Query method and device, server and storage medium

Abstract

The embodiment of the invention discloses a query method and device, a server and a storage medium, and belongs to the technical field of information retrieval. The method comprises the steps that under the condition that a query request of a client side is received, a second query text corresponding to a first query text in the query request is determined, and the second query text is obtained through rewriting based on the first query text; returning a first query result corresponding to the first query text and a second query result corresponding to the second query text to the client, wherein the client is used for displaying the first query result and the second query result; based on first conversion data of the first query result and second conversion data of the second query result, rewriting accuracy of the second query text is checked, a rewriting accuracy check result of the second query text is obtained, the first conversion data is used for representing the conversion condition of the first query result at the client, and the second conversion data is used for representing the conversion condition of the second query result at the client. The second conversion data is used for representing the conversion condition of the second query result at the client; according to the technical scheme, the query accuracy can be improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of information retrieval technology, and in particular to a query method, device, server and storage medium. Background Art

[0002] During the song search process, the query text entered by the user may contain spelling errors, inaccurate song names, vague singer names, etc., making it difficult for the search engine to return the search results required by the user.

[0003] Search engines rewrite the query text entered by users according to preset rules or machine learning models, trying to return search results that meet user needs. However, if the rewriting is inappropriate, the accuracy of the search results returned by the search engine will be reduced. Summary of the invention

[0004] The embodiments of the present application provide a query method, device, server and storage medium, including the following technical contents.

[0005] On the one hand, an embodiment of the present application provides a query method, the method comprising:

[0006] In the case of receiving a query request from a client, determining a second query text corresponding to a first query text in the query request, where the second query text is rewritten based on the first query text;

[0007] Returning a first query result corresponding to the first query text and a second query result corresponding to the second query text to the client, and the client is used to display the first query result and the second query result;

[0008] Based on the first conversion data of the first query result and the second conversion data of the second query result, the rewriting accuracy of the second query text is verified to obtain the rewriting accuracy verification result of the second query text, the first conversion data is used to characterize the conversion status of the first query result at the client, and the second conversion data is used to characterize the conversion status of the second query result at the client.

[0009] On the other hand, an embodiment of the present application provides a query device, the device comprising:

[0010] A text determination module, configured to, upon receiving a query request from a client, determine a second query text corresponding to a first query text in the query request, wherein the second query text is rewritten based on the first query text;

[0011] A return module, used to return a first query result corresponding to the first query text and a second query result corresponding to the second query text to the client, and the client is used to display the first query result and the second query result;

[0012] A verification module is used to verify the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result, so as to obtain the rewriting accuracy verification result of the second query text, wherein the first conversion data is used to characterize the conversion status of the first query result at the client, and the second conversion data is used to characterize the conversion status of the second query result at the client.

[0013] On the other hand, an embodiment of the present application provides a server, which includes a processor and a memory, wherein the memory stores at least one computer instruction, and the at least one computer instruction is loaded and executed by the processor to implement the query method described in the above aspects.

[0014] On the other hand, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one computer instruction, and the at least one computer instruction is loaded and executed by a processor to implement the query method as described in the above aspects.

[0015] On the other hand, an embodiment of the present application provides a computer program product, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium; the processor of the server reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the server implements the query method described in the above aspects.

[0016] In the related art, the query text input by the user is rewritten by preset rules or using a machine learning model to try to return search results that meet the user's needs. However, if the rewriting is inappropriate, the accuracy of the search results fed back by the search engine will be reduced.

[0017] In the embodiment of the present application, after determining the second query text corresponding to the first query text in the query request, the server does not return the second query result corresponding to the second query text, but returns the first query result corresponding to the first query text and the second query result corresponding to the second query text, thereby avoiding the difficulty for the user to obtain the required query content from the client due to inappropriate rewriting of the second query text, ensuring that the server can return the correct query result, and is conducive to improving the user's query experience. By checking the rewriting accuracy of the second query text, the server can determine whether the second query text is rewritten correctly, which is conducive to the server's subsequent adjustment of the query strategy, thereby improving the query accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a schematic diagram of an implementation environment of a query method provided by an exemplary embodiment of the present application;

[0019] Figure 2 is a flowchart of a query method provided by an exemplary embodiment of the present application;

[0020] Figure 3 is a schematic diagram of a client screen showing a first query result and a second query result provided by an exemplary embodiment of the present application;

[0021] Figure 4 is a flowchart of a query method provided by another exemplary embodiment of the present application;

[0022] Figure 5 is a flowchart of a process for determining display priority provided by an exemplary embodiment of the present application;

[0023] Figure 6 This is a schematic diagram of a client screen displaying a first query result and a second query result based on display priority provided by an exemplary embodiment of the present application;

[0024] Figure 7 is a flowchart of a query method provided by another exemplary embodiment of the present application;

[0025] Figure 8 is a flowchart of a song query method provided by an exemplary embodiment of the present application;

[0026] Fig. 9 is a structural block diagram of a query device provided by an exemplary embodiment of the present application;

[0027] Fig.10 It is a schematic diagram of the structure of a server provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0029] During the song search process, the query text entered by the user may contain spelling errors, inaccurate song names, vague singer names, etc., making it difficult for the search engine to return the search results required by the user.

[0030] The search engine rewrites the query text entered by the user according to preset rules or using a machine learning model, and then the server returns the search results determined based on the rewritten query text. However, if the rewriting is inappropriate, the accuracy of the search results fed back by the search engine will be reduced.

[0031] The following first describes the implementation environment of the query method provided in the embodiment of the present application.

[0032] See also Figure 1 , Figure 1 1 is a schematic diagram of an implementation environment of a query method provided by an exemplary embodiment of the present application. The implementation environment includes a terminal 110 and a server 120.

[0033] The terminal 110 is provided with a client application for content querying. The terminal 110 receives the query text entered by the user in the search box of the client application, and then sends a query request containing the query text to the server 120, so as to receive and display the query results fed back by the server 120 later.

[0034] The server 120 is used to provide query services. The server receives the query request sent by the terminal 110, searches for content that matches the query text according to the query text in the query request and the attribute information of the searched content, and then feeds back the query result to the terminal 110.

[0035] The following embodiments are described by taking the query method applied to a server as an example.

[0036] See also Figure 2 , Figure 2 The following is a flowchart of a query method provided by an exemplary embodiment of the present application. The method includes the following steps.

[0037] Step 201: upon receiving a query request from a client, determining a second query text corresponding to a first query text in the query request, wherein the second query text is rewritten based on the first query text.

[0038] In some embodiments, the first query text is the query text entered by the user at the client, that is, the first query text is the original query text without rewriting, which may contain spelling errors, inaccurate descriptions, ambiguous text content, etc., or may accurately describe the user's query requirements. The server extracts the first query text from the query request of the client.

[0039] In some embodiments, the second query text is determined by the server rewriting the first query text. Optionally, the server can rewrite the first query text by using a machine learning model, or rewrite the first query text according to a set rewriting rule. The embodiment of the present application does not limit the way in which the server rewrites the first query text.

[0040] Exemplarily, the set rewriting rules may include the server counting the total number of searches for the first query text entered by the user, and the number of searches for the third query text entered by the user again after entering the first query text, to obtain the secondary search rates corresponding to different third query texts, and based on the secondary search rates, determining the second query text from the different third query texts.

[0041] Step 202: Return a first query result corresponding to the first query text and a second query result corresponding to the second query text to the client, and the client is used to display the first query result and the second query result.

[0042] Optionally, the first query result and the second query result can be determined based on the semantic relevance between the query text and the query content, or based on the text matching between the query text and the query content, or based on the association relationship between the query text and the query content. The embodiment of the present application does not limit the manner in which the server determines the first query result and the second query result. The first query result is determined based on the first query text query, and the second query result is determined based on the second query text query.

[0043] In the related art, after determining the second query text, the server feeds back the second query result corresponding to the second query text to the client, and when the user chooses to search by the first query text, the server feeds back the first query result corresponding to the first query text to the client. In the embodiment of the present application, the first query result and the second query result are returned to the client together.

[0044] Optionally, the number of first query results and second query results returned by the client may be the same or different. Optionally, the number of first query results may be greater than the number of second query results to avoid rewriting errors causing the client to display too many second query results that do not meet the user's query requirements, resulting in a reduced user query experience. Exemplarily, the server feeds back a single second query result and at least two first query results to the client.

[0045] In some possible display scenarios, after receiving the first query result and the second query result fed back by the server, the client arranges and displays the first query result and the second query result on the user interface for the user to view.

[0046] Indicative, such as Figure 3 As shown, the first query text 300 corresponds to 9 first query results 301, and the second query text corresponds to 1 second query result 302. The second query result 302 and the first query result 301 are arranged and displayed on the client user interface.

[0047] Step 203, based on the first conversion data of the first query result and the second conversion data of the second query result, verify the rewriting accuracy of the second query text to obtain the rewriting accuracy verification result of the second query text, the first conversion data is used to characterize the conversion status of the first query result at the client, and the second conversion data is used to characterize the conversion status of the second query result at the client.

[0048] In some embodiments, conversion data refers to data generated by conversion operations such as clicking, collecting, downloading, and forwarding by the user on the query results displayed by the client after the user enters the first query text. Optionally, conversion data may refer to different types of conversion numbers, such as the user's click volume, collection volume, download volume, forwarding volume, etc. Optionally, conversion data may also include different types of conversion rates, such as the user's click rate, collection rate, download rate, forwarding rate, etc. The conversion data may be determined by the server through data statistics. Among them, the conversion rate of the query result = the number of conversions of the query result ÷ the number of exposures of the query result. Among them, the number of exposures of the query result is equivalent to the number of queries of the first query text.

[0049] Optionally, the conversion data may also include data determined by weighted calculation of different types of conversion data. For example, the server uses the number of user clicks and favorites of a single query result as a conversion data item after weighted calculation.

[0050] Among them, the rewriting accuracy test results include accurate rewriting and inaccurate rewriting.

[0051] Optionally, the server verifies the rewriting accuracy of the second query text by comparing the first conversion data of the first query result and the second conversion data of the second query result.

[0052] Optionally, the server determines an indicator that can be used to measure the rewriting accuracy of the second query text according to the first conversion data of the first query result and the second conversion data of the second query result, thereby verifying the rewriting accuracy of the second query text.

[0053] In the embodiment of the present application, after determining the second query text corresponding to the first query text in the query request, the server does not return the second query result corresponding to the second query text, but returns the first query result corresponding to the first query text and the second query result corresponding to the second query text, thereby avoiding the difficulty for the user to obtain the required query content from the client due to inappropriate rewriting of the second query text, ensuring that the server can return the correct query result, and is conducive to improving the user's query experience. By checking the rewriting accuracy of the second query text, the server can determine whether the second query text is rewritten correctly, which is conducive to the server's subsequent adjustment of the query strategy, thereby improving the query accuracy.

[0054] In some embodiments, the server verifies the rewriting accuracy of the second query text by comparing the first conversion data with the second conversion data, or by comparing the actual conversion index determined based on the first conversion data and the second conversion data. The following is a detailed description of the manner in which the server verifies the second query text.

[0055] Mode 1: Determine a rewriting accuracy check result of the second query text based on a conversion data comparison result of the first conversion data of the first query result and the second conversion data of the second query result.

[0056] In some embodiments, the conversion data comparison result refers to a comparison result between the first conversion data and the second conversion data. The rewriting accuracy verification result includes that the second query text is rewritten accurately and the second query text is rewritten inaccurately.

[0057] In a possible implementation, the server obtains a conversion data comparison result by comparing the same data item in the first conversion data and the second conversion data, and determines the rewriting accuracy test result of the second query text. For example, the server can compare the number of clicks corresponding to the first query result with the number of clicks corresponding to the second query result. For another example, the server can compare the download rate corresponding to the first query result with the download rate corresponding to the second query result.

[0058] In another possible implementation, the server determines the rewriting accuracy test result of the second query text by comparing the comprehensive index of the first conversion data and the comprehensive index of the second conversion data. For example, when the conversion data includes clicks, favorites, and downloads, the server can compare the comprehensive index obtained by weighted average of the clicks, favorites, and downloads.

[0059] Optionally, the rewriting accuracy check result of the second query text may be determined in the following two situations:

[0060] Case 1: when the second conversion data is higher than or equal to the first conversion data, it is determined that the rewriting accuracy test result of the second query text is accurate in rewriting.

[0061] Case 2: when the second conversion data is lower than the first conversion data, it is determined that the rewriting accuracy test result of the second query text is inaccurate rewriting.

[0062] When the second conversion data is higher than the first conversion data, the second query result corresponding to the second query text is more in line with the user's needs than the first query result corresponding to the first query text. Since the query effect of the second query text is better than the first query text, the use of the second query text can improve the query accuracy. The server determines that the rewriting accuracy test result of the second query text is rewriting accurate.

[0063] When the second conversion data is equal to the first conversion data, the conversion situation of the second query result is the same as that of the first query result. It can be seen that the query effect of the second query text has reached the query effect of the first query text. Whether to use the second query text has little effect on the query accuracy. The server determines that the rewriting accuracy test result of the second query text is accurate.

[0064] When the second conversion data is lower than the first conversion data, the conversion status of the second query result is worse than that of the first query result, the query effect of the second query text fails to reach the query effect of the first query text, and the second query text cannot replace the first query text. Otherwise, the query accuracy rate will be reduced, and the server determines that the rewriting accuracy check result of the second query text is inaccurate.

[0065] In an illustrative example, the server returns query result A and query result B to the client, wherein query result A is the first query result and query result B is the second query result. The conversion rate of query result A is 8%. When the conversion rate of query result B is 5%, the rewriting accuracy test result of the second query text is inaccurate rewriting; when the conversion rate of query result B is 10%, the rewriting accuracy test result of the second query text is accurate rewriting.

[0066] In another exemplary example, the server returns query results A, B, and C to the client, wherein query results A and B are the first query results, and query result C is the second query result. The conversion rate of query result A is 10%, the conversion rate of query result B is 20%, and the conversion rate of query result C is 30%. The conversion rate of the first query result can be the average value (average conversion rate) of the conversion rate of query result A and the conversion rate of query result B, i.e., 15%, which is less than the conversion rate of the second query result. The server determines that the rewriting accuracy test result of the second query text is accurate in rewriting; conversely, when the average conversion rate of each query result in the first query result is greater than the conversion rate of the second query result, the rewriting of the second query text is inaccurate.

[0067] In each of the above exemplary examples, the server may also determine the rewriting accuracy test result of the second query text by comparing the conversion data. Since the principle is similar to the conversion rate, it will not be described in detail in the embodiments of the present application.

[0068] Method 2: determining an actual conversion index of the second query result based on the first conversion data of the first query result and the second conversion data of the second query result; and determining a rewriting accuracy test result of the second query text based on the actual conversion index and the conversion index threshold.

[0069] In some embodiments, each item of data in the conversion data is determined for a single query result. The conversion data of different first query results is different, and the conversion data of different second query results is also different, and the number of first query results and the number of second query results may also be different. In order to measure the actual query effect of the first query result and the second query result returned together, the server combines the first conversion data of the first query result and the second conversion data of the second query result to determine the actual conversion index.

[0070] The actual conversion index is used to measure the actual query effect of the second query result. Unlike method 1, the actual conversion index of the second query text is not used to compare with the first query text, but is compared with the conversion index threshold. The accuracy test result determined by method 2 is no longer affected by the first query result. When the query effect of the first query result is poor, the server can still accurately determine the rewriting accuracy.

[0071] For example, in mode 1, when the query effect of the first query result is poor, even if the second conversion data is higher than or equal to the first conversion data, the query effect of the second query result may still be poor. For example, the conversion rate of the first query result is 5%, and the conversion rate of the second query result is 8%. When the conversion rate is lower than 10%, the query effect is poor.

[0072] Optionally, the actual conversion index may include a real-time conversion rate index and a real-time conversion ratio index. The real-time conversion ratio index refers to the ratio of conversion data to total conversion data. For example, if the number of conversions of query result A is 100, the number of conversions of query result B is 200, and the number of conversions of query result C is 300, then the total number of conversions is 600, the conversion ratio of query result A is 16%, the conversion ratio of query result B is 33%, and the conversion ratio of query result C is 50%.

[0073] The following describes in detail the process of determining the rewriting accuracy test result of the second query text by taking the actual conversion indicators as the real-time conversion proportion indicator and the real-time conversion rate indicator as examples.

[0074] In some embodiments, the server determines a real-time conversion rate indicator and a real-time conversion proportion indicator of the second query result based on the first conversion data and the second conversion data. The real-time conversion proportion indicator is used to characterize the proportion of the second conversion data to the total conversion data. The total conversion data is the sum of the first conversion data and the second conversion data.

[0075] The real-time conversion rate index and the real-time conversion ratio index are real-time changing indicators. Optionally, the server may determine the real-time conversion rate index and the real-time conversion ratio index of the second query result based on the first conversion data and the second conversion data between the historical time point and the current time point. The historical time may be a fixed time point or a historical time point with a fixed time interval from the current time point.

[0076] The real-time conversion rate indicator is used to represent the ratio of the number of conversions of the second query result to the number of exposures of the second query result.

[0077] Taking the conversion number as the click volume as an example, at the current time of 18:00, in the historical time period from 17:30 to 18:00, the first query result includes query result A and query result B, where the click volume of query result A is 1000, and the click volume of query result B is 2000. The second query result includes query result C and query result D, where the click volume of query result C is 1000 and the number of exposures is 2000, and the click volume of query result D is 1000 and the number of exposures is 4000. The real-time conversion rate index of query result C is 50%, and the real-time conversion rate index of query result D is 25%. The total click volume of the first query result and the second query result is 5000, the real-time conversion ratio index of query result C is 20%, and the real-time conversion ratio index of query result D is 20%.

[0078] Among them, the real-time conversion rate indicator and the real-time conversion ratio indicator have the following two situations:

[0079] In case 1, when the real-time conversion rate index of the second query result is greater than or equal to the conversion rate index threshold, and the real-time conversion share index of the second query result is greater than or equal to the conversion share index threshold, the rewriting accuracy test result of the second query text is determined to be accurate rewriting.

[0080] In case 2, when the real-time conversion rate index of the second query result is less than the conversion rate index threshold, and / or the real-time conversion share index of the second query result is less than the conversion share index threshold, the rewriting accuracy test result of the second query text is determined to be inaccurate rewriting.

[0081] In some embodiments, the conversion rate indicator threshold is used to measure whether the real-time conversion rate indicator of the second query result meets the conditions, and the conversion share indicator threshold is used to measure whether the real-time conversion share indicator of the second query result meets the conditions.

[0082] In order to ensure the accuracy of the rewriting accuracy test results, the server determines that the rewriting accuracy test result of the second query text is an accurate rewrite when both the real-time conversion rate indicator and the real-time conversion ratio indicator meet the conditions; if at least one of the real-time conversion rate indicator and the real-time conversion ratio indicator does not meet the conditions, the server determines that the rewriting accuracy test result of the second query text is an inaccurate rewrite.

[0083] Optionally, the second query result may include a single query result or at least two query results. When it includes at least two query results, the server determines the rewriting accuracy test result of the second query text based on the real-time conversion rate indicators and the real-time conversion share indicators of the at least two query results, for example, based on the average of the actual conversion rate indicators and the average of the actual conversion share indicators in the at least two query results.

[0084] The following is an example in which the conversion rate indicator threshold is 10% and the conversion proportion indicator threshold is 10%.

[0085] In an illustrative example, the second query result includes a single query result. The real-time conversion rate index of the second query result is 10%, and the real-time conversion share index is 15%. The server determines that the conversion rate index threshold and the conversion share index threshold are reached, and the rewriting accuracy test result of the second query text is that the rewriting is accurate.

[0086] In another illustrative example, the second query result includes query result C, query result D, and query result E. Among them, the real-time conversion rate index of query result C is 10%, the real-time conversion rate index of query result D is 15%, and the real-time conversion rate index of query result E is 20%. The server determines that the average conversion rate of the second query result (the average of the actual conversion rate index) is 15%, reaching the conversion rate index threshold. The real-time conversion share index of query result C is 5%, the real-time conversion share index of query result D is 10%, and the real-time conversion share index of query result E is 15%. The server determines that the average conversion share (the average of the actual conversion share index) of the second query result is 10%, reaching the conversion share index threshold. The server determines that the rewriting accuracy test result of the second query text is accurate rewriting.

[0087] In case 2, when the conversion rate index threshold is less than 10% and / or the conversion ratio index threshold is less than 10%, the server determines that the rewriting accuracy test result of the second query text is inaccurate rewriting. This example is not repeated here.

[0088] Optionally, in addition to determining the average conversion rate and the average conversion ratio, the server may also measure the conversion status of at least two second query results in other ways. For example, the server determines the number of results of the second query results whose real-time conversion rate index is greater than or equal to the conversion rate index threshold, and whose real-time conversion ratio index is greater than or equal to the conversion ratio index threshold. If the number of results reaches the number threshold, the rewriting accuracy test result of the second query text is determined to be accurate rewriting; otherwise, the rewriting accuracy test result of the second query text is determined to be inaccurate rewriting.

[0089] Optionally, the server may also determine the rewriting accuracy test result of the second query text when the conversion data comparison results and the real-time conversion indicators in the above two methods meet the conditions at the same time, which is helpful to improve the accuracy of the rewriting accuracy test result.

[0090] In an embodiment of the present application, the server determines the rewriting accuracy test result of the second query text based on the conversion data comparison result and / or the real-time conversion index, enriches the test method of the second query text, and is conducive to improving the server's test capability for the second query text.

[0091] Among them, the rewriting accuracy is determined by comparing the conversion data of the first query result and the second query result, so that when the first conversion data and the second conversion data are difficult to represent the actual conversion situation due to problems such as too little data, the server can still determine the rewriting accuracy of the second query text, thereby improving the ability to determine the rewriting accuracy test results.

[0092] By measuring the real-time conversion rate indicator and the real-time conversion ratio indicator of the second query result, the server can determine the actual rewriting accuracy of the second query text. For example, when the amount of conversion data is sufficient, the query effect of the second query result is more realistic, which is conducive to improving the accuracy of the rewriting accuracy test results. In addition, when the real-time conversion rate indicator and the real-time conversion ratio indicator both meet the conditions, the server determines that the rewriting accuracy test result is accurate, which further improves the accuracy of the rewriting accuracy test results.

[0093] In some embodiments, the display position of the second query result has an impact on the verification accuracy and user experience. On the one hand, if the display position of the second query result is too far back, the user may not pay attention to the second query result. Even if the second query result meets the user's needs, the conversion data of the second query result will be reduced, resulting in a decrease in the accuracy of the rewriting accuracy test result determined by the server. On the other hand, if the display position of the second query result is too far forward and the rewriting of the second query text is inaccurate, the inaccurate second query result will occupy the front display position, making it difficult for the user to obtain valid query results in the first time, resulting in a decrease in user experience.

[0094] In order to improve verification accuracy and user experience, the server needs to determine a suitable display position for the second query result.

[0095] See also Figure 4 , Figure 4 FIG. 1 is a flowchart of a query method provided by another exemplary embodiment of the present application. The method includes the following steps.

[0096] Step 401: upon receiving a query request from a client, determining a second query text corresponding to a first query text in the query request, wherein the second query text is rewritten based on the first query text.

[0097] Among them, the specific description of step 401 can refer to step 201, and will not be repeated in the embodiment of the present application.

[0098] Step 402: Determine a display priority of a second query result corresponding to the second query text. The display priority is used to represent a display order of the second query result when it is displayed at the client.

[0099] Optionally, the display priority may be a discrete value, for example, a display priority with 10 display levels. When the second query result is a single query result, the display priority may characterize the display position of the second query result when it is displayed on the client. When the second query result is at least two query results, the display priority of each second query result may be used to determine the display position of each second query result when it is displayed on the client.

[0100] Optionally, the number of display priorities can be the same as the number of query results displayed on the first screen of the user interface. For example, if 10 query results are displayed on the first screen of the user interface, the 10 levels of display priority correspond to 10 query result display positions. The first screen display refers to displaying a fixed number of query results after the user triggers the query of the first query text, and the client further displays other query results when receiving an operation to increase the display of query results.

[0101] Optionally, the display priority may be a continuous value, for example, the display priority value is between 0 and 1. In a possible implementation, the server may determine the display order of the first query result and the second query result by comparing the display priority of the first query result and the display priority of the second query result.

[0102] Optionally, the server may determine the display priority of the second query result according to the correlation between the first query result and the second query result. For example, when the first query result is strongly correlated with the second query result, the display priority of the second query result is high, that is, the display position of the second query result is ranked higher.

[0103] Optionally, the display priority may also be determined based on the text relevance between the second query text and the first query text, such as Figure 5 As shown, the following steps are included.

[0104] Step 402A: determine the text relevance between the second query text and the first query text.

[0105] Optionally, the text relevance may include at least one of a semantic relevance and a text matching degree between the first query text and the second query text.

[0106] In a possible implementation, the server vectorizes the first query text and the second query text, and determines the text correlation between the second query text and the first query text according to the vector similarity between the first query text and the second query text.

[0107] In another possible implementation, the server determines the text relevance between the second query text and the first query text by counting the text repetition rate between the first query text and the second query text, or the letter matching degree between the spelling letters corresponding to the first query text and the spelling letters corresponding to the second query text.

[0108] Optionally, the text relevance between the second query text and the first query text may also be determined based on a secondary search rate or a text content relationship.

[0109] Among them, the secondary search rate refers to the probability that the user will search for the second query text for the second time after entering the first query text. The secondary search rate is positively correlated with the text relevance. For example, there are 100 queries for the first query text in a single day, and within 3 minutes after 20 queries for the first query text, the user searches for the second query text. The server determines that the secondary search rate is 20%.

[0110] The text content relationship refers to the relationship between the content represented by the first query text and the content represented by the second query text. For example, when the first query text and the second query text are used to describe the same content, or the content represented by the first query text and the content represented by the second query text belong to the same content set (corresponding to the same content object), the text relevance is a strong correlation.

[0111] Step 402B: determine the display priority of the second query result based on the text relevance, and the display priority is positively correlated with the text relevance.

[0112] In a possible implementation, the server stores a correspondence between text relevance and display priority, and determines the display priority of the second query result according to the correspondence and the text relevance.

[0113] Exemplarily, when the text relevance is greater than the first relevance threshold, the display priority is determined to be the first display level, when the text relevance is less than the first relevance threshold and greater than the second relevance threshold, the display priority is determined to be the second display priority, and when the text relevance is less than the second relevance threshold, the display priority is determined to be the third display priority. Wherein, the first relevance threshold is greater than the second relevance threshold, the first display priority is higher than the second display priority, and the second display priority is higher than the third display priority.

[0114] Step 403: Return the first query result corresponding to the first query text, the second query result corresponding to the second query text, and the display priority of the second query result to the client.

[0115] In a possible implementation, the client inserts the second query result between different first query results according to the display priority of the second query result.

[0116] In an illustrative example, Figure 6 As shown, after the user inputs the first query text 600 and triggers the query, the client first displays 10 query results based on the first query result, the second query result, and the display priority of the second query result returned by the server. When the user triggers "more content", the client further displays other query results. The display priority includes 10 display levels, and each display level corresponds to one of the 10 query result display positions. When the display priority of the second query result 601 is level 3, the client displays the second query result 601 as the third query result; when the display priority of the second query result 602 is level 5, the client displays the second query result 602 as the fifth query result; when the display priority of the second query result 603 is level 8, the client displays the second query result 603 as the eighth query result.

[0117] Optionally, the server may also first determine the display order of the first query result and the second query result based on the first query result corresponding to the query text, the second query result corresponding to the second query text, and the display priority of the second query result. The display order can be represented by the display serial number of the query result, and then return the query result and the display serial number corresponding to the query result to the client.

[0118] Step 404, based on the first conversion data of the first query result and the second conversion data of the second query result, verify the rewriting accuracy of the second query text to obtain the rewriting accuracy verification result of the second query text, the first conversion data is used to characterize the conversion status of the first query result at the client, and the second conversion data is used to characterize the conversion status of the second query result at the client.

[0119] Among them, the specific description of step 404 can refer to the above embodiments, and the embodiments of the present application will not be repeated here.

[0120] In the above-mentioned application embodiment, the server determines the display priority of the second query result, and returns the display priority together with the first query result and the second query result to the client, so that the client can determine the appropriate display order for the second query result, thereby ensuring that the second query result can be fully exposed to improve the inspection efficiency, and avoiding giving priority to the display of inaccurate second query results, which may cause user disgust.

[0121] Among them, the display priority can be determined according to the text relevance between the second query text and the first query text, which is conducive to realizing the fusion sorting of the first query result and the second query result, so that all query results can be sorted according to the degree of compliance with user needs, which is conducive to improving query efficiency.

[0122] In some embodiments, the server may count the number of queries of the user for the first query text to measure the reliability of the first conversion data and the second conversion data. For example, when the number of queries is small, the reliability of the first conversion data and the second conversion data is low, and when the number of queries is high, the reliability of the first conversion data and the second conversion data is high.

[0123] See also Figure 7 , Figure 7 It is a flowchart of a query method provided by another exemplary embodiment of the present application.

[0124] Step 701: upon receiving a query request from a client, increase the query count corresponding to the first query text included in the query request by one.

[0125] In some embodiments, the query number refers to the total number of times different users input the first query text on the client and trigger the query. The query number is the same as the exposure number of the query result.

[0126] Optionally, there are two stages for the client to display the second query result corresponding to the second query text, namely, an application stage and a verification stage. In the verification stage, the client displays the first query result and the second query result, and the server verifies the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result; in the application stage, the server determines whether to return the second query result corresponding to the second query text to the client based on the rewriting accuracy verification result.

[0127] Optionally, during the verification phase, the server counts query requests from clients on different terminal devices, and upon receiving a query request from each client, increases the query times corresponding to the first query text included in the query request by one, wherein different first query texts correspond to different query times.

[0128] Step 702, when the number of queries corresponding to the first query text reaches the number threshold, based on the first conversion data of the first query result and the second conversion data of the second query result, check the rewriting accuracy of the second query text to obtain the rewriting accuracy check result of the second query text.

[0129] When the number of queries corresponding to the first query text reaches the number threshold, the server determines the rewriting accuracy verification result and ends the verification phase.

[0130] Since the first conversion data of the first query result and the second conversion data of the second query result are determined based on the number of queries, when the number of queries corresponding to the first query text reaches the number threshold, the data reliability of the first conversion data and the second conversion data is relatively high. When the number of queries corresponding to the first query text reaches the number threshold, the server verifies the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result, thereby improving the accuracy of the rewriting accuracy verification results.

[0131] Step 703, when the number of queries corresponding to the first query text does not reach the number threshold, the first n query results with the highest relevance to the second query text in all query results of the second query text are determined as the second query results corresponding to the second query text, where n is a positive integer.

[0132] In some embodiments, a number threshold is used to determine the timing at which the server starts to verify the rewriting accuracy of the second query text, wherein when the number of queries corresponding to the first query text reaches the number threshold, the server starts to verify the rewriting accuracy of the second query text, and when the number of queries corresponding to the first query text does not reach the number threshold, the server continues to collect first conversion data of the first query text and second conversion data of the second query text.

[0133] In some embodiments, the full program results refer to all query results related to the second query text. The server selects the top n query results with the highest relevance to the second query text from all query results related to the second query text and returns them to the client, and returns the first query result to the client, thereby collecting the first conversion data and the second conversion data.

[0134] Optionally, when the rewriting accuracy check result indicates that the second query text is rewritten accurately, the server determines first m query results with the highest relevance to the second query text among all query results of the second query text as second query results corresponding to the second query text, where m is an integer greater than n.

[0135] In the case where the rewriting accuracy test result indicates that the second query text is accurately rewritten, the server determines that the second query text can be applied to the subsequent query process of the first query text, that is, the verification phase of the second query text ends and the application phase begins. By increasing the number of second query results displayed in the application phase, the query accuracy can be improved.

[0136] In the embodiment of the present application, the server specifies the timing for starting verification of the second query text by setting a threshold condition for the number of queries, which is conducive to fully collecting conversion data and improving verification accuracy.

[0137] Wherein, when the number of queries does not reach the number threshold, the server returns a small number of second query results that are most relevant to the second query text to the client, so as to be used to subsequently verify the accuracy of the rewriting of the second query text, and avoid returning too many inappropriate second query results, which leads to a decrease in query accuracy. When the second query text is rewritten correctly, the number of second query results returned is increased to improve the query accuracy.

[0138] Optionally, in the technical solutions provided in the above embodiments, after the server verifies the rewriting accuracy of the second query text and obtains the rewriting accuracy verification result of the second query text, the server can use the rewriting accuracy verification result to further improve the query accuracy.

[0139] Considering that the server may rewrite the first query text according to a preset rule or may use a machine learning model to rewrite the first query text, the server may take the following two measures to improve query efficiency after determining the rewriting accuracy test result:

[0140] Measure 1: when the rewriting accuracy check result indicates that the rewriting of the second query text is inaccurate, stop determining the second query text as the rewritten text of the first query text.

[0141] In one possible implementation, the server first determines at least two second query texts according to a preset rule or a machine learning model, and then checks the rewriting accuracy of the at least two second query texts, and screens the accurate second query text by stopping determining the inaccurately rewritten second query text as the rewritten text of the first query text.

[0142] In another possible implementation, the server stops determining the second query text as the rewritten text of the first query text, and then returns a first query result corresponding to the first query text to the client upon receiving a query request including the first query text.

[0143] Measure 2: When the rewriting accuracy test result indicates that the rewriting of the second query text is inaccurate, the first query text and the second query text are determined as a negative sample pair, and the negative sample pair is used to train a text rewriting model, and the text rewriting model is used to output a rewritten text according to the input text.

[0144] In some embodiments, the text rewriting model is used to output a second query text based on an input first query text. A negative sample pair refers to a first query text and a second query text that is an incorrect rewrite of the first query text, and a positive sample pair refers to a first query text and a second query text that is a correct rewrite of the first query text. Both the positive sample pair and the negative sample pair can be used to train the text rewriting model.

[0145] In a possible application scenario, the server inputs the first query text into the text rewriting model to obtain the second query text, and then verifies the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result. If the rewriting of the second query text is accurate, in the subsequent query process of the first query text, the second query result corresponding to the second query text is added and displayed. If the rewriting of the second query text is inaccurate, the first query text and the second query text are used as negative sample pairs for the subsequent training process of the text rewriting model to further improve the rewriting accuracy of the second query text output by the model.

[0146] In an embodiment of the present application, when the second query text is rewritten inaccurately, the server can stop using the second query text to reduce the recall of query results that do not meet user needs and improve the accuracy of the query results, or determine the first query text and the second query text as a negative sample pair for training a text rewriting model to improve the accuracy of text rewriting, thereby improving the accuracy of the query results.

[0147] The following embodiments are described by applying the query method to a song query. When the query request is a song query request, the first query result and the second query result are recalled songs, and the first conversion data and the second conversion data are determined based on at least one of the playback data, download data, and collection data of the recalled songs. Exemplarily, the playback data may include playback volume and playback rate, the download data may include download volume and download rate, and the collection data may include collection volume and collection rate.

[0148] See also Figure 8 , Figure 8 1 is a flowchart of a song query method provided by an exemplary embodiment of the present application. The method includes the following steps.

[0149] Step 801, upon receiving a song query request from a song client, determining a second song query text corresponding to a first song query text in the song query request, wherein the second song query text is rewritten based on the first song query text.

[0150] Step 802, returning the first recalled song corresponding to the first song query text and the second recalled song corresponding to the second song query text to the song client, and the song client is used to display the first recalled song and the second recalled song.

[0151] Optionally, the server determines the song display priority of the second recalled song corresponding to the second song query text, and the song display priority is used to characterize the display order of the second recalled song when it is displayed at the song client; then, the server returns the first recalled song corresponding to the first song query text, the second recalled song corresponding to the second song query text, and the song display priority of the second recalled song to the song client.

[0152] Optionally, in the process of the server determining the display priority of the second recalled song corresponding to the second song query text, the server first determines the text correlation between the second song query text and the first song query text; then, based on the text correlation, the server determines the song display priority of the second recalled song, and the display priority is positively correlated with the text correlation.

[0153] Step 803, based on the first song conversion data of the first recalled song and the second song conversion data of the second recalled song, check the rewriting accuracy of the second song query text to obtain the rewriting accuracy check result of the second song query text, the first song conversion data is used to characterize the conversion status of the first recalled song at the song client, and the second song conversion data is used to characterize the conversion status of the second recalled song at the song client.

[0154] Optionally, the verification method of the rewriting accuracy of the second song query text includes at least one of the following two methods:

[0155] Method 1, based on the conversion data comparison result of the first song conversion data of the first recalled song and the second song conversion data of the second recalled song, determining the rewriting accuracy test result of the second song query text.

[0156] Among them, when the second song conversion data is higher than or equal to the first song conversion data, the server determines that the rewriting accuracy test result of the second song query text is accurate rewriting; when the second song conversion data is lower than the first song conversion data, the server determines that the rewriting accuracy test result of the second song query text is inaccurate rewriting.

[0157] Method 2, based on the first song conversion data of the first recalled song and the second song conversion data of the second recalled song, determine the actual conversion index of the second recalled song; based on the actual conversion index and the conversion index threshold, determine the rewriting accuracy test result of the second song query text.

[0158] Among them, the server first determines the real-time conversion rate indicator and the real-time conversion proportion indicator of the second recalled song based on the first song conversion data and the second song conversion data. The real-time conversion proportion indicator is used to characterize the proportion of the second song conversion data to the total song conversion data. The total song conversion data is the sum of the first song conversion data and the second song conversion data.

[0159] Then, when the real-time conversion rate index of the second recalled song is greater than or equal to the conversion rate index threshold, and the real-time conversion share index of the second recalled song is greater than or equal to the conversion share index threshold, the server determines that the rewriting accuracy test result of the second song query text is accurate rewriting; when the real-time conversion rate index of the second recalled song is less than the conversion rate index threshold, and / or the real-time conversion share index of the second recalled song is less than the conversion share index threshold, the server determines that the rewriting accuracy test result of the second song query text is inaccurate rewriting.

[0160] Optionally, when the server receives a song query request from a song client, the server may increase the number of queries corresponding to the first song query text contained in the song query request by one; then, when the number of queries corresponding to the first song query text reaches a number threshold, based on the first song conversion data of the first recalled song and the second song conversion data of the second recalled song, the server may check the rewriting accuracy of the second song query text to obtain a rewriting accuracy check result of the second song query text.

[0161] Optionally, when the number of queries corresponding to the first song query text does not reach the number threshold, the server may also determine the first n query results with the highest relevance to the second song query text among all query results of the second song query text as the second recalled songs corresponding to the second song query text, where n is a positive integer.

[0162] Optionally, when the rewriting accuracy check result indicates that the second song query text is rewritten accurately, the server determines the first m query results with the highest relevance to the second song query text among all query results of the second song query text as the second recalled songs corresponding to the second song query text, where m is an integer greater than n.

[0163] Optionally, when the rewriting accuracy check result indicates that the rewriting of the second song query text is inaccurate, the server may take at least one of the following measures:

[0164] Measure 1: when the rewriting accuracy test result indicates that the rewriting of the second song query text is inaccurate, stop determining the second song query text as the rewritten text of the first song query text.

[0165] Measure 2, when the rewriting accuracy test result indicates that the rewriting of the second song query text is inaccurate, the first song query text and the second song query text are determined as a negative sample pair, and the negative sample pair is used to train a text rewriting model, and the text rewriting model is used to output a rewritten text according to the input text.

[0166] In the embodiment of the present application, after determining the second song query text corresponding to the first song query text in the song query request, the server does not return the second recalled song corresponding to the second song query text, but returns the first recalled song corresponding to the first song query text and the second recalled song corresponding to the second song query text, thereby avoiding the difficulty for the user to obtain the required song from the client due to the inappropriate second recalled song corresponding to the second song query text, ensuring that the server can return the correct song query result, and is conducive to improving the user's query experience and satisfaction. By checking the correctness of the rewriting of the second song query text, the server can determine whether the second song query text is rewritten correctly, which is conducive to the server's subsequent adjustment of the query strategy, thereby improving the accuracy of the song query.

[0167] See also Fig. 9 , Fig. 9 This is a structural block diagram of a query device provided by an exemplary embodiment of the present application. The device includes the following modules.

[0168] A text determination module 901 is used to determine, upon receiving a query request from a client, a second query text corresponding to a first query text in the query request, where the second query text is rewritten based on the first query text;

[0169] A return module 902 is used to return a first query result corresponding to the first query text and a second query result corresponding to the second query text to the client, and the client is used to display the first query result and the second query result;

[0170] Verification module 903 is used to verify the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result, and obtain the rewriting accuracy verification result of the second query text, wherein the first conversion data is used to characterize the conversion status of the first query result at the client, and the second conversion data is used to characterize the conversion status of the second query result at the client.

[0171] Optionally, the inspection module 903 is further used for at least one of the following:

[0172] determining the rewriting accuracy inspection result of the second query text based on a conversion data comparison result of the first conversion data of the first query result and the second conversion data of the second query result;

[0173] Based on the first conversion data of the first query result and the second conversion data of the second query result, an actual conversion index of the second query result is determined; based on the actual conversion index and a conversion index threshold, the rewriting accuracy test result of the second query text is determined.

[0174] Optionally, the inspection module 903 is further used to:

[0175] When the second conversion data is higher than or equal to the first conversion data, determining that the rewriting accuracy test result of the second query text is that the rewriting is accurate;

[0176] When the second conversion data is lower than the first conversion data, it is determined that the rewriting accuracy check result of the second query text is inaccurate rewriting.

[0177] Optionally, the inspection module 903 is further used to:

[0178] Based on the first conversion data and the second conversion data, determine a real-time conversion rate index and a real-time conversion ratio index of the second query result, where the real-time conversion ratio index is used to represent the ratio of the second conversion data to the total conversion data, and the total conversion data is the sum of the first conversion data and the second conversion data;

[0179] When the real-time conversion rate index of the second query result is greater than or equal to the conversion rate index threshold, and the real-time conversion ratio index of the second query result is greater than or equal to the conversion ratio index threshold, determining that the rewriting accuracy test result of the second query text is rewriting accurate;

[0180] When the real-time conversion rate index of the second query result is less than the conversion rate index threshold, and / or the real-time conversion share index of the second query result is less than the conversion share index threshold, it is determined that the rewriting accuracy test result of the second query text is inaccurate rewriting.

[0181] Optionally, the returning module 902 is further configured to:

[0182] Determine a display priority of the second query result corresponding to the second query text, where the display priority is used to represent a display order of the second query result when it is displayed on the client;

[0183] The first query result corresponding to the first query text, the second query result corresponding to the second query text, and the display priority of the second query result are returned to the client.

[0184] Optionally, the returning module 902 is further configured to:

[0185] Determining a text relevance between the second query text and the first query text;

[0186] The display priority of the second query result is determined based on the text relevance, and the display priority is positively correlated with the text relevance.

[0187] Optionally, the device further includes a query times increasing module, which is used to:

[0188] When receiving the query request from the client, increasing the query count corresponding to the first query text included in the query request by one;

[0189] Optionally, the inspection module 903 is further used to:

[0190] When the number of queries corresponding to the first query text reaches a threshold number, the rewriting accuracy of the second query text is verified based on the first conversion data of the first query result and the second conversion data of the second query result to obtain the rewriting accuracy verification result of the second query text.

[0191] Optionally, the device further includes a query result determination module, configured to:

[0192] When the number of queries corresponding to the first query text does not reach the number threshold, the first n query results with the highest relevance to the second query text in all query results of the second query text are determined as the second query results corresponding to the second query text, where n is a positive integer.

[0193] Optionally, the inspection module 903 is further used to:

[0194] When the rewriting accuracy check result indicates that the second query text is rewritten accurately, the first m query results with the highest relevance to the second query text among the full query results of the second query text are determined as the second query results corresponding to the second query text, where m is an integer greater than n.

[0195] Optionally, the device further includes at least one of the following:

[0196] Stop module for:

[0197] If the rewriting accuracy check result indicates that the rewriting of the second query text is inaccurate, stop determining the second query text as the rewritten text of the first query text;

[0198] Negative sample determination module, used to:

[0199] In the case where the rewriting accuracy check result indicates that the rewriting of the second query text is inaccurate, the first query text and the second query text are determined as a negative sample pair, and the negative sample pair is used to train a text rewriting model, and the text rewriting model is used to output a rewritten text according to the input text.

[0200] Optionally, when the query request is a song query request, the first query result and the second query result are recalled songs, and the first conversion data and the second conversion data are determined based on at least one of the playback data, download data and collection data of the recalled songs.

[0201] In the embodiment of the present application, after determining the second query text corresponding to the first query text in the query request, the server does not return the second query result corresponding to the second query text, but returns the first query result corresponding to the first query text and the second query result corresponding to the second query text, thereby avoiding the difficulty for the user to obtain the required query content from the client due to the inappropriate second query result corresponding to the second query text, ensuring that the server can return the correct query result, and is conducive to improving the user's query experience and satisfaction. By checking the correctness of the rewriting of the second query text, the server can determine whether the second query text is rewritten correctly, which is conducive to the server's subsequent adjustment of the query strategy, thereby improving the query accuracy.

[0202] See also Fig.10 , Fig.10 1 is a schematic diagram of the structure of a server provided by an exemplary embodiment of the present application. Specifically, the server 1000 includes a central processing unit (CPU) 1001, a system memory 1004 including a random access memory 1002 and a read-only memory 1003, and a system bus 1005 connecting the system memory 1004 and the central processing unit 1001. The server 1000 may also include a basic input / output system (I / O system) 1006 that helps transmit information between various devices in the computer, and a large-capacity storage device 1007 for storing an operating system 1013, an application program 1014, and other program modules 1015, wherein the application program 1014 may include an application program with content query.

[0203] In some embodiments, the basic input / output system 1006 includes a display 1008 for displaying information and an input device 1009 such as a mouse and a keyboard for user inputting information. The display 1008 and the input device 1009 are connected to the central processing unit 1001 through an input / output controller 1010 connected to the system bus 1005. The basic input / output system 1006 may also include an input / output controller 1010 for receiving and processing inputs from a plurality of other devices such as a keyboard, a mouse, or an electronic stylus. Similarly, the input / output controller 1010 also provides output to a display screen, a printer, or other types of output devices.

[0204] The mass storage device 1007 is connected to the central processing unit 1001 via a mass storage controller (not shown) connected to the system bus 1005. The mass storage device 1007 and its associated computer-readable medium provide non-volatile storage for the server 1000. That is, the mass storage device 1007 may include a computer-readable medium (not shown) such as a hard disk or drive.

[0205] Without loss of generality, the computer readable medium may include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented by any method or technology for storing information such as computer readable instructions, data structures, program modules or other data. Computer storage media include random access memory (RAM), read-only memory (ROM), flash memory or other solid-state storage technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, cassette, magnetic tape, disk storage or other magnetic storage devices. Of course, those skilled in the art will know that the computer storage medium is not limited to the above. The above-mentioned system memory 1004 and mass storage device 1007 can be collectively referred to as memory.

[0206] The memory stores one or more programs, and the one or more programs are configured to be executed by one or more central processing units 1001. The one or more programs contain computer instructions for implementing the above-mentioned method. The central processing unit 1001 executes the one or more programs to implement the query method provided by the above-mentioned various method embodiments.

[0207] According to various embodiments of the present application, the server 1000 can also be connected to a remote computer on the network through a network such as the Internet. That is, the server 1000 can be connected to the network 1011 through the network interface unit 1012 connected to the system bus 1005, or the network interface unit 1012 can be used to connect to other types of networks or remote computer systems (not shown).

[0208] In addition, those skilled in the art will appreciate that the structure of the server shown in the above figures does not constitute a limitation on the server, and the server may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0209] An embodiment of the present application also provides a computer-readable storage medium, which stores at least one computer instruction, and the at least one computer instruction is loaded and executed by a processor to implement the query method provided in the above embodiments.

[0210] Optionally, the computer readable storage medium may include: ROM, RAM, solid state drives (SSD) or optical disks, etc. Among them, RAM may include resistance random access memory (ReRAM) and dynamic random access memory (DRAM).

[0211] An embodiment of the present application also provides a computer program product, which includes computer instructions stored in a computer-readable storage medium; a processor of a server reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the server implements the query method provided in the above-mentioned embodiments.

[0212] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or may be accomplished by a program to guide the relevant hardware. The program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0213] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A query method, characterized in that: The method comprises: In the case of receiving a query request from a client, determining a second query text corresponding to a first query text in the query request, where the second query text is rewritten based on the first query text; Returning a first query result corresponding to the first query text and a second query result corresponding to the second query text to the client, and the client is used to display the first query result and the second query result; Based on the first conversion data of the first query result and the second conversion data of the second query result, the rewriting accuracy of the second query text is verified to obtain the rewriting accuracy verification result of the second query text, the first conversion data is used to characterize the conversion status of the first query result at the client, and the second conversion data is used to characterize the conversion status of the second query result at the client.

2. The method according to claim 1, characterized in that The checking of the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result to obtain the rewriting accuracy checking result of the second query text includes at least one of the following: determining the rewriting accuracy inspection result of the second query text based on a conversion data comparison result of the first conversion data of the first query result and the second conversion data of the second query result; determining an actual conversion index of the second query result based on the first conversion data of the first query result and the second conversion data of the second query result; Based on the actual conversion index and the conversion index threshold, the rewriting accuracy test result of the second query text is determined.

3. The method according to claim 2, characterized in that The determining the rewriting accuracy inspection result of the second query text based on the conversion data comparison result of the first conversion data of the first query result and the second conversion data of the second query result includes: When the second conversion data is higher than or equal to the first conversion data, determining that the rewriting accuracy test result of the second query text is that the rewriting is accurate; When the second conversion data is lower than the first conversion data, it is determined that the rewriting accuracy check result of the second query text is inaccurate rewriting.

4. The method according to claim 2, characterized in that: The determining, based on the first conversion data of the first query result and the second conversion data of the second query result, an actual conversion index of the second query result includes: Based on the first conversion data and the second conversion data, determine a real-time conversion rate index and a real-time conversion ratio index of the second query result, where the real-time conversion ratio index is used to represent the ratio of the second conversion data to the total conversion data, and the total conversion data is the sum of the first conversion data and the second conversion data; The determining the rewriting accuracy test result of the second query text based on the actual conversion index and the conversion index threshold includes: When the real-time conversion rate index of the second query result is greater than or equal to the conversion rate index threshold, and the real-time conversion ratio index of the second query result is greater than or equal to the conversion ratio index threshold, determining that the rewriting accuracy test result of the second query text is rewriting accurate; When the real-time conversion rate index of the second query result is less than the conversion rate index threshold, and / or the real-time conversion share index of the second query result is less than the conversion share index threshold, it is determined that the rewriting accuracy test result of the second query text is inaccurate rewriting.

5. The method according to any one of claims 1 to 4, characterized in that: The returning to the client a first query result corresponding to the first query text and a second query result corresponding to the second query text, comprises: Determine a display priority of the second query result corresponding to the second query text, where the display priority is used to represent a display order of the second query result when it is displayed on the client; The first query result corresponding to the first query text, the second query result corresponding to the second query text, and the display priority of the second query result are returned to the client.

6. The method according to claim 5, characterized in that The determining the display priority of the second query result corresponding to the second query text includes: Determining a text relevance between the second query text and the first query text; The display priority of the second query result is determined based on the text relevance, and the display priority is positively correlated with the text relevance.

7. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: When receiving the query request from the client, increasing the query count corresponding to the first query text included in the query request by one; The step of verifying the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result to obtain the rewriting accuracy verification result of the second query text includes: When the number of queries corresponding to the first query text reaches a threshold number, the rewriting accuracy of the second query text is verified based on the first conversion data of the first query result and the second conversion data of the second query result to obtain the rewriting accuracy verification result of the second query text.

8. The method according to claim 7, characterized in that The method further comprises: When the number of queries corresponding to the first query text does not reach the number threshold, the first n query results with the highest relevance to the second query text in all query results of the second query text are determined as the second query results corresponding to the second query text, where n is a positive integer.

9. The method according to claim 8, characterized in that After verifying the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result and obtaining the rewriting accuracy verification result of the second query text, the method further includes: When the rewriting accuracy check result indicates that the second query text is rewritten accurately, the first m query results with the highest relevance to the second query text among the full query results of the second query text are determined as the second query results corresponding to the second query text, where m is an integer greater than n.

10. The method according to any one of claims 1 to 4, characterized in that: The method further comprises at least one of the following: If the rewriting accuracy check result indicates that the rewriting of the second query text is inaccurate, stop determining the second query text as the rewritten text of the first query text; In the case where the rewriting accuracy check result indicates that the rewriting of the second query text is inaccurate, the first query text and the second query text are determined as a negative sample pair, and the negative sample pair is used to train a text rewriting model, and the text rewriting model is used to output a rewritten text according to the input text.

11. The method according to any one of claims 1 to 4, characterized in that: In the case where the query request is a song query request, the first query result and the second query result are recalled songs, and the first conversion data and the second conversion data are determined based on at least one of the playback data, download data and collection data of the recalled songs.

12. A query device, characterized in that: The device comprises: A text determination module, configured to, upon receiving a query request from a client, determine a second query text corresponding to a first query text in the query request, wherein the second query text is rewritten based on the first query text; A return module, used to return a first query result corresponding to the first query text and a second query result corresponding to the second query text to the client, and the client is used to display the first query result and the second query result; A verification module is used to verify the rewriting accuracy of the second query text based on the first conversion data of the first query result and the second conversion data of the second query result, and obtain the rewriting accuracy verification result of the second query text, wherein the first conversion data is used to characterize the conversion status of the first query result at the client, and the second conversion data is used to characterize the conversion status of the second query result at the client.

13. A server, characterized in that: The server includes a processor and a memory, wherein the memory stores at least one computer instruction, and the at least one computer instruction is loaded and executed by the processor to implement the query method according to any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer instruction, and the at least one computer instruction is loaded and executed by a processor to implement the query method according to any one of claims 1 to 11.

15. A computer program product, characterized in that The computer program product includes computer instructions, which are stored in a computer-readable storage medium; a processor of a server reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the server implements the query method as described in any one of claims 1 to 11.

Images