Search service prefetching method and computing device

By judging supportive prefetching operations on the server side and selecting high-rated associative words for pre-search in advance, the problem of prefetching in the existing technology affects user experience, and improves the performance and user experience of search services.

CN118916556APending Publication Date: 2024-11-08CHEZHI HULIAN BEIJING SCI & TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410962237.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

When existing prefetching technology performs too many prefetching operations on the server, it will affect the user's normal search behavior and cannot 100% cache the content that the user really wants to search, resulting in poor user experience.

Method used

It is determined on the server side whether prefetching is currently supported, and prefetching is performed only when it is supported. By determining the search possibility score, select prefetching associative words for prefetching, and store the results on the client to avoid prefetching requests occupying server resources.

Benefits of technology

It improves the performance and user experience of search services, reduces the use of server resources, and ensures that users' normal search requests can be processed in a timely manner.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118916556A_ABST
    Figure CN118916556A_ABST
Patent Text Reader

Abstract

The invention relates to the field of search services, and particularly discloses a search service prefetching method and computing device.The method comprises the steps that in response to a received search word collected from a client side, a plurality of associational words are determined according to the search word, and the associational words are sent to the client side; in response to a received pre-fetching request sent by a client, whether the server supports a pre-fetching operation currently is judged, and the pre-fetching request comprises one or more pre-fetching associational words determined from the multiple associational words; if the pre-fetching operation is supported, pre-searching is carried out according to each pre-fetching associational word, and a pre-searching result of each predicted associational word is obtained; and sending one or more pre-search results of one or more pre-fetched associational words to the client for storage, so that when the client is triggered to perform search operation according to the pre-fetched associational words, the pre-stored pre-search results are acquired and loaded, and the pre-search results are displayed. The search performance can be effectively improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of search services, and in particular to a search service pre-fetching method and computing device. Background Art

[0002] Prefetching is a pre-search technology. Specifically, before the user triggers a search request, the server sends the content that the user wants to search and browse to the user in advance. In this way, when the user actually triggers the search behavior on the user side, the search request will no longer be sent to the server. The content pre-cached on the user side will be directly retrieved and presented to the user. Prefetching saves the time of generating search requests and waiting for the server to search and return. It can minimize the delay affected by the network and equipment, thereby shortening the user's waiting time and improving the response efficiency of the search service.

[0003] However, the current pre-fetching technology still has many shortcomings. For example, when too many pre-fetching operations are performed on the server side, it will affect the normal search behavior of users. Specifically, the pre-fetching operation cannot cache 100% of all the content that the user really wants to search. When the content that the user really needs to search is not cached on the user side, the user side still needs to send a search request to the server side. If the server-side processing of the pre-fetching operation affects the response to this part of the search request, it will reduce the service efficiency of the search engine, resulting in a poor user experience.

[0004] Therefore, a new search service pre-fetching method is urgently needed to solve the above technical problems. Summary of the invention

[0005] To this end, the present invention provides a search service pre-fetching method in an effort to solve or at least alleviate the above problems.

[0006] According to a first aspect of the present invention, a search service pre-fetching method is provided, which runs in a server, and the server is communicatively connected to one or more clients, and the method comprises: in response to receiving a search term collected from a client, determining a plurality of associated terms according to the search term, and sending the associated terms to the client; in response to receiving a pre-fetching request sent by the client, determining whether the server currently supports a pre-fetching operation, the pre-fetching request comprising one or more pre-fetched associated terms determined from a plurality of associated terms; if the pre-fetching operation is supported, performing a pre-search according to each pre-fetched associated term to obtain a pre-search result of each predicted associated term; and sending one or more pre-search results of the one or more pre-fetched associated terms to the client for storage, so that when the client is triggered to perform a search operation according to the pre-fetched associated terms, the pre-stored pre-search results are obtained and loaded, and the pre-search results are displayed.

[0007] Optionally, in the search service prefetching method according to the present invention, in response to receiving a prefetching request sent by a client, determining whether the server currently supports the prefetching operation includes: determining whether the prefetching switch of the server is turned on; if the prefetching switch is turned on, the server supports the prefetching operation.

[0008] Optionally, in the search service prefetching method according to the present invention, the prefetching switch is suitable for closing upon receiving a fuse signal, and the method also includes: determining the service status of each service node providing search service, the service status including resource utilization and service parameters of one or more resources; if it is determined that the resource utilization or service parameters of a service node reaches a preset threshold; then sending a fuse signal to the prefetching switch.

[0009] Optionally, in the search service pre-fetching method according to the present invention, the one or more resources of the service node include one or more of memory, CPU, and disk, and the service parameter includes the service node response time.

[0010] Optionally, in the search service prefetching method according to the present invention, if the prefetching operation is supported, a pre-search is performed based on each pre-fetching associated word, including: determining whether the service node can obtain a token from a token bucket; if the token can be obtained from the token bucket, a pre-search is performed based on the pre-fetching associated word, and the number of tokens in the token bucket is set to be reduced by 1; if the token cannot be obtained from the token bucket, a pre-fetching failure message is returned.

[0011] Optionally, the search service pre-fetching method according to the present invention further includes: determining a token adding speed of the token bucket, and adding tokens to the token bucket according to the token adding speed.

[0012] Optionally, in the search service pre-fetching method according to the present invention, determining the token addition speed of the token bucket includes: obtaining the search request historical processing peak value and the current search request processing speed of the service node; if the current search request processing speed is greater than or equal to the search request historical processing peak value, setting the token addition speed to 0; if the current search request processing speed is less than the search request historical processing peak value, determining the token addition speed based on the theoretical processing capacity of the service node and the current search request processing speed.

[0013] According to another aspect of the present invention, a search service pre-fetching method is provided, which is executed in a client, and the client is communicatively connected to the server, and the method includes: in response to receiving a search term input by a user, sending the search term to the server; receiving multiple associative terms returned by the server, and displaying the multiple associative terms so that the user can select one from the associative terms for search; determining one or more pre-fetched associative terms from the multiple associative terms, and sending the one or more pre-fetched associative terms to the server; if the server supports the pre-fetching operation, receiving and storing one or more pre-search results returned by the server; in response to receiving a selection operation for one of the pre-fetched associative terms among the multiple associative terms, acquiring and loading a pre-stored pre-search result according to the selection operation, and displaying the pre-search result.

[0014] Optionally, in the search service prefetching method according to the present invention, in response to receiving a selection operation on an associative word that has not been prefetched among multiple associative words, a search request is generated according to the selection operation; and the search request is sent to the server to search for the associative word that has not been prefetched.

[0015] Optionally, in the search service prefetching method according to the present invention, determining one or more prefetched associative words from the multiple associative words includes: determining a search possibility score for each associative word; and selecting one or more associative words with the highest search possibility score as prefetched associative words.

[0016] Optionally, in the search service pre-fetching method according to the present invention, determining the search possibility score for each associated word includes: inputting each associated word and the search term into a pre-trained multi-task learning model to obtain the search possibility score of each associated word.

[0017] According to another aspect of the present invention, a computing device is provided, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing the search service prefetching method according to the present invention.

[0018] According to another aspect of the present invention, a computer-readable storage medium storing one or more programs is provided. The one or more programs include instructions that, when executed by a computing device, cause the computing device to execute the search service prefetching method according to the present invention.

[0019] According to another aspect of the present invention, a computer program product is provided, comprising a computer program / instruction, wherein the computer program / instruction implements the search service pre-fetching method of the present invention when executed by a processor.

[0020] According to the search service pre-fetching method of the present invention, after the server receives the pre-fetching request sent by the client, it first determines whether the server currently supports the pre-fetching operation. If the server supports the pre-fetching operation, it searches according to the pre-fetching associated words to obtain the pre-search results. It can be seen that when the present invention performs pre-fetching, it does not directly process the pre-fetching request, but first determines the state of the server, and only performs pre-fetching when the current state of the server supports pre-fetching. In this way, it can avoid the situation where the user's normal search request cannot be processed due to pre-fetching, thereby improving the performance of the search service and further improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] To achieve the above and related purposes, certain illustrative aspects are described herein in conjunction with the following description and accompanying drawings, which indicate various ways in which the principles disclosed herein can be practiced, and all aspects and their equivalents are intended to fall within the scope of the claimed subject matter. The above and other purposes, features and advantages of the present disclosure will become more apparent by reading the following detailed description in conjunction with the accompanying drawings. Throughout the present disclosure, the same reference numerals generally refer to the same parts or elements.

[0022] Figure 1 A schematic diagram of a server 120 according to an embodiment of the present invention is shown;

[0023] Figure 2 A schematic diagram showing a communication connection between a search server and a client according to an embodiment of the present invention is shown;

[0024] Figure 3 A flowchart of a search service pre-fetching method 300 according to an embodiment of the present invention is shown;

[0025] Figure 4 A structural block diagram of a computing device 400 according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0026] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0027] The present invention proposes a search service pre-fetching method, and each step of the method can be executed in a server and a client in a distributed manner.

[0028] According to one embodiment of the present invention, a server can specifically provide multiple services. The server can be specifically implemented as a device in a service cluster composed of multiple devices. The service cluster uniformly deploys and integrates the resources of each device to obtain multiple service nodes. Each service node can provide one or more services independently, and can also change the service type provided by the service node. The present invention does not limit the specific service type provided by each service node. In the service cluster, each service node can be specifically implemented as a pod, which is the smallest deployable unit created and managed in the Kubernetes service cluster.

[0029] Figure 1 FIG. 1 is a schematic diagram of a server 120 according to an embodiment of the present invention. Figure 1 As shown, the server 120 can provide search engine services, advertising services, recall services, sorting services, etc. Multiple service nodes can be deployed under each service. The present invention does not limit the types and quantities of services provided by the server. The present invention does not limit the number of service nodes specifically deployed under each service.

[0030] According to an embodiment of the present invention, the server provides a search service and processes pre-fetch requests and search requests. The server can be specifically implemented as a search server.

[0031] Figure 2 FIG. 2 is a schematic diagram showing a communication connection between a search server and a client according to an embodiment of the present invention. Figure 2 As shown, the search server 120 is in communication connection with the clients 111 to 113. The present invention does not limit the number of clients connected to the search server 120 and the specific connection method.

[0032] The search server 120 is in communication with the data server 140 and the configuration server 130. The search server 120 is adapted to provide a search service, and obtains pre-search results or search results from the data server 140 according to a received pre-fetch request or search request. The configuration server 130 is adapted to receive and store configuration information, and configure the search server 120 according to the configuration information.

[0033] According to an embodiment of the present invention, the configuration server 130 may be implemented as an Apollo server, and the data server 140 may be implemented as a Redis server. The present invention does not limit the specific implementation of the configuration server 130 and the data server 140.

[0034] Figure 3 FIG. 3 is a flowchart of a search service pre-fetching method 300 according to an embodiment of the present invention. Figure 3As shown, first, the client executes step 301, and in response to receiving the search word input by the user, sends the search word to the server. The server can be specifically implemented as a search server.

[0035] According to one embodiment of the present invention, a user can perform operations on the client, such as browsing information, inputting information, or clicking a "search button" to search. The client provides a form or input box to the user so that the user can input information. When the information input by the user is a query word, it means that the user wants to query relevant information based on the query word. The query word can be specifically implemented as a short sentence, a phrase, a word, a character, a character, etc. The present invention does not limit the specific form of the query word.

[0036] The present invention sends the search term to the search server when the user inputs the search term and has not clicked the "search button" (ie when the search operation is triggered on the client).

[0037] Subsequently, the search server executes step 302, in response to receiving the search terms collected from the client, determines multiple associative terms (sug) according to the search terms, and sends the associative terms to the client. Determine multiple associative terms according to the search terms, specifically, the search terms can be set to the content that should be included in the associative terms to query the associative terms, or directly query the words close to the search terms. The present invention does not limit the specific method of determining associative terms according to the search terms. According to one embodiment of the present invention, the present invention can set the number of associative terms, such as setting the number of associative terms for each search term to 10, and the present invention does not limit the number of associative terms corresponding to each search term.

[0038] According to an embodiment of the present invention, when the search term is "technology", the associated terms may be determined as "technology pictures", "technology essays" and the like.

[0039] Subsequently, the client executes step 303 to receive the multiple associative words returned by the server and display the multiple associative words so that the user can select one from the associative words for search. When the client displays the multiple associative words, the multiple associative words can be displayed in the form of a list below or beside the search word input by the user. The present invention does not limit the specific display method of the multiple associative words.

[0040] After the client displays the associated words, the user can directly search based on the associated word by selecting or clicking on the associated word, without having to complete the search word to the length of the associated word, that is, no longer having to enter the part of the search word that is missing relative to the associated word.

[0041] Subsequently, the client performs step 304 to determine one or more pre-fetched associative words from the multiple associative words, and sends the one or more pre-fetched associative words to the server. Determining the pre-fetched associative words includes: determining a search possibility score for each associative word; selecting one or more associative words with the highest search possibility score as the pre-fetched associative words. When determining the search possibility score for each associative word, each associative word and the search word are input into a pre-trained multi-task learning model to obtain a search possibility score for each associative word.

[0042] According to one embodiment of the present invention, when training a multi-task learning model, a PLE model can be used for training, multiple search terms are obtained in advance, and one or more associated terms of each search term are set to obtain a training set, which is input into the multi-task learning model for training. When training a multi-task learning model, the user's behavioral characteristics can also be considered, such as adding the user's historical search records to the training set for training. The generated multi-task learning model generates a search possibility score for each associated term when receiving a search term and one or more corresponding associated terms. The search possibility score can reflect the correlation between the associated term and the search term, such as the probability of co-occurrence or the degree of similarity of the word concept, so the search possibility score can characterize the probability that the user selects the associated term for search. Among them, each client can receive the multi-task learning model that has been trained by the server in advance, and then perform further training, or it can perform training by itself after receiving the multi-task learning model. The multi-task learning model of each client can be different, and the present invention does not limit the specific training method of the multi-task learning model.

[0043] In order to reduce the pre-fetching pressure of the search server and save the resources of the search server, one or more associative words with high scores can be selected from multiple associative words as pre-fetched associative words, and associative words with low scores can be eliminated. Further, when selecting the pre-fetched associative words, it can also be determined whether the associative words have been pre-fetched and the corresponding pre-search results are stored. If they have been pre-fetched and the corresponding pre-search results are stored, other associative words are set as pre-fetched associative words.

[0044] Subsequently, the search server executes step 305, in response to receiving the pre-fetch request sent by the client, to determine whether the server currently supports the pre-fetch operation, the pre-fetch request including one or more pre-fetch associated words determined from the multiple associated words, specifically, one pre-fetch request including one pre-fetch associated word. Determining whether the server currently supports the pre-fetch operation includes: determining whether the pre-fetch switch of the server is turned on; if the pre-fetch switch is turned on, the server supports the pre-fetch operation.

[0045] The present invention sets a pre-fetch switch in the search server, and the pre-fetch switch is used to control and indicate whether the search server is allowed to perform a pre-fetch operation. The pre-fetch switch can be turned on and off manually or automatically. The pre-fetch switch can be turned on or off by setting an identification value for the pre-fetch switch. When the pre-fetch switch is turned on, it indicates that the search server can currently perform a pre-fetch operation; when the pre-fetch switch is turned off, it indicates that the search server cannot currently perform a pre-fetch operation.

[0046] The search server may also stop the pre-fetch operation when a fuse signal is detected. According to one embodiment of the present invention, the pre-fetch switch is adapted to be closed when a fuse signal is received, and the fuse signal is a signal for controlling the pre-fetch switch to be closed. The fuse signal is triggered when it is detected that the conditions for providing the pre-fetch service are not met. When the pre-fetch switch is in the closed state, the fuse signal may not be issued. When the pre-fetch switch is in the open state, the fuse signal may be issued to close the pre-fetch switch.

[0047] According to one embodiment of the present invention, the step of controlling the issuance of a fuse signal includes: determining the service status of each service node providing search services, the service status including resource utilization and service parameters of one or more resources; if it is determined that the resource utilization or service parameters of a service node reaches a preset threshold, sending a fuse signal to the pre-fetch switch.

[0048] According to one embodiment of the present invention, one or more service nodes may be deployed in the search server, and the present invention does not limit the number of service nodes deployed in the search server. One or more service nodes in the search server provide search services for responding to and processing pre-fetch requests and search requests. When determining the service status of a service node, if the service status of a service node does not meet the conditions, that is, one or more resource utilization rates or service parameters reach a preset threshold, a fuse signal is sent to the pre-fetch switch.

[0049] The one or more resources of the service node include one or more of memory, CPU, and disk. The resources managed by each service node are determined and allocated by the service cluster, and the present invention does not limit the resource types included in the service node.

[0050] The resource utilization rate of each resource is the current occupancy of the resource of the service node. For example, if the memory usage of the current service node is 80%, the resource utilization rate is 80%. The preset threshold of the resource utilization rate of each resource can be set to be the same or different. The present invention does not limit the specific setting value and setting method of the preset threshold of the resource utilization rate of the resource.

[0051] According to an embodiment of the present invention, the service parameters include service node response time, service node network status, etc. The present invention does not limit the specific type of service parameters. The service node response time can be implemented as the response time to a pre-fetch request or a search request.

[0052] According to an embodiment of the present invention, the setting information of the pre-fetch switch and the setting information of the fuse signal can be set in the configuration server, and the configuration server synchronizes the above setting information to the search server. The setting information of the pre-fetch switch includes whether the pre-fetch switch controls whether pre-fetching is allowed, and the configuration information of the fuse information includes the resources that need to be monitored in the service node, and the preset threshold value set for each resource.

[0053] According to one embodiment of the present invention, the service of the service node can be determined once at a preset time interval to determine whether a fuse signal is to be issued, such as determining the service status of the service node in the previous minute every one minute. When the service pressure of the service node is reduced, the fuse signal can be turned off, the pre-fetch switch can be turned on, and the pre-fetch service of the service node can be restored. Regularly detecting whether a fuse signal is issued at a preset time interval can shorten the response time of the pre-fetch request and improve the response efficiency compared to determining whether the search server can perform pre-fetching when receiving the pre-fetch request.

[0054] According to one embodiment of the present invention, after sending a fuse signal to close the pre-fetch switch, an email alarm or SMS alarm can be issued based on pre-stored email or phone information to inform of abnormal conditions, resource usage logs exceeding preset thresholds, etc.

[0055] The present invention can prevent the service node providing the search service from experiencing an avalanche in the upstream and downstream due to excessive pressure on the traffic by setting a mechanism for sending and closing a fuse signal. If the pressure is too high during monitoring, the service node can be downgraded in time, a fuse signal can be sent, and the pre-fetch service can be turned off. Furthermore, since the present invention directly monitors the resources of the service node, it can also warn the service in time when the current traffic data is calculated incorrectly.

[0056] Subsequently, the search server executes step 306, and if the pre-fetch operation is supported, a pre-search is performed according to each pre-fetched associative word to obtain a pre-search result of each predicted associative word. When the pre-search is performed according to the pre-fetched associative word, the service node can call the search interface according to the pre-fetched associative word, search the pre-fetched associative word, and obtain the pre-search result. The pre-search result is the same as the result obtained by searching when the predicted associative word is used as the search word. The present invention searches the pre-fetched associative word in advance, and sends the pre-search result to the client, which can speed up the search and display.

[0057] According to one embodiment of the present invention, the present invention also limits the execution of pre-fetch operations by setting a token bucket for each service node, and sets the speed at which the service node executes the pre-fetch operation through the token bucket, thereby avoiding occupying too many resources when executing the pre-fetch operation and affecting the service node's response to the search request.

[0058] According to an embodiment of the present invention, a pre-search is performed based on each pre-fetched associative word, including: judging whether the service node can obtain a token from a token bucket, that is, whether there are tokens in the token bucket provided to the service node for acquisition. If there are tokens in the token bucket, the service node can obtain the token from the token bucket; if there are no tokens in the token bucket, the service node cannot obtain the token from the token bucket.

[0059] If the token can be obtained from the token bucket, a pre-search is performed according to the pre-fetch associated word, and the number of tokens in the token bucket is set to be reduced by 1; if the token cannot be obtained from the token bucket, a pre-fetch failure message is returned.

[0060] The present invention also determines the token adding speed of the token bucket, and adds tokens to the token bucket according to the token adding speed. According to one embodiment of the present invention, the token adding speed is the number of tokens added to the token bucket per second.

[0061] The steps to determine the token addition rate of the token bucket include:

[0062] Obtain the historical processing peak value of search requests and the current processing speed of search requests of the service node; the historical processing peak value of search requests is the highest processing speed of search requests by the service node in the past preset time period, such as the highest processing speed of search requests per second in the past 7 days. Search requests are different from pre-fetch requests. In the client, when the user does not click on the pre-fetched associated words to search, but clicks on other associated words that are not pre-fetched or directly searches for the search words, the search request is triggered. The search request is the request actually triggered by the client, and the pre-fetch request is the request that the client predicts that the corresponding associated words may be retrieved and searches in advance. The user does not necessarily search for the pre-fetched associated words. The current search request processing speed is the processing speed of the search request by the current service node, such as taking the number of search requests processed in the last second as the current search request processing speed. The current search request processing speed can also be determined by the total number of search requests processed in the last minute, such as dividing the total number of search requests processed in the last minute by 60 to obtain the search request processing speed. When counting the total number of search request processing per minute, you can set the redis key of each search request, and then count the key value to get the total number of search requests. After setting the key value, you can set the expiration time of the key value to facilitate statistics.

[0063] If the current search request processing speed is greater than or equal to the historical processing peak value of the search request, the token adding speed is set to 0. If the current search request processing speed is already greater than or equal to the historical processing peak value, it means that the service node is already under great pressure to process the search request and needs to suspend processing of pre-fetch requests. The token adding speed can be set to 0. According to one embodiment of the present invention, it can also be set that if the current search request processing speed is greater than or equal to the ratio value of the historical processing peak value of the search request, such as 70% or 80%, the processing of pre-fetch requests is suspended.

[0064] If the current search request processing speed is less than the historical search request processing peak, the token adding speed is determined based on the theoretical processing capacity of the service node and the current search request processing speed. The theoretical processing capacity of a service node is the maximum number of requests that the service node can process per second, including pre-fetch requests and search requests. The value of the theoretical processing capacity can be set in advance or according to the resources managed by the service node and the network conditions. The present invention does not limit the specific method for determining the theoretical processing capacity. Setting the theoretical processing capacity can prevent over-issuance of tokens and avoid data errors in the program.

[0065] When determining the token adding speed based on the theoretical processing capacity of the service node and the current search request processing speed, the token adding speed is obtained by subtracting the current request processing speed from the theoretical processing capacity, that is, how many pre-fetch requests can be processed per second.

[0066] According to an embodiment of the present invention, in order to ensure the validity of the tokens in the token bucket, the number of tokens in the token bucket may be set to be reset to zero at preset intervals, such as at intervals of 1 minute.

[0067] According to an embodiment of the present invention, if the token adding speed is set to 0, a fuse signal may be sent to the pre-fetch switch to close the pre-fetch switch.

[0068] Subsequently, the search server executes step 307 to send one or more pre-search results of one or more pre-fetched associative words to the client for storage, so that when the client is triggered to perform a search operation based on the pre-fetched associative words, it can obtain and load the pre-stored pre-search results and display the pre-search results.

[0069] Then, the client executes step 308, and if the server supports the pre-fetch operation, receives and stores one or more pre-search results returned by the server.

[0070] According to one embodiment of the present invention, the client may include a network library, and the pre-search results are stored in the network library. The network library can be specifically implemented as an abstract layer for managing network requests in the application running in the client, and the network requests include pre-fetch requests, search requests, etc. According to one embodiment of the present invention, the present invention can also set a cache expiration time of the network library, such as automatically deleting the pre-search results in the network library when the storage time exceeds the cache expiration time, so as to avoid excessive occupation of client resources by the pre-search result cache.

[0071] When the client obtains pre-search results from the network according to the pre-fetched associative words, if the cache has expired and been deleted, the client sends a search request to the server according to the pre-fetched associative words to obtain the search results.

[0072] Subsequently, the client executes step 309, and in response to receiving a selection operation for a pre-fetched associative word among the multiple associative words, obtains and loads the pre-stored pre-search results according to the selection operation, and displays the pre-search results. When obtaining the pre-search results, the pre-search results corresponding to the pre-fetched associative words can be matched and obtained from the network library, and displayed. The selection operation of the pre-fetched associative word indicates that the user wants to search according to the pre-fetched associative word, and at this time, the stored pre-search results are taken out and displayed.

[0073] The present invention sets a pre-fetch strategy through the above steps, and stores the pre-fetched search results of multiple association words with high scores in advance in the network library of the client. In this way, after the user selects the pre-fetched association word, the corresponding pre-fetched search result can be directly obtained from the network library for rendering, which saves the time of requesting the server and the time of the server responding to the search request. The search time is the time of client rendering, which effectively improves the search performance.

[0074] Among them, the size of search traffic has a certain regularity, and the peaks and troughs of search traffic are correlated with time. Therefore, the present invention can increase the search volume at the traffic trough by setting a pre-fetching strategy. In addition, after setting the pre-fetching strategy, the response speed of the user's search behavior will be improved, and the rendering time from the server to the client will be reduced, so that the number of times the user can operate in the same unit time will increase, thereby improving PV, distribution, and per capita data viewing. After the search pre-fetching strategy is implemented, the search traffic can keep the main search traffic from growing during the peak period, increase the pre-fetched traffic during the off-peak period, and significantly increase the search volume, achieving the theoretical effect of peak shaving and valley flattening. To ensure that the traffic does not grow during the peak period, it is necessary to use specific rules to limit the flow in different time periods of the high and low peak periods, such as setting a token bucket algorithm for each service node to limit the flow.

[0075] After implementing the pre-fetch strategy, the traffic data is compared and the improvement effect is as follows:

[0076] pv_ctr: +10.47% pt

[0077] uv_ctr: +2.92% pt

[0078] Server to Client Time: -9.68% pt

[0079] Percentage of searches with clicks: +1.82% pt

[0080] Average number of data viewed per searcher: 4.72% pt

[0081] Search unique number conversion rate: 7.15% pt

[0082] Global lead conversion rate: 6.79% pt

[0083] Global unique number conversion rate: 6.40% pt

[0084] The number of materials or products selected per capita increased: 3.45% pt

[0085] According to one embodiment of the present invention, when in a specific business field, applying the pre-fetching strategy of the present invention can obtain better technical effects. In a specific business field, or a specific vertical field, the probability that the words that the user finally retrieves are associative words is high. For example, when selecting a car or viewing car information, statistics show that the number of users searching based on associative words over a period of time reached 1.4 million, accounting for 40% of the number of all search behaviors. At this time, applying the pre-fetching strategy of the present invention can greatly improve the response efficiency to the search.

[0086] According to one embodiment of the present invention, when multiple associative words are displayed to a user, the user can select a pre-fetched associative word for search, or can select an associative word that is not used as a pre-fetched associative word for search, or can directly search according to the search word. When the user selects an associative word that is not used as a pre-fetched associative word for search, the client generates a search request according to the selection operation in response to receiving a selection operation on an associative word that is not pre-fetched among multiple associative words; sends the search request to the server to search for the associative word that is not pre-fetched. When the user searches directly according to the search word, the client generates a search request according to the search operation in response to the search operation on the search word; sends the search request to the server to search for the search word, and after the search, returns the search result to the client for display to the user.

[0087] According to the search service pre-fetching method of the present invention, after the server receives the pre-fetching request sent by the client, it first determines whether the server currently supports the pre-fetching operation. If the server supports the pre-fetching operation, it searches according to the pre-fetching associated words to obtain the pre-search results. It can be seen that when the present invention performs pre-fetching, it does not directly process the pre-fetching request, but first determines the state of the server, and only performs pre-fetching when the current state of the server supports pre-fetching. In this way, it can avoid the situation where the user's normal search request cannot be processed due to pre-fetching, thereby improving the performance of the search service and further improving the user experience.

[0088] The client, configuration server, data server, server or search server of the present invention can be specifically implemented as a computing device. Figure 4 4 shows a block diagram of a computing device according to an embodiment of the present invention. In a basic configuration, computing device 400 includes at least one processing unit 420 and system memory 410. According to one aspect, depending on the configuration and type of computing device, system memory 410 includes, but is not limited to, volatile storage (e.g., random access memory), non-volatile storage (e.g., read-only memory), flash memory, or any combination of such memories. According to one aspect, system memory 410 includes operating system 411.

[0089] According to one aspect, operating system 411, for example, is suitable for controlling the operation of computing device 400. Furthermore, examples may be practiced in conjunction with graphics libraries, other operating systems, or any other application programs, and are not limited to any particular application or system. Figure 4 This basic configuration is illustrated in FIG by those components within dashed line 415. According to one aspect, computing device 400 has additional features or functionality. For example, according to one aspect, computing device 400 includes additional data storage devices (removable and / or non-removable), such as magnetic disks, optical disks, or tapes.

[0090] As stated above, according to one aspect, the program module 412 is stored in the system memory 410. According to one aspect, the program module 412 can be implemented as one or more computer program products, and the present application does not limit the type of computer program products, for example, it can include: email, word processing application, spreadsheet application, database application, slide show application, painting or computer-aided application, web browser, etc. In some embodiments according to the present application, the computer program / instructions related to the search service pre-fetching method 300 are packaged into a computer program product, and when these computer programs / instructions are executed by the processor (i.e., the processing unit 420), the search service pre-fetching method 300 according to the present application is implemented.

[0091] According to one aspect, the examples may be practiced on a circuit comprising discrete electronic components, a packaged or integrated electronic chip containing logic gates, a circuit utilizing a microprocessor, or a single chip containing electronic components or a microprocessor. Figure 4 Each or many components shown in can be integrated into a system on chip (SOC) on a single integrated circuit to practice examples. According to one aspect, such a SOC device may include one or more processing units, a graphics unit, a communication unit, a system virtualization unit, and various application functions, all of which are integrated (or "burned") into a chip substrate as a single integrated circuit. When operated via SOC, the functions described in this article can be operated via a dedicated logic integrated with other components of the computing device 400 on a single integrated circuit (chip). Other technologies that can perform logical operations (such as AND, OR, and NOT) can also be used to practice embodiments of the present invention, including but not limited to mechanical, optical, fluid, and quantum technologies. In addition, embodiments of the present invention can be practiced in a general-purpose computer or in any other circuit or system.

[0092] According to one aspect, the computing device 400 may also have one or more input devices 431, such as a keyboard, a mouse, a pen, a voice input device, a touch input device, etc. Output devices 432 may also be included, such as a display, a speaker, a printer, etc. The aforementioned devices are examples and other devices may also be used. The computing device 400 may include one or more communication connections 433 that allow communication with other computing devices 440. Examples of suitable communication connections 433 include, but are not limited to: RF transmitters, receivers and / or transceiver circuits; Universal Serial Bus (USB), parallel and / or serial ports. The computing device 400 can be communicatively connected to other computing devices 440 via the communication connection 433.

[0093] The embodiment of the present invention also provides a non-transitory readable storage medium, which stores instructions, and the instructions are used to make the computing device perform a method according to an embodiment of the present invention. The readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be a computer-readable instruction, a data structure, a module of a program, or other data. Examples of readable storage media include, but are not limited to: phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, tape disk storage or other magnetic storage device or any other non-transitory readable storage medium.

[0094] According to one aspect, communication media is implemented by computer readable instructions, data structures, program modules, or other data in a modulated data signal (e.g., a carrier wave or other transport mechanism), and includes any information delivery media. According to one aspect, the term "modulated data signal" describes a signal that has one or more characteristics set or changed in a manner that encodes information in the signal. By way of example and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, radio frequency (RF), infrared, and other wireless media.

[0095] It should be noted that, although the computing device shown above only includes a processing unit 420, a system memory 410, an input device 431, an output device 432, and a communication connection 433, in a specific implementation, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the device may only include components necessary for implementing the embodiments of the present specification, and need not include all the components shown in the figure.

[0096] In addition, embodiments of the present invention also include: A10, a method as described in A8 or A9, wherein the determining one or more pre-fetched associative words from the multiple associative words includes: determining a search possibility score for each associative word; selecting one or more associative words with the highest search possibility score as pre-fetched associative words. A11, a method as described in A10, wherein the determining a search possibility score for each associative word includes: inputting each associative word and the search term into a pre-trained multi-task learning model to obtain a search possibility score for each associative word. A13, a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by a computing device, the computing device executes a method according to any one of A1-A11. A14, a computer program product, including a computer program / instruction, wherein when the computer program / instruction is executed by a processor, the method as described in any one of A1-A11 is implemented.

[0097] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0098] Similarly, it should be understood that in order to streamline the present disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof.

[0099] In the case where the program code is executed on a programmable computer, the computing device generally includes a processor, a storage medium readable by the processor (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device. The memory is configured to store the program code; the processor is configured to execute the search service pre-fetching method of the present invention according to the instructions in the program code stored in the memory.

[0100] By way of example and not limitation, computer readable media include computer storage media and communication media. Computer readable media include computer storage media and communication media. Computer storage media stores information such as computer readable instructions, data structures, program modules or other data. Communication media generally embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and includes any information delivery medium. Any combination of the above is also included within the scope of computer readable media.

[0101] Although the present invention has been described according to a limited number of embodiments, it will be apparent to those skilled in the art, with the benefit of the above description, that other embodiments may be envisioned within the scope of the invention thus described. In addition, it should be noted that the language used in this specification is selected primarily for readability and instructional purposes, rather than for explaining or defining the subject matter of the present invention. Therefore, many modifications and variations will be apparent to those skilled in the art. The disclosure of the present invention is illustrative, rather than restrictive, with respect to the scope of the present invention.

Claims

1. A search service pre-fetching method, executed in a server, wherein the server is in communication connection with one or more clients, the method comprising: In response to receiving a search term collected from a client, determining a plurality of associated terms according to the search term, and sending the associated terms to the client; In response to receiving a pre-fetch request sent by a client, determining whether the server currently supports a pre-fetch operation, the pre-fetch request including one or more pre-fetched associative words determined from a plurality of associative words; If the pre-fetch operation is supported, a pre-search is performed based on each pre-fetched associated word to obtain a pre-search result for each predicted associated word; One or more pre-search results of one or more pre-fetched associative words are sent to the client for storage, so that when the client is triggered to perform a search operation according to the pre-fetched associative words, the pre-stored pre-search results are obtained and loaded, and the pre-search results are displayed.

2. The method of claim 1, wherein: In response to receiving the prefetch request sent by the client, determining whether the server currently supports the prefetch operation includes: Determining whether a pre-fetch switch of the server is turned on; If the pre-fetch switch is turned on, the server supports pre-fetch operations.

3. The method of claim 2, wherein: The pre-fetch switch is adapted to be closed upon receiving a fuse signal, and the method further comprises: Determine a service status of each service node providing a search service, wherein the service status includes resource usage and service parameters of one or more resources; If it is determined that the resource usage rate or service parameter of a service node reaches a preset threshold, a fuse signal is sent to the pre-fetch switch.

4. The method of claim 3, wherein: The one or more resources of the service node include one or more of memory, CPU, and disk, and the service parameter includes service node response time.

5. The method according to any one of claims 1 to 4, wherein: If the pre-fetch operation is supported, a pre-search is performed according to each pre-fetched associated word, including: Determine whether the service node can obtain a token from the token bucket; If a token can be obtained from the token bucket, a pre-search is performed according to the pre-fetched associative words, and the number of tokens in the token bucket is set to be reduced by 1; If a token cannot be obtained from the token bucket, a pre-fetch failure message is returned.

6. The method of claim 5, further comprising: A token adding speed of the token bucket is determined, and tokens are added to the token bucket according to the token adding speed.

7. The method of claim 6, wherein: Determining the token addition speed of the token bucket includes: Obtaining a historical search request processing peak value and a current search request processing speed of the service node; If the current search request processing speed is greater than or equal to the search request historical processing peak value, set the token adding speed to 0; If the current search request processing speed is lower than the search request historical processing peak value, the token adding speed is determined according to the theoretical processing capacity of the service node and the current search request processing speed.

8. A search service pre-fetching method, executed in a client, the client being in communication connection with the server, the method comprising: In response to receiving a search term input by a user, sending the search term to the server; receiving multiple associated words returned by the server, and displaying the multiple associated words so that the user can select one from the associated words for search; Determine one or more pre-fetched associative words from the plurality of associative words, and send the one or more pre-fetched associative words to the server; If the server supports pre-fetching operation, receiving and storing one or more pre-search results returned by the server; In response to receiving a selection operation of a pre-fetched associative word among a plurality of associative words, a pre-stored pre-search result is acquired and loaded according to the selection operation, and the pre-search result is displayed.

9. The method of claim 8, further comprising: In response to receiving a selection operation on an associative word that has not been pre-fetched among the plurality of associative words, generating a search request according to the selection operation; The search request is sent to the server to search for the associated words that have not been pre-fetched.

10. A computing device comprising: one or more processors; Memory; as well as One or more programs, wherein the one or more programs are stored in a memory and configured to be executed by one or more processors, and the one or more programs include instructions for executing the method according to any one of claims 1-9.