A method, device and medium for off-site remote testing of a vehicle machine

Abstract

The application provides a method, device and medium for remotely testing a vehicle machine in a different place, and relates to the technical field of data processing. The vehicle machine is connected with a client through a control device. The control device is used for sending a current screen of the vehicle machine to the client in real time, and the client is used for displaying a target interface. The target interface is always the same as the current screen. In the method, the button information corresponding to each click operation performed by a target user on the target interface in a specific time period is processed to obtain a first description text by combining an LLM model. Based on the semantic matching result of the first description text and a second description text corresponding to a candidate test case, a plurality of target test cases are obtained, and the plurality of target test cases are sent to the vehicle machine through the control device to remotely test the vehicle machine. The method reduces manual intervention, improves the reusability and test efficiency of the test cases, and enhances the accuracy of the test results.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a method, device and medium for remotely testing a vehicle machine. BACKGROUND

[0002] The functional integrity and system stability of a vehicle infotainment system (vehicle machine) directly affect driving safety and user experience. Testing the vehicle machine is a key link to ensure vehicle safety, reliability and user experience, and is crucial. Existing vehicle machine testing methods generally rely on manual operation, i.e., a tester operates on a real vehicle or a test bench according to a preset test case, verifies whether the vehicle machine functions normally through clicking the screen, buttons or voice instructions, and records the test results.

[0003] However, the above method also has the following technical problems:

[0004] With the increasing complexity of the vehicle machine functions, the user interface interaction logic of the vehicle machine is also increasingly rich. Relying on manual operation to test the vehicle machine requires manual pre-setting of a plurality of different test cases according to the user interface interaction logic of the vehicle machine, and testing the vehicle machine according to the preset test cases. The reuse of test cases cannot be realized, which is time-consuming, labor-intensive, low in efficiency, and prone to missing steps or recording errors due to human negligence, affecting the accuracy of the test results. SUMMARY

[0005] To solve the above technical problems, the technical solution adopted by the present application is as follows:

[0006] According to a first aspect of the present application, a method for remotely testing a vehicle machine is provided. The vehicle machine is connected to a client through a control device. The control device is used to send the current screen of the vehicle machine to the client in real time, and the client is used to display a target interface. The target interface is always the same as the current screen. The method comprises the following steps:

[0007] S1, obtaining a target button information set A; A includes button information corresponding to each click operation performed by a target user on a target interface within a specific time period T; the starting time point of T is the time point when a pre-edit button in the client is clicked; the ending time point of T is the time point when a confirmation execution button in the client is clicked; the button information includes the button label corresponding to the click operation, the click time point and the click pressing duration.

[0008] S2, obtaining a first description text list B=(B1, B2, …, B i , …, B n ) based on A and an LLM model, where B i is the i-th preset behavior characteristic description mode X iThe first description text obtained by performing the description task on A; 1≤i≤n, where n is the number of preset behavioral characteristic description methods.

[0009] S3. Determine a candidate test case set C = {C1, C2, ..., C...} from several historical test cases. j C m}, C j Let be the j-th candidate test case, 1 ≤ j ≤ m, where m is the number of candidate test cases; where the target interface includes C. j Each click operation corresponds to a button.

[0010] S4, according to B i semantic feature vectors and E ji Based on the vector similarity between them, several target test cases are obtained, and these target test cases are sent to the vehicle system via a control device to enable remote testing of the vehicle system; E ji C j Corresponding to X i The semantic feature vector of the relevant second descriptive text.

[0011] According to a second aspect of the present invention, a non-transitory computer-readable storage medium is provided, wherein a computer program is stored in the storage medium, and the computer program is loaded and executed by a processor to implement the aforementioned method.

[0012] According to a third aspect of the present invention, an electronic device is provided, comprising: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aforementioned method.

[0013] The present invention has at least the following beneficial effects:

[0014] The application provides a method, device and medium for remotely testing a car machine in a different place, wherein the car machine establishes a connection with a client through a control device; the control device is used for sending a current screen of the car machine to the client in real time, and the client is used for displaying a target interface; the target interface is always the same as the current screen; in the method, a target button information set is obtained, the target button information set including button information corresponding to each click operation of a target user on the target interface in a specific time period; a plurality of first description texts are obtained according to a LLM model and the target button information set; a plurality of candidate test cases are determined from a plurality of historical test cases; a plurality of target test cases are obtained based on the similarity between a semantic feature vector of the first description text and a semantic feature vector of a second description text corresponding to each candidate test case; and the plurality of target test cases are sent to the car machine through the control device to remotely test the car machine. It can be known that the car machine screen is transmitted to the client in real time through the control device, remote testing is supported, and the testing cost is reduced; the target button information set is processed by combining the LLM model to obtain the first description text, a plurality of target test cases are obtained based on the semantic matching result between the first description text and the second description text corresponding to the candidate test case, and the plurality of target test cases are sent to the car machine through the control device to remotely test the car machine; the manual intervention is reduced, the reusability and testing efficiency of the test case are improved, and the accuracy of the testing result is improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description only show some of the embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0016] Figure 1 A flowchart of a method for remotely testing a car machine in a different place is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0018] It should be noted that the terms "first", "second", and the like in the description and in the claims of the present application and in the above-described drawings are used to distinguish like elements and are not necessarily used to describe a particular sequential or chronological order. It is to be understood that the terms so used are interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in other sequences than the one illustrated or other than the one explicitly described herein. Furthermore, the terms "comprise", "comprising", "include", "including", and the like are intended to cover non-exclusive inclusions, such that a process, method, system, product, or server that comprises a list of steps or units are not necessarily limited to those steps or units that are clearly listed, but can include other steps or units not expressly listed or inherent to such process, method, product, or device.

[0019] Embodiments of the present application provide a method for testing a car machine remotely, wherein the car machine is connected with a client through a control device; the control device is used to send a current screen of the car machine to the client in real time, and the client is used to display a target interface; the target interface is always the same as the current screen; the method comprises the following steps, as shown in the following figure: Figure 1

[0020] S1, obtaining a target button information set A; the A includes button information corresponding to each click operation performed by a target user on a target interface within a specific time period T; the starting time point of T is the time point when a pre-edit button in the client is clicked; the ending time point of T is the time point when a confirmation execution button in the client is clicked; the button information includes the button label corresponding to the click operation, the click time point and the click pressing time length.

[0021] Specifically, the control device has a remote communication function.

[0022] Specifically, the control device is connected with the car machine.

[0023] Specifically, the control device is connected with the client.

[0024] Specifically, the car machine and the client are located in different geographical positions and support remote testing.

[0025] Specifically, the target user can be understood as a tester who tests the car machine remotely through the client.

[0026] Specifically, the target user can click the button in the client through a mouse.

[0027] Further, the target user can click the button in the target interface through a mouse.

[0028] ​Specifically, the button tag corresponding to the click operation is the button tag of the button clicked when the click operation is performed; the click time point corresponding to the click operation is the time point when the click operation occurs; and the click press duration corresponding to the click operation is the time interval between the time when the button clicked corresponding to the click operation is pressed and the time when the button is released; for example, the button clicked corresponding to the click operation is pressed at 10:05:05 and released at 10:05:08, and the click press duration corresponding to the click operation is 3 seconds.

[0029] Specifically, the button clicked corresponding to the click operation is the button clicked when the click operation is performed.

[0030] Specifically, the client records only the click operation of the target user in the current target interface in the pre-editing mode (from the time when the pre-editing button is clicked to the time when the confirmation execution button is clicked), and does not collect operations across interfaces, so as to ensure the completeness and executability of the test case.

[0031] In one specific embodiment, after the target user performs the click operation on the target interface, the client generates the touch instruction corresponding to the click operation according to the button information corresponding to the click operation and the position information of the button clicked corresponding to the click operation in the target interface.

[0032] Specifically, the touch instruction can be understood as an instruction capable of enabling the car machine to automatically perform the click operation.

[0033] Optionally, the position information of the button clicked corresponding to the click operation in the target interface is the position coordinates of the center point of the button clicked in the target interface, and the position coordinates are obtained based on a preset coordinate system, the center point of the preset coordinate system is the center point of the target interface, and the x-axis and the y-axis of the preset coordinate system are perpendicular to each other and pass through the center point of the target interface.

[0034] In one specific embodiment, the method comprises the following steps:

[0035] After the client detects the click operation, the client generates the touch instruction corresponding to the click operation according to the button information corresponding to the click operation and the position information of the button clicked corresponding to the click operation in the target interface, and sends the touch instruction to the control device.

[0036] After the control device receives the touch instruction, the control device sends the touch instruction to the car machine, so as to perform remote testing on the car machine.

[0037] In the embodiment, through the above steps, the car machine picture is transmitted to the client in real time by the control device, the target user can click the button in the target interface displayed in the client through the mouse; a touch instruction is generated and sent to the car machine by the control device to remotely test the car machine when the client detects the click operation; the remote test is supported, and the test cost is reduced.

[0038] In one specific embodiment, the method comprises the following steps:

[0039] The client constructs a touch instruction list when detecting that the pre-edit button is clicked, and the touch instruction list is initially NULL.

[0040] The client generates a touch instruction corresponding to the click operation according to the button information corresponding to the click operation and the position information of the clicked button in the target interface corresponding to the click operation, and adds the touch instruction to the touch instruction list after detecting the click operation; the touch instructions in the touch instruction list are arranged from early to late according to the click time points of the corresponding click operations.

[0041] The client sends the touch instruction list to the control device when detecting that the confirmation execution button is clicked.

[0042] Specifically, the control device sends each touch instruction to the car machine in turn according to the order of each touch instruction in the touch instruction list and the time interval between the click time points of the click operations corresponding to adjacent touch instructions after receiving the touch instruction list, so as to remotely test the car machine; for example: the time interval between the click time point of the click operation corresponding to the first touch instruction in the touch instruction list and the click time point of the click operation corresponding to the second touch instruction in the touch instruction list is 2 minutes, so the control device sends the first touch instruction to the car machine first after receiving the touch instruction list, and then sends the second touch instruction to the car machine after 2 minutes.

[0043] Specifically, the touch instruction list can be understood as a test case.

[0044] Specifically, the car machine automatically executes the click operation corresponding to the touch instruction after receiving the touch instruction.

[0045] In the last embodiment of the embodiment, when the target user clicks the same or different buttons in the target interface in a specific frequency and order, due to the existence of network delay, inconsistent delay may be caused in the transmission of the generated several touch instructions, so that the interval time of the execution of the adjacent two touch instructions in the vehicle machine is inconsistent with the interval time of the two click operations corresponding to the adjacent two touch instructions executed by the target user; the timing consistency of the operation sequence is destroyed, which further affects the test effect, so that the original test purpose of the target user cannot be achieved.

[0046] In the embodiment, through the above steps, when the client detects that the pre-edit button is clicked, a touch instruction list initially set as NULL is constructed; the client generates a touch instruction corresponding to the click operation after detecting the click operation and adds the touch instruction to the touch instruction list; the client sends the touch instruction list to the control device when detecting that the confirmation execution button is clicked; the touch instructions are transmitted to the control device in batches, and the control device sends each touch instruction to the vehicle machine in turn according to the order of each touch instruction in the touch instruction list and the time interval between the click time points of the adjacent touch instructions corresponding to the click operation, so as to remotely test the vehicle machine; the influence of network delay on operation timing is effectively alleviated, and the interval time of the execution of the adjacent two touch instructions in the vehicle machine is highly consistent with the interval time of the two click operations corresponding to the adjacent two touch instructions executed by the target user; the test effect is avoided.

[0047] S2, obtaining a first description text list B=(B1, B2, …, B i , …, B n ) based on the A and the LLM model, B i is the first description text obtained by the LLM model in the ith preset behavior characteristic description manner X i to A performing a description task; 1≤i≤n, n is the number of preset behavior characteristic description manners.

[0048] Specifically, the first description text can be understood as a description text generated and output by the LLM model in the preset behavior characteristic description manner to the target button information set performing a description task.

[0049] Specifically, step S2 includes: inputting A and a preset prompt word corresponding to each preset behavior characteristic description manner into the LLM model to obtain B.

[0050] Specifically, the preset prompt word corresponding to the preset behavior characteristic description manner is used to guide the LLM model to describe the button information set in the preset behavior characteristic description manner, and output the corresponding description text.

[0051] Specifically, the preset behavioral characteristic description methods and their corresponding preset prompts are behavioral characteristic description methods and prompts preset by those skilled in the art according to actual needs, and will not be elaborated here.

[0052] Specifically, the description methods for any two preset behavioral characteristics are different, and the corresponding preset prompts are also different. For example, the first preset behavioral characteristic description method is a chronological description method, and its corresponding preset prompt is: "Requires a detailed description of each operation and waiting interval in strict chronological order." For example: "First, click button 1; wait 5 seconds, then click button 2; wait 3 seconds again, then click button 1 a second time." The second preset behavioral characteristic description method is a statistical inductive description method, and its corresponding preset prompt is: "Requires ignoring the specific order and summarizing from the overall behavioral pattern." For example: "During the operation, button 1 and button 2 were clicked alternately, and a 5-second waiting interval was maintained after each click."

[0053] S3. Determine a candidate test case set C = {C1, C2, ..., C...} from several historical test cases. j C m}, C j Let be the j-th candidate test case, 1 ≤ j ≤ m, where m is the number of candidate test cases; where the target interface includes C. j Each click operation in C corresponds to a button; that is, for C j For each click operation, there is a corresponding button on the target interface, and the button label of this button is the same as the button label of the clicked button; for example: if C j In the target interface, there is a click operation, and the clicked button is button 1 with the button label "volume up button". If there is a button 2 in the target interface with the button label "volume up button", then button 2 is the button in the target interface that corresponds to the click operation.

[0054] Specifically, the aforementioned historical test cases are stored in a database; the database also includes the button label of the clicked button corresponding to each click operation in each historical test case, and the semantic feature vector of the second descriptive text related to each preset behavior characteristic description method corresponding to each historical test case.

[0055] Specifically, historical test cases are test cases that have been recorded, saved, and executed and verified during past tests of the vehicle's infotainment system. They have complete operation procedures and expected results.

[0056] Specifically, the database is connected to the client.

[0057] Specifically, for each historical test case, the historical test case is processed according to a preset processing manner to obtain a button label of a clicked button corresponding to each click operation in the historical test case, and a semantic feature vector of a second description text corresponding to each preset behavior characteristic description manner of the historical test case; wherein the preset processing manner comprises the following steps S01-S03:

[0058] S01, for the clicked button corresponding to each click operation in the historical test case, the function description text of the clicked button is obtained.

[0059] Specifically, the function description text of the clicked button is a text used to describe the function of the clicked button, for example: increase volume; reduce volume.

[0060] S02, the preset button label corresponding to the function description text of the clicked button in the preset button label mapping list is taken as the button label of the clicked button; wherein the preset button label mapping list comprises a plurality of preset button labels and a preset function description text corresponding to each preset button label.

[0061] Specifically, the preset button label mapping list is a list preset by a person skilled in the art according to actual needs, which will not be described here.

[0062] In one specific embodiment, if the function description text of the clicked button is completely consistent with the preset function description text, that is, if the length of the function description text of the clicked button is consistent with the length of the preset function description text, and each character of the function description text of the clicked button is also consistent with the character at the corresponding position in the preset function description text, it is determined that the function description text of the clicked button is the same as the preset function description text.

[0063] In one specific embodiment, if the vector similarity between the semantic feature vector of the function description text of the clicked button and the semantic feature vector of the preset function description text is not less than a preset threshold, it is determined that the function description text of the clicked button is the same as the preset function description text; wherein the greater the vector similarity, the more similar the semantic feature vector of the function description text of the clicked button and the semantic feature vector of the preset function description text.

[0064] Specifically, the value range of the vector similarity is 0 to 1, and the preset threshold is set by a person skilled in the art according to actual needs, for example: 0.8, 0.9, which will not be described here.

[0065] In a specific embodiment, the vector distance between the semantic feature vector of the function description text of the clicked button and the semantic feature vector of the preset function description text is converted into a value between 0 and 1, and the value is taken as the vector similarity, wherein the smaller the vector distance, the closer the vector similarity converted from the vector distance to 1, and the larger the vector distance, the closer the vector similarity converted from the vector distance to 0. Any method in the prior art for converting a vector distance into a value between 0 and 1 and making the smaller the vector distance, the closer the vector similarity converted from the vector distance to 1, and the larger the vector distance, the closer the vector similarity converted from the vector distance to 0, falls within the protection scope of the present application, and will not be described here.

[0066] S03, obtaining an LLM model to X i The second description text obtained by performing the description task on the historical button information set corresponding to the historical test case, and the semantic feature vector of the second description text is taken as the semantic feature vector of the second description text corresponding to the historical test case and X i related. The historical button information set includes button information corresponding to each click operation in the historical test case.

[0067] Specifically, the second description text can be understood as a description text generated and output by the LLM model performing a description task on the historical button information set in a preset behavior characteristic description manner.

[0068] Specifically, the historical button information set and the preset prompt word corresponding to each preset behavior characteristic description manner are input into the LLM model to obtain the second description text corresponding to the historical test case corresponding to the historical button information set.

[0069] Through the above steps, the button label of each clicked button is determined according to the function description text of the clicked button and the preset function description text in the preset button mapping list; it can be understood as determining the standardized label of each clicked button; the large language model is used to perform a description task on the historical button information set based on a plurality of preset behavior characteristic description manners to generate a plurality of second description texts, and the semantic feature vector of the second description text is extracted; the standardized label and the semantic feature vector of the second description text are stored in the database, which is convenient for management and can be directly called in subsequent use, realizes efficient reuse, and is conducive to improving the response speed.

[0070] S4, obtaining a plurality of target test cases according to the vector similarity between the semantic feature vector of B i and the semantic feature vector of E ji ; and sending the plurality of target test cases to the car machine through the control device to remotely test the car machine; E ji is C jThe corresponding semantic feature vector of the second description text related to X i The corresponding semantic feature vector of the second description text related to X

[0071] In one specific embodiment, sending the target test cases to the car machine through the control device comprises: the client sends the target test cases to the control device, and the control device sends each target test case to the car machine in sequence after receiving the target test cases.

[0072] Through the above steps, the target button information set is obtained, the target button information set including button information corresponding to each click operation of the target user on the target interface in a specific time period; a plurality of first description texts are obtained according to the LLM model and the target button information set; a plurality of candidate test cases are determined from a plurality of historical test cases; a plurality of target test cases are obtained based on the similarity between the semantic feature vector of the first description text and the semantic feature vector of the second description text corresponding to each candidate test case; the plurality of target test cases are sent to the car machine through the control device to remotely test the car machine; the remote test is supported, and the test cost is reduced; the first description text is obtained by processing the target button information set in combination with the LLM model, the plurality of target test cases are obtained based on the semantic matching result between the first description text and the second description text corresponding to the candidate test case, and the plurality of target test cases are sent to the car machine through the control device to remotely test the car machine; the manual intervention is reduced, the reusability and test efficiency of the test cases are improved, and the accuracy of the test result is enhanced.

[0073] Specifically, the following steps are further included before step S1:

[0074] S001, for each target button, obtaining a function description text of the target button; the target button is a button in a target interface.

[0075] S002, mapping, in a preset button label mapping list, a preset button label corresponding to a preset function description text same as the function description text of the target button, as a button label of the target button.

[0076] Specifically, the function description text of the target button is a text used to describe the function of the target button, for example: increasing the volume; reducing the volume.

[0077] In one specific embodiment, the way of determining whether the function description text of the target button is same as the preset function description text is consistent with the way of determining whether the function description text of the clicked button is same as the preset function description text, which will not be described here.

[0078] Through the above steps, the button label of each target button is determined according to the function description text of the target button and the preset function description text in the preset button mapping list; it can be understood that the standardized label of each target button is determined, the unified representation of the button label is realized, and subsequent processing is facilitated.

[0079] Specifically, the step S3 includes the following steps:

[0080] S31, obtain a button label set D corresponding to the target interface, the button label set D including the button label of each target button.

[0081] S32, for each historical test case, if the button label set corresponding to the historical test case is a subset of the button label set D, the historical test case is taken as a candidate test case; wherein the button label set corresponding to the historical test case includes the button label of the clicked button corresponding to each click operation in the historical test case.

[0082] Through the above steps, if the button label set corresponding to the historical test case is a subset of the button label set corresponding to the target interface, it means that for each clicked button corresponding to each click operation in the historical test case, there is a button in the target interface, which is the same as the clicked button, or the function of the button is the same as the function of the clicked button; it can be understood that the target interface has all the function buttons required to execute the historical test case; on this basis, if it is further determined that the logic of each click operation is consistent, it means that the historical test case can be migrated and applied to the car machine corresponding to the target interface for testing. Therefore, for each historical test case, if the button label set corresponding to the historical test case is a subset of the button label set corresponding to the target interface, the historical test case is taken as a candidate test case; candidate test cases that can be migrated and applied to the car machine corresponding to the target interface for testing can be quickly screened out; which is beneficial to improve the test efficiency.

[0083] Specifically, the step S4 includes the following steps:

[0084] S41, obtain C j a corresponding vector similarity list F j =(F j1 , F j2 , …, F ji , …, F jn ), F ji is the vector similarity between the semantic feature vector of B i and E ji ; the greater F ji is, the more similar the semantic feature vector of B i is to E ji .

[0085] Specifically, 0≤Fji ≤1.

[0086] Specifically, the method for obtaining the vector similarity between the semantic feature vector of B i and the semantic feature vector of E ji is consistent with the method for obtaining the vector similarity between the semantic feature vector of the clicked button function description text and the semantic feature vector of the preset function description text, and will not be described here.

[0087] S42, if F j exists in F ji ≥ F 0 , C j is selected as a test case; F 0 is a first preset similarity threshold; it can be understood that: if at least one of F j1 , F j2 , …, F ji , …, F jn is not less than the first preset similarity threshold, C j is selected as a test case.

[0088] Specifically, the value range of F 0 is 0.8 to 1.

[0089] S43, replace the position information of the clicked button corresponding to each click operation in the selected test case with the position information of the clicked button in the target interface to obtain a target test case; wherein the position information of the clicked button in the target interface is the position information of the target button in the target interface whose button label is consistent with the button label of the clicked button; for example: the clicked button is button 1, the button label of button 1 is "increase volume button"; the target button with the button label "increase volume button" in the target interface is button 2, then the position information corresponding to button 1 in the target interface is the position information of button 2 in the target interface.

[0090] Optionally, the position information of the target button in the target interface is the position coordinates of the center point of the target button in the target interface, and the position coordinates are obtained based on a preset coordinate system.

[0091] In one specific embodiment, the position information of the clicked button corresponding to the click operation in the selected test case is the position coordinates of the center point of the clicked button in the historical in-vehicle interface corresponding to the selected test case, and the historical in-vehicle interface corresponding to the selected test case is the interface presented by the in-vehicle machine for testing using the selected test case.

[0092] The greater the vector similarity between the semantic feature vector of the first description text obtained by the LLM model performing the description task on the target button information set in the i-th preset behavior characteristic description manner and the semantic feature vector of the second description text corresponding to the candidate test case and related to the i-th preset behavior characteristic description manner, the more similar the first description text and the second description text are; it is indicated that the behavior characteristics described in the first description text are closer to the behavior characteristics described in the second description text, and further, it is indicated that the operation logic corresponding to the candidate test case is closer to the operation logic of each click operation of the target user on the target interface in a specific time period. Therefore, if there is at least one vector similarity not less than the first preset similarity threshold in the vector similarity list corresponding to the candidate test case, it can be considered that the operation logic corresponding to the candidate test case is highly consistent with the operation logic of each click operation of the target user on the target interface in a specific time period, and has migration feasibility. At this time, the candidate test case is used as a selected test case, the position information of the clicked button corresponding to each click operation in the selected test case is replaced by the corresponding position information of the clicked button in the target interface, and a target test case suitable for the target interface is generated. The reuse of test cases can improve the test reuse rate and test efficiency, while reducing manual intervention, which can enhance the accuracy of test results.

[0093] In a specific embodiment, after step S41 and before step S43, the following steps are included:

[0094] S411, input the knowledge graph corresponding to A into the feature extraction model to obtain the feature vector H corresponding to A; the knowledge graph corresponding to A includes a plurality of first nodes and edges connecting any two first nodes; the first node represents the clicked button corresponding to the button information; the edge connecting any two first nodes has a first weight, and the first weight represents the straight line distance between the two clicked buttons corresponding to the two first nodes at both ends of the edge; the straight line distance can be understood as the real physical distance.

[0095] Specifically, the straight line distance between two clicked buttons can be understood as the straight line distance between the center points of the two clicked buttons, and the straight line distance can be obtained according to the position information of the center points of the two clicked buttons.

[0096] Specifically, the number of first nodes in the knowledge graph corresponding to A is the same as the number of button information in A, and the first node and the button information are one-to-one corresponding.

[0097] Specifically, the feature extraction model is a graph neural network model capable of encoding a knowledge graph and outputting a feature vector, for example: a GAT model.

[0098] S412, obtaining C j corresponding feature vector R j feature similarity K between R and H j ; wherein R j according to C j corresponding knowledge graph and feature extraction model, C j corresponding knowledge graph includes a plurality of second nodes and edges connecting any two second nodes; the second node represents the clicked button corresponding to the click operation in the candidate test case; the edge connecting any two second nodes has a second weight, and the second weight represents the straight line distance between the clicked buttons corresponding to the two second nodes at both ends of the corresponding edge.

[0099] Specifically, 0≤K j ≤1; K j is greater, R j is more similar to H.

[0100] Specifically, the vector similarity between R j and H is taken as K j ; wherein the method for obtaining the vector similarity between R j and H is consistent with the method for obtaining the vector similarity between the semantic feature vector of the clicked button function description text and the semantic feature vector of the preset function description text, which will not be repeated here.

[0101] Specifically, the C j corresponding knowledge graph is input into the feature extraction model to obtain R j and R j is stored in the database.

[0102] S413, obtaining C j corresponding comprehensive vector similarity list P j =(P j1 , P j2 , …, P ji , …, P jn ), P ji is the i-th comprehensive vector similarity corresponding to C j , P ji meets the following conditions:

[0103] P ji =α×F ji +β×K j , α is a preset first adjustment weight, and β is a preset second adjustment weight.

[0104] Specifically, α and β meet the following conditions:

[0105] 0

[0106] Optional, α=β=0.5.

[0107] S414, if P j P exists in ji ≥P 0 Then C j As a candidate test case; P 0 The second preset similarity threshold can be understood as: if P j1 P j2 , ..., P ji , ..., P jn If at least one of the vectors has a comprehensive vector similarity not less than the second preset similarity threshold, then C will be... j As candidate test cases.

[0108] Specifically, P 0 The value ranges from 0.8 to 1.

[0109] Through the above steps, based on the knowledge graph corresponding to the target button information set, the feature vector corresponding to the target button information set is obtained; the feature similarity between the feature vector corresponding to the candidate test case and the feature vector corresponding to the target button information set is calculated. The greater the feature similarity, the more similar the knowledge graph corresponding to the candidate test case is to the knowledge graph corresponding to the target button information set. In the knowledge graph corresponding to the target button information set, the first node represents the clicked button corresponding to the button information in the target button information set, and the first weight on the edge represents the straight-line distance between the two clicked buttons corresponding to the two first nodes at both ends of the edge. In the knowledge graph corresponding to the candidate test case, the second node represents the clicked button corresponding to each click operation in the candidate test case, and the second weight on the edge represents the straight-line distance between the two clicked buttons corresponding to the two second nodes at both ends of the edge. Therefore, the more similar the knowledge graph corresponding to the candidate test case is to the knowledge graph corresponding to the target button information set, the more similar the layout of the clicked buttons corresponding to each click operation in the candidate test case is to the layout of the clicked buttons corresponding to each button information in the target button information set. This further indicates that the testing purpose of the candidate test case is closer to the testing purpose of the click operations performed by the target user on the target interface within a specific time period.

[0110] Further, based on the vector similarity between the semantic feature vector of the first description text and the semantic feature vector of the second description text corresponding to the candidate test case, and the feature similarity between the feature vector corresponding to the candidate test case and the feature vector corresponding to the target button information set, a comprehensive vector similarity corresponding to the candidate test case is obtained. If there is at least one comprehensive vector similarity not less than the second preset similarity threshold in the comprehensive vector similarity list corresponding to the candidate test case, it can be considered that the operation logic corresponding to the candidate test case is highly consistent with the operation logic of each click operation performed by the target user on the target interface in a specific time period, and the layout of each clicked button corresponding to the candidate test case is highly consistent with the layout of each clicked button corresponding to each click operation performed by the target user on the target interface in a specific time period, and the test purpose of the candidate test case is highly consistent with the test purpose of the click operation performed by the target user on the target interface in a specific time period. Therefore, the candidate test case is taken as a test case to be selected, and a target test case is further obtained. Not only the consistency of the operation logic is considered, but also the consistency of the test purpose and the button layout is considered, which can effectively exclude physically unfeasible operations such as continuous clicking of remote buttons in a short time, and significantly improve the accuracy and practical feasibility of the test case to be selected.

[0111] Specifically, in all embodiments in the present application, the button is an interactive control in the user interface; it can be understood as a visually visible and clickable area in the user interface; the user can trigger a specific function through clicking or other operations; the user interface includes the target interface and the current screen of the car machine.

[0112] The embodiments of the present application also provide a non-transitory computer readable storage medium, which can be arranged in an electronic device to save a computer program related to a method in the method embodiments, and the computer program is loaded and executed by the processor to implement the method provided by the above embodiments.

[0113] The embodiments of the present application also provide an electronic device, which includes a processor, a memory and a computer program stored in the memory and executable on the processor, and the processor implements the method provided by the above embodiments when executing the computer program.

[0114] The embodiments of the present application also provide a computer program product, which includes program codes, and when the program product is executed on an electronic device, the program codes are used to make the electronic device execute the steps in the method according to various exemplary embodiments of the present application described in the specification.

[0115] The application provides a method, device and medium for remotely testing a vehicle machine, wherein the vehicle machine is connected with a client through a control device; the control device is used for sending a current screen of the vehicle machine to the client in real time, and the client is used for displaying a target interface; the target interface is always the same as the current screen; in the method, a target button information set is obtained, the target button information set including button information corresponding to each click operation of a target user on the target interface in a specific time period; a plurality of first description texts are obtained according to a LLM model and the target button information set; a plurality of candidate test cases are determined from a plurality of historical test cases; a plurality of target test cases are obtained based on the similarity between a semantic feature vector of the first description texts and a semantic feature vector of a second description text corresponding to each candidate test case; and the plurality of target test cases are sent to the vehicle machine through the control device to remotely test the vehicle machine. It can be known that the application realizes real-time transmission of the vehicle machine screen to the client through the control device, supports remote testing, and reduces testing cost; the target button information set is processed in combination with the LLM model to obtain the first description texts, a plurality of target test cases are obtained based on the semantic matching result of the first description texts and a second description text corresponding to each candidate test case, and the plurality of target test cases are sent to the vehicle machine through the control device to remotely test the vehicle machine; manual intervention is reduced, the reusability and testing efficiency of the test cases are improved, and the accuracy of the testing result is improved.

[0116] Although some specific embodiments of the application have been described in detail above, those skilled in the art should understand that the above examples are only for illustration, and are not intended to limit the scope of the application. Those skilled in the art should also understand that various modifications can be made to the embodiments without departing from the scope and spirit of the application.

Claims

1. A method for testing a car machine remotely at a different location, characterized by, The vehicle machine establishes a connection with a client through a control device; the control device is configured to send a current screen of the vehicle machine to the client in real time, and the client is configured to display a target interface; the target interface is always the same as the current screen; the method comprises the following steps: S1, obtaining a target button information set A; A includes button information corresponding to each click operation performed by a target user on a target interface within a specific time period T; the starting time point of T is the time point when a pre-edit button in the client is clicked; the ending time point of T is the time point when a confirmation execution button in the client is clicked; the button information includes a button label corresponding to the click operation, a click time point and a click pressing duration; S2, obtaining a first description text list B=(B1, B2, …, Bi, …, Bn) based on the A and the LLM model, wherein Bi is a first description text obtained by the LLM model in an ith preset behavior characteristic description manner X i , …, B n ), B i is a first description text obtained by the LLM model in an ith preset behavior characteristic description manner X i for A to perform a description task; 1≤i≤n, and n is a number of preset behavior characteristic description manners. S3. Determine a candidate test case set C = {C1, C2, ..., C...} from several historical test cases. j C m }, C j Let be the j-th candidate test case, 1 ≤ j ≤ m, where m is the number of candidate test cases; where the target interface includes C. j Each click operation corresponds to a button in the text; S4、according to B i the semantic feature vector of the second description text corresponding to X ji and E ji between the semantic feature vectors of the second description text corresponding to X j and the semantic feature vector of the second description text corresponding to X i related to X 2. The method for testing a car machine remotely at a different place according to claim 1, wherein, The database also includes the button label of the clicked button corresponding to each click operation in each historical test case, and the semantic feature vector of the second description text corresponding to each historical test case and related to each preset behavior characteristic description method.

3. The method for testing a car machine remotely at a different place according to claim 2, wherein, For each historical test case, the historical test case is processed according to a preset processing method to obtain the button label of the clicked button corresponding to each click operation in the historical test case, and the semantic feature vector of the second description text corresponding to the historical test case and related to each preset behavior characteristic description method; wherein the preset processing method comprises the following steps: S01, for each clicked button corresponding to each click operation in the historical test case, obtaining the function description text of the clicked button; S02, mapping the preset button label in the preset button label mapping list corresponding to the function description text of the clicked button to the button label of the clicked button; wherein the preset button label mapping list includes a plurality of preset button labels and a preset function description text corresponding to each preset button label; S03, obtaining the LLM model to X i performing a description task on a second description text obtained by describing the historical button information set corresponding to the historical test case, and taking a semantic feature vector of the second description text as a semantic feature vector of a second description text corresponding to the historical test case and related to X i The historical button information set includes button information corresponding to each click operation in the historical test case.

4. The method for testing a car machine remotely at a different place according to claim 3, wherein, Before step S1, the following steps are further included: S001, for each target button, obtaining the function description text of the target button; the target button is a button in the target interface; S002, mapping the preset button label in the preset button label mapping list corresponding to the function description text of the target button to the button label of the target button.

5. The method for testing a car machine remotely at a different place according to claim 4, wherein, In step S3, the following steps are included: S31, obtaining a button label set D corresponding to the target interface, wherein D includes the button label of each target button; S32, for each historical test case, if the button label set corresponding to the historical test case is a subset of D, the historical test case is regarded as a candidate test case; wherein the button label set corresponding to the historical test case includes the button label of the clicked button corresponding to each click operation in the historical test case.

6. The method for testing a car machine remotely at a different place according to claim 4, wherein, In step S4, the following steps are included: S41, obtaining C j corresponding vector similarity list F j = (F j1 , F j2 , …, F ji , …, F jn ), F ji is the vector similarity between the semantic feature vector of B i and E ji ; the larger F ji is, the more similar the semantic feature vector of B i is to E ji ; S42, if F j F exists in the middle ji ≥ F 0 C j is a test case to be selected; F 0 is a first preset similarity threshold; S43, replace the position information of the clicked button corresponding to each click operation in the candidate test case with the position information of the clicked button in the target interface to obtain a target test case; wherein the position information of the clicked button in the target interface is the position information of a target button with a button label consistent with the button label of the clicked button in the target interface.

7. The method for testing a car machine remotely at a different place according to claim 6, wherein, After step S41 and before step S43, the following step is included: S411, input the knowledge graph corresponding to A into the feature extraction model to obtain a feature vector H corresponding to A; the knowledge graph corresponding to A includes a plurality of first nodes and edges connecting any two first nodes; the first node represents a clicked button corresponding to button information; the edge connecting any two first nodes has a first weight, and the first weight represents the straight-line distance between the two clicked buttons corresponding to the two first nodes at both ends of the corresponding edge; S412, acquire C j corresponding feature vector R j characteristic similarity K between H j ; wherein; R j According to C j corresponding knowledge graph and feature extraction model, C j The corresponding knowledge graph comprises a plurality of second nodes and edges connecting any two second nodes. the second node represents a clicked button corresponding to a click operation in the candidate test case; the edge connecting any two second nodes has a second weight, and the second weight represents the straight-line distance between the two clicked buttons corresponding to the two second nodes at both ends of the corresponding edge; S413、get C j corresponding comprehensive vector similarity list P j = (P j1 , P j2 , …, P ji , …, P jn ), P ji is C j corresponding comprehensive vector similarity, P ji satisfies the following conditions: P ji =α×F ji +β×K j , α is a preset first adjustment weight, and β is a preset second adjustment weight. S414、if P j there is P ji ≥ P 0 , C j is selected as a test case to be tested; P 0 is a second preset similarity threshold.

8. The method for testing a car machine remotely at a different place according to claim 7, wherein, α and β satisfy the following conditions: 0 < α < 1, 0 < β < 1, and α + β = 1.

9. A non-transitory computer-readable storage medium, comprising: The storage medium has a computer program stored therein, the computer program is loaded by the processor and executed to implement the method for remotely testing a car machine in a different place according to any one of claims 1-8.

10. An electronic device comprising: The processor, the memory, and the computer program stored on the memory and executable on the processor, wherein the processor implements the method for remotely testing a car machine in a different place according to any one of claims 1-8 when executing the computer program.

Images