Test method and system and storage medium

Abstract

The invention relates to a test method and system and a storage medium. The method comprises the steps that a task overall planning agent decomposes a target test task to obtain a task execution plan, task execution agents corresponding to all sub-tasks in the task execution plan are determined respectively, and the task execution plan is sent to all the task execution agents and a task evaluation agent; each task execution agent sequentially controls the terminal corresponding to the task execution agent to execute the corresponding subtask based on the task execution plan and the task evaluation information, and stores the task processing record of the subtask to a shared storage space; the task evaluation agent obtains the task processing record of the subtask from the shared storage space, evaluates the completion condition of the subtask in combination with the task execution plan, and obtains task evaluation information; the task evaluation information is used for indicating whether the execution process of the subtask is wrong. The purpose of automatically testing the multi-device interaction function can be achieved.

Description

Technical Field

[0001] The present disclosure relates to the field of testing technology, and in particular to a testing method, system and storage medium. Background Art

[0002] With the popularity of mobile devices and social networks, multi-device interaction functions have gradually become an indispensable function in various applications. In the process of product development, most functional tests can be completed automatically through automated testing tools, but there is currently no mature automated solution for multi-device interaction function tests that require multiple devices to interact in a specific order. Summary of the invention

[0003] In order to solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a testing method, a system and a storage medium.

[0004] In a first aspect, the present disclosure provides a testing method, comprising:

[0005] The task coordinating agent decomposes the target test task to obtain a task execution plan; the target test task is a test task for multi-device interaction functions; the task execution plan includes multiple subtasks and the execution order of the multiple subtasks;

[0006] The task coordinating agent determines the task execution agents corresponding to the subtasks respectively, and sends the task execution plan to each of the task execution agents and the task evaluation agent; according to the execution order, the task execution agents corresponding to two adjacent subtasks are different; each of the task execution agents has a terminal corresponding to it;

[0007] Each of the task execution agents controls the corresponding terminal to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and stores the task processing record of the subtask in the shared storage space;

[0008] The task evaluation agent obtains the task processing record of the subtask from the shared storage space, evaluates the completion status of the subtask in combination with the task execution plan, and obtains the task evaluation information; the task evaluation information is used to indicate whether there is an error in the execution process of the subtask.

[0009] In a second aspect, the present disclosure further provides a test system, the test system comprising a task coordination agent, a task execution agent, and a task evaluation agent;

[0010] The task coordinating agent is used to decompose the target test task to obtain a task execution plan; the target test task is a test task for multi-device interaction functions; the task execution plan includes multiple subtasks and the execution order of the multiple subtasks;

[0011] The task coordinating agent is also used to determine the task execution agents corresponding to the subtasks respectively, and send the task execution plan to each of the task execution agents and the task evaluation agent; according to the execution order, the task execution agents corresponding to two adjacent subtasks are different; each of the task execution agents has a terminal corresponding to it;

[0012] Each of the task execution agents is used to control the corresponding terminal to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and store the task processing record of the subtask in the shared storage space;

[0013] The task evaluation agent is used to obtain the task processing record of the subtask from the shared storage space, and evaluate the completion status of the subtask in combination with the task execution plan to obtain the task evaluation information; the task evaluation information is used to indicate whether there is an error in the execution process of the subtask.

[0014] In a third aspect, the present disclosure further provides a computer-readable storage medium having a computer program stored thereon, which implements the method described above when the program is executed by a processor.

[0015] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:

[0016] The technical solution provided by the disclosed embodiment decomposes the target test task by setting a task coordination agent to obtain a task execution plan; the target test task is a test task for a multi-device interactive function; the task execution plan includes multiple subtasks and the execution order of multiple subtasks; the task coordination agent determines the task execution agent corresponding to the subtask respectively, and sends the task execution plan to each task execution agent and the task evaluation agent; according to the execution order, the task execution agents corresponding to two adjacent subtasks are different; each task execution agent has a terminal corresponding to it; each task execution agent controls the terminal corresponding to it to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and stores the task processing record of the subtask in a shared storage space; the task evaluation agent obtains the task processing record of the subtask from the shared storage space, evaluates the completion of the subtask in combination with the task execution plan, and obtains the task evaluation information; the task evaluation information is used to indicate whether the execution process of the subtask is wrong. Its essence is to use the role-based agents to perform automated testing on multi-device interactive functions, which can fill the gap in this regard and meet the testing requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] Figure 1 A schematic diagram of a test system provided by an embodiment of the present disclosure;

[0020] Figure 2 A flow chart of a testing method provided in an embodiment of the present disclosure;

[0021] Figure 3 A method provided in the embodiment of the present disclosure Figure 2 Flow chart of the implementation method of S130. DETAILED DESCRIPTION

[0022] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0023] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0024] Figure 1 A schematic diagram of a test system provided for an embodiment of the present disclosure, which is used to perform automated testing on multi-device interaction functions of an application. Multi-device interaction functions may be, for example, functions that require data exchange and interaction between multiple devices, and in the process of data exchange and interaction, multiple devices need to be triggered sequentially. Multi-device interaction functions generally involve collaborative work between different devices. This function can be applied to a variety of scenarios, such as personal communication, video conferencing, team collaboration, remote office, etc. Exemplarily, the multi-device interaction functions may be, for example, video calls, remote control, collaborative editing, etc.

[0025] See also Figure 1 , the test system includes a task coordination agent, a set of candidate agents, and a task evaluation agent. Optionally, the number of task coordination agents and task evaluation agents is one. The candidate agent set includes multiple independent candidate agents. During the actual test process, the candidate agents in the candidate agent set will be determined by the task coordination agent as task execution agents and participate in the actual test process. However, which specific candidate agent will be determined as the task execution agent and the number of agents determined as task execution agents are determined by the task coordination agent according to actual needs. For example, assuming that in a certain scenario, Figure 1 Candidate Agent 2 and Candidate Agent 3 in are used as task execution agents.

[0026] Each candidate agent corresponds to a terminal and can control the running state of the corresponding terminal to complete the specified test task. When a candidate agent is determined as a task execution agent, the task execution agent can control the corresponding terminal to execute the corresponding subtask.

[0027] The task coordination agent can directly transfer information with the candidate agents in the candidate agent set and the task evaluation agent. The candidate agents in the candidate agent set and the task evaluation agent, as well as different candidate agents in the candidate agent set, can transfer information by storing data in a shared storage space and reading data from a shared storage space.

[0028] Figure 2 A flow chart of a testing method provided by an embodiment of the present disclosure. Figure 2 As shown, the method may specifically include:

[0029] S110, the task coordination intelligent agent decomposes the target test task to obtain a task execution plan; the target test task is a test task for multi-device interaction functions; the task execution plan includes multiple subtasks and the execution order of the multiple subtasks.

[0030] The target test task may be, for example, a task for performing automated testing on multi-device interaction functions of an application.

[0031] In practice, the method for determining the target test task may include: obtaining a test instruction input by a tester, which may be a voice instruction or a text instruction; and semantically understanding the test instruction to obtain the target test task. For example, the tester inputs the test instruction "perform a video call function test". Based on the test instruction, the target test task is determined to be a test of the video call function.

[0032] In the field of artificial intelligence and computer science, an agent is a digital entity that can perceive the environment, make decisions, and take actions to achieve goals. An agent is used to simulate human thinking and operation. Alternatively, an agent can be a software program.

[0033] The essence of this step is to treat the tasks that a single task execution agent needs to complete as a subtask, determine the execution order of each subtask, and then obtain the task execution plan.

[0034] For example, if the target test task is to test the video call function, the target test task is decomposed, and it can be found that in order to complete the target test task, two task execution agents are required to cooperate to complete it. The first task execution agent is used to make a video call, and the second task execution agent is used to answer the video call. That is, the target test task includes two subtasks; one subtask is to make a video call, and the other subtask is to answer the video call. The execution order of the subtasks is to make a video call first, and then answer the video call.

[0035] Optionally, the implementation method of this step may include: the task coordination agent uses the target test task as prompt information, inputs the task decomposition model, and obtains the task execution plan. The task decomposition model can be, for example, a pre-trained model for splitting and planning the target test task. It can specifically be a neural network model, such as a large language model.

[0036] S120, the task coordination agent determines the task execution agents corresponding to the subtasks respectively, and sends the task execution plan to each task execution agent and the task evaluation agent; according to the execution order, the task execution agents corresponding to two adjacent subtasks are different; each task execution agent has a corresponding terminal.

[0037] The essence of this step is to select some or all of the candidate agents from the candidate agent set as task execution agents and assign specific subtasks to them.

[0038] It should be noted that when splitting the target test task, it is split according to the roles required to complete the target test task, rather than splitting according to the actions. In other words, if multiple actions need to be completed in order to complete the target test task. Among these actions, some actions are continuous, and these actions need to be performed by the same task execution agent, and these actions are treated as a subtask, rather than treating each action as a subtask. For example, if the target test task is to test the video call function, in order to perform the target test task, two task execution agents are set, namely agent 1 and agent 2. These two agents represent two different users. Agent 1 needs to complete two actions, one is to open a chat window to communicate with agent 2, and the other is to initiate a video call through the video call option in the chat window communicating with agent 2. These two actions need to be completed by the same agent, and will be assigned to the same task processing device as the same subtask. Therefore, according to the execution order, the task execution agents corresponding to two adjacent subtasks are different. It should also be noted that, in practice, when the task coordinating agent decomposes the target test task, it does not decompose the target task into very detailed actions based on the changes in the terminal screen interface. The "opening a chat window to communicate with agent 2, and using the video call option in the chat window" in the above example is a very detailed action and is not considered a subtask. Instead, it is roughly divided according to the final effect that each role needs to achieve. In the above example, "initiating a video call" is a rough division result and can be used as a subtask. In other words, when decomposing the target test task, there is no strict limit, estimate, or setting of how many actions are required to complete a subtask.

[0039] S130. Each task execution agent controls the corresponding terminal to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and stores the task processing record of the subtask in the shared storage space.

[0040] S140, the task evaluation agent obtains the task processing record of the subtask from the shared storage space, evaluates the completion status of the subtask in combination with the task execution plan, and obtains task evaluation information; the task evaluation information is used to indicate whether there is an error in the execution process of the subtask.

[0041] The above technical scheme decomposes the target test task by setting a task coordination agent to obtain a task execution plan; the target test task is a test task for multi-device interaction functions; the task execution plan includes multiple subtasks and the execution order of multiple subtasks; the task coordination agent determines the task execution agents corresponding to the subtasks respectively, and sends the task execution plan to each task execution agent and the task evaluation agent; according to the execution order, the task execution agents corresponding to two adjacent subtasks are different; each task execution agent has a terminal corresponding to it; each task execution agent controls the terminal corresponding to it to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and stores the task processing record of the subtask in a shared storage space; the task evaluation agent obtains the task processing record of the subtask from the shared storage space, and evaluates the completion of the subtask in combination with the task execution plan to obtain the task evaluation information; the task evaluation information is used to indicate whether the execution process of the subtask is wrong. Its essence is to use the role-based agents to perform automated testing on multi-device interaction functions, which can fill the gap in this area and meet the testing needs.

[0042] In the above technical solution, there are many specific implementation methods of S130, which are not limited in this application. Exemplarily, the implementation method of this step may include:

[0043] S131. For any task execution agent, the task execution agent periodically obtains task processing records related to the target test task from the shared storage space.

[0044] The shared storage space can be, for example, a storage space accessible to all task execution agents and task evaluation agents. After completing its corresponding subtask, each task execution agent forms a task processing record and stores the task processing record in the shared storage space for access by other task execution agents and task evaluation agents.

[0045] Optionally, for any task execution agent, when it is determined that its corresponding subtask has been completed and the subtask is the last subtask in the task execution plan, a task processing record including a first prompt information is generated; the first prompt information is used to prompt that the target test task has been completed; when it is determined that its corresponding subtask has been completed and the subtask is not the last subtask in the task execution plan, a task processing record including a second prompt information is generated; the second prompt information is used to prompt the switching of the task execution agent. The task execution agent stores the task processing records it generates in the shared storage space. In this way, any task execution agent can obtain the task processing records of other subtasks related to the target test task stored by other task execution agents from the shared storage space.

[0046] S132. The task execution agent determines whether the execution node of the subtask corresponding to the task execution agent has been reached based on the task execution plan and the task processing record.

[0047] Arriving at the execution node of the subtask corresponding to the task execution agent means that it is the turn of the task execution agent to execute its corresponding subtask according to the execution order of the subtasks.

[0048] Exemplarily, any task execution agent, based on the task execution plan and the task processing records formed by other task execution agents, determines that in the execution order, the task processing record of the previous subtask corresponding to the task execution agent includes the second prompt information. Since the second prompt information is used to prompt the switching task execution agent, this means that the execution node of the subtask corresponding to the task execution agent has been reached.

[0049] S133. If the execution node of the subtask corresponding to the task execution agent is reached, the task execution agent controls the corresponding terminal to execute the corresponding subtask based on the task execution plan and task evaluation information.

[0050] By setting any task execution agent, the task execution agent periodically obtains the task processing records related to the target test task from the shared storage space; the task execution agent determines whether the execution node of the subtask corresponding to the task execution agent is currently reached based on the task execution plan and the task processing records; if the execution node of the subtask corresponding to the task execution agent is currently reached, the task execution agent controls the corresponding terminal to execute its corresponding subtask based on the task execution plan and task evaluation information. The advantage is that the task coordination agent does not need to monitor the execution progress of the subtask at all times, but the task execution agent determines whether it has reached the node for executing its subtask, which can reduce the load of the task coordination agent, reduce the design difficulty of the task coordination agent, and also reduce the network delay of frequent communication between the task execution agent and the task coordination agent.

[0051] There are many implementation methods of S133, and this application does not limit this. Exemplarily, the implementation method of this step may include: if the execution node of the subtask corresponding to the task execution agent is currently reached, repeat the subtask execution step until the end condition is reached; the end condition includes: the subtask is completed; or, the number of repetitions of the subtask execution step reaches a set number threshold. The subtask execution step includes: obtaining the current screen interface information and task evaluation information of the terminal corresponding to the task execution agent; based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan, determine the target component to be triggered and the trigger action; use the trigger action to trigger the target component; store the updated screen interface information, target component and trigger action of the terminal corresponding to the task execution agent after the trigger as a task processing record in the shared storage space.

[0052] The screen interface information may be, for example, information about the page displayed on the screen of the terminal corresponding to the task execution agent. When the screen of the terminal corresponding to the task execution agent displays different pages, the screen interface information is different. In other words, with the help of the screen interface information, it is possible to clearly determine which page is displayed on the screen of the terminal corresponding to the task execution agent. Exemplarily, the current screen interface information of the terminal corresponding to the task execution agent may include at least one of the following: the page ID, page route, page type or page name currently displayed on the screen of the terminal controlled by the task execution agent.

[0053] In practice, the task evaluation agent can periodically determine whether the subtask currently being executed is wrong while the task execution agent is executing the corresponding subtask, form task evaluation information, and send the task evaluation information to the task execution agent. The task evaluation information can reflect whether the execution process of the subtask currently being executed is wrong.

[0054] The page displayed on the screen of the terminal corresponding to the task execution agent at the current moment includes one or more components. A component is a control or area in a page with a specific function and appearance. The target component is the component that needs to be triggered, which is determined from the components included in the page displayed on the screen of the terminal corresponding to the task execution agent at the current moment. The triggering of the target component will cause the terminal screen interface information to change. The trigger action is a specific way to trigger the target component. Exemplarily, the trigger action can be determined from a preset trigger action set. The trigger action set may include, for example, clicking, inputting text, inputting images, long pressing controls, dragging controls, and sliding the screen in a preset direction.

[0055] Since in the present application an intelligent agent is used to perform subtasks on behalf of the user, triggering the target component by using a trigger action means that the intelligent agent simulates the user to trigger the target component by using a trigger action.

[0056] It should be noted that, in practice, optionally, "determining the target component to be triggered and the triggering action based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan" may include: judging whether the subtask corresponding to the current task execution agent is completed based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan; if the subtask corresponding to the current task execution agent has been completed, and the subtask corresponding to the current task execution agent is the last subtask in the task execution plan, generating a task processing record including a first prompt information; the first prompt information is used to prompt that the target test task has been completed; if the subtask corresponding to the current task execution agent has been completed, and the subtask corresponding to the current task execution agent is not the last subtask in the task execution plan, generating a task processing record including a second prompt information; the second prompt information is used to prompt the switching of the task execution agent; if the subtask corresponding to the current task execution agent is not completed, determining the target component to be triggered and the triggering action based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan.

[0057] Furthermore, based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan, the target component to be triggered and the trigger action are determined, including: obtaining page jump link information related to the current screen interface information of the terminal corresponding to the task execution agent; based on the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan, the target component to be triggered and the trigger action are determined.

[0058] Page jump link information, for example, can be information describing how the page is converted after which component in the page is triggered. In some embodiments, the page jump link information can be queried in the application interface conversion graph based on the current screen interface information of the terminal corresponding to the task execution agent. The application interface conversion graph describes the navigation path and interaction method of the application. Specifically, the application interface conversion graph can be regarded as a collection of multiple page jump links in the application. The storage structure of the application interface conversion graph is a graph structure, which stores all trigger actions and page nodes in an application. When a user performs an action on a page, the process of going to the next page will be described as a path in the graph. According to the current screen interface information of the task execution agent, query in the application interface conversion graph, that is, find all optional actions on the current screen in the application interface conversion graph, and which page these actions will cause to switch to.

[0059] By introducing page jump link information, the difficulty of determining the target component to be triggered and the triggering action can be reduced, thereby improving test efficiency.

[0060] Furthermore, based on the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan, the target component to be triggered and the trigger action are determined, which may include: using the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan as prompt information, inputting the trigger operation determination model, and obtaining the target component to be triggered and the trigger action. The trigger operation determination model is a pre-trained model for determining the target component to be triggered and the trigger action. It can specifically be a neural network model, such as a large language model.

[0061] For example, suppose that a certain application has an email sending function and an instant messaging function. Among them, the instant messaging function includes a video messaging function. If the subtask a corresponding to the task execution agent A is to initiate a video call request to the task execution agent B. The subtask b corresponding to the task execution agent B is to accept the video call request initiated by the task execution agent A.

[0062] At a certain moment, the execution node of subtask a corresponding to task execution agent A is reached, and task evaluation information about subtask a is received from the task evaluation agent (optionally, when the execution step of subtask a is executed for the first time, the default task evaluation information is that the execution process of subtask a is correct). Task execution agent A determines that the target component to be triggered is the return control and the trigger action is click based on the current screen interface information of its corresponding terminal, task evaluation information and task execution plan. After task execution agent A simulates the user clicking the return control, the page displayed on the screen of the terminal controlled by task execution agent A will be updated. It is assumed that the updated page is the homepage of the application. Task execution agent A stores the updated screen interface information of its corresponding terminal (i.e., the homepage of the application), the target component (i.e., the return control) and the trigger action (i.e., the click action) as task processing records in the shared storage space. The task evaluation agent obtains the task processing record from the shared storage space, determines whether the execution process of subtask a is incorrect, and forms new task evaluation information about subtask a. At this time, the task evaluation information of subtask a is that the execution process of subtask a is correct. Task execution agent A determines the target component to be triggered as the icon representing task execution agent B, and the trigger action as click, based on the current screen interface information of the corresponding terminal (i.e., the homepage of the application), task evaluation information, and task execution plan. After task execution agent A simulates the user clicking the icon representing task execution agent B, the page displayed on the screen of the terminal corresponding to task execution agent A will be updated. It is assumed that the updated page is a chat window for communicating with task execution agent C. Task execution agent A stores the updated screen interface information (i.e., the chat window for communicating with task execution agent C), target component (i.e., the icon of task execution agent B), and trigger action (i.e., click action) as task processing records in the shared storage space. The task evaluation agent obtains the task processing record from the shared storage space, determines whether the execution process of subtask a is wrong, and forms new task evaluation information about subtask a. At this time, the task evaluation information of subtask a is that the execution process of subtask a is wrong. If the chat window communicating with task execution agent C includes an icon representing task execution agent B, task execution agent A determines the target component to be triggered as the icon representing task execution agent B based on its current screen interface information (i.e., the chat window communicating with task execution agent C), task evaluation information, and task execution plan, and the triggering action is click. After task execution agent A simulates a user clicking on the icon representing task execution agent B, the page displayed on the terminal screen corresponding to task execution agent A will be updated. It is assumed that the updated page is the chat window communicating with task execution agent B.Task execution agent A stores the updated screen interface information (i.e., the chat window for communicating with task execution agent B), target component (i.e., the icon of task execution agent B) and trigger action (i.e., click action) as task processing records in the shared storage space. The task evaluation agent obtains the task processing record from the shared storage space, determines whether the execution process of subtask a is wrong, and forms new task evaluation information about subtask a. At this time, the task evaluation information of subtask a is that the execution process of subtask a is correct. Based on the current screen interface information of its corresponding terminal (i.e., the chat window for communicating with task execution agent B), task evaluation information and task execution plan, task execution agent A determines that the target component to be triggered is the video communication control and the trigger action is click. After task execution agent A simulates the user clicking the video call option, the page displayed on the screen of the terminal corresponding to task execution agent A will be updated. It is assumed that the updated page is the page waiting for task execution agent B to answer the video call. The task execution agent A stores the updated screen interface information (i.e., the page waiting for the task execution agent B to answer the video call), the target component (i.e., the video call option) and the trigger action (i.e., the click action) as a task processing record in the shared storage space. The task evaluation agent obtains the task processing record from the shared storage space, determines whether the execution process of the subtask a is wrong, and forms new task evaluation information about the subtask a. At this time, the task evaluation information of the subtask a is that the execution process of the subtask a is correct. The task execution agent A determines that the subtask is completed based on the current screen interface information of the corresponding terminal (i.e., the page waiting for the task execution agent B to answer the video call), the task evaluation information and the task execution plan, and generates a task processing record including the first prompt information, and the first prompt information is used to prompt the switching device. The task processing record is stored in the shared storage space. When the task execution agent B detects that the task processing record includes the first prompt information, it determines that the execution node of the subtask b corresponding to the task execution agent B has been reached, and executes the subtask b.

[0063] It should be noted that, with the technical solution provided by this application, before the task execution agent controls the corresponding terminal to execute its corresponding subtask, it does not limit the interface type displayed by the controlled terminal or the page jump link to complete the subtask. On the contrary, during the task execution process, the task execution agent gradually analyzes and selects the appropriate operation method based on the currently displayed page, so that the subtask is gradually approaching completion. This solution makes the link selection of the subtask random to a certain extent, which helps to discover problems in the multi-device interaction function.

[0064] It should be noted that in the above technical solution, the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan are used as prompt information, and the trigger operation determination model is input to obtain the target component to be triggered and the trigger action, which may include: preprocessing the current screen interface information and / or the page jump link information of the terminal corresponding to the task execution agent; using the preprocessed current screen interface information of the terminal corresponding to the task execution agent, the preprocessed page jump link information, the task evaluation information and the task execution plan as prompt information, inputting the trigger operation determination model, and obtaining the target component to be triggered and the trigger action. The purpose of the preprocessing here is to eliminate the information that the trigger operation determination model cannot understand. Optionally, the preprocessing may specifically include: eliminating the information that is not related to the control in the screen interface information and / or the page jump link information. Exemplarily, the preprocessed information includes: the ID, position, size, content description, purpose description, type, whether it can be clicked, whether it can be scrolled, etc. of the control or area.

[0065] It should also be noted that for any subtask, if the number of repetitions of the subtask execution step reaches the set number threshold, it usually means that the multi-device interaction function has a more serious defect, resulting in the inability to complete the target test task. Setting the end condition includes the number of repetitions of the subtask execution step reaching the set number threshold, which can prevent the task execution agent from falling into an infinite loop. Here, the set number threshold is a pre-specified threshold, and its specific value is not limited in this application.

[0066] Based on the above technical solution, there are multiple implementation methods of S140, which are not limited in this application. Exemplarily, the implementation method of S140 may include: the task evaluation agent periodically obtains the task processing record of the subtask from the shared storage space, and the task evaluation agent uses the task execution plan and the task processing record as prompt information, inputs the task evaluation model, and obtains the task evaluation information. Exemplarily, the task evaluation model may be, for example, a large language model.

[0067] Furthermore, the task execution plan and the task processing record are used as prompt information and input into the task evaluation model to obtain the task evaluation information, which may include: preprocessing the screen interface information in the task processing record; using the preprocessed task processing record and the task execution plan as prompt information and inputting into the task evaluation model to obtain the task evaluation information. The purpose of the preprocessing here is to eliminate information that cannot be understood by the task evaluation model. Optionally, the preprocessing may specifically include: eliminating information in the screen interface information that is not related to the control. Exemplarily, the preprocessed screen interface information includes: the ID, position, size, content description, usage description, type, whether it can be clicked, whether it can be scrolled, etc. of the control or area.

[0068] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0069] The present application also provides a testing system. Figure 1 The test system includes a task coordination agent, a task execution agent, and a task evaluation agent. Figure 1 , the test system specifically includes:

[0070] The task coordinating agent is used to decompose the target test task to obtain a task execution plan; the target test task is a test task for multi-device interaction functions; the task execution plan includes multiple subtasks and the execution order of the multiple subtasks;

[0071] The task coordination agent is also used to determine the task execution agents corresponding to the subtasks respectively, and send the task execution plan to each of the task execution agents and the task evaluation agent; according to the execution order, the task execution agents corresponding to two adjacent subtasks are different;

[0072] Each of the task execution agents is used to execute the corresponding subtasks in sequence based on the task execution plan and task evaluation information, and store the task processing records of the subtasks in a shared storage space;

[0073] The task evaluation agent is used to obtain the task processing record of the subtask from the shared storage space, and evaluate the completion status of the subtask in combination with the task execution plan to obtain the task evaluation information; the task evaluation information is used to indicate whether there is an error in the execution process of the subtask.

[0074] Further, for any of the task execution agents, the task execution agent periodically obtains the task processing records related to the target test task from the shared storage space;

[0075] The task execution agent determines whether the execution node of the subtask corresponding to the task execution agent has been reached based on the task execution plan and the task processing record;

[0076] If the execution node of the subtask corresponding to the task execution agent is reached, the task execution agent executes the subtask corresponding to the task execution agent based on the task execution plan and task evaluation information.

[0077] Furthermore, if the execution node of the subtask corresponding to the task execution agent is currently reached, the task execution agent executes the subtask corresponding to the task execution agent based on the task execution plan and the task evaluation information, including:

[0078] If the execution node of the subtask corresponding to the task execution agent is reached, the subtask execution steps are repeated until the end condition is reached; the end condition includes: the subtask is completed; or the number of repetitions of the subtask execution steps reaches a set number threshold;

[0079] The subtask execution steps include:

[0080] Obtaining the current screen interface information of the terminal corresponding to the task execution agent and the task evaluation information;

[0081] Based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan, determine the target component to be triggered and the triggering action;

[0082] Triggering the target component using the trigger action;

[0083] The updated screen interface information of the terminal corresponding to the task execution agent after being triggered, the target component and the triggering action are stored in the shared storage space as task processing records.

[0084] Further, the target component to be triggered and the triggering action to be triggered are determined based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan, including:

[0085] Obtaining page jump link information related to the current screen interface information of the terminal corresponding to the task execution agent;

[0086] Based on the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan, the target component to be triggered and the triggering action are determined.

[0087] Further, the target component to be triggered and the triggering action to be triggered are determined based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan, including:

[0088] Based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan, determine whether the subtask corresponding to the current task execution agent is completed;

[0089] If the subtask corresponding to the current task execution agent has been completed, and the subtask corresponding to the task execution agent is the last subtask in the task execution plan, a task processing record including first prompt information is generated; the first prompt information is used to prompt that the target test task has been completed;

[0090] If the subtask corresponding to the current task execution agent has been completed, and the subtask corresponding to the current task execution agent is not the last subtask in the task execution plan, a task processing record including second prompt information is generated; the second prompt information is used to prompt the switching of the task execution agent;

[0091] If the subtask corresponding to the current task execution agent is not completed, the target component to be triggered and the triggering action are determined based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan.

[0092] Further, the target component to be triggered and the triggering action to be triggered are determined based on the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan, including:

[0093] The current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan are used as prompt information, and the trigger operation determination model is input to obtain the target component to be triggered and the trigger action.

[0094] Furthermore, the task coordinating agent decomposes the target test task to obtain a task execution plan, including:

[0095] The task coordination agent uses the target test task as prompt information, inputs it into the task decomposition model, and obtains the task execution plan.

[0096] Furthermore, the task evaluation agent obtains the task processing record of the subtask from the shared storage space, evaluates the completion of the subtask in combination with the task execution plan, and obtains the task evaluation information, including:

[0097] The task evaluation agent periodically obtains the task processing record of the subtask from the shared storage space,

[0098] The task evaluation agent uses the task execution plan and task processing record as prompt information, inputs them into a task evaluation model, and obtains the task evaluation information.

[0099] The test system provided in the embodiment of the present disclosure can execute the steps included in the test method provided in the embodiment of the method of the present disclosure, and has the same or corresponding beneficial effects as the steps included, which will not be repeated here.

[0100] According to one or more embodiments of the present disclosure, the present disclosure provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the program implements any of the testing methods provided by the present disclosure.

[0101] The embodiments of the present disclosure further provide a computer program product, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the testing method described above is implemented.

[0102] The steps of the method described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0103] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0104] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A testing method, characterized in that: include: The task coordinating agent decomposes the target test task to obtain a task execution plan; the target test task is a test task for multi-device interactive functions; The task execution plan includes a plurality of subtasks and the execution order of the plurality of subtasks; The task coordinating agent determines the task execution agents corresponding to the subtasks respectively, and sends the task execution plan to each of the task execution agents and the task evaluation agent; According to the execution order, two adjacent subtasks correspond to different task execution agents; each task execution agent has a terminal corresponding to it; Each of the task execution agents controls the corresponding terminal to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and stores the task processing record of the subtask in the shared storage space; The task evaluation agent obtains the task processing record of the subtask from the shared storage space, evaluates the completion status of the subtask in combination with the task execution plan, and obtains the task evaluation information; The task evaluation information is used to indicate whether there is an error in the execution process of the subtask.

2. The method according to claim 1, characterized in that Each of the task execution agents controls the corresponding terminal to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and stores the task processing record of the subtask in a shared storage space, including: For any of the task execution agents, the task execution agent periodically obtains the task processing records related to the target test task from the shared storage space; The task execution agent determines whether the execution node of the subtask corresponding to the task execution agent has been reached based on the task execution plan and the task processing record; If the execution node of the subtask corresponding to the task execution agent is reached, the task execution agent controls the corresponding terminal to execute the corresponding subtask based on the task execution plan and task evaluation information.

3. The method according to claim 2, characterized in that If the execution node of the subtask corresponding to the task execution agent is reached, the task execution agent controls the corresponding terminal to execute the corresponding subtask in sequence based on the task execution plan and task evaluation information, including: If the execution node of the subtask corresponding to the task execution agent is reached, the subtask execution steps are repeated until the end condition is reached; the end condition includes: the subtask is completed; or the number of repetitions of the subtask execution steps reaches a set number threshold; The subtask execution steps include: Obtaining the current screen interface information of the terminal corresponding to the task execution agent and the task evaluation information; Determine the target component to be triggered and the triggering action based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan; Triggering the target component using the trigger action; The updated screen interface information of the terminal corresponding to the task execution agent after being triggered, the target component and the triggering action are stored in the shared storage space as task processing records.

4. The method according to claim 3, characterized in that The determining of the target component to be triggered and the triggering action based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan includes: Obtaining page jump link information related to the current screen interface information of the terminal corresponding to the task execution agent; Based on the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan, the target component to be triggered and the triggering action are determined.

5. The method according to claim 3, characterized in that: The determining of the target component to be triggered and the triggering action based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan includes: Based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan, determine whether the subtask corresponding to the current task execution agent is completed; If the subtask corresponding to the current task execution agent has been completed, and the subtask corresponding to the task execution agent is the last subtask in the task execution plan, a task processing record including first prompt information is generated; the first prompt information is used to prompt that the target test task has been completed; If the subtask corresponding to the current task execution agent has been completed, and the subtask corresponding to the current task execution agent is not the last subtask in the task execution plan, a task processing record including second prompt information is generated; the second prompt information is used to prompt the switching of the task execution agent; If the subtask corresponding to the current task execution agent is not completed, the target component to be triggered and the triggering action are determined based on the current screen interface information of the terminal corresponding to the task execution agent, the task evaluation information and the task execution plan.

6. The method according to claim 4, characterized in that The determining of the target component to be triggered and the triggering action based on the current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan includes: The current screen interface information of the terminal corresponding to the task execution agent, the page jump link information, the task evaluation information and the task execution plan are used as prompt information, and the trigger operation determination model is input to obtain the target component to be triggered and the trigger action.

7. The method according to claim 1, characterized in that The task coordinating agent decomposes the target test task to obtain a task execution plan, including: The task coordination agent uses the target test task as prompt information, inputs it into the task decomposition model, and obtains the task execution plan.

8. The method according to claim 1, characterized in that: The task evaluation agent obtains the task processing record of the subtask from the shared storage space, evaluates the completion of the subtask in combination with the task execution plan, and obtains the task evaluation information, including: The task evaluation agent periodically obtains the task processing record of the subtask from the shared storage space, The task evaluation agent uses the task execution plan and task processing record as prompt information, inputs them into a task evaluation model, and obtains the task evaluation information.

9. A testing system, characterized in that: The test system includes a task coordination agent, a task execution agent and a task evaluation agent; The task coordinating agent is used to decompose the target test task and obtain a task execution plan; the target test task is a test task for multi-device interaction functions; The task execution plan includes a plurality of subtasks and the execution order of the plurality of subtasks; The task coordination agent is also used to determine the task execution agents corresponding to the subtasks respectively, and send the task execution plan to each of the task execution agents and the task evaluation agent; According to the execution order, two adjacent subtasks correspond to different task execution agents; each task execution agent has a terminal corresponding to it; Each of the task execution agents is used to control the corresponding terminal to execute its corresponding subtask in turn based on the task execution plan and task evaluation information, and store the task processing record of the subtask in the shared storage space; The task evaluation agent is used to obtain the task processing record of the subtask from the shared storage space, and evaluate the completion status of the subtask in combination with the task execution plan to obtain the task evaluation information; The task evaluation information is used to indicate whether there is an error in the execution process of the subtask.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images