Session message processing method and device, electronic equipment and storage medium

Abstract

The invention provides a session message processing method and device, electronic equipment and a storage medium, according to the embodiment of the invention, at least one session message is acquired from a session client, and the session message corresponds to one or more information formats; secondly, generating a cue word used for requesting analysis, and sending the cue word to the large language model to generate content analysis data of the session message, the cue word carrying the session message; and finally, returning the content analysis data to the session client, displaying an interaction control for triggering content analysis by the session client, and triggering execution of content analysis operation after the interaction control is accessed. By adopting the scheme, the session message can be intelligently analyzed and summarized, the problems of more messages, complicated processing, inconvenient operation and the like are solved for a user of the session client, and the processing efficiency of the session message is further improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, electronic device, and storage medium for processing session messages. Background Technology

[0002] With the advancement of computer technology and the development of artificial intelligence and machine learning, enterprises face the need to improve efficiency and reduce costs, leading to the shift of more and more office tasks and conversations online. Taking the handling of work-related chat messages as an example, the content or type of these messages may involve multiple aspects of the company's work. Different messages have varying degrees of importance or involve different office tasks, requiring responses or processing based on priority. Alternatively, long messages may need to be read quickly. Using traditional message processing methods might require manual differentiation of key points and follow-up, resulting in complex and inconvenient operations that consume significant human time.

[0003] Therefore, there is an urgent need for a solution to handle online chat messages, which can help employees break free from the daily grind of online chat messages, help them distinguish the importance of chat messages or identify key matters in their work, and follow up on them. Summary of the Invention

[0004] This application provides a method, apparatus, electronic device, and storage medium for processing session messages to solve one or more of the above-mentioned technical problems.

[0005] In a first aspect, embodiments of this application provide a method for processing conversation messages, comprising: obtaining at least one conversation message from a conversation client, the conversation message corresponding to one or more information formats; generating a prompt word for requesting content analysis, generating content analysis data of the conversation message by sending the prompt word to a large language model, the prompt word carrying the conversation message; returning the content analysis data to the conversation client; the conversation client displaying an interactive control for triggering content analysis, the interactive control triggering the execution of content analysis operations after being accessed.

[0006] Secondly, embodiments of this application provide a method for processing conversation messages, including: in response to the access of an interactive control of a conversation client, obtaining at least one conversation message, the conversation message corresponding to one or more information formats, the interactive control triggering the execution of a content analysis operation after being accessed; obtaining content analysis data of the conversation message generated by calling a large language model based on the conversation message; and displaying the content analysis data.

[0007] Thirdly, embodiments of this application provide a session message processing apparatus, comprising: a session message acquisition module, configured to acquire at least one session message from a session client, the session message corresponding to one or more information formats; a prompt word generation module, configured to generate prompt words for requesting content analysis, and generate content analysis data of the session message by sending the prompt words to a large language model, the prompt words carrying the session message; and a content analysis data return module, configured to return the content analysis data to the session client; the session client displays an interactive control that triggers content analysis, and the interactive control triggers the execution of a content analysis operation after being accessed.

[0008] Fourthly, embodiments of this application provide a session message processing apparatus, comprising: a session message acquisition module, configured to acquire at least one session message in response to the access of an interactive control of a session client, wherein the session message corresponds to one or more information formats, and the access of the interactive control triggers the execution of a content analysis operation; a content analysis data acquisition module, configured to acquire content analysis data of the session message generated by calling a large language model based on the session message; and a content analysis data display module, configured to display the content analysis data.

[0009] Fifthly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor, when executing the computer program, implements the method described in any of the above-mentioned embodiments.

[0010] Sixthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any of the preceding claims.

[0011] According to the embodiments of this application, at least one conversation message is first obtained from the conversation client, wherein the conversation message corresponds to one or more information formats; secondly, a prompt word for requesting content analysis is generated, and content analysis data of the conversation message is generated by sending the prompt word to a large language model, wherein the prompt word carries the conversation message; finally, the content analysis data is returned to the conversation client, and the conversation client displays an interactive control that triggers the content analysis data. After the interactive control is accessed, the content analysis operation is triggered. Using the above scheme, conversation messages can be intelligently analyzed and summarized, solving problems such as numerous messages, complex processing, and inconvenient operation for users of the conversation client, thereby improving the processing efficiency of conversation messages.

[0012] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application, it can be implemented according to the contents of the specification. In order to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0013] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the various drawings denote the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments according to this application and should not be construed as limiting the scope of this application.

[0014] Figure 1 A flowchart illustrating a session message processing scheme provided in an embodiment of this application is shown.

[0015] Figure 2 A schematic diagram illustrating an application example of a session message processing scheme provided in an embodiment of this application is shown.

[0016] Figure 3 This illustration shows an application example diagram of another session message processing scheme provided in the embodiments of this application;

[0017] Figure 4 This illustration shows an application example diagram of another session message processing scheme provided in the embodiments of this application;

[0018] Figure 5 A flowchart illustrating a session message processing method provided in an embodiment of this application is shown;

[0019] Figure 6 A flowchart illustrating another method for processing session messages provided in an embodiment of this application is shown;

[0020] Figure 7 A structural block diagram of a session message processing apparatus provided in an embodiment of this application is shown;

[0021] Figure 8 This invention illustrates a structural block diagram of another session message processing apparatus provided in an embodiment of this application; and

[0022] Figure 9 A block diagram of an electronic device used to implement embodiments of this application is shown. Detailed Implementation

[0023] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the concept or scope of this application. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0024] To facilitate understanding of the technical solutions of the embodiments of this application, the relevant technologies of the embodiments of this application are described below. The following relevant technologies are optional solutions and can be combined with the technical solutions of the embodiments of this application in any way, and all of them fall within the protection scope of the embodiments of this application.

[0025] Figure 1 The diagram illustrates a flowchart of a session message processing scheme provided in an embodiment of this application. The session message processing scheme provided in this embodiment can be implemented through five steps: information collection, initial information processing, recall, fine information processing, and scenario integration.

[0026] First, the server can obtain at least one session message from the session client (1. Information Collection). This session message can correspond to one or more information formats, such as IM session text (instant messaging messages), voice, images, links, files, or IM documents (instant messaging documents). Furthermore, it can identify external data (the internet) linked to by the session message, including database data, web page data, and server data. By collecting information from the session client's session messages, it can connect various types of IM session messages, such as text, images, files, and documents, gathering the necessary materials for subsequent analysis or responses.

[0027] Furthermore, the collected conversation messages can be formatted and swapped or structured (2. Preliminary information processing). For example, the collected conversation messages can be preliminarily processed by using pre-set conversation message templates (metadata) to identify keywords and extract key content from the conversation messages (information extraction).

[0028] Then, the information after initial processing can be recalled (3. Recall). If it is a document or file type message, such as a document in external data (Internet) linked to by the session message, the external data can be divided into multiple data segments through intelligent slicing (for example, segmenting a large document in a large file to obtain part of the content in the large file); or the indexing capability can be used to retrieve data from the collected session messages, store and / or query the relevant content after data retrieval, such as document retrieval content, internal database retrieval content, content in the session message related to a certain topic or project, etc.

[0029] Secondly, prompt words for requesting content analysis can be generated based on the data after initial information processing and recall (or the original collected conversation messages). Content analysis data for the conversation messages is then generated by sending these prompt words to a large language model (4. Information Refinement). Further processing can be performed on the data after initial information processing and recall (or the original collected conversation messages), such as summarizing large files, summarizing and analyzing icon / image content (slice summaries or data combinations), to generate prompt words for requesting content analysis. For example, when the conversation message also carries the target content analysis dimensions used for content analysis, such as message read / unread status, message time range, followed users, etc., prompt words carrying both the conversation message and the target content analysis dimensions can be generated; when external data is segmented into multiple data fragments using intelligent slicing, prompt words carrying both the conversation message and at least one data fragment can be generated; when a reply is required, prompt words for requesting content analysis and a reply can be generated. By sending prompt words requesting content analysis to a large language model to generate content analysis data for conversational messages, the large language model can be used to further refine the content of the acquired conversational messages, generating corresponding responses, summaries, or to-do items to support the summary analysis of conversational messages, identify important matters, and follow-up actions.

[0030] Finally, content analysis data can be returned to the session client. The session client can display interactive controls that trigger content analysis at relevant locations within the session message or other locations. Accessing these interactive controls can trigger the execution of content analysis operations (5. Scenario Integration). The content analysis data can be a summary of the session messages, unread messages, or weekly reports, etc.

[0031] By integrating with the IM intelligent entry point (instant messaging intelligent entry point), which involves placing entry points through interactive controls displayed in the instant messaging client (such as floating windows or web links), the display of content analysis data can be seamlessly integrated with existing instant messaging communication scenarios. This allows for WYSIWYG (What You See Is What You Get) responses, summaries, and to-do information for conversation messages to be displayed through multiple entry points. For example, unread messages can be accessed through multiple channels such as preset anchor points, shortcuts, and voice commands. Message content can be intelligently read and analyzed, supporting various information formats such as text, links, and images. By differentiating conversation messages based on multiple dimensions such as read / unread status, message time range, and followed individuals, it can support summarizing conversation messages and analyzing and identifying important matters and follow-up actions. It can also intelligently follow up on conversation messages or matters related to conversation messages. Through conversation message analysis, it can identify whether subsequent actions are reply-type or action-type. Reply-type actions can support intelligent replies, and action-type actions can support intelligent follow-ups. For example, it can identify content follow-up matters corresponding to the content analysis data, including to-do items, schedule items, or meeting items.

[0032] Correspondingly, the session message processing scheme provided in this application embodiment can also be to respond to the access of the interactive control of the session client, obtain at least one session message, wherein the session message corresponds to one or more information formats, and the access of the interactive control can trigger the execution of content analysis operations; obtain the content analysis data of the session message generated based on the session message by calling a large language model; and display the content analysis data.

[0033] Specifically, Figure 2 This illustration shows an application example of a session message processing scheme provided in this application. The interactive controls can be directly displayed or "hidden." They can be directly displayed on the session page of the session client, for example, displayed at the relevant location of the session message, such as... Figure 2 The "Quick Read, Intelligent Image Recognition" button displayed below the image in the conversation message is a directly displayed interactive control. Users can click on this control to trigger the subsequent process of retrieving conversation messages for intelligent analysis. The interactive control can also be "hidden," meaning it can be accessed through other entry points (such as wake-up controls) displayed on the conversation page of the conversation client. After users enter the interface that allows access to the interactive control through this entry point, they can click on the interactive control to trigger the subsequent process of retrieving conversation messages for intelligent analysis.

[0034] Before the interactive controls of the session client are accessed and at least one session message is retrieved, the user can also access the wake-up control of the session client, allowing the session client to provide an interactive interface displaying the interactive controls (which may include voice input controls or text input controls). Taking the application scenario of a user waking up the AI ​​assistant of the session client as an example, such as... Figure 2 As shown, the conversation client can be in the form of a chat window, and may include a message input dialog box (which may display "Please enter a message" and icons for sending emoticons, screenshots, etc.), a chat interface (which may display chat content, such as images, text, and notifications of whether messages have been read or not), a chat username (e.g., "Jingjing"), icons for search tools, dedicated services, video and voice call tools, folder tools, friend adding, settings tools, and other small tools. This application does not impose any restrictions on the style of the conversation client. At the top of the conversation page (chat window) of the conversation client, an AI icon may be displayed. The wake-up control can be any form of wake-up control, and this application does not impose any restrictions on it. When the user clicks the wake-up control, in response to the access of the wake-up control in the session client, an interactive interface displaying interactive controls can be provided to the user in the session client, such as the session interaction page between Jingjing and the "AI Assistant". In this interaction page, the AI ​​Assistant can display interactive controls, such as input boxes in the form of voice input controls or text input controls. In the input box, a prompt can be displayed: "Ask me any questions". The user can access the interactive controls in the session client, such as entering voice, text, or images in the interactive control (input box) to trigger the subsequent process of obtaining at least one session message. In addition, this interactive page can also display other prompts for the AI ​​assistant to guide users on how to use interactive controls and to input content for these controls. For example, it could prompt, "Hello, I am your AI assistant. I am still a novice assistant, constantly upgrading and improving. Please pay attention! Try asking questions using the following methods: Use AI to write a song for me; Imagine the mode of transportation in 2030; View my work summary; Open my disk; Visit the [AI Community] for learning and exchange." The content of user interactions with the AI ​​assistant can be used as training samples for training large-scale language models, making the solutions provided by the AI ​​assistant more intelligent and more tailored to the user's personalized needs.

[0035] Furthermore, in response to the access of the interactive controls of the conversation client, the system can determine whether to trigger a content analysis operation by recognizing the voice data or text data (or image data, etc.) submitted based on the interactive controls, and retrieve at least one conversation message in response to the triggering of the content analysis operation. For example, a user can enter "Summarize the conversation messages for me" in the input box (i.e., the interactive control) on the dialogue page with the AI ​​assistant (the interactive control is accessed), and then determine whether to trigger a content analysis operation by recognizing the text data "Summarize the conversation messages for me" submitted based on the interactive control, and retrieve at least one conversation message in response to the triggering of the content analysis operation.

[0036] In some embodiments, in response to access to the interactive controls of the session client, at least one content analysis dimension can be displayed to obtain the target content analysis dimension selected by the user. By obtaining at least one session message under the target content analysis dimension, interaction with a large language model is performed to obtain content analysis data of the session message generated by the large language model based on the target content analysis dimension. The content analysis data may also carry corresponding content keywords. Figure 2 As shown in the example above, after the user inputs "Summarize the conversation messages for me" through the interactive control (the interactive control is accessed), multiple content analysis dimensions can be displayed on the conversation page between the conversation client and the AI ​​assistant. These dimensions can include "Generated by conversation," "Generated by group speaker," and "Generated by time," etc. The content analysis dimensions can be preset, and this application does not impose any restrictions on this. Simultaneously, the page displaying the content analysis dimensions can also prompt the user with messages such as "I understand you need to summarize your chat content in a specific conversation today," and "Hello, you can generate summaries or conclusions in the following ways: by conversation, by group speaker, or by time, with content generated by AI." Furthermore, if the user selects a specific content analysis dimension (the target content analysis dimension), such as "Generated by conversation," at least one conversation message from the user-specified conversation can be obtained. Then, content analysis data based on the "Generated by conversation" conversation message can be obtained through interaction with a large language model. In addition, users can limit the scope of conversation messages they receive by selecting a specific conversation window. For example, if the user selects the conversation window of the group chat "Network Family", the interactive interface (the conversation page with the AI ​​assistant) can provide the user with the corresponding selection entry. This application does not impose any restrictions on this.

[0037] If users are not satisfied with the generated content analysis data, or the content analysis dimensions and dialogue window selection entries provided by the interactive interface (conversation page with the AI ​​assistant), they can also input more specific needs (text, voice, or image format, etc.) through interactive controls (input boxes for dialogue with the AI ​​assistant). Through multiple interactions with the large language model, the content analysis data required by the user can be obtained more accurately and in a more personalized way.

[0038] Figure 3 This illustration shows an application example of another session message processing scheme provided in this application embodiment. After generating and displaying the content analysis data of the session message, the corresponding content follow-up items can be obtained. After the interactive control is triggered, a follow-up control for the content follow-up function can be configured in the display interface of the content analysis data. The operation interface of the content follow-up function is displayed after the follow-up control is triggered.

[0039] like Figure 3 As shown, taking the intelligent generation of chat history summaries as an example, firstly, user "Sally" wants to obtain a summary of 24 chat records from the "**Family" group chat page within the last two days. At this time, the conversation client can add an interactive interface for dialogue with the AI ​​assistant at the relevant location of the conversation messages (e.g., in the "**Family" group chat page of "**Technology Co., Ltd."). In this interactive interface, interactive controls for content analysis data can be displayed (e.g., the dialogue input box with the "AI assistant" in the previous example). The user can request the "AI assistant" to generate a summary of these 24 conversation messages in the dialogue input box (the interactive controls are accessed). Then, through interaction with a large language model, the content analysis data of these 24 conversation messages is obtained and displayed. The displayed content analysis data (conversation message summary) can specifically include prompts, topics, progress, details, action suggestions, etc. For example, the system may display messages such as "Based on the summary of 24 chat records from the '**Family' group chat over the past two days, the following content is presented"; the topic "Rich text reply function optimization"; the progress "Avatar update issue to be resolved"; details "Zhang Mingming suggested that the color of the rich text editing button needs optimization"; and action suggestions "Follow up on the button color issue." It may also display entry points such as "Content generated by AI, click to learn more." In this embodiment, the dimensions, style, and word count of the displayed content analysis data can be pre-configured, and this application does not impose any limitations on this.

[0040] Furthermore, after displaying the aforementioned content analysis data, the corresponding content follow-up items can also be obtained. These follow-up items can include to-do items, schedule items, or meeting items. After the interactive control is triggered, a follow-up control for the content follow-up function can be configured in the content analysis data display interface. For example... Figure 3As shown, the follow-up control for the configured content follow-up function can be "Create To-Do". There can be one or more follow-up controls. Users can trigger the follow-up control by clicking "Create To-Do" or by talking to the "AI Assistant" (entering text, voice, or images in the dialogue input box). For example, if you enter "Help me create a to-do for the follow-up button color problem" in the dialogue with the "AI Assistant", the conversation client can display the operation interface of the content follow-up function after the follow-up control is triggered based on the conversation message. The interface for the content follow-up function can include response messages to the dialogue between the user and the "AI assistant," such as "Okay, the following information is required to create a to-do. Please confirm and click Create." It can also include entry points or controls for selecting the executor, adding a to-do title, selecting a deadline, selecting a reminder time, and creating the to-do. For example, you can select the executor "Sally" (which can be done by clicking the "Add" control under Select Executor), add the to-do title "Follow-up Button Color Issue," select the deadline "2023-12-15 14:00," and select the reminder time "15 minutes before the deadline." When the user clicks "Create To-Do" on the content follow-up function's interface, the above-mentioned to-do items can be created, and the user will be reminded at the set time.

[0041] Figure 4 This illustration shows an application example of another session message processing scheme provided in this application embodiment. Besides summarizing the session message itself and creating content follow-up items, the server can also identify external data or content in images linked to by the session message. For example, after a user clicks the interactive control "Quick Read: Intelligent Image Recognition," content analysis data (summary) about the content in the image can be generated (see [link to relevant documentation]). Figure 2 Alternatively, users can generate content analysis data (summaries) about the content in the external data by clicking the interactive control below the link that says "Quick Read: Click here to read in 3 seconds." External data can include database data, web page data, server data, etc. This can be achieved by segmenting the external data into multiple data fragments, generating prompts carrying conversation messages and at least one data fragment, and then sending a request for content analysis to a large language model, along with the prompts carrying conversation messages and at least one data fragment, to generate content analysis data for the conversation message, which is then displayed on the conversation client.

[0042] like Figure 4As shown, the conversation message in the conversation client contains a link to webpage data titled "Comprehensive Survey on the Development and Application of Artificial Intelligence Agents (AIAgent) in 2023: Concepts, Principles, etc." The solution provided in this application embodiment can display an interactive control below the link that says "Quick Read: Click here to read in 3 seconds." Users can click on this interactive control to generate prompts carrying conversation messages and at least one data fragment by dividing external data into multiple data segments. Content analysis data about the webpage content in the webpage data is then generated through a large language model and displayed in the conversation client. For example, content analysis data could be a summary of the content in the webpage data: "The content of this webpage link is summarized as follows: This article introduces the concept, principles, development, applications, challenges, and prospects of AI agents. AI agents work in a self-directed, cyclical manner, setting tasks for artificial intelligence, prioritizing tasks, and reprioritizing tasks until the overall goal is achieved. AI agents can be driven by large language models, enabling language interaction, decision-making capabilities, flexible adaptation, and collaborative interaction. They can be applied to various scenarios such as personal assistants, multi-agent systems, human-machine collaboration, and professional fields. The article mentions some of the best open-source AI agents and multi-agent frameworks, such as LangChain, AutoGen, and PromptAppGPT. In addition, the article explores the role of AI agents in research and data collection, content generation, web crawling, document and spreadsheet summarization, language translation, virtual assistants for creative tasks, and automated management tasks. However, AI agents also face challenges, such as the problem of relying on large language models and the illusion problem. With the iteration and development of future artificial intelligence models, AI agents will have a profound impact on productivity and workflows."

[0043] The execution entity in this application embodiment can be an application, service, instance, functional module in software form, virtual machine (VM), container, or cloud server, or a hardware device (such as a server or terminal device) or hardware chip (such as a CPU, GPU, FPGA, NPU, AI accelerator card, or DPU) with data processing capabilities. The device for implementing session message processing can be deployed on the computing device of the application providing the corresponding service or on a cloud computing platform providing computing power, storage, and network resources. The cloud computing platform can provide services in the following modes: IaaS (Infrastructure as a Service), PaaS (Platform as a Service), SaaS (Software as a Service), or DaaS (Data as a Service). Taking the platform providing SaaS (Software as a Service) as an example, the cloud computing platform can utilize its own computing resources to provide training for the session message processing model or execution of the session message processing module. The specific application architecture can be built according to service requirements. For example, the platform can provide building services based on the above model to application parties or individuals using platform resources, and further invoke the above model and implement online or offline session message processing functions based on session message processing requests submitted by relevant client or server devices.

[0044] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0045] The technical solution of this application and how it solves the aforementioned technical problems are described in detail below with specific embodiments. The listed specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0046] This application provides a method for processing session messages, such as... Figure 5 The diagram shown is a flowchart of a session message processing method 500 according to an embodiment of this application. The method 500 may include:

[0047] In step S501, at least one session message is obtained from the session client, and the session message corresponds to one or more information formats.

[0048] The conversation messages involved in this application embodiment may include chat information generated or received by the conversation client, web page links, plugins, or certain program entry points, etc., as long as they can be obtained from the conversation client. This application does not impose any restrictions on this. The information format corresponding to the conversation messages may be IM conversation text (instant messaging information), voice, images, files or documents, links, etc., and this application does not impose any restrictions on this.

[0049] In step S502, a prompt word for requesting content analysis is generated, and the content analysis data of the conversation message is generated by sending the prompt word to a large language model. The prompt word carries the conversation message.

[0050] For example, generating prompts for requesting content analysis can be achieved by retrieving prompt templates associated with the current session message from a prompt database and then modifying the session message according to these templates. In other words, when generating prompts for requesting content analysis, if the session message deviates significantly from the pre-trained templates of a large language model, the session message can be corrected. This involves searching the database of prompt templates for the large language model, finding templates related to the content of the session message, correcting the issues according to the templates, generating prompts for requesting content analysis, and then sending the prompts to the large language model.

[0051] The large-scale language models involved in this application refer to deep learning models or neural network models trained using large amounts of text data. These models work by analyzing large amounts of text data and learning language usage patterns, generating natural language text or understanding the meaning of language text. The generated text closely resembles what humans say or write. They are mainly categorized into autoencoder-based models, sequence-to-sequence models, Transformer-based models, recurrent neural network models, and hierarchical models, etc. Different large-scale language models excel at handling different types of tasks. For example, autoencoder-based models encode the input text into a low-dimensional representation and then generate new text from that representation, making them suitable for tasks such as text summarization or content generation, and can be used for generating various content types related to checklists. Transformer-based models employ a neural network architecture, excelling at understanding long-distance dependencies in text data, and are suitable for handling various language tasks such as text generation, language translation, and question answering, and can also be used for generating various content types related to checklists. In specific implementations, the appropriate large-scale language model can be selected as needed, and this application does not impose any restrictions on this.

[0052] Large-scale language models can simulate human dialogue, obtaining output results through input questions. The input question format can include input prompt words. Different prompt words cover different amounts of information and express different ways, significantly impacting the output results. Creating prompt word templates can effectively address these issues, but the appropriate template for different input content needs to be determined. Therefore, this embodiment retrieves prompt word templates associated with the acquired conversational message from a prompt word database and modifies the conversational message according to the prompt word templates. This embodiment finds prompt word templates that match the conversational message, ensuring compatibility between the prompt word templates and the conversational message. This allows for better questioning methods given a fixed amount of information in the acquired conversational message, leading to higher-quality output results from the large-scale language model. After obtaining a matching prompt word template, the conversational message and the prompt word template can be combined to generate target prompt words for input into the large-scale language model, further modifying the conversational message.

[0053] The prompt word database involved in this application embodiment can be a database storing various prompt word templates. These prompt word templates can be created based on different conversation scenarios, different conversation content, or different conversation objects, and can be stored in a cloud-based database. This application does not impose any restrictions on the type or storage format of the prompt word database. Conversation messages of different content types or different information formats can each have corresponding prompt word template libraries. The prompt word database storing prompt word templates can be pre-configured and used to generate prompt words for large language model input by combining the acquired conversation messages. Reply messages or content analysis data of conversation messages of different content types can be generated by calling the same or different large language models, which can be configured according to actual needs.

[0054] In one possible implementation, the session message also carries the target content analysis dimension used for content analysis. The above-mentioned method of generating prompt words for requesting content analysis can be to generate prompt words that carry both the session message and the target content analysis dimension.

[0055] Target content analysis dimensions can be based on analysis dimensions related to the session message, such as the sender of the session message, the content of the session message, and the time of the session message. Figure 2The target content analysis dimensions, such as "generated by conversation," "generated by group speaker," or "generated by time," can be pre-configured, or multiple dimensions can be pre-set for users to choose from. This application does not impose any restrictions on the specific dimension settings and determination methods for the target content analysis dimensions. After generating prompt words carrying conversation messages and target content analysis dimensions, the prompt words can be sent to a large language model so that the large language model can generate content analysis data for the conversation messages from the target analysis dimensions.

[0056] For example, before generating the prompt for requesting content analysis, external data linked to by the session message can be identified and segmented into multiple data fragments, wherein the external data may include at least one of database data, web page data, and server data; correspondingly, generating the prompt for requesting content analysis may be generating a prompt carrying the session message and at least one data fragment.

[0057] Because external data often contains large documents or numerous files, segmenting it into multiple data fragments can reduce redundant content (e.g., removing some fragments), extract most of the content, and improve the efficiency of generating corresponding prompts. For example, by segmenting documents in web page data and extracting prompts from different paragraphs, if the keyword "AI technology" is extracted from multiple paragraphs, the prompt "AI technology," carrying conversational messages and at least one data fragment, can be input into a large language model to obtain content analysis data generated by the large language model.

[0058] In one possible implementation, after generating content analysis data for conversational messages by sending prompt words to a large language model, content recognition can be performed on the content analysis data, and content keywords can be added to identify the various types of content included in the content analysis data.

[0059] The content keywords can be pre-defined business keywords based on requirements, used to help users of the session client quickly distinguish key matters, matter types, etc., and follow up accordingly. For example, content keywords can involve keywords related to projects, risks, progress, etc., and can include "finance," "human resources," "database," "maintenance," "AI," etc. This application does not impose any restrictions on this.

[0060] In step S503, the content analysis data is returned to the session client; the session client displays an interactive control that triggers content analysis, and the content analysis operation is triggered when the interactive control is accessed.

[0061] Interactive controls can be displayed in the relevant locations of conversation messages in the conversation client. For example, different topics in conversation messages may correspond to control buttons for creating to-dos, creating schedules or initiating meetings, initiating processes or following up on matters, and smart reply control buttons. This application does not impose any restrictions on this. Interactive controls can also be displayed in the corresponding interactive interface after the user clicks a specific wake-up control in the conversation client. For example, by clicking "AI" (wake-up control) in the conversation client, in response to the wake-up control being accessed, the conversation client can provide an interactive interface for conversing with the "AI assistant". Interactive controls (such as input boxes for conversing with the "AI assistant") can be displayed in the interactive interface. Interactive controls may include voice input controls or text input controls.

[0062] Within the same session client's display interface, one or more interactive controls can be displayed. Each interactive control can correspond to different topics or have different functions; this application does not impose any restrictions on this. The interactive controls can be triggered by clicking control buttons or by matching keywords in session messages, etc.; this application does not impose any restrictions on this.

[0063] In one possible implementation, the method of generating content analysis data for conversational messages by sending prompt words to a large language model can be to generate content analysis data for conversational messages and corresponding reply messages by sending prompt words requesting content analysis and responses to the large language model; correspondingly, returning content analysis data to the conversational client can be to return content analysis data and reply messages to the conversational client so that the reply messages are loaded in the content analysis data display interface.

[0064] The response message can be generated temporarily or generated along with the prompt word. This application does not impose any restrictions on the generation time of the response message. The prompt words for request content analysis and response can be generated based on the conversation messages, such as by calling a large language model or a pre-set conversation message template, prompt word database, etc. This application does not impose any restrictions on this.

[0065] For example, when a user asks for a to-do item with the "follow-up button color issue," the system can send a prompt word requesting content analysis and a response to a large language model. This generates content analysis data for the conversation message and a corresponding response message. For example, the response message could be "Okay, creating a to-do item requires the following information. Please confirm and click 'create'." The response message is then loaded into the content analysis data display interface.

[0066] For example, based on conversational messages such as "Please generate a summary of the conversation" entered by the user in the input box (interactive control) of the conversational client and the "AI assistant", prompt words requesting content analysis and response can be sent to a large language model to generate a response message such as "Hello, you can generate a summary or conclusion in the following way".

[0067] In one possible implementation, after generating content analysis data for conversational messages by sending prompt words to a large language model, it is also possible to identify and obtain content follow-up items corresponding to the content analysis data. These content follow-up items may include to-do items, schedule items, or meeting items. Accordingly, the aforementioned return of content analysis data to the conversational client may be the return of the content analysis data and the corresponding content follow-up items to the conversational client.

[0068] Content follow-up information can be obtained based on further analysis of content analysis data, either simultaneously with or after the generation of content analysis data, and this application does not impose any restrictions on this.

[0069] In some embodiments, in response to the interactive control being triggered, a follow-up control for the content follow-up function can be added to the page source file of the display interface of the content analysis data corresponding to the interactive control, and returned to the session client.

[0070] The page source file of the display interface can be composed of metadata corresponding to the content analysis data generated after the interactive controls are triggered. Furthermore, by adding a follow-up control for the content follow-up function to the page source file of the display interface corresponding to the content analysis data of the interactive controls and returning it to the session client, users can create to-do items, schedules, or meeting items by operating the follow-up control of the content follow-up function displayed on the session client.

[0071] Furthermore, in response to the triggering of the follow control, the page source file corresponding to the operation interface of the content follow function can be generated and returned to the session client.

[0072] The page source file corresponding to the operation interface can be composed of metadata corresponding to the operation interface of the content follow-up function generated after the follow-up control is triggered. One or more follow-up controls can be displayed in the display interface of the same session client. Each follow-up control can correspond to a different topic or have different functions; this application does not impose any restrictions on this.

[0073] The follow-up control can be triggered by clicking the control button or by matching keywords in the conversation message, and this application does not impose any restrictions on this. For example, users can trigger the follow-up control by clicking "Create To-Do" on the display interface of the conversation client, or by talking to the "AI Assistant".

[0074] This application also provides a method for processing session messages, such as... Figure 6 The diagram shown is a flowchart of a session message processing method 600 according to an embodiment of this application. The method 600 may include:

[0075] In step S601, in response to the access of the interactive control of the session client, at least one session message is obtained, the session message corresponds to one or more information formats, and the access of the interactive control triggers the execution of content analysis operation.

[0076] In step S602, the content analysis data of the session message generated by calling the large language model based on the session message is obtained.

[0077] In one possible implementation, before obtaining the content analysis data of the conversational message generated by calling the large language model based on the conversational message, at least one content analysis dimension can be displayed in response to the access of the interactive control of the conversational client to obtain the target content analysis dimension selected by the user; correspondingly, the way to obtain the content analysis data of the conversational message generated by calling the large language model based on the conversational message can be to obtain the content analysis data of the conversational message generated by calling the large language model according to the target content analysis dimension, wherein the content analysis data can also carry corresponding content keywords.

[0078] In step S603, the content analysis data is displayed.

[0079] In one possible implementation, the content follow-up items corresponding to the content analysis data can also be obtained. Then, after the interactive control is triggered, the follow-up control of the content follow-up function can be configured in the display interface of the content analysis data. After the follow-up control is triggered, the operation interface of the content follow-up function can be displayed.

[0080] In one possible implementation, the response message generated by calling a large language model can be obtained. Then, after the interactive control is triggered, the response message can be loaded in the content analysis data display interface, and after a confirmation operation for the response message is detected, the response message can be entered into the conversation window of the conversation message.

[0081] In one possible implementation, before retrieving at least one session message in response to the access of the interactive control of the session client, an interactive interface can be provided in response to the access of the wake-up control of the session client. The interactive interface displays interactive controls, which may include voice input controls or text input controls. Accordingly, the method of retrieving at least one session message in response to the access of the interactive control of the session client can be determined by identifying the voice data or text data submitted based on the interactive control to trigger the execution of a content analysis operation, and in response to triggering the execution of the content analysis operation, at least one session message can be retrieved.

[0082] The wake-up control can take any form, such as an icon, input box, or button. Once accessed, the wake-up control opens a dialogue interface between the user and the robot (e.g., an "AI assistant"), which can be the provided interactive interface. The interactive controls displayed on the interactive page can be in the form of dialogue input boxes or control buttons. These controls can include voice input controls, text input controls, and image input controls, all of which can be pre-configured. This application does not impose any restrictions on the display method or type of the wake-up control or interactive controls.

[0083] Corresponding to the examples and method embodiments provided in this application, this application also provides a session message processing apparatus. For example... Figure 7 The diagram shown is a structural block diagram of a session message processing apparatus 700 according to an embodiment of this application. The apparatus 700 may include:

[0084] A session message acquisition module 701 is used to acquire at least one session message from a session client, wherein the session message corresponds to one or more information formats. A prompt word generation module 702 is used to generate prompt words for requesting content analysis, and generates content analysis data for the session message by sending the prompt words to a large language model; the prompt words carry the session message. A content analysis data return module 703 is used to return the content analysis data to the session client; the session client displays an interactive control that triggers content analysis, and the content analysis operation is triggered when the interactive control is accessed.

[0085] In one possible implementation, the session message also carries the target content analysis dimension used for content analysis; the above-mentioned prompt word generation module 702 may include: a prompt word generation submodule, used to generate prompt words carrying the session message and the target content analysis dimension.

[0086] In one possible implementation, after generating the content analysis data of the conversation message by sending the prompt words to a large language model, the above-mentioned apparatus may further include: a content recognition module, used to perform content recognition on the content analysis data and add content keyword identifiers to the various types of content included in the content analysis data.

[0087] In one possible implementation, before generating the prompt word for requesting content analysis, the above-mentioned device may further include: an external data identification module for identifying external data linked to by the session message, the external data including at least one of database data, web page data, and server data; a data fragment segmentation module for segmenting the external data into multiple data fragments; the above-mentioned prompt word generation module 702 may include: a data fragment prompt word generation submodule for generating prompt words carrying the session message and at least one data fragment.

[0088] In one possible implementation, after generating the content analysis data of the conversation message by inputting the prompt words into the large language model, the above-mentioned device may further include: a content follow-up item identification module, used to identify the content follow-up item corresponding to the obtained content analysis data, the content follow-up item including to-do items, schedule items or meeting items; the above-mentioned content analysis data return module 703 may include: a content follow-up item return submodule, used to return the content analysis data and the content follow-up item corresponding to the content analysis data to the conversation client.

[0089] In some embodiments, the above-described apparatus may further include: a follow-up control display module, configured to add a follow-up control for content follow-up function to the page source file of the display interface of the content analysis data corresponding to the interactive control in response to the triggering of the interactive control, and return it to the session client.

[0090] For example, the above-described apparatus may further include: an operation interface generation module, configured to generate a page source file corresponding to the operation interface of the content follow-up function and return it to the session client in response to the triggering of the follow-up control.

[0091] In one possible implementation, the prompt word generation module 702 may include: a reply information generation submodule, used to generate content analysis data of the conversation message and a reply message corresponding to the conversation message by sending prompt words requesting content analysis and reply to a large language model; the content analysis data return module 703 may include: a reply message return submodule, used to return the content analysis data and reply message to the conversation client so as to load the reply message in the display interface of the content analysis data.

[0092] Corresponding to the examples and method embodiments provided in this application, this application also provides a session message processing apparatus. For example... Figure 8 The diagram shown is a structural block diagram of a session message processing apparatus 800 according to an embodiment of this application. The apparatus 800 may include:

[0093] The session message acquisition module 801 is used to acquire at least one session message in response to the access of the interactive control of the session client. The session message corresponds to one or more information formats. After the interactive control is accessed, a content analysis operation is triggered. The content analysis data acquisition module 802 is used to acquire content analysis data of the session message generated by calling a large language model based on the session message. The content analysis data display module 803 is used to display the content analysis data.

[0094] In one possible implementation, the above-mentioned device may further include: a content follow-up item acquisition module, used to acquire content follow-up items corresponding to the content analysis data; and an operation page display module, used to configure a follow-up control for the content follow-up function in the display interface of the content analysis data after the interactive control is triggered, and to display the operation interface of the content follow-up function after the follow-up control is triggered.

[0095] In one possible implementation, the above apparatus may further include: a reply message acquisition module, used to acquire reply messages generated by calling a large language model; and a reply message loading module, used to load the reply message in the display interface of the content analysis data after the interactive control is triggered, and input the reply message into the session window of the session message after detecting a confirmation operation for the reply message.

[0096] In one possible implementation, before acquiring the content analysis data of the session message generated by calling a large language model based on the session message, the above apparatus may further include: a target content analysis dimension acquisition module, used to display at least one content analysis dimension in response to the access of an interactive control of the session client, so as to acquire the target content analysis dimension selected by the user; the content analysis data acquisition module 802 may include: a content analysis data acquisition submodule, used to acquire the content analysis data of the session message generated by calling a large language model according to the target content analysis dimension, wherein the content analysis data also carries corresponding content keywords.

[0097] In one possible implementation, before the interactive control of the session client is accessed and at least one session message is obtained, the apparatus may further include: an interactive interface providing module, configured to provide an interactive interface in response to the access of the wake-up control of the session client, wherein the interactive interface displays the interactive control, the interactive control including a voice input control or a text input control; the session message obtaining module 801 may include: a trigger execution content analysis operation determining submodule, configured to determine the trigger execution content analysis operation by recognizing the voice data or text data submitted based on the interactive control; and a session message obtaining submodule, configured to obtain at least one session message in response to the trigger execution content analysis operation.

[0098] The functions of each module in each device in the embodiments of this application can be found in the corresponding description in the above method, and they have corresponding beneficial effects, which will not be repeated here.

[0099] Figure 9 This is a block diagram of an electronic device used to implement embodiments of this application. For example... Figure 9 As shown, the electronic device includes a memory 901 and a processor 902. The memory 901 stores a computer program that can run on the processor 902. When the processor 902 executes the computer program, it implements the method described in the above embodiments. The number of memories 901 and processors 902 can be one or more.

[0100] The electronic device also includes:

[0101] The communication interface 903 is used to communicate with external devices and exchange and transmit data.

[0102] If the memory 901, processor 902, and communication interface 903 are implemented independently, they can be interconnected via a bus to communicate with each other. This bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0103] Optionally, in a specific implementation, if the memory 901, processor 902, and communication interface 903 are integrated on a single chip, then the memory 901, processor 902, and communication interface 903 can communicate with each other through an internal interface.

[0104] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method provided in this application.

[0105] This application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the methods provided in any embodiment of this application.

[0106] This application also provides a chip including a processor for calling and executing instructions stored in a memory, causing a communication device with the chip installed to perform the method provided in this application.

[0107] This application also provides a chip, including: an input interface, an output interface, a processor, and a memory. The input interface, output interface, processor, and memory are connected through an internal connection path. The processor is used to execute code in the memory. When the code is executed, the processor is used to execute the method provided in this application.

[0108] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor can be a processor supporting Advanced Reduced Instruction Set Machines (ARM) architecture.

[0109] Further, optionally, the aforementioned memory may include read-only memory and random access memory. The memory may be volatile memory or non-volatile memory, or may include both. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available. Examples include Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0110] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0111] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0112] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0113] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process. Furthermore, the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functionality involved.

[0114] The logic and / or steps described in the flowchart or otherwise herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0115] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware, the program being stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiments.

[0116] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. This storage medium can be a read-only memory, a disk, or an optical disk, etc.

[0117] The above description is merely an exemplary embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope described in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for processing session messages, comprising: Obtain at least one session message from the session client, wherein the session message corresponds to one or more information formats; Generate prompt words for requesting content analysis, and generate content analysis data for the session message by sending the prompt words to a large language model, wherein the prompt words carry the session message; The content analysis data is returned to the session client; the session client displays an interactive control that triggers content analysis, and the content analysis operation is executed after the interactive control is accessed.

2. The method according to claim 1, wherein, The session message also carries the target content analysis dimensions used for content analysis; The generated prompt words for requesting content analysis include: Generate prompt words carrying the session message and the target content analysis dimension.

3. The method according to claim 1, wherein, After generating the content analysis data of the conversational message by sending the cue words to a large language model, the method further includes: Content identification is performed on the content analysis data, and content keywords are added to identify the various types of content included in the content analysis data.

4. The method according to claim 1, wherein, Before generating the prompt words for request content analysis, the method further includes: Identify the external data linked to by the session message, wherein the external data includes at least one of database data, web page data, and server data; The external data is divided into multiple data segments; The generated prompt words for requesting content analysis include: Generate a prompt word carrying the session message and at least one data fragment.

5. The method according to claim 1, wherein, After generating the content analysis data of the conversational message by sending the cue words to a large language model, the method further includes: Identify and acquire the content follow-up items corresponding to the content analysis data, including to-do items, schedule items, or meeting items; The step of returning the content analysis data to the session client includes: The content analysis data and the corresponding content follow-up items are returned to the session client.

6. The method according to claim 5, wherein, The method further includes: In response to the triggering of the interactive control, a follow-up control for the content follow-up function is added to the page source file of the display interface of the content analysis data corresponding to the interactive control, and returned to the session client.

7. The method according to claim 6, wherein, The method further includes: In response to the triggering of the follow-up control, the page source file corresponding to the operation interface of the content follow-up function is generated and returned to the session client.

8. The method according to claim 1, wherein, The content analysis data generated by sending the prompt words to a large language model to produce the conversational message includes: By sending prompt words requesting content analysis and responses to a large language model, content analysis data of the conversation message and corresponding response messages are generated. The step of returning the content analysis data to the session client includes: The content analysis data and reply message are returned to the session client so that the reply message is loaded in the display interface of the content analysis data.

9. A method for processing session messages, comprising: In response to the access of the interactive control of the session client, at least one session message is obtained, the session message corresponds to one or more information formats, and the access of the interactive control triggers the execution of content analysis operations; Obtain content analysis data of the conversation messages generated by a large language model based on the conversation messages; Display the content analysis data.

10. The method according to claim 9, wherein, The method further includes: Obtain the content follow-up items corresponding to the content analysis data; After the interactive control is triggered, a follow-up control for the content follow-up function is configured in the display interface of the content analysis data. After the follow-up control is triggered, the operation interface of the content follow-up function is displayed.

11. The method according to claim 9, wherein, The method further includes: Retrieve the response message generated by calling a large language model; After the interactive control is triggered, the reply message is loaded into the display interface of the content analysis data, and after a confirmation operation for the reply message is detected, the reply message is entered into the session window of the session message.

12. The method according to claim 9, wherein, Before obtaining the content analysis data of the session message generated based on the session message by the large language model, the method further includes: In response to the access of interactive controls on the session client, at least one content analysis dimension is displayed to obtain the target content analysis dimension selected by the user; The process of obtaining content analysis data of the session messages generated by the large language model based on the session messages includes: The system acquires content analysis data of conversational messages generated by calling a large language model based on the target content analysis dimensions. The content analysis data also carries corresponding content keywords.

13. The method according to claim 9, wherein, Before the interactive control responding to the session client is accessed and at least one session message is obtained, the method further includes: In response to the access of the wake-up control of the session client, an interactive interface is provided, in which the interactive control is displayed, including a voice input control or a text input control; The process of responding to the access of an interactive control in the session client and obtaining at least one session message includes: The content analysis operation is triggered by identifying the voice or text data submitted based on the interactive control. In response to triggering a content analysis operation, retrieve at least one session message.

14. A session message processing apparatus, comprising: The session message acquisition module is used to acquire at least one session message from the session client, wherein the session message corresponds to one or more information formats; The prompt word generation module is used to generate prompt words for requesting content analysis, and to generate content analysis data of the conversation message by sending the prompt words to a large language model, wherein the prompt words carry the conversation message; The content analysis data return module is used to return the content analysis data to the session client; the session client displays an interactive control that triggers content analysis, and the content analysis operation is triggered when the interactive control is accessed.

15. A session message processing apparatus, comprising: The session message acquisition module is used to acquire at least one session message in response to the access of the interactive control of the session client. The session message corresponds to one or more information formats. After the interactive control is accessed, a content analysis operation is triggered. The content analysis data acquisition module is used to acquire content analysis data of the conversation messages generated by a large language model based on the conversation messages. The content analysis data display module is used to display the content analysis data.

16. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein the processor, when executing the computer program, implements the method of any one of claims 1-13.

17. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of any one of claims 1-13.

Images