Information recording and retrieval methods, devices, computer equipment, readable storage media, and program products

Abstract

This application relates to an information record search method, apparatus, computer device, computer-readable storage medium, and computer program product. The method includes: upon receiving a search request, extracting search keywords and object information from the search request; constructing expanded suggestion words based on the object information and search keywords to obtain expanded suggestion words adapted to a large language model; and inputting the expanded suggestion words into the large language model for keyword expansion processing to obtain multiple sets of expanded keywords. This application generates a search request by responding to a search operation where the target object inputs search keywords, and then submits the search request to the server, thereby calling the large language model to expand keywords and obtaining broader expanded search results through the expanded keywords, thus improving search accuracy.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to an information record search method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] With the development of computer technology, search engine technology emerged. This technology retrieves specific information from the internet based on user needs and certain algorithms, using specific strategies to provide the information to the user. Building upon search engines, intelligent search technology has also been developed. Intelligent search technology utilizes artificial intelligence, especially machine learning and natural language processing, to improve the accuracy, relevance, and user experience of search engines.

[0003] For example, while intelligent search technology can achieve efficient searching of information records, it currently typically employs word segmentation-based intelligent search, which segments the user's search keywords and uses these segments to match information records. However, this word segmentation approach is relatively fixed and rigid, unable to be adjusted according to the actual situation of the enterprise, making it difficult to guarantee the accuracy of intelligent search. Summary of the Invention

[0004] Therefore, it is necessary to provide an information record search method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can improve search accuracy in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a method for searching information records, including:

[0006] Upon receiving a search request, extract the search keywords and object information from the search request;

[0007] Based on the object information and the search keywords, expanded suggestion words are constructed to obtain expanded suggestion words adapted to the large language model;

[0008] The extended prompt words are input into a large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. The extended keywords are used to perform information record search processing and obtain corresponding extended search results.

[0009] Secondly, this application also provides an information recording and searching device, configured on a server side, comprising:

[0010] The information extraction module is used to extract search keywords and object information from a search request upon receiving such a request.

[0011] The prompt word construction module is used to construct extended prompt words based on the object information and the search keywords to obtain extended prompt words adapted to the large language model;

[0012] The keyword expansion module is used to input the expansion prompts into a large language model for keyword expansion processing, thereby obtaining multiple sets of expanded keywords corresponding to the expansion prompts. The expanded keywords are used to perform information record search processing and obtain corresponding expanded search results.

[0013] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0014] Upon receiving a search request, extract the search keywords and object information from the search request;

[0015] Based on the object information and the search keywords, expanded suggestion words are constructed to obtain expanded suggestion words adapted to the large language model;

[0016] The extended prompt words are input into a large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. The extended keywords are used to perform information record search processing and obtain corresponding extended search results.

[0017] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0018] Upon receiving a search request, extract the search keywords and object information from the search request;

[0019] Based on the object information and the search keywords, expanded suggestion words are constructed to obtain expanded suggestion words adapted to the large language model;

[0020] The extended prompt words are input into a large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. The extended keywords are used to perform information record search processing and obtain corresponding extended search results.

[0021] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0022] Upon receiving a search request, extract the search keywords and object information from the search request;

[0023] Based on the object information and the search keywords, expanded suggestion words are constructed to obtain expanded suggestion words adapted to the large language model;

[0024] The extended prompt words are input into a large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. The extended keywords are used to perform information record search processing and obtain corresponding extended search results.

[0025] The aforementioned information record search method, apparatus, computer equipment, computer-readable storage medium, and computer program product, upon receiving a search request, extract search keywords and object information from the search request to obtain the basic data for the search expansion process. Based on the object information and search keywords, expansion suggestion words are constructed to obtain expansion suggestion words adapted to a large language model, thereby expressing the search intent. The expansion suggestion words are input into the large language model for keyword expansion processing, resulting in multiple sets of expansion keywords corresponding to the expansion suggestion words, thus expanding new search content; subsequently, information record search processing is performed using these expansion keywords to obtain expanded search results, completing the search process expansion. This application, by expanding keywords based on the keywords in the search request after receiving the search request through a large language model to obtain multiple sets of expansion keywords, can obtain broader expanded search results and improve search accuracy.

[0026] Sixthly, this application provides an information record retrieval method, applied to a client, including:

[0027] Responding to the search operation of the target object by inputting search keywords, a search request is generated based on the search keywords and the object information of the target object;

[0028] Send a search request to the server;

[0029] Receive expanded keywords from the server based on the search request. The expanded keywords are obtained by expanding the search keywords using a large language model.

[0030] Based on the extended keywords, information records are searched locally on the client side to obtain extended search results, which are then displayed.

[0031] Seventhly, this application also provides an information recording and searching device, installed on a client side, comprising:

[0032] The request generation module is used to respond to the search operation of the target object by inputting search keywords, and to generate a search request based on the search keywords and the object information of the target object;

[0033] The request sending module is used to send search requests to the server.

[0034] The keyword receiving module is used to receive expanded keywords from the server based on the search request. The expanded keywords are obtained by expanding the search keywords through a large language model.

[0035] The record search module is used to perform information record search processing on the client's local machine based on extended keywords, obtain extended search results, and display the extended search results.

[0036] Eighthly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0037] Responding to the search operation of the target object by inputting search keywords, a search request is generated based on the search keywords and the object information of the target object;

[0038] Send a search request to the server;

[0039] Receive expanded keywords from the server based on the search request. The expanded keywords are obtained by expanding the search keywords using a large language model.

[0040] Based on the extended keywords, information records are searched locally on the client side to obtain extended search results, which are then displayed.

[0041] Ninthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0042] Responding to the search operation of the target object by inputting search keywords, a search request is generated based on the search keywords and the object information of the target object;

[0043] Send a search request to the server;

[0044] Receive expanded keywords from the server based on the search request. The expanded keywords are obtained by expanding the search keywords using a large language model.

[0045] Based on the extended keywords, information records are searched locally on the client side to obtain extended search results, which are then displayed.

[0046] In a tenth aspect, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0047] Responding to the search operation of the target object by inputting search keywords, a search request is generated based on the search keywords and the object information of the target object;

[0048] Send a search request to the server;

[0049] Receive expanded keywords from the server based on the search request. The expanded keywords are obtained by expanding the search keywords using a large language model.

[0050] Based on the extended keywords, information records are searched locally on the client side to obtain extended search results, which are then displayed.

[0051] The aforementioned information recording search method, apparatus, computer equipment, computer-readable storage medium, and computer program product respond to a search operation where the target object inputs search keywords, generate a search request based on the search keywords and the target object's information, and then send the search request to the server; receive expanded keywords from the server based on the search request, the expanded keywords being obtained by expanding the search keywords through a large language model, thus realizing model invocation processing in the search process. Based on the expanded keywords, information recording search processing is performed locally on the client to obtain expanded search results, which are then displayed. This application generates a search request by responding to a search operation where the target object inputs search keywords, and then submits the search request to the server, thereby invoking a large language model to expand the keywords, obtaining multiple sets of expanded keywords, and thus obtaining expanded search results with broader coverage through the expanded keywords, improving search accuracy. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is an application environment diagram of the information record search method in one embodiment;

[0054] Figure 2 This is a flowchart illustrating an information record search method in one embodiment;

[0055] Figure 3 This is a schematic diagram illustrating the prediction of expanded keywords in one embodiment;

[0056] Figure 4 This is a schematic diagram of the instant messaging software recording the search page in one embodiment;

[0057] Figure 5 This is a schematic diagram of a page in one embodiment where a full-word matching search is performed.

[0058] Figure 6This is a schematic diagram of the page where search results are obtained through a precise search in one embodiment;

[0059] Figure 7 This is a schematic diagram of a page for launching a fuzzy search page in one embodiment;

[0060] Figure 8 A schematic diagram of a permission settings page in one embodiment;

[0061] Figure 9 This is a flowchart illustrating the information record search method in another embodiment;

[0062] Figure 10 This is a schematic diagram of a page showing the search results obtained by combining precise search with expanded fuzzy search in one embodiment;

[0063] Figure 11 This is a schematic diagram of a page where expanded keywords in search results are highlighted in one embodiment;

[0064] Figure 12 This is a timing diagram of an information record search method in one embodiment;

[0065] Figure 13 This is a structural block diagram of an information recording and searching device in one embodiment;

[0066] Figure 14 This is a structural block diagram of the information recording and searching device in another embodiment;

[0067] Figure 15 This is an internal structural diagram of a computer device in one embodiment;

[0068] Figure 16 This is a diagram of the internal structure of a computer device in another embodiment. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0070] 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, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0071] The information record search method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, client 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located in the cloud or on another network server. When a target object on client 102 wants to search for specific content from information records, it may need to search as many records as possible to improve the search success rate. At this time, the target object can enter search keywords on client 102 to trigger a search operation. Client 102 responds to the search operation initiated by the target object by generating a search request based on the search keywords and the target object's information; it then sends the search request to server 104. Upon receiving a search request from client 102, server 104 extracts the search keywords and object information from the search request; based on the object information and search keywords, it constructs expanded suggestion words to obtain expanded suggestion words adapted to the large language model; it inputs the expanded suggestion words into the large language model for keyword expansion processing to obtain multiple sets of expanded keywords corresponding to the expanded suggestion words; it feeds back the expanded keywords to client 102, which receives the expanded keywords fed back by the server based on the search request; and it performs information record search processing locally on the client based on the expanded keywords to obtain expanded search results. The target object can browse the expanded search results to obtain the required information record information. Client 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can include virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides cloud computing services.

[0072] In one exemplary embodiment, such as Figure 2 As shown, an information record search method is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 201 to 205. Wherein:

[0073] Step 201: Upon receiving a search request, extract the search keywords and object information from the search request.

[0074] Among them, for Figure 1In the application environment described, when a target object wishes to search for specific content data in an information record, it can submit a corresponding search request through client 102 to achieve search processing. The search request is sent by client 102 to server 104, requesting the server to perform search-related processing. The search request includes search keywords and object information. The search keywords are input by the target object operating client 102 and can be a single word or a sentence composed of words. The object information is used to identify the specific identity of the target object, such as a user identifier.

[0075] For example, after logging into a social application, a target user can chat with other users on the application, and these chat logs are saved. When the target user wants to find specific chat logs, they can browse the logs chronologically to obtain the relevant information. However, if there are too many chat logs and the target user cannot remember which day the chat log was from, the target user can search for specific content. They can input search keywords through client 102 to search for content related to those keywords in the chat logs. However, the search keywords entered by the target user from memory may not be accurate, and searching chat logs based on those keywords may not yield satisfactory results. In this case, the chat log search method of this application can be used to expand the search keywords. Therefore, client 102 can submit the search keywords and user information together to server 104 for search keyword expansion processing. In one embodiment, this application is specifically applied to the information record search processing related to enterprise instant messaging (IM) applications. When a user wants to search for relevant content information in the application's groups, they can click the "Open Message Records" button in the application bar to enter the information record search page, then enter search keywords in the search field, and click the confirm button to start the search. At this time, the local client of the enterprise instant messaging application will search for information in the local information records based on the search keywords. Simultaneously, it will generate a search request based on the user's identification information and search keywords in the application, and submit the search request to the application's server for extended processing to obtain more search keywords and ensure search effectiveness.

[0076] Step 203: Based on object information and search keywords, construct extended suggestion words to obtain extended suggestion words adapted to the large language model.

[0077] A prompt is an injected instruction that directs an AI system to think about a problem and output content according to a pre-defined framework. Specifically, it's an input text paragraph or phrase that serves as the starting point or guide for the model's output. It can be a question, a text description, a dialogue, or any form of text input. The model generates the corresponding output text based on the context and semantic information provided by the prompt. In the field of large language models, the prompt is the user's input, and the model uses this input to predict the probability of the next word appearing, thus generating the following text word by word. Prompt construction processing refers to the process of constructing input prompts based on object information and search keywords.

[0078] For example, this application can specifically use a large language model to expand search keywords, constructing multiple sets of expanded keywords that are semantically the same as or similar to the search keywords, thereby improving search coverage and ensuring success rate. To call the large language model, corresponding prompt words need to be constructed. This can be done by combining object information and search keywords with a prompt word template to construct expanded prompt words. Expanded prompt words are constructed by filling the object information and search keywords into designated positions in the template. In one embodiment, the constructed expanded keywords are: "Hello, my identity is XXX. Below, I will provide search keywords for information record search. You need to help me replace the search keywords with other possible words. You can combine my identity and provide replacement keywords from the following perspectives: 1. Simplify my keywords, replacing them with shorter words; 2. Help me segment words, using only one or more words after segmentation; 3. Help me correct errors, for example, changing 'comb' to 'come'; 4. Replace my words with synonyms. Next, my input is the keywords, and you need to output at least 20 different keyword combinations, connected by '&', and formatted with numbered lines."

[0079] Step 205: Input the extended prompt words into the large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. The extended keywords are used to perform information record search processing and obtain corresponding extended search results.

[0080] Large Language Models (LLMs) are deep learning models trained on large amounts of text data, designed to understand and generate human language. Through deep learning techniques, particularly the Transformer architecture and self-attention mechanism, LLMs can be trained on large amounts of unlabeled text data to learn language patterns and structures, thereby simulating the human language cognition and generation process. The LLM model in this application is a model fine-tuned based on a keyword expansion task for information record searching, and can be effectively applied to keyword expansion processing in the information record search process.

[0081] For example, this application specifically expands search keywords using a large language model. First, a pre-trained large language model is obtained by pre-training the model on a large corpus. Then, training data for keywords and expanded keywords is constructed using search data related to information records. The initial keywords are those entered by the user, while the expanded keywords are those changed by the user. The pre-trained model is then fine-tuned using the training data to obtain a large language model that can be used to expand keywords. When performing information record search tasks, this model can process the input expanded suggestion words to generate multiple sets of expanded keywords. It is worth noting that these expanded keywords are generated continuously as the model runs, rather than simultaneously.

[0082] After obtaining the expanded keywords, the server can send them back to the client. The client then uses these expanded keywords to perform information record search processing and obtain expanded search results. For example, after obtaining the expanded keywords, the server 104 can sequentially send them back to the client 102 via the network. The client 102 can then use these expanded keywords as the basis for its search to perform information record search processing and obtain the corresponding expanded search results for each keyword. The expanded search results are then displayed on the client 102's display module. The target audience can view the displayed content to confirm the search results and can click on the search results to enter the specific information record context page for more detailed information.

[0083] In another embodiment, this application can also be directly applied to the client. In this case, the large language model is directly deployed on the client. When the user operates the client to enter the search page of the information record, the client can be regarded as the user submitting a search request. At this time, the search keywords and object information in the search request are directly extracted, that is, the keywords entered by the user and the object information used by the user when logging into the client application. Then, on the client, based on the object information and search keywords, the extended prompt words are constructed to obtain extended prompt words adapted to the large language model. The extended prompt words are then directly input into the large language model deployed on the client, and the large language model performs keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. Then, for each extended keyword, a line information record search processing can be performed once to obtain the corresponding extended search results for each extended keyword.

[0084] The aforementioned information record search method, upon receiving a search request, extracts the search keywords and object information from the request to obtain the basic data for the search expansion process. Based on the object information and search keywords, it constructs expansion prompts to obtain expansion prompts adapted to a large language model, thus expressing the search intent. The expansion prompts are then input into the large language model for keyword expansion processing, resulting in multiple sets of expansion keywords corresponding to the expansion prompts, thereby expanding new search content. Finally, information record search processing is performed using these expansion keywords to obtain expanded search results, completing the search expansion process. This application, by expanding keywords based on the search request's keywords using a large language model after receiving the search request, obtains multiple sets of expansion keywords, thereby achieving broader coverage of expanded search results and improving search accuracy.

[0085] In an exemplary embodiment, step 203 includes: determining personalized background information of the search request based on object information; filling the personalized background information and search keywords into the prompt word template for extended prompt word construction processing to obtain extended prompt words adapted to the large language model.

[0086] Personalized background information refers to the unique background information possessed by the target audience, such as job type, hobbies, company affiliation, and company type. Cue word templates, on the other hand, are template information used to guide the generation of expanded cue words; standardized expanded cue words can be generated using these templates.

[0087] For example, this application specifically combines the personalized background information of the target object to generate extended keywords. This personalized background information can be obtained with the authorization of the target object, thereby constructing extended prompts that are more suitable for the target object, thus ensuring the efficiency and accuracy of extended keyword recognition. During the prompt generation process, the personalized background information and search keywords can be filled into the corresponding positions of the prompt template to construct extended prompts, thereby ensuring the efficiency and accuracy of extended prompt generation. In one embodiment, this application is specifically applied to the search processing of information records related to enterprise instant messaging applications. Assuming a user works in a computer industry company as a programmer, and the search keyword is "group chat bug," the target object's job type and company type can be considered when generating extended prompts. Therefore, after obtaining authorization, the personalized background information corresponding to the search request can be used as the basic information of the target object. This information, along with the search keywords, is then filled into the suggestion template to obtain the expanded suggestion: "Hello, I work in a computer industry company, and my position is programmer. Below, I will provide several keywords connected by '&' for information record search. You need to help me replace some of these keywords with other possible words. You can provide replacement keywords based on my identity from the following perspectives: 1. Simplify my keywords, replacing them with shorter words; 2. Help me segment words, using only one or more words after segmentation; 3. Help me correct errors, for example, changing 'comb' to 'come'; 4. Replace my words with synonyms. Next, my input is the keywords. You need to output at least 20 different keyword combinations, always connected by '&', and formatted with numerical sequence lines." Then, this expanded suggestion is input into the large language model, which can then perform prediction processing based on the information provided by the suggestion to obtain expanded keywords. The expanded keywords predicted by the large language model can be referenced... Figure 3 As shown in the figure. In this embodiment, personalized background information and search keywords are combined to construct expanded suggestion words, thereby ensuring that the constructed expanded suggestion words are relevant to the target object and improving search accuracy.

[0088] In one exemplary embodiment, determining personalized background information for a search request based on object information includes: finding authorization information corresponding to the object information; determining authorization item information based on the authorization information; and using the authorization item information as personalized background information for the search request.

[0089] For example, each object can be pre-assigned corresponding authorization information, which is then stored on the server. When a target object wants to search for extended prompts, the target object's information can be used as the query basis to search for authorization information on the server. The authorization information then determines which data of the target object has been authorized for personalized search configuration; this information is the authorization item information. For example, the target object can authorize one or more of its job type, hobbies, or company to be used by the server. The server then directly searches for the authorization items based on the authorization information and uses this information as personalized background information. In this embodiment, authorization item information is determined by querying the object's authorization information, and then this information is used as personalized background information to construct extended prompts and expand keywords, thereby ensuring the security of the target object's information and the compliance of the query process.

[0090] In an exemplary embodiment, the method further includes: upon receiving a permission setting request, providing a permission setting interface to the client; receiving authorization information based on the permission setting interface; and storing the object information and authorization information in the permission setting request together.

[0091] For example, this application also includes a process for setting permission information. When applied to an interactive system between a server and a client, the target object can trigger a permission setting request on the client, and the server can, upon receiving the permission setting request submitted by the client, provide a permission setting interface to the client. The target object on the client can set authorization items on the permission setting interface and then submit the completed interface to the server. The server receives the authorization information based on the permission setting interface, sets the association between the authorization information and the object information, and stores the two together. Subsequently, the authorization information can be directly retrieved through the object information. When applied to a client-side system, the user can directly operate the client to submit a permission setting request and directly fill in the authorization information on the client. The client application will then store the object information and authorization information in the permission setting request together. In one embodiment, this application is applied to the information record query of instant messaging software, where the user's chat page can refer to... Figure 4 At the top, there's a search box; clicking it allows you to enter keywords. The search box contains an "AI" icon; a lighter color indicates it's not enabled. When AI is disabled, the user's search will use a normal full-word matching search, such as... Figure 5 As shown, searching for "group chat bug" will return conversations containing these two words, aggregated by conversation. The first conversation in the image is a one-on-one chat between the user and user A. Clicking on user A's conversation in the search results will then... Figure 6The screen lists all matching records for this session, with matching keywords highlighted in blue. Users can adjust permissions by clicking the AI ​​button. When the target user clicks the light-colored AI button, the process of enabling AI fuzzy search will begin, as shown below. Figure 7 As shown, the icon will expand into a floating window, asking the user whether to enable it. If the user clicks to enable it, it is considered that the user has submitted a permission setting request through the client. The permission setting interface displayed at this time can be as follows: Figure 8 As shown, the AI ​​icon changes from light to dark, reminding the user that AI fuzzy search is enabled. Simultaneously, a floating window displays configuration options for user customization: such as authorizing company information, organizational structure (after authorization, the constructed prompts will include industry information, company name, user's department, etc.), authorizing personal identity information (after authorization, the prompts will include user name, gender, position, etc.), and disabling intelligent correction (when disabled, the prompts will instruct the model not to generate corrective keywords). In this embodiment, permission information is received through a permission settings interface via a client-side permission setting request, ensuring the efficiency and accuracy of the authorization process.

[0092] In an exemplary embodiment, the object information includes an object identifier. Step 201 includes: upon receiving a search request submitted by the target object, parsing the search request to obtain search input information and the object identifier; and performing word segmentation on the search input information to obtain search keywords.

[0093] For example, the target object can be identified by an object identifier. For instance, when applying this application to information record searching in instant messaging software, the target object's object information can be the UID information of the target object logging into the instant messaging software. As for search keywords, they can be constructed by segmenting the directly input search information. The target object directly inputs the complete query statement into the search bar. After these query statements are transmitted to the server, the server can use a word segmentation algorithm to segment the search input information for processing, obtaining multiple words as search keywords. In this embodiment, constructing search keywords by segmenting the search input information and then performing the query can effectively improve the efficiency and accuracy of the search.

[0094] In an exemplary embodiment, the method is applied to the server side, and step 207 includes: upon receiving a search request submitted by the client, extracting search keywords and object information from the search request; constructing extended suggestion words based on the object information and search keywords to obtain extended suggestion words adapted to the large language model; inputting the extended suggestion words into the large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended suggestion words; using the extended keywords to perform information record search processing and obtain corresponding extended search results; and feeding back the newly generated extended keywords to the client in the order of generation during the sequential generation of extended keywords.

[0095] For example, since large language model inference is time-consuming and the model returns results word by word, the server-side backend can send newly generated keywords to the client for local search as soon as it receives a complete keyword. Therefore, in the specific feedback process, newly generated expanded keywords can be fed back to the client in the order of generation during the sequential generation of expanded keywords. Thus, with the model outputting words one by one, the client can immediately use the first keyword to search after it is generated. The first keyword has a higher priority than the subsequent ones, so the search result will always be at the top, ensuring that the user UI will not change the sorting even after subsequent keywords are generated. This priority definition ensures that the search results are stable and continuously updated. In this embodiment, by sequentially feeding back expanded keywords, the ranking of the query process is kept stable, and the efficiency of query processing is improved.

[0096] In one exemplary embodiment, such as Figure 9 As shown, an information record search method is provided, which can be applied to... Figure 1 Taking client 102 as an example, the explanation includes the following steps 902 to 908. Wherein:

[0097] Step 902: Respond to the search operation of the target object by inputting search keywords, and generate a search request based on the search keywords and the object information of the target object.

[0098] Step 904: Send a search request to the server.

[0099] For example, a search operation refers to the search processing triggered when the target object enters search keywords on client 102. For example... Figure 4As shown, the target can enter "group chat bug" as a search keyword in the search bar to trigger a search operation. Client 102 generates a search request based on the search keyword entered by the target and the target's existing information. This search request is then sent to server 104, which performs corresponding search term expansion processing. In one embodiment, this application is specifically applied to information record search processing related to enterprise instant messaging (IM) applications. When a user wants to search for relevant content information in the application's groups, they can click the "Open Message Records" button in the application bar to enter the information record search page, then enter the search keyword in the search bar and click the confirmation button to start the search. At this time, the client generates a search request based on the user's identification information and search keywords in the application and submits the search request to the application's server for expansion processing to obtain more search keywords and ensure search effectiveness.

[0100] Step 906: Receive expanded keywords from the server based on the search request. The expanded keywords are obtained by expanding the search keywords using a large language model.

[0101] For example, this application specifically expands search keywords using a large language model. First, the model is pre-trained using a large corpus to obtain a pre-trained large language model. When performing information record search tasks, this model can process the input expanded prompts to generate multiple sets of expanded keywords. It is worth noting that these expanded keywords are continuously generated as the model runs, rather than simultaneously. The expanded keywords generated by the server (104) are continuously submitted to the client after generation, thereby helping the client perform corresponding fuzzy search processing, ensuring search coverage, and improving search accuracy.

[0102] Step 908: Based on the extended keywords, perform information record search processing on the client's local machine to obtain extended search results and display them.

[0103] For example, after obtaining the expanded keywords, the server 104 can sequentially transmit the obtained expanded keywords to the client 102 via the network. The client 102 can then use the expanded keywords as the basic data for information recording and search processing to obtain the expanded search results corresponding to each expanded keyword. The expanded search results are then displayed on the display module of the client 102. The target audience can confirm the search results by browsing the displayed content of the display module, and can also enter the specific information recording context page by clicking on the search results to obtain more detailed content information.

[0104] The aforementioned information recording and search method responds to the search operation of the target object inputting search keywords, generates a search request based on the search keywords and the target object's information, and then sends the search request to the server. It receives expanded keywords from the server based on the search request; these expanded keywords are obtained by expanding the search keywords using a large language model, thus realizing model invocation processing in the search process. Based on the expanded keywords, information recording and search processing is performed locally on the client side to obtain expanded search results, which are then displayed. This application generates a search request by responding to the search operation of the target object inputting search keywords, and then submits the search request to the server, thereby invoking a large language model to expand the keywords, obtaining multiple sets of expanded keywords. This allows for searching using expanded keywords to obtain broader expanded search results, improving search accuracy.

[0105] In an exemplary embodiment, step 904 is followed by: performing information record search processing on the client's local machine based on search keywords to obtain accurate search results and displaying the accurate search results.

[0106] Step 908 includes: asynchronously performing information record search processing on the client's local machine based on extended keywords to obtain extended search results, and displaying the extended search results in order after the precise search results.

[0107] For example, the search process in this application includes two processes: precise search and fuzzy search. Precise search is the process of searching based on the search keywords input by the target object. Precise matches are always given the highest priority, so the client can display the precise match results before fuzzy keywords are generated, without affecting historical logic. Therefore, after sending a search request to the server, the client can also directly perform information record search processing on its local machine based on the search keywords to obtain precise search results, which will be directly displayed on the client. Subsequent expanded search results are displayed sequentially after the precise search results. In one embodiment, such as... Figure 5 As shown, a search for the keyword "group chat bug" yields one record related to user A and one record related to user B; these are precise search results. After expansion, searching for the expanded keyword "group bug" results in the client, the number of results for user A is increased by 1. The displayed page then looks like this. Figure 10 As shown. After clicking on the small A session in the search results, the fuzzy search results will appear after the precise results. Further expanded search results will also be displayed in the list in order. Furthermore, for keywords in the search results, such as... Figure 11As shown, extended keywords in the expanded search results can be identified during the display process; then, these extended keywords are highlighted; finally, the highlighted extended search results are displayed sequentially after the precise search results. Simultaneously, the upper right corner of the fuzzy search results can also be labeled with an AI icon, indicating to the user that the result is a match based on AI-generated keywords. In this embodiment, local keyword searches ensure basic search effectiveness, and the final search results are pushed and displayed based on search result priority, ensuring the accuracy of the displayed results.

[0108] In an exemplary embodiment, the method further includes: responding to a permission setting operation of a target object, generating a permission setting request based on the object information of the target object; sending the permission setting request to the server; displaying a permission setting interface based on the feedback of the permission setting request; and responding to the target object's permission information entry operation on the permission setting interface, sending authorization information back to the server.

[0109] For example, this application also includes a permission information setting process. The target object can trigger a permission setting request on the client side. The terminal responds to the target object's permission setting operation, generates a permission setting request based on the target object's object information, and sends the permission setting request to the server. The server, upon receiving the permission setting request submitted by the client, can provide the client with a permission setting interface. The target object on the client side can set authorization items and fill in permission information on the permission setting interface, and then submit the completed interface to the server. The server receives the authorization information based on the permission setting interface, sets the association between the authorization information and the object information, and stores the two together. Subsequently, the authorization information can be directly retrieved through the object information. In this embodiment, by receiving permission information through the client's permission setting request and a permission setting interface, the efficiency and accuracy of the authorization process are ensured.

[0110] This application also provides an application scenario, which is illustrated by taking the above-mentioned information record search as an example. The information record search method specifically includes:

[0111] In the field of instant messaging software, user information records are generally stored on the local client. When users wish to query historical information records, they can do so through the software's record query page. However, users may not be able to recall the specific content of the chat at that time, and entering keywords from memory may not yield the desired results. In this case, users can use the information record search method described in this application to expand their search keywords, thereby retrieving more record information and improving search accuracy.

[0112] The complete timeline flowchart for the information record search in this application can be found in [reference]. Figure 12As shown, users can select the AI ​​search function on the client to search for information records and enter search keywords. The server will submit a query request to the backend interface based on the user's search keywords, using the keywords as request parameters. Upon receiving the query request, the backend server will first asynchronously return a confirmation message to the client indicating that the request has been received. Simultaneously, the client will also perform a local search based on the user's entered keywords, prioritizing the display of search results obtained from directly entering keywords, and polling the server for newly generated extended keywords. The backend server will construct extended suggestions based on the search keywords and the requested object information, and then call a large language model based on the extended suggestions. The large language model will perform text generation processing to generate extended keywords corresponding to the keywords, and return the generated extended keywords to the backend server. The backend server will then return the obtained extended keywords to the terminal sequentially. After receiving the extended keywords, the terminal will perform a search using the extended keywords and display the search results sequentially until the search result for the last keyword is generated. Furthermore, the extended keywords in the search results can be highlighted.

[0113] Furthermore, this application also includes content related to permission settings. Users can initiate permission setting requests on the client to configure permissions during the keyword search process. The backend server sends a permission setting interface to the terminal based on the user's request. On the terminal page, users fill in the permission information that can be accessed during the search process, such as personal and company information. Then, when generating suggestion words, the server can construct suggestion words based on the authorized information, ensuring the effectiveness of expanded keyword generation.

[0114] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0115] Based on the same inventive concept, this application also provides an information record search device for implementing the information record search method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more information record search device embodiments provided below can be found in the limitations of the information record search method described above, and will not be repeated here.

[0116] In one exemplary embodiment, such as Figure 13 As shown, an information recording and searching device is provided, comprising:

[0117] The information extraction module 1302 is used to extract search keywords and object information from a search request when a search request is received.

[0118] The prompt word construction module 1304 is used to construct extended prompt words based on object information and search keywords, so as to obtain extended prompt words adapted to the large language model.

[0119] The keyword expansion module 1306 is used to input expansion prompts into the large language model for keyword expansion processing, and obtain multiple sets of expansion keywords corresponding to the expansion prompts. The expansion keywords are used to perform information record search processing and obtain corresponding expansion search results.

[0120] In one embodiment, the prompt word construction module 1304 is specifically used to: determine the personalized background information of the search request based on the object information; fill the personalized background information and search keywords into the prompt word template, perform extended prompt word construction processing, and obtain extended prompt words adapted to the large language model.

[0121] In one embodiment, the prompt word construction module 1304 is specifically used to: find the authorization information corresponding to the object information; determine the authorization item information based on the authorization information, and use the authorization item information as personalized background information for the search request.

[0122] In one embodiment, the system further includes an authorization processing module, configured to: upon receiving a permission setting request submitted by a client, provide a permission setting interface to the client; receive authorization information based on the permission setting interface; and associate and store the object information and authorization information in the permission setting request.

[0123] In one embodiment, the object information includes an object identifier. The information extraction module 1302 is specifically configured to: upon receiving a search request submitted by the target object, parse the search request to obtain search input information and the object identifier; and perform word segmentation on the search input information to obtain search keywords.

[0124] In one embodiment, the keyword feedback module 1308 is specifically used to: upon receiving a search request submitted by the client, extract the search keywords and object information from the search request; construct extended suggestion words based on the object information and search keywords to obtain extended suggestion words adapted to the large language model; input the extended suggestion words into the large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended suggestion words; the extended keywords are used to perform information record search processing and obtain corresponding extended search results. During the sequential generation of extended keywords, the newly generated extended keywords are fed back to the client in the order of generation.

[0125] In one exemplary embodiment, such as Figure 14 As shown, an information recording and search device installed on a terminal is provided, comprising:

[0126] The request generation module 1401 is used to respond to the search operation of the target object by inputting search keywords and to generate a search request based on the search keywords and the object information of the target object.

[0127] The request sending module 1403 is used to send search requests to the server.

[0128] The keyword receiving module 1405 is used to receive extended keywords from the server based on the search request. The extended keywords are obtained by expanding the search keywords through a large language model.

[0129] The record search module 1407 is used to perform information record search processing on the client's local machine based on extended keywords, obtain extended search results, and display the extended search results.

[0130] In one embodiment, the system further includes a precision search module, which is used to: perform information record search processing on the client's local machine based on search keywords to obtain precision search results and display the precision search results.

[0131] The record search module 1407 is specifically used to: asynchronously perform information record search processing on the local client based on extended keywords, obtain extended search results, and display the extended search results in order after the accurate search results.

[0132] In one embodiment, the record search module 1407 is further configured to: identify extended keywords in the extended search results; highlight the extended keywords in the extended search results; and display the highlighted extended search results in order after the accurate search results.

[0133] In one embodiment, the system further includes a permission setting module, configured to: respond to a permission setting operation of a target object, generate a permission setting request based on the object information of the target object; send the permission setting request to the server; display a permission setting interface based on the feedback of the permission setting request; and respond to the target object's permission information entry operation on the permission setting interface, and send authorization information back to the server.

[0134] Each module in the aforementioned information recording and search device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0135] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 15 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data related to information record retrieval. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements an information record retrieval method applied to a server.

[0136] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 16As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements an information recording and searching method applied to the terminal. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0137] Those skilled in the art will understand that Figure 15 and Figure 16 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0138] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0139] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0140] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0141] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0142] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0143] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An information record retrieval method, characterized in that, The method includes: Upon receiving a search request, extract the search keywords and object information from the search request; Based on the object information and the search keywords, expanded suggestion words are constructed to obtain expanded suggestion words adapted to the large language model; The extended prompt words are input into a large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. The extended keywords are used to perform information record search processing and obtain corresponding extended search results.

2. The method according to claim 1, characterized in that, The process of constructing expanded suggestion words based on the object information and the search keywords, resulting in expanded suggestion words adapted to the large language model, includes: Based on the object information, determine the personalized background information for the search request; The personalized background information and search keywords are filled into the prompt word template, and the extended prompt word construction process is carried out to obtain extended prompt words adapted to the large language model.

3. The method according to claim 2, characterized in that, The process of determining the personalized background information for the search request based on the object information includes: Find the authorization information corresponding to the object information; Based on the authorization information, authorization item information is determined, and the authorization item information is used as personalized background information for the search request.

4. The method according to claim 3, characterized in that, The method further includes: Upon receiving a permission setting request, generate and display the permission setting interface; Receive authorization information based on the permission settings interface; The object information in the permission setting request and the authorization information are associated and stored.

5. The method according to claim 1, characterized in that, The object information includes an object identifier; Upon receiving a search request submitted by the target object, extracting the search keywords and object information from the search request includes: Upon receiving a search request submitted by the target object, the search request is parsed to obtain search input information and object identifier; The search input information is segmented into words to obtain search keywords.

6. The method according to claim 1, characterized in that, The method is applied to the server side, and the method includes: Upon receiving a search request submitted by a client, extract the search keywords and object information from the search request; Based on the object information and the search keywords, expanded suggestion words are constructed to obtain expanded suggestion words adapted to the large language model; The extended prompt words are input into a large language model for keyword expansion processing to obtain multiple sets of extended keywords corresponding to the extended prompt words. The extended keywords are used to perform information record search processing and obtain corresponding extended search results. The method further includes: During the sequential generation of the extended keywords, the newly generated extended keywords are fed back to the client in the order of generation.

7. An information record search method, applied to a client, characterized in that, The method includes: In response to a search operation initiated by the target object using search keywords, a search request is generated based on the search keywords and the object information of the target object. Send the search request to the server; The server receives expanded keywords based on the search request, and the expanded keywords are obtained by expanding the search keywords using a large language model. Based on the extended keywords, information records are searched locally on the client side to obtain extended search results, which are then displayed.

8. The method according to claim 7, characterized in that, After sending the search request to the server, the following steps are included: Based on the search keywords, information records are searched and processed locally on the client side to obtain accurate search results, which are then displayed. The process of performing information record search on the client's local machine based on the extended keywords to obtain extended search results includes: Based on the extended keywords, information recording and search processing is performed asynchronously on the local client to obtain extended search results, which are then displayed sequentially after the accurate search results.

9. The method according to claim 8, characterized in that, The step of displaying the expanded search results sequentially after the precise search results includes: Identify the expanded keywords in the expanded search results; The extended keywords in the extended search results are highlighted. The annotated extended search results will be displayed sequentially after the precise search results.

10. The method according to claim 7, characterized in that, The method further includes: In response to a permission setting operation of the target object, a permission setting request is generated based on the object information of the target object; Send the permission setting request to the server; Displays the permission settings interface based on the feedback from the permission setting request; In response to the target object's permission information input operation on the permission settings interface, authorization information is sent back to the server.

11. An information recording and searching device, characterized in that, The device includes: The information extraction module is used to extract search keywords and object information from a search request upon receiving such a request. The prompt word construction module is used to construct extended prompt words based on the object information and the search keywords to obtain extended prompt words adapted to the large language model; The keyword expansion module is used to input the expansion prompts into a large language model for keyword expansion processing, thereby obtaining multiple sets of expanded keywords corresponding to the expansion prompts. The expanded keywords are used to perform information record search processing and obtain corresponding expanded search results.

12. An information recording and searching device, installed on a client side, characterized in that, The device includes: The request generation module is used to respond to the search operation of the target object by inputting search keywords, and to generate a search request based on the search keywords and the object information of the target object; The request sending module is used to send the search request to the server. The keyword receiving module is used to receive the expanded keywords fed back by the server based on the search request. The expanded keywords are obtained by expanding the search keywords through a large language model. The record search module is used to perform information record search processing on the client's local machine based on the extended keywords, obtain extended search results, and display the extended search results.

13. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 10.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 10.

15. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 10.

Images