

Description

Technical Field

[0001] The present invention relates to the field of operation and maintenance, and in particular to an interface interaction method, device, electronic device and storage medium. Background Art

[0002] In the field of operations and maintenance technology, the user interfaces of common operations and maintenance software typically have a fixed layout, with various charts and graphs fixed within the interface. As the number of operations and maintenance projects continues to increase, the user interfaces of traditional operations and maintenance software are becoming increasingly complex, significantly increasing the difficulty for operations and maintenance personnel. Summary of the Invention

[0003] The purpose of the present invention is to provide an interface interaction method, device, electronic device and storage medium, which can provide users with a dialogue interaction mode, and can automatically generate an operation and maintenance window to be displayed based on the content of the dialogue with the user, and can dynamically display the operation and maintenance window to be displayed in the display interface, thereby improving the convenience of use for operation and maintenance personnel.

[0004] To solve the above technical problems, the present invention provides an interface interaction method, comprising:

[0005] When receiving the operation and maintenance dialogue text input by the user, the operation and maintenance dialogue text is parsed using the pre-trained language model to determine the information to be retrieved and the user's viewing intention information;

[0006] Searching a database for target operation and maintenance data according to the information to be retrieved, determining a target rendering template from a plurality of candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information, and rendering the target operation and maintenance data into an operation and maintenance window to be displayed based on the target rendering template;

[0007] The operation and maintenance window currently displayed in the display interface is destroyed, and the operation and maintenance window to be displayed is output to the display interface.

[0008] Optionally, the display interface includes a window display area and a dialogue display area;

[0009] The method further comprises:

[0010] Parsing the operation and maintenance dialogue text using a pre-trained language model to generate dialogue response text;

[0011] The destroying the operation and maintenance window currently displayed in the display interface and outputting the operation and maintenance window to be displayed to the display interface includes:

[0012] The operation and maintenance window currently displayed in the window display area is destroyed, the operation and maintenance window to be displayed is output to the window display area, and the dialogue reply text is output to the dialogue display area.

[0013] Optionally, it also includes:

[0014] When a window fixation command input by a user is received, determining a target operation and maintenance window corresponding to the window fixation command in the display interface, and setting the target operation and maintenance window as a fixed operation and maintenance window;

[0015] When receiving a window unpinning command input by the user, determining a target fixed operation and maintenance window corresponding to the window pinning command in the display interface, and setting the target fixed operation and maintenance window to a non-fixed operation and maintenance window;

[0016] Destroying the operation and maintenance window currently displayed in the display interface includes:

[0017] Destroy the non-fixed operation and maintenance window in the display interface.

[0018] Optionally, determining a target rendering template from a plurality of candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information includes:

[0019] Obtaining preset matching rule information corresponding to each candidate rendering template;

[0020] The data type, the preset matching rule information, and the user viewing intention information are integrated into a first input text using a first preset input template; the first preset input template includes preset matching guide information;

[0021] inputting the first input text into the pre-trained language model, so that the pre-trained language model matches the data type and the user viewing intention information with the preset matching rule information according to the preset matching guidance information to obtain template selection information;

[0022] The target rendering template is determined according to the template selection information.

[0023] Optionally, after determining a target rendering template from the candidate rendering templates according to the template selection information, the method further includes:

[0024] Determine a default rendering template corresponding to the data type, and determine whether the target rendering template exists in the default rendering template;

[0025] If the target rendering template does not exist in the default rendering template, the target operation and maintenance data is rendered into an operation and maintenance window to be displayed based on the default rendering template.

[0026] Optionally, the using a pre-trained language model to parse the operation and maintenance dialogue text to determine the information to be retrieved and the user's viewing intention information includes:

[0027] Obtain the historical operation and maintenance conversation text of the user;

[0028] The historical operation and maintenance dialogue text and the operation and maintenance dialogue text are parsed using the pre-trained language model to determine the information to be retrieved and the user viewing intention information.

[0029] Optionally, it also includes:

[0030] Receive raw operation and maintenance data;

[0031] Extracting data structure features of the original operation and maintenance data, and matching the data structure features with preset rules of multiple candidate data types to determine the data type corresponding to the original operation and maintenance data;

[0032] The raw operation and maintenance data and the data type are integrated into a second input text using a second preset input template; the second preset input template includes a preset tag to generate guidance information;

[0033] Inputting the second input text into a pre-trained language model, so that the pre-trained language model parses the original operation and maintenance data and the data type according to preset label generation guidance information, and generates a semantic enhancement label corresponding to the original operation and maintenance data;

[0034] Adding a type tag corresponding to the data type and the semantic enhancement tag to the original operation and maintenance data, and saving the data to a database;

[0035] The step of searching a database for target operation and maintenance data according to the information to be retrieved, and determining a target rendering template from a plurality of candidate rendering templates according to a data type of the target operation and maintenance data and the user viewing intention information, includes:

[0036] Matching the information to be retrieved with each operation and maintenance data in the database and with the semantically enhanced tags of each operation and maintenance data to obtain the target operation and maintenance data;

[0037] A target rendering template is determined from a plurality of candidate rendering templates according to the type tag of the target operation and maintenance data and the user viewing intention information.

[0038] Optionally, extracting data structure features of the original operation and maintenance data, and matching the data structure features with preset rules of multiple candidate data types to determine the data type corresponding to the original operation and maintenance data includes:

[0039] Performing statistical processing on the original operation and maintenance data, and determining data structure characteristics of the original operation and maintenance data based on the statistical processing results;

[0040] Matching the data structure feature with the preset rule of the candidate data type to determine the target preset rule satisfied by the data structure feature;

[0041] Summing the preset weight values ​​corresponding to the target preset rules to obtain a matching value corresponding to the original operation and maintenance data;

[0042] When the matching value is greater than a preset threshold, the candidate data type is used as the data type corresponding to the original operation and maintenance data.

[0043] The present invention also provides an interface interaction device, comprising:

[0044] The model processing module is used to parse the operation and maintenance dialogue text input by the user using the pre-trained language model to determine the information to be retrieved and the user's viewing intention information;

[0045] a rendering processing module, configured to search a database for target operation and maintenance data according to the information to be retrieved, determine a target rendering template from a plurality of candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information, and render the target operation and maintenance data into an operation and maintenance window to be displayed based on the target rendering template;

[0046] The output module is used to destroy the operation and maintenance window currently displayed in the display interface and output the operation and maintenance window to be displayed to the display interface.

[0047] The present invention further provides an electronic device, comprising:

[0048] memory for storing computer programs;

[0049] A processor is used to implement the interface interaction method described above when executing the computer program.

[0050] The present invention also provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are loaded and executed by a processor, the interface interaction method described above is implemented.

[0051] The present invention provides an interface interaction method, comprising: when receiving an operation and maintenance dialogue text input by a user, using a pre-trained language model to parse the operation and maintenance dialogue text to determine information to be retrieved and user viewing intention information; searching a database for target operation and maintenance data according to the information to be retrieved, determining a target rendering template from multiple candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information, and rendering the target operation and maintenance data into an operation and maintenance window to be displayed based on the target rendering template; destroying the operation and maintenance window currently displayed in the display interface, and outputting the operation and maintenance window to be displayed to the display interface.

[0052] The beneficial effect of the present invention is that upon receiving user-entered operation and maintenance dialogue text, the present invention can first utilize a pre-trained language model to parse the operation and maintenance dialogue text to determine the information to be retrieved and the user's viewing intent. In other words, the present invention can utilize a pre-trained language model to automatically parse the target operation and maintenance data that the user wishes to retrieve and the desired viewing method for the operation and maintenance data. Subsequently, the present invention can search the database for the target operation and maintenance data based on the information to be retrieved, and can determine a target rendering template from multiple candidate rendering templates based on the data type of the target operation and maintenance data and the user's viewing intent. In other words, the present invention can automatically select the most suitable target rendering template based on the data type of the operation and maintenance data and the user's viewing intent, and render the target operation and maintenance data into the operation and maintenance window to be displayed based on the target rendering template. Subsequently, the present invention can destroy the operation and maintenance window currently displayed in the display interface and output the operation and maintenance window to be displayed to the display interface, that is, it can dynamically output the operation and maintenance window to be displayed to the display interface. In this way, the present invention can automatically generate the operation and maintenance window to be displayed based on the content of the conversation with the user and dynamically display the operation and maintenance window to be displayed in the display interface, thereby improving the user convenience of the operation and maintenance personnel.

[0053] The present invention also provides an interface interaction device, an electronic device and a storage medium, which have the above-mentioned beneficial effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0055] Figure 1 A flowchart of an interface interaction method provided by an embodiment of the present invention;

[0056] Figure 2 A schematic diagram of a first display interface provided by an embodiment of the present invention;

[0057] Figure 3 A schematic diagram of a second display interface provided by an embodiment of the present invention;

[0058] Figure 4 A schematic diagram of a third display interface provided by an embodiment of the present invention;

[0059] Figure 5 A structural block diagram of an interface interaction device provided by an embodiment of the present invention;

[0060] Figure 6This is a structural block diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0061] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0062] In the field of operations and maintenance technology, the user interfaces of common operations and maintenance software typically have a fixed layout, with various charts and graphs fixed within the interface. However, with the increasing number of operations and maintenance projects and data, the user interfaces of traditional operations and maintenance software have become increasingly complex, significantly increasing the difficulty for operations and maintenance personnel.

[0063] In view of this, in order to solve the technical problem of how to improve the usability of the user interface of operation and maintenance software, the present invention can provide an interface interaction method. First, it can provide users with a dialogue interaction mode. In addition, based on the capabilities of a pre-trained language model, it can automatically query the operation and maintenance data to be displayed according to the content of the dialogue with the user, automatically generate a suitable operation and maintenance window to be displayed based on the operation and maintenance data, and dynamically display the operation and maintenance window to be displayed in the display interface. This can avoid the problem of poor usability caused by the display interface only being able to statically set the operation and maintenance window. It can improve the convenience of use for operation and maintenance personnel.

[0064] For easier understanding, please refer to Figure 1 , Figure 1 This is a flow chart of an interface interaction method provided by an embodiment of the present invention. The method may include:

[0065] S101. When an operation and maintenance dialogue text input by a user is received, the operation and maintenance dialogue text is parsed using a pre-trained language model to determine information to be retrieved and user viewing intention information.

[0066] In this step, the operation and maintenance dialogue text refers to the natural language text entered by the user that contains the user's operation and maintenance intentions. By analyzing the operation and maintenance dialogue text, we can understand both the operation and maintenance data the user wishes to view and the way the user prefers to display the data. The user can enter this operation and maintenance dialogue text through the front-end display interface.

[0067] A pre-trained language model (UI-Agent) is a neural network model pre-trained using text corpus. It has the ability to automatically parse and generate text and can be fine-tuned based on actual application needs. A pre-trained language model can be a Large Language Model (LLM), a machine learning model trained using large amounts of text data that can generate natural language text or understand the meaning of text. Because LLMs differ significantly from typical language models in terms of model size (e.g., number of parameters), amount of training data, and complexity, they offer significant performance advantages in semantic understanding and text generation. Common pre-trained language models include ChatGLM-6B (Chat General Language Model with 6 billion parameters) and GLM-130B (General Language Model with 130 billion parameters).

[0068] The information to be retrieved is text information used to retrieve operation and maintenance data, which can be obtained by parsing the operation and maintenance dialogue text using a pre-trained language model.

[0069] The user viewing intention information is text information that represents the user's desired display method for operation and maintenance data, which can also be obtained by parsing the operation and maintenance dialogue text by the pre-trained language model.

[0070] As can be seen in this step, in order to dynamically query operation and maintenance data and dynamically generate and display operation and maintenance windows based on user intent, a pre-trained language model can be introduced to automatically parse the operation and maintenance dialogue text entered by the user, thereby automatically determining the information to be retrieved and the user's viewing intention information. For example, if the user enters the operation and maintenance dialogue text as: "Can you show me the changes in the number of Nginx Https requests over the past week?", entering this operation and maintenance dialogue text into the pre-trained language model can determine that the user wishes to query the number of Nginx Https requests over the past week and wants to display the changes in the number of Nginx Https requests in the form of time series changes. Furthermore, the pre-trained language model can generate the information to be retrieved for querying the number of Nginx Https requests over the past week, as well as the user's viewing intention information that represents the time series display.

[0071] Furthermore, since users may engage in multiple rounds of conversations with the pre-trained language model, and user intent may be hidden in the user's previously entered operation and maintenance conversation text, this embodiment can also obtain the user's historical operation and maintenance conversation text and input both the current operation and maintenance conversation text and the historical conversation text into the pre-trained language model, allowing the latter to parse the historical operation and maintenance conversation text and the operation and maintenance conversation text. This can then correct the user's intent based on the historical operation and maintenance conversation text, thereby improving the accuracy of the information to be retrieved and the user's viewing intention information.

[0072] Based on this, the use of a pre-trained language model to parse the operation and maintenance dialogue text to determine the information to be retrieved and the user's viewing intention information may include:

[0073] Step 11: Obtain the user's historical operation and maintenance conversation text;

[0074] Step 12: Utilize the pre-trained language model to parse the historical operation and maintenance dialogue text and the operation and maintenance dialogue text to determine the information to be retrieved and the user's viewing intention information.

[0075] Furthermore, to further enhance the user experience, the pre-trained language model can generate not only the information to be retrieved and the user's viewing intent, but also the response text for the user's operation and maintenance conversation text, which can be output and displayed. This can provide users with a smoother conversation experience.

[0076] Based on this, the method may further include:

[0077] Step 21: Use the pre-trained language model to parse the operation and maintenance dialogue text to generate dialogue response text.

[0078] Furthermore, to prevent users from maliciously manipulating the monitored system through the pre-trained language model and to prevent direct output of conversation replies containing harmful content to users, this embodiment also incorporates model protection measures. Upon receiving user-inputted operation and maintenance conversation information, a pre-trained malware analysis model can be used to analyze the information for malicious intent. The malware analysis model can also be a language model. If malicious content is determined to be present in the operation and maintenance conversation information, an error message can be directly returned to the user without being fed into the pre-trained language model for processing. Conversely, if the information is determined to be free of malicious content, it can be fed into the pre-trained language model for processing. Furthermore, upon receiving conversation replies output by the pre-trained language model, the same malware analysis model can be used to analyze the text for malicious intent. If malicious content is determined to be present in the conversation reply information, an error message can be directly returned to the user. Otherwise, the text can be fed into the pre-trained language model for processing.

[0079] This embodiment does not limit the specific malicious analysis model, and can be set according to actual application requirements, for example, it can be NeMo Guardrails, Guardrails AI, LangChain, etc.

[0080] S102: searching a database for target operation and maintenance data according to the information to be retrieved, determining a target rendering template from a plurality of candidate rendering templates according to a data type of the target operation and maintenance data and the user viewing intention information, and rendering the target operation and maintenance data into an operation and maintenance window to be displayed based on the target rendering template.

[0081] In this step, the target operation and maintenance data refers to the operation and maintenance data related to the information to be retrieved. Operation and maintenance data refers to data information related to operation and maintenance, which can correspond to four sources: 1. Anomaly detection data generated by the operation and maintenance software when performing anomaly detection on the detected system; 2. Log data generated by the detected system; 3. Operation data generated by the detected system; 4. Operation and maintenance knowledge of the detected system. Among them, the operation and maintenance knowledge can further include: 1. User guides and user manuals of the detected system; 2. Guides and notes used for anomaly troubleshooting in the past; 3. System and network diagrams; 4. Component relationship diagrams; 5. System configuration documents; 6. Real-time or periodic data feedback from external systems (such as fault ticket logs, historical alarms from other systems, and known interruption logs). In this embodiment, different databases can be used to store different operation and maintenance data, such as using an anomaly detection database to store anomaly detection data, using a log database to store log data, using an operation database to store operation data, and using an operation and maintenance knowledge base to store operation and maintenance knowledge.

[0082] Furthermore, in order to improve the flexibility and convenience of searching for operation and maintenance data, this embodiment can set different database tags for different databases. The database tag can include a database description, a database usage method, a data category of the data stored in the database (anomaly detection data, log data, operation data, operation and maintenance knowledge), a data type (time series operation and maintenance data, distributed operation and maintenance data, relational operation and maintenance data, text operation and maintenance data), and a data direction (characterizing the actual content of the operation and maintenance data). Furthermore, this embodiment can control the pre-trained language model to determine the target database to be searched in multiple databases based on the information to be retrieved and the database tag, and retrieve the required operation and maintenance data in the target database. Alternatively, this embodiment can set a machine learning model dedicated to performing data retrieval, and control the machine learning model to determine the target database to be searched in multiple databases based on the information to be retrieved and the database tag, and retrieve the required operation and maintenance data in the target database.

[0083] Furthermore, it can be understood that the data structure and data distribution of each type of operation and maintenance data are different, and thus the appropriate display method for each type of operation and maintenance data may also be different. For example, the operation and maintenance data type of numerical type may include time-series operation and maintenance data (Time-series), distribution operation and maintenance data (Distribution), relational operation and maintenance data (Distribution), text-type operation and maintenance data, etc. For time-series operation and maintenance data, it is suitable to use time-series display methods, such as line charts, area charts, scatter plots, heat maps, etc. For distribution-type operation and maintenance data, it is suitable to use display methods that can display data distribution, such as histograms, box plots, violin plots, and bar charts. For relational operation and maintenance data, it is suitable to use display methods that can display the relationship between data, such as scatter plots, network charts, bubble charts, Sankey diagrams, etc. For text-type operation and maintenance data (such as logs, codes), they can be directly displayed in text form. Therefore, this embodiment needs to additionally determine the data type of the operation and maintenance data to determine the appropriate display method based on the data type.

[0084] A candidate rendering template refers to a preset template used to render operation and maintenance data into an operation and maintenance window. For example, it can include chart templates such as line charts, area charts, scatter charts, and heat maps, as well as text templates that can include rendered text.

[0085] The Artifact window is a window that visually displays charts and text. It should be noted that the shape, size, and display position of the Artifact window in the display interface can be dynamically adjusted according to the situation.

[0086] It can be seen that in this step, first, the operation and maintenance data that meets the user's intention can be searched in the database based on the information to be retrieved. Subsequently, in order to ensure that the rendering of the operation and maintenance window meets the user's expectations and the characteristics of the data structure, the target rendering template can be determined from multiple candidate rendering templates based on the data type of the operation and maintenance data and the user's viewing intention information. Finally, the operation and maintenance data can be rendered into an operation and maintenance window to be displayed based on the target rendering template. It is worth noting that operation and maintenance data can correspond to one or more data types. For example, for an operation log with timestamp information, it can be regarded as either time-series operation and maintenance data or text-type operation and maintenance data. In other words, operation and maintenance data can be displayed in a variety of forms. Therefore, for operation and maintenance data with multiple data types, the user's viewing intention information plays a role in determining the final rendering and display form of the operation and maintenance data, and can dynamically select a suitable form to render and display the operation and maintenance data based on the user's viewing intention.

[0087] S103: Destroy the operation and maintenance window currently displayed in the display interface, and output the operation and maintenance window to be displayed to the display interface.

[0088] In this step, the display interface refers to the user interface (UI). Unlike related technologies, the display interface in this embodiment does not set a fixed operation and maintenance window by default, but can dynamically display the operation and maintenance window according to the user's intention and needs (on-demand). In addition, after the user enters the new operation and maintenance dialogue text, this embodiment can destroy the operation and maintenance window currently displayed in the display interface and output the newly generated operation and maintenance window to be displayed to the display interface, that is, it can dynamically update the operation and maintenance window in the display interface according to the new operation and maintenance dialogue text entered by the user. In this way, the display interface in this embodiment is more flexible in displaying the operation and maintenance window and is more in line with the user's intuitive use, thereby improving the user experience.

[0089] Furthermore, the display interface in this embodiment may include a window display area and a dialogue display area. The window display area is a display area for displaying the operation and maintenance window, while the dialogue display area is used to display the user's operation and maintenance dialogue text and the dialogue reply text generated by the pre-trained language model.

[0090] Based on this, destroying the operation and maintenance window currently displayed in the display interface and outputting the operation and maintenance window to be displayed to the display interface may include:

[0091] Step 31: destroying the operation and maintenance window currently displayed in the window display area, outputting the operation and maintenance window to be displayed to the window display area, and outputting the dialogue reply text to the dialogue display area.

[0092] To understand the dynamic display of the operation and maintenance window, please refer to Figure 2 and Figure 3 , Figure 2 A schematic diagram of a first display interface provided by an embodiment of the present invention, Figure 3 This is a schematic diagram of a second display interface provided by an embodiment of the present invention. Figure 2 In the example, the display interface may include a window display area 2 and a dialogue display area 1. The dialogue display area 1 includes a dialog box 11, a dialog box 12, and an input box 13. The dialog box 11 displays the user's operation and maintenance dialogue text "How is the system energy consumption today?", and the dialog box 12 displays the dialogue reply text generated by the pre-trained language model "It is slightly higher than usual, please see this picture. It seems to be related to these software processes, please see these logs". The input box 13 is used to enter the operation and maintenance dialogue text. The window display area 2 currently includes an operation and maintenance window 21 and an operation and maintenance window 22. The operation and maintenance window 21 displays the system energy consumption, and the operation and maintenance window 22 displays the system log. It can be seen that the operation and maintenance window 21 and the operation and maintenance window 22 are generated based on the operation and maintenance dialogue text.

[0093] Further, in Figure 3In the example, the user further enters the question "Is the configuration update last week related to the change in energy consumption?" in dialogue display area 1, and the pre-trained language model responds, "Please look at the four charts; they appear to be related to the increase in energy consumption. Furthermore, it appears to be related to the configuration set last week. Please see these logs." At this point, the O&M windows in window display area 2 change to O&M windows 23-27. O&M window 23 displays the system log, and O&M windows 24-27 display the changes in four system detection values. This shows that this embodiment can automatically analyze the user's intent based on the pre-trained language model, search for relevant O&M data based on that intent, and dynamically update the O&M windows, thereby effectively improving user convenience. Furthermore, given that users may need to monitor certain O&M data over a long period of time, they may wish to pin the O&M windows in the display interface. Furthermore, when their attention needs change, users may wish to unpin a pinned O&M window. Therefore, to improve user convenience, this embodiment also provides the ability to pin and unpin O&M windows.

[0094] Specifically, upon receiving a window pinning command input by the user, this embodiment may determine the target operation and maintenance window corresponding to the window pinning command in the display interface, and set the target operation and maintenance window as the fixed operation and maintenance window. The window pinning command is an instruction command for pinning an operation and maintenance window. A fixed operation and maintenance window refers to a fixed operation and maintenance window that will not be destroyed when the operation and maintenance window is updated.

[0095] Additionally, upon receiving a window unpinning command from the user, the target fixed operation and maintenance window corresponding to the window pinning command in the display interface is determined, and the target fixed operation and maintenance window is set as an unpinned operation and maintenance window. The window pinning touch command is an instruction command for touching a fixed operation and maintenance window, and an unpinned operation and maintenance window refers to an unpinned operation and maintenance window, which will be destroyed when the operation and maintenance window is updated.

[0096] Based on this, the method may further include:

[0097] Step 41: upon receiving a window fixation command input by a user, determining a target operation and maintenance window corresponding to the window fixation command in the display interface, and setting the target operation and maintenance window as a fixed operation and maintenance window;

[0098] Step 42: When a window pinning unpinning command input by the user is received, a target fixed operation and maintenance window corresponding to the window pinning command in the display interface is determined, and the target fixed operation and maintenance window is set as a non-fixed operation and maintenance window.

[0099] Destroying the operation and maintenance window currently displayed in the display interface may include:

[0100] Step 51: Destroy the non-fixed operation and maintenance window in the display interface.

[0101] It can be seen that this embodiment provides a dynamic update and destruction mechanism and a fixed retention mechanism for the operation and maintenance window. Compared with the method in the related art where the display interface can only fix the operation and maintenance window, the operation and maintenance window in the display interface of this embodiment can be dynamically generated and destroyed according to the user's current needs, and can also be fixed / unfixed in the display interface under the user's operation. This can not only meet the user's immediate needs for temporary reference to operation and maintenance data, but also meet the user's needs for continuous attention to a certain operation and maintenance data, and also meet the user's needs for continuous monitoring of changes. It can effectively solve the defects in the related art that the display interface is too complex, there is information overload, and it is not conducive to technical personnel to view, and it is more helpful for technical personnel to view important operation and maintenance data, thereby effectively improving the user experience of the operation and maintenance software.

[0102] Further, the following will introduce the embodiment of the display interface provided by this embodiment in the network system operation and maintenance management. Figure 4 , Figure 4 This is a schematic diagram of the third display interface provided by an embodiment of the present invention. Figure 4 In the scenario shown, the user requires to view the overall network health status, view the changes in network quality over the past seven days, and analyze the causes of the changes in network quality. Figure 2 、 3 akin, Figure 4 The display interface shown also includes window display area 2 and dialogue display area 1. Window display area 2 includes operation and maintenance windows 28 and 29. Operation and maintenance window 28 provides an overall view of network health, including scores for network quality, connection quality, coverage quality, and throughput quality. Operation and maintenance window 29 displays changes in network quality over the past seven days. Dialogue display area 1 displays a textual analysis of the causes of network quality changes.

[0103] Combine Figure 2 、 3 As can be seen from Figure 4, this embodiment can display different types of operation and maintenance data in different formats. For example, time-series operation and maintenance data can be displayed in a time-series format; relational operation and maintenance data can be displayed in a relational format; and text-based operation and maintenance data can be displayed in a text format.

[0104] Based on the above embodiments, upon receiving user input for an operation and maintenance conversation, the present invention can first utilize a pre-trained language model to parse the conversation text to determine the information to be retrieved and the user's viewing intent. In other words, the present invention can utilize a pre-trained language model to automatically parse the operation and maintenance data the user wishes to retrieve and the method the user intends to use to view the data. Subsequently, the present invention can search a database for operation and maintenance data based on the information to be retrieved and, based on the data type of the operation and maintenance data and the user's viewing intent, determine a target rendering template from multiple candidate rendering templates. This automatically selects the most suitable target rendering template based on the data type of the operation and maintenance data and the user's viewing intent, and renders the operation and maintenance data into a display window based on the target rendering template. Subsequently, the present invention can destroy the currently displayed operation and maintenance window in the display interface and output the display window to be displayed to the display interface, i.e., dynamically output the display window to be displayed to the display interface. In this way, the present invention can automatically generate a display window based on the content of the conversation with the user and dynamically display it in the display interface, thereby improving user convenience for operation and maintenance personnel.

[0105] Based on the above embodiment, the matching determination process of the target rendering template is introduced below. In one possible scenario, determining the target rendering template from multiple candidate rendering templates based on the data type of the target operation and maintenance data and the user viewing intention information may include:

[0106] S201: Obtain preset matching rule information corresponding to each candidate rendering template.

[0107] In this step, each candidate rendering template is assigned corresponding preset matching rule information. This preset matching rule information is a set of rules used to match a specific data type and user viewing intent with a specific rendering template. The pre-trained language model can semantically match the data type of the operation and maintenance data and the user viewing intent information with the preset matching rule information to select the target rendering template from the candidate rendering templates.

[0108] It should be noted that this embodiment does not limit the specific preset matching rule information, which can be set according to actual application needs. In one possible scenario, the preset matching rule information can include data feature requirements and context rules, as follows:

[0109] For time series data, the available rendering templates are:

[0110] Preferred templates: line chart, area chart.

[0111] Secondary templates: scatter plot, heat map.

[0112] Data feature requirements include:

[0113] Necessary requirements: Include timestamp columns and numeric columns.

[0114] Optional requirements: multiple data series, fixed time interval.

[0115] Context rules include:

[0116] For continuous measurements, a line chart is recommended; for cumulative values, an area chart is suitable. For high-density time series data, a heat map can be used.

[0117] For distributed operation and maintenance data (Distribution), the available rendering templates are:

[0118] Preferred templates: Histogram, Boxplot.

[0119] Secondary templates: violin chart, bar chart.

[0120] Data feature requirements include:

[0121] Necessary requirements: Numerical or categorical data with a numerical range

[0122] Optional requirements: multi-category data, outliers

[0123] Context rules include:

[0124] A histogram is recommended for continuously distributed data, a box plot is recommended for comparing distributions, and a bar chart can be used for discrete categorical data.

[0125] For relational operation and maintenance data (Relationship), the available rendering templates are:

[0126] Preferred templates: scatter plot, network diagram.

[0127] Secondary templates: Bubble chart, Sankey chart.

[0128] Data feature requirements include:

[0129] Necessary requirements: multivariate dataset, corresponding associated entities.

[0130] Optional requirements: weighted connection relationship and hierarchical structure.

[0131] Context rules include:

[0132] Scatter plots are recommended for correlation analysis, network diagrams are recommended for entity relationships, and Sankey diagrams can be used for traffic visualization.

[0133] S202. Integrate the data type, the preset matching rule information, and the user viewing intention information into a first input text using a first preset input template; the first preset input template includes preset matching guide information.

[0134] S203. Input the first input text into the pre-trained language model, so that the pre-trained language model matches the data type and the user viewing intention information with the preset matching rule information according to the preset matching guidance information to obtain template selection information.

[0135] In steps S202 and S203, to facilitate parsing by the pre-trained language model, this embodiment can use a first preset input template to integrate the data type, preset matching rule information, and user viewing intent information into a first input text (Prompt). The first preset input template should contain corresponding fill-in spaces for the data type, preset matching rule information, and user viewing intent information. Furthermore, the first preset input template must also include preset matching guidance information, which instructs the pre-trained language model on how to match the data type and user viewing intent information with the preset matching rule information.

[0136] It should be noted that this embodiment does not limit the specific content of the preset matching guidance information. For example, the preset matching guidance information may include a matching step of matching the data type and the user viewing intention information using the preset matching rule information, so that the model performs matching according to the matching step.

[0137] For another example, the preset matching guidance information may include output rules, and the output rules may be used to constrain the form of the template selection information output by the model.

[0138] For example, in order to constrain the model's selection of rendering templates, the preset matching guidance information can also include model constraints (Matching Rules), and add the model constraints to the input text template, so that the pre-trained language model can perform semantic matching on the data type, preset matching rule information, and user viewing intention information according to the constraints to obtain template selection information.

[0139] For example, to prevent the model from making mistakes when generating template selection information and failing to select the expected rendering template, the preset matching guidance information can also include fallback and validation rules, and add the fallback and validation rules to the input text template, so that the pre-trained language model verifies the template selection information according to the fallback and validation rules when generating template selection information, and performs rollback processing according to the fallback and validation rules if the verification fails.

[0140] It should be noted that the specific model constraints in this embodiment are not limited and can be set according to actual application requirements. For example, one possible model constraint is:

[0141] 1. Require the model to only access operation and maintenance data of pre - marked data types;

[0142] 2. Only use pre - set candidate rendering templates;

[0143] 3. Must verify the template selection information based on predefined preset matching rule information;

[0144] 4. Cannot create new rendering types.

[0145] It should also be noted that the specific rollback and verification rules in this embodiment are not limited and can be set according to actual application requirements. For example, one possible rollback and verification rule is: set a default rendering template for each data type. If the target rendering template selected by the template selection information is incorrect, the default rendering template can be used as the target rendering template.

[0146] In one possible case, the first preset input template can be:

[0147] Instructions:

[0148] Based on the following input elements and preset rules, strictly follow the verification logic to match the optimal target rendering template for operation and maintenance data:

[0149] <Input elements>

[0150] 1. Data type (required):

[0151] [Time Series / Distribution / Relationship] (only accept the three preset data types)

[0152] 2. User viewing intention (required):

[0153] [Intention keywords / phrases] (example: continuous measurement values, comparison distribution, entity relationship, etc.)

[0154] 3. Data characteristics: 4]

[0155] Necessary fields: [column name / data type]

[0156] Optional fields: [column name / data type]

[0157] < / Input elements>

[0158] <Matching rules>

[0159] Execute the following steps in sequence:

[0160] Step 1: Data type verification

[0161] If the input data type is not one of the three preset data types, the process is terminated and an error is returned: "Only time series, distributed, and relational data are supported."

[0162] If the type is valid, proceed to the next step.

[0163] Step 2: Matching data features and user intent

[0164] Match the data features and context rules corresponding to the data type with the input data features and user viewing intent, select a rendering template, and perform selection verification:

[0165] 1. Time Series Operational Data

[0166] - Optional rendering templates:

[0167] Preferred templates: line chart, area chart.

[0168] Secondary templates: scatter plot, heat map.

[0169] Data feature requirements include:

[0170] Necessary requirements: Include timestamp columns and numeric columns.

[0171] Optional requirements: multiple data series, fixed time interval.

[0172] - Context rules include:

[0173] For continuous measurements, a line chart is recommended; for cumulative values, an area chart is suitable. For high-density time series data, a heat map can be used.

[0174] 2. Distributed Operations Data (Distribution)

[0175] - Optional rendering templates:

[0176] Preferred templates: Histogram, Boxplot.

[0177] Secondary templates: violin chart, bar chart.

[0178] Data feature requirements include:

[0179] Necessary requirements: Numerical or categorical data with a numerical range

[0180] Optional requirements: multi-category data, outliers

[0181] - Contextual rules:

[0182] For continuously distributed data, a histogram is recommended. For comparing distributions, a box plot is recommended. For discrete categorical data, it can be switched to a bar chart.

[0183] 3. Relational operation and maintenance data (Relationship)

[0184] - Rendering templates available for selection:

[0185] Preferred templates: Scatter plot, network diagram.

[0186] Alternative templates: Bubble chart, Sankey diagram.

[0187] Data feature requirements include:

[0188] Necessary requirements: Multivariate data set, corresponding associated entities.

[0189] Optional requirements: Weighted connection relationship, hierarchical structure.

[0190] - Context rules include:

[0191] For correlation analysis, a scatter plot is recommended. For entity relationships, a network diagram is recommended. For traffic visualization, a Sankey diagram can be used.

[0192] Step 3: Template priority selection

[0193] Preferred template: Directly match the intention and pass the necessary verification.

[0194] Alternative template: If the preferred template is not satisfied, it can be selected (e.g., for time series data without a fixed interval, the alternative is a heat map).

[0195] Forced rollback: If all templates are not satisfied, select the default rendering template and output a prompt: "No available template. Please optimize the data or adjust the intention."

[0196] < / Matching rules> [[ID=???]]

[0197] <Task rules>

[0198] 1. Require the model to only access data of pre-labeled data types; [[ID=???]]

[0199] 2. Only use pre-set candidate rendering templates;

[0200] 3. Must verify the template selection information based on pre-defined preset matching rule information;

[0201] 4. Cannot create new rendering types.

[0202] < / Task rules>

[0203] <Output format>

[0204] { <o It should be noted that there are some "???" in the original text where the tags seem to be incorrect or incomplete. I have translated based on what was provided as accurately as possible.

[0205] "data_type": "[Time Series / Distribution / Relationship]",

[0206] "user_intent": "[Description of intent]",

[0207] "selected_template": "[Template name]",

[0208] "validation_result": {

[0209] "required_features": "Meet",

[0210] "optional_features": "[Meet / Not meet]",

[0211] "fallback_triggered": "[Yes / No]"

[0212] }

[0213] }

[0214] < / Output format>

[0215] It can be seen that by setting the corresponding preset matching rule information for the candidate rendering template, integrating the data type, user viewing intent information, and preset matching rule information into the first input text, and inputting the first input text into the pre-trained language model for parsing and matching, this embodiment can combine the knowledge in the operation and maintenance field and the user intent to select a suitable rendering form for the multi-modal operation and maintenance data, thereby improving the user's usage convenience.

[0216] S203. Determine the target rendering template according to the template selection information.

[0217] In this step, the target rendering template can be determined according to the template selection information.

[0218] Furthermore, considering that the model may make mistakes when generating the template selection information, and thus fail to select the expected template, for example, the selected target rendering template cannot adapt to the data type of the operation and maintenance data. Therefore, this embodiment can also set a default rendering template for each data type. After selecting the target rendering template, it can be judged whether the target rendering template exists in the default rendering template. If it exists, it can be determined that the selection is correct. Otherwise, it can be determined that the target rendering template selected by the model is incorrect. At this time, the default rendering template can be used to render the operation and maintenance data into the operation and maintenance window to be displayed, for example, rendering with the preferred template in the default rendering template.

[0219] Based on this, after determining the target rendering template from the candidate rendering templates according to the template selection information, the method further includes:

[0220] Step 61: Determine a default rendering template corresponding to the data type, and determine whether the target rendering template exists in the default rendering template;

[0221] Step 62: If the target rendering template does not exist in the default rendering template, the operation and maintenance data is rendered into an operation and maintenance window to be displayed based on the default rendering template.

[0222] Based on the above embodiment, the following describes the automatic marking process of operation and maintenance data in this embodiment. In one possible scenario, this method may also include:

[0223] S301: Receive original operation and maintenance data.

[0224] In this step, raw operation and maintenance data refers to unlabeled operation and maintenance data that can be obtained from the monitored system. The data type of raw operation and maintenance data has not been confirmed.

[0225] S302: Extract data structure features of the original operation and maintenance data, and match the data structure features with preset rules of multiple candidate data types to determine the data type corresponding to the original operation and maintenance data.

[0226] In this step, the data structure characteristics of the original operation and maintenance data are first extracted to determine the data type of the original operation and maintenance data. In this embodiment, statistical means can be used to determine the data structure characteristics of the original operation and maintenance data to determine the data type of the original operation and maintenance data. For example, statistical means can be used to determine the following data structure characteristics:

[0227] 1. When the original operation and maintenance data has time information, the periodic changes of the original operation and maintenance data (time series characteristic) can be counted to confirm whether the original operation and maintenance data has time series characteristics.

[0228] 2. The distribution shapes of the original operation and maintenance data can be statistically analyzed to determine whether the original operation and maintenance data is distributed data.

[0229] 3. The cardinality of values ​​corresponding to the original operation and maintenance data can be counted, that is, the number of different values ​​in a column or attribute in the data set can be determined.

[0230] 4. The data type and value range corresponding to the original operation and maintenance data can be counted.

[0231] 5. It can count whether the original operation and maintenance data is missing data patterns.

[0232] Furthermore, this embodiment can determine the data structure characteristics of the original operation and maintenance data based on the statistical processing results, such as determining the data structure of the original operation and maintenance data in terms of level, data relationship, and time series, determining the update frequency of the original operation and maintenance data, and determining whether the original operation and maintenance data has value constraints, etc.

[0233] Furthermore, the candidate data types in this step are multiple data types pre-set for operation and maintenance data, such as time-series operation and maintenance data, distributed operation and maintenance data, and relational operation and maintenance data. In this embodiment, corresponding preset rules can be set for each candidate data type, and the data structure characteristics of the raw operation and maintenance data can be matched with the various preset rules to determine the data type actually corresponding to the raw operation and maintenance data.

[0234] Specifically, each candidate data type can correspond to multiple preset rules, and each preset rule can be assigned a corresponding weight value based on its importance. Furthermore, this embodiment can determine the preset rules that the data structure characteristics satisfy for each candidate data type, and sum the preset weight values ​​of the preset rules satisfied by the same candidate data type to obtain a matching value corresponding to the original operation and maintenance data. Finally, if the matching value is greater than a preset threshold, the candidate data type can be used as the data type corresponding to the original operation and maintenance data.

[0235] Based on this, matching the data structure features with preset rules of multiple candidate data types to determine the data type corresponding to the original operation and maintenance data includes:

[0236] Step 71: Match the data structure feature with the preset rule of the candidate data type to determine the target preset rule satisfied by the data structure feature;

[0237] Step 72: summing the preset weight values ​​corresponding to the target preset rules to obtain a matching value corresponding to the original operation and maintenance data;

[0238] Step 73: When the matching value is greater than a preset threshold, the candidate data type is used as the data type corresponding to the original operation and maintenance data.

[0239] It should be noted that this embodiment does not limit the preset rules corresponding to each candidate data type, and can be set according to actual application requirements. For example:

[0240] For time series operation and maintenance data, there are three preset rules:

[0241] 1. Has a timestamp column with a preset weight of 5;

[0242] 2. Regular time intervals with a preset weight of 3;

[0243] 3. Has continuous values, with a preset weight of 3.

[0244] For distributed operation and maintenance data, there are three preset rules:

[0245] 1. It has a data category with a preset weight of 5;

[0246] 2. It has a limited range of values ​​and a preset weight of 3;

[0247] 3. It has multiple discrete values, with a preset weight of 4;

[0248] For relational operation and maintenance data, there are three preset rules:

[0249] 1. There are multiple numerical columns with a preset weight of 4;

[0250] 2. Variables with multiple associations have a preset weight of 3;

[0251] 3. It has a hierarchical structure with a preset weight of 3.

[0252] Of course, step S302 can also be performed using a pre-trained machine learning model specifically for data classification. The input text (prompt) can be generated using the raw operation and maintenance data and preset rules for the candidate data types. A data analysis requirement text can be added to the input text. This allows the pre-trained machine learning model to automatically parse the raw operation and maintenance data based on the data analysis requirement text, determine the data features of the raw operation and maintenance data, and then match the data features of the raw operation and maintenance data with the preset rules for the candidate data types to determine the data type of the raw operation and maintenance data.

[0253] S303: Integrate the original operation and maintenance data and the data type into a second input text using a second preset input template; the second preset input template includes preset tags to generate guidance information.

[0254] S304: Input the second input text into a pre-trained language model, so that the pre-trained language model generates guidance information according to preset tags to parse the original operation and maintenance data and the data type, and generates semantically enhanced tags corresponding to the original operation and maintenance data.

[0255] In step S303 and step S304, in order to facilitate the query of the original operation and maintenance data in the database, semantic enhancement tags can also be generated for the original operation and maintenance data. The semantic enhancement tags are used to record the key semantic information of the original operation and maintenance data. For example, the semantic enhancement tags may include: data category (Category), data classification level (Taxonomy), data topic (Topic), and data context (Context). Among them, the data category is the basic category label that reflects the content of the original operation and maintenance data, such as hardware failure, software anomaly, etc. The data classification level is a multi-level label that reflects the content of the original operation and maintenance data, such as hardware failure> storage device> hard disk. The data topic is the subject summary of the original operation and maintenance data. The data context is the context information related to the original operation and maintenance data.

[0256] To generate the semantically enhanced labels, steps S303-S304 first integrate the raw operation and maintenance data and data types into a second input text using a second preset input template. This second preset input template should include fill-in spaces for the raw operation and maintenance data and data types, as well as preset label generation guidance information that instructs the model on how to generate semantically enhanced labels based on the raw operation and maintenance data and data types.

[0257] It should be noted that this embodiment does not limit the specific content of the preset tag generation guidance information. For example, the preset tag generation guidance information may include a parsing step of parsing the original operation and maintenance data and the data type, so that the model is matched according to the parsing step.

[0258] For another example, the preset tag generation guidance information may include output rules, and the output rules may be used to constrain the form of the template selection information output by the model.

[0259] For example, in order to constrain the model's selection of rendering templates, the preset label generation guidance information can also include model constraints, and add the model constraints to the input text template, so that the pre-trained language model understands under what constraints to parse the original operation and maintenance data and data types.

[0260] Furthermore, after completing the integration of the second input text, the second input text can be input into the pre-trained language model so that the pre-trained language model generates guidance information according to the preset labels to parse the original operation and maintenance data and data types, and generate semantically enhanced labels corresponding to the original operation and maintenance data.

[0261] S305: Add a type tag corresponding to the data type and the semantic enhancement tag to the original operation and maintenance data, and save the data to a database.

[0262] In this step, after determining the data type and semantic enhancement label corresponding to the original operation and maintenance data, you can add type tags and semantic enhancement labels to it and save it to the database. This completes the automatic labeling of the operation and maintenance data and facilitates the pre-trained language model to retrieve the labeled operation and maintenance data from the database.

[0263] Furthermore, after obtaining the semantic enhancement label, the data type of the original operation and maintenance data can be further corrected according to the semantic enhancement label to improve the accuracy of the data type.

[0264] Finally, for ease of understanding, the following is the process of data annotation, template selection, and automatic rendering:

[0265] Data acquisition → statistical analysis → data structure verification → matching template rules → language model-assisted template selection (including template verification) → visualization generation.

[0266] Furthermore, after generating the type tag and the semantic enhancement model, searching the database for target operation and maintenance data according to the information to be retrieved, and determining the target rendering template from multiple candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information, may include:

[0267] S401: Match the information to be retrieved with each operation and maintenance data in the database and with the semantically enhanced tags of each operation and maintenance data to obtain the target operation and maintenance data.

[0268] In this step, the search information can be matched with the operation and maintenance data and the semantic enhancement tags of each operation and maintenance data to improve the matching effect.

[0269] S402: Determine a target rendering template from a plurality of candidate rendering templates according to the type tag of the target operation and maintenance data and the user viewing intention information.

[0270] In this step, the data type of the target operation and maintenance data can be determined according to the type tag to perform target rendering template matching.

[0271] The interface interaction device, electronic device and computer-readable storage medium provided in the embodiments of the present invention are introduced below. The interface interaction device, electronic device and computer-readable storage medium described below can be referenced to the interface interaction method described above.

[0272] Please refer to Figure 5 , Figure 5 This is a structural block diagram of an interface interaction device provided by an embodiment of the present invention, which may include:

[0273] The model processing module 501 is used to parse the operation and maintenance dialogue text input by the user using the pre-trained language model to determine the information to be retrieved and the user's viewing intention information;

[0274] A rendering processing module 502 is configured to search a database for target operation and maintenance data based on the information to be retrieved, determine a target rendering template from a plurality of candidate rendering templates based on the data type of the target operation and maintenance data and the user's viewing intention information, and render the target operation and maintenance data into an operation and maintenance window to be displayed based on the target rendering template;

[0275] The output module 503 is configured to destroy the operation and maintenance window currently displayed in the display interface, and output the operation and maintenance window to be displayed to the display interface.

[0276] Optionally, the display interface includes a window display area and a dialogue display area; the device may further include:

[0277] A dialogue reply text generation module is used to parse the operation and maintenance dialogue text using a pre-trained language model to generate a dialogue reply text;

[0278] The output module 503 can be used to:

[0279] The operation and maintenance window currently displayed in the window display area is destroyed, the operation and maintenance window to be displayed is output to the window display area, and the dialogue reply text is output to the dialogue display area.

[0280] Optionally, the device may further include:

[0281] a window fixing module, configured to, upon receiving a window fixing command input by a user, determine a target operation and maintenance window corresponding to the window fixing command in the display interface, and set the target operation and maintenance window as a fixed operation and maintenance window;

[0282] a window unpinning module configured to, upon receiving a window unpinning command input by a user, determine a target fixed operation and maintenance window corresponding to the window pinning command in the display interface, and set the target fixed operation and maintenance window to a non-fixed operation and maintenance window;

[0283] The output module 503 can be used to:

[0284] Destroy the non-fixed operation and maintenance window in the display interface.

[0285] Optionally, the rendering processing module 502 may include:

[0286] An information acquisition submodule, configured to acquire preset matching rule information corresponding to each candidate rendering template;

[0287] A first input text integration module is configured to integrate the data type, the preset matching rule information, and the user viewing intention information into a first input text using a first preset input template; the first preset input template includes preset matching guidance information;

[0288] a template selection submodule, configured to input the first input text into the pre-trained language model, so that the pre-trained language model matches the data type and the user viewing intention information with the preset matching rule information according to the preset matching guidance information, and obtain template selection information;

[0289] The template determination submodule is configured to determine the target rendering template according to the template selection information.

[0290] Optionally, the device may further include:

[0291] A template verification module is used to determine a default rendering template corresponding to the data type and to judge whether the target rendering template exists in the default rendering template; if the target rendering template does not exist in the default rendering template, the operation and maintenance data is rendered into an operation and maintenance window to be displayed based on the default rendering template.

[0292] Optionally, the model processing module 501 may include:

[0293] A historical text acquisition submodule is used to obtain the historical operation and maintenance conversation texts of the user;

[0294] The model parsing submodule is used to parse the historical operation and maintenance dialogue text and the operation and maintenance dialogue text using the pre-trained language model to determine the information to be retrieved and the user's viewing intention information.

[0295] Optionally, the device may further include:

[0296] Receiving module, used to receive original operation and maintenance data;

[0297] A type determination module is used to extract data structure features of the original operation and maintenance data, and match the data structure features with preset rules of multiple candidate data types to determine the data type corresponding to the original operation and maintenance data;

[0298] A second text integration module is configured to integrate the original operation and maintenance data and the data type into a second input text using a second preset input template; the second preset input template includes a preset tag to generate guidance information;

[0299] a label generation module, configured to input the second input text into a pre-trained language model, so that the pre-trained language model parses the raw operation and maintenance data and the data type according to preset label generation guidance information, and generates a semantically enhanced label corresponding to the raw operation and maintenance data;

[0300] A label setting module is used to add a type tag corresponding to the data type to the original operation and maintenance data and save it to the database;

[0301] The rendering processing module 502 includes:

[0302] A first matching submodule is configured to match the information to be retrieved with each operation and maintenance data in the database and with a semantically enhanced label of each operation and maintenance data to obtain the target operation and maintenance data;

[0303] The second matching submodule is configured to determine a target rendering template from a plurality of candidate rendering templates according to the type tag of the target operation and maintenance data and the user viewing intention information.

[0304] Optionally, the type determination module may include:

[0305] A feature extraction submodule is used to perform statistical processing on the raw operation and maintenance data, and determine the data structure features of the raw operation and maintenance data based on the statistical processing results;

[0306] A matching submodule, configured to match the data structure feature with the preset rule of the candidate data type, and determine the target preset rule satisfied by the data structure feature;

[0307] A matching value calculation submodule, configured to sum the preset weight values ​​corresponding to the target preset rules to obtain a matching value corresponding to the original operation and maintenance data;

[0308] The type setting submodule is used to use the candidate data type as the data type corresponding to the original operation and maintenance data when the matching value is greater than a preset threshold.

[0309] Please refer to Figure 6 , Figure 6 This is a structural block diagram of an electronic device provided by an embodiment of the present invention. The embodiment of the present invention provides an electronic device 60, including a processor 61 and a memory 62; wherein the memory 62 is used to store a computer program; and the processor 61 is used to execute the interface interaction method provided by the aforementioned embodiment when executing the computer program.

[0310] For the specific process of the above-mentioned interface interaction method, please refer to the corresponding content provided in the aforementioned embodiments, which will not be repeated here.

[0311] Furthermore, the memory 62 as a carrier for resource storage may be a read-only memory, a random access memory, a magnetic disk or an optical disk, etc., and the storage method may be temporary storage or permanent storage.

[0312] In addition, the electronic device 60 also includes a power supply 63, a communication interface 64, an input / output interface 65 and a communication bus 66; wherein, the power supply 63 is used to provide an operating voltage for each hardware device on the electronic device 60; the communication interface 64 can create a data transmission channel between the electronic device 60 and an external device, and the communication protocol it follows is any communication protocol that can be applied to the technical solution of the present invention, and is not specifically limited here; the input / output interface 65 is used to obtain external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0313] An embodiment of the present invention further provides a computer program product, including a computer program / instruction, which implements the interface interaction method described in the above embodiment when executed by a processor.

[0314] Since the embodiments of the computer program product part correspond to the embodiments of the interface interaction method part, please refer to the description of the embodiments of the interface interaction method part for the embodiments of the computer program product part, and will not be repeated here.

[0315] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the interface interaction method described in the above embodiment is implemented.

[0316] Since the embodiments of the computer-readable storage medium part correspond to the embodiments of the interface interaction method part, the embodiments of the storage medium part refer to the description of the embodiments of the interface interaction method part and will not be repeated here.

[0317] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0318] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0319] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0320] The above is a detailed introduction to an interface interaction method, device, electronic device and storage medium provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present invention, the present invention can also be improved and modified, and these improvements and modifications also fall within the scope of protection of the present invention.

Claims

1. An interface interaction method, characterized in that: include: When receiving the operation and maintenance dialogue text input by the user, the operation and maintenance dialogue text is parsed using the pre-trained language model to determine the information to be retrieved and the user's viewing intention information; Searching a database for target operation and maintenance data according to the information to be retrieved, determining a target rendering template from a plurality of candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information, and rendering the target operation and maintenance data into an operation and maintenance window to be displayed based on the target rendering template; The operation and maintenance window currently displayed in the display interface is destroyed, and the operation and maintenance window to be displayed is output to the display interface.

2. The interface interaction method according to claim 1, characterized in that: The display interface includes a window display area and a dialogue display area; The method further comprises: Parsing the operation and maintenance dialogue text using a pre-trained language model to generate dialogue response text; The destroying the operation and maintenance window currently displayed in the display interface and outputting the operation and maintenance window to be displayed to the display interface includes: The operation and maintenance window currently displayed in the window display area is destroyed, the operation and maintenance window to be displayed is output to the window display area, and the dialogue reply text is output to the dialogue display area.

3. The interface interaction method according to claim 1, characterized in that: Also includes: When a window fixation command input by a user is received, determining a target operation and maintenance window corresponding to the window fixation command in the display interface, and setting the target operation and maintenance window as a fixed operation and maintenance window; When receiving a window unpinning command input by the user, determining a target fixed operation and maintenance window corresponding to the window pinning command in the display interface, and setting the target fixed operation and maintenance window to a non-fixed operation and maintenance window; Destroying the operation and maintenance window currently displayed in the display interface includes: Destroy the non-fixed operation and maintenance window in the display interface.

4. The interface interaction method according to claim 1, characterized in that: The determining a target rendering template from a plurality of candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information includes: Obtaining preset matching rule information corresponding to each candidate rendering template; The data type, the preset matching rule information, and the user viewing intention information are integrated into a first input text using a first preset input template; the first preset input template includes preset matching guide information; inputting the first input text into the pre-trained language model, so that the pre-trained language model matches the data type and the user viewing intention information with the preset matching rule information according to the preset matching guidance information to obtain template selection information; The target rendering template is determined according to the template selection information.

5. The interface interaction method according to claim 4, characterized in that: After determining a target rendering template from the candidate rendering templates according to the template selection information, the method further includes: Determine a default rendering template corresponding to the data type, and determine whether the target rendering template exists in the default rendering template; If the target rendering template does not exist in the default rendering template, the target operation and maintenance data is rendered into an operation and maintenance window to be displayed based on the default rendering template.

6. The interface interaction method according to claim 1, characterized in that: The use of a pre-trained language model to parse the operation and maintenance dialogue text to determine the information to be retrieved and the user's viewing intention information includes: Obtain the historical operation and maintenance conversation text of the user; The historical operation and maintenance dialogue text and the operation and maintenance dialogue text are parsed using the pre-trained language model to determine the information to be retrieved and the user viewing intention information.

7. The interface interaction method according to any one of claims 1 to 6, characterized in that: Also includes: Receive raw operation and maintenance data; Extracting data structure features of the original operation and maintenance data, and matching the data structure features with preset rules of multiple candidate data types to determine the data type corresponding to the original operation and maintenance data; Integrating the original operation and maintenance data and the data type into a second input text using a second preset input template; The second preset input template includes preset label generation guidance information; Inputting the second input text into a pre-trained language model, so that the pre-trained language model parses the original operation and maintenance data and the data type according to preset label generation guidance information, and generates a semantic enhancement label corresponding to the original operation and maintenance data; Adding a type tag corresponding to the data type and the semantic enhancement tag to the original operation and maintenance data, and saving the data to a database; The step of searching a database for target operation and maintenance data according to the information to be retrieved, and determining a target rendering template from a plurality of candidate rendering templates according to a data type of the target operation and maintenance data and the user viewing intention information, includes: Matching the information to be retrieved with each operation and maintenance data in the database and with the semantically enhanced tags of each operation and maintenance data to obtain the target operation and maintenance data; A target rendering template is determined from a plurality of candidate rendering templates according to the type tag of the target operation and maintenance data and the user viewing intention information.

8. The interface interaction method according to claim 7, characterized in that: The extracting data structure features of the original operation and maintenance data, and matching the data structure features with preset rules of multiple candidate data types to determine the data type corresponding to the original operation and maintenance data includes: Performing statistical processing on the original operation and maintenance data, and determining data structure characteristics of the original operation and maintenance data based on the statistical processing results; Matching the data structure feature with the preset rule of the candidate data type to determine the target preset rule satisfied by the data structure feature; Summing the preset weight values ​​corresponding to the target preset rules to obtain a matching value corresponding to the original operation and maintenance data; When the matching value is greater than a preset threshold, the candidate data type is used as the data type corresponding to the original operation and maintenance data.

9. An interface interaction device, characterized in that: include: The model processing module is used to parse the operation and maintenance dialogue text input by the user using the pre-trained language model to determine the information to be retrieved and the user's viewing intention information; a rendering processing module, configured to search a database for target operation and maintenance data according to the information to be retrieved, determine a target rendering template from a plurality of candidate rendering templates according to the data type of the target operation and maintenance data and the user viewing intention information, and render the target operation and maintenance data into an operation and maintenance window to be displayed based on the target rendering template; The output module is used to destroy the operation and maintenance window currently displayed in the display interface and output the operation and maintenance window to be displayed to the display interface.

10. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the interface interaction method according to any one of claims 1 to 8 when executing the computer program.

Images