Dialogue robot detection method, device, electronic device and storage medium

Abstract

The present application provides a method, device, electronic device and storage medium for detecting a conversation robot, the method comprising: obtaining the conversation configuration information of the conversation robot, the conversation configuration information comprising one or more conversation trees, the conversation tree comprising multiple types of nodes; performing static verification on the conversation configuration information to verify the legality of the conversation configuration information; if the verification result is that the conversation configuration information is legal, traversing all nodes of the conversation tree to generate multiple test cases; the test case is composed of multiple conversation combinations; the conversation combination comprises conversation content generated according to the configuration information of at least two different types of nodes; based on multiple test cases, simulating a user to interact with the conversation robot to obtain multiple test results; generating a detection report according to the multiple test results. The technical solution of the embodiment of the present application realizes a dynamic and static combined conversation robot detection method, reduces the time consumed by the conversation robot detection, and improves the detection efficiency.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a detection method, device, electronic device and storage medium for a conversational robot. Background Art

[0002] With the rapid development of Internet technology, online business processing has involved all walks of life, resulting in the emergence of conversational robots that handle online question-and-answer business. In order to improve the conversational accuracy of conversational robots in the actual business processing process, they are usually tested before they go online to determine whether the conversational robots can perform normal or correct question-and-answer processing.

[0003] Currently, the testing of conversational robots is mainly done manually, that is, manually writing test cases to test them. This testing method is time-consuming and laborious. In addition, for complex intelligent customer robots that include different types of dialogue flows such as question-and-answer dialogues, multi-language dialogues, and greeting dialogues, manual testing may encounter runtime problems such as corpus conflicts and user API call exceptions after going online, which are difficult to detect. Summary of the invention

[0004] One purpose of the present application is to propose a detection method, device, electronic device and storage medium for a conversational robot, so as to realize dynamic and static combined intelligent detection before the conversational robot goes online, and solve the problem of low detection efficiency of the conversational robot.

[0005] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0006] According to one aspect of an embodiment of the present application, a method for detecting a conversational robot is provided, comprising:

[0007] Acquire dialogue configuration information of the dialogue robot, wherein the dialogue configuration information includes one or more dialogue trees, and the dialogue tree includes multiple types of nodes;

[0008] Performing static verification on the dialogue configuration information to verify the legitimacy of the dialogue configuration information;

[0009] If the verification result shows that the dialog configuration information is legal, all nodes of the dialog tree are traversed to generate multiple test cases; the test cases are composed of multiple dialog combinations; the dialog combination includes dialog content generated according to the configuration information of at least two different types of nodes;

[0010] Simulating a conversation between a user and the conversation robot based on the multiple test cases to obtain multiple test results;

[0011] A test report is generated according to the plurality of test results.

[0012] According to one aspect of an embodiment of the present application, a detection device for a conversational robot is provided, comprising:

[0013] An information acquisition module, used to acquire dialogue configuration information of the dialogue robot, wherein the dialogue configuration information includes one or more dialogue trees, and the dialogue tree includes multiple types of nodes;

[0014] A static verification module, used to perform static verification on the dialogue configuration information to verify the legitimacy of the dialogue configuration information;

[0015] A test case generation module, for traversing all nodes of the dialog tree to generate a plurality of test cases if the verification result shows that the dialog configuration information is legal; the test cases are composed of a plurality of dialog combinations; the dialog combinations include dialog contents generated according to the configuration information of at least two different types of nodes;

[0016] A testing module, used to simulate a conversation interaction between a user and the conversation robot based on a plurality of the test cases to obtain a plurality of test results;

[0017] The test report generating module is used to generate a test report according to the multiple test results.

[0018] In one embodiment of the present application, the node types of the dialogue tree include at least two of a user question node, a robot inquiry node, a robot judgment node, and a robot answer node; and the dialogue combination includes:

[0019] The dialogue content generated based on the configuration information of the user question node and the robot inquiry node; or

[0020] The dialogue content generated based on the configuration information of the user question node and the robot answer node; or

[0021] The dialogue content is generated according to the configuration information of the first robot inquiry node, the robot judgment node and the second robot inquiry node; or

[0022] The dialogue content is generated based on the configuration information of the robot inquiry node, robot judgment node, and robot answer node.

[0023] In one embodiment of the present application, the test case generation module is specifically used to:

[0024] Starting from the user question node of the dialogue tree, traversing all nodes of the dialogue tree by depth-first traversal method;

[0025] When traversing to a robot answer node, a test case is generated according to multiple dialogue combinations formed from the user question node to the robot answer node.

[0026] In one embodiment of the present application, the test case generation module is further used to:

[0027] When traversing to a robot judgment node, traverse the dialogue tree branches derived from the robot judgment node by using a breadth-first traversal method;

[0028] For each dialogue tree branch, continue to traverse the remaining nodes using the depth-first traversal method.

[0029] In one embodiment of the present application, the dialogue content in the test case includes simulated input information and expected robot responses, and the device further includes:

[0030] A similar input information generating module, used for generating similar input information of the simulated input information by using a semantic similarity algorithm;

[0031] The extended test case generation module is used to generate an extended test case according to the similar input information and the expected response of the robot.

[0032] In one embodiment of the present application, the dialogue content in the test case includes simulated input information and expected responses of the robot; the test module is specifically used for:

[0033] Conducting a dialogue interaction with the dialogue robot based on the simulated input information to obtain a response result fed back by the dialogue robot;

[0034] If the response result is consistent with the expected response of the robot, the test result obtained is that the test passed;

[0035] If the response result is inconsistent with the expected response of the robot, the test result obtained is a test failure.

[0036] In one embodiment of the present application, the information acquisition module is specifically used to:

[0037] According to the dialogue robot detection request, the task management server creates a robot detection task;

[0038] Adding the robot detection task to the task queue of the message middleware;

[0039] The task execution server reads the robot detection task from the task queue to obtain the dialogue configuration information of the dialogue robot corresponding to the robot detection task.

[0040] According to one aspect of an embodiment of the present application, there is provided an electronic device, including:

[0041] a memory storing computer-readable instructions;

[0042] The processor reads the computer-readable instructions stored in the memory to execute the detection method of the conversational robot in the above technical solution.

[0043] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer executes a method for detecting a conversational robot as in the above technical solution.

[0044] According to one aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the detection method of a conversational robot as in the above technical solution.

[0045] In the technical solution provided in the embodiment of the present application, a dynamic and static combined dialogue robot detection method is implemented before the dialogue robot goes online. The legality of the dialogue configuration information is initially detected through static verification, and then test cases are automatically generated according to the dialogue configuration information, and dynamic detection of the dialogue robot is realized through test cases. In this way, the detection scope of the dialogue robot is more comprehensive, and more complex intelligent customer service robots with multi-round dialogue configurations can also be automatically detected; test cases do not require manual input, which greatly reduces the time spent on dialogue robot detection, improves detection efficiency, and thus improves the dialogue interaction experience between users and dialogue robots.

[0046] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 This is a block diagram of an exemplary system architecture for applying the technical solution of the present application.

[0049] Figure 2 A flowchart of a method for detecting a conversational robot provided in accordance with an embodiment of the present application.

[0050] Figure 3 A structural diagram of a dialogue tree provided for one embodiment of the present application.

[0051] Figure 4 A specific flow chart of step 230 of a method for detecting a conversational robot provided in one embodiment of the present application.

[0052] Figure 5 A specific flow chart of step 230 of a method for detecting a conversational robot provided in another embodiment of the present application.

[0053] Figure 6 A specific flow chart of step 210 of a method for detecting a conversational robot provided in one embodiment of the present application.

[0054] Figure 7 An example diagram of a system architecture for applying the technical solution of the present application is provided for one embodiment of the present application.

[0055] Figure 8 A structural block diagram of a detection device for a conversational robot provided in one embodiment of the present application.

[0056] Fig. 9 A schematic diagram of the structure of an electronic device provided for one embodiment of the present application. DETAILED DESCRIPTION

[0057] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present application will be more comprehensive and complete and the concepts of the example embodiments will be fully conveyed to those skilled in the art. The accompanying drawings are only schematic illustrations of the present application and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted.

[0058] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present application. However, those skilled in the art will appreciate that the technical solution of the present application may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring the present application and making the various aspects of the present application obscure.

[0059] Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0060] Figure 1 The exemplary system architecture block diagram applying the technical solution of the present application is schematically shown.

[0061] like Figure 1 As shown, the system architecture 100 may include a terminal device 110, a network 120, and a server 130. The terminal device 110 may include various electronic devices such as a smart phone, a tablet computer, a laptop computer, and a desktop computer. The server 130 may be an independent physical server, or a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The network 120 may be a communication medium of various connection types that can provide a communication link between the terminal device 110 and the server 130, such as a wired communication link or a wireless communication link.

[0062] According to the implementation requirements, the system architecture in the embodiment of the present application can have any number of terminal devices, networks and servers. For example, the server 130 can be a server group composed of multiple server devices. In addition, the technical solution provided in the embodiment of the present application can be applied to the terminal device 110, can also be applied to the server 130, or can be implemented by the terminal device 110 and the server 130 together, and the present application does not make any special restrictions on this.

[0063] The detection method of the conversational robot provided in this application is described in detail below in conjunction with specific implementation methods.

[0064] Please refer to Figure 2 . Figure 2 A flow chart of a method for detecting a conversational robot provided for one embodiment of the present application. The method for detecting a conversational robot can be implemented through any platform, system or server that uses a conversational robot, such as a terminal device configured with a customer service robot configuration platform, a terminal device configured with an intelligent customer service system, etc.

[0065] like Figure 2 As shown, the detection method of the conversational robot at least includes steps 210 to 250, and the specific contents are as follows:

[0066] Step 210: Obtain the dialogue configuration information of the dialogue robot, where the dialogue configuration information includes one or more dialogue trees, and the dialogue tree includes multiple types of nodes.

[0067] In one embodiment of the present application, the conversation robot can be any robot that can have a conversation with a user, such as an intelligent customer service robot. In general, the conversation robot is constructed through artificial intelligence technology. Artificial Intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that the machines have the functions of perception, reasoning and decision-making.

[0068] Artificial intelligence technology is a comprehensive discipline that covers a wide range of fields, including both hardware-level and software-level technologies. The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0069] In one embodiment of the present application, the dialogue configuration information refers to the dialogue content information when the dialogue robot interacts with the user. The dialogue content between the dialogue robot and the user is the question-and-answer process between the dialogue robot and the user, which is expressed in the form of a dialogue tree. Then the dialogue configuration information generally includes one or more dialogue trees, and the dialogue tree includes multiple types of nodes.

[0070] In one embodiment of the present application, the dialogue configuration information is pre-configured before the dialogue robot goes online, so that the dialogue robot can realize dialogue interaction with the user according to the configured dialogue configuration information after going online. The dialogue robot will use speech technology and natural language processing technology in the process of interaction with the user. The key technologies of speech technology (Speech Technology) include automatic speech recognition technology (ASR), speech synthesis technology (TTS) and voiceprint recognition technology. Allowing computers to listen, see, speak and feel is the future development direction of human-computer interaction, among which speech has become one of the most promising human-computer interaction methods in the future.

[0071] Natural language processing (NLP) is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that can achieve effective communication between people and computers using natural language. Natural language processing is a science that integrates linguistics, computer science, and mathematics. Therefore, research in this field will involve natural language, that is, the language people use in daily life, so it is closely related to the study of linguistics. Natural language processing technology usually includes text processing, semantic understanding, machine translation, robot question answering, knowledge graph and other technologies.

[0072] In order to detect whether the conversational robot can realize conversational interaction with the user according to the conversational configuration information, the conversational robot needs to be tested according to the conversational configuration information before it goes online.

[0073] In one embodiment of the present application, the node types of the dialogue tree are: user question node, robot inquiry node, robot judgment node and robot answer node. The user question node is the starting node of the dialogue tree, and the configuration information of the node is the user question information that starts the dialogue interaction between the dialogue robot and the user. The robot inquiry node is a follow-up node, and the configuration information of the node is the follow-up information of the user. The follow-up information is used to ask the user specific word slot information (it also means that at the robot inquiry node, the user also needs to provide information to the dialogue robot). The follow-up information can also be regarded as the reply provided by the dialogue robot to the user during the dialogue interaction. The robot judgment node is used to control the direction of the dialogue according to the specific word slot information value given by the user, that is, to select a specific dialogue tree branch. The robot answer node is the end node of the dialogue tree, and the configuration information of the node is the final reply provided by the dialogue robot to the user.

[0074] In one embodiment of the present application, a dialog tree should include at least two types of nodes among user question nodes, robot inquiry nodes, robot judgment nodes and robot answer nodes.

[0075] In one embodiment of the present application, in a dialogue tree, the number of nodes of the node type contained therein may be multiple. For example, a dialogue tree includes a user question node and a robot answer node; for another example, a dialogue tree includes a user question node, a robot judgment node and two robot answer nodes; for another example, a dialogue tree includes a user question node, a robot inquiry node, a robot judgment node and three robot answer nodes.

[0076] For example, the structure of a dialogue tree is as follows: Figure 3As shown. The dialogue tree includes a user question node 31, a robot inquiry node 32, a robot judgment node 33 and three robot answer nodes. For the convenience of distinction and description, the three robot answer nodes are respectively recorded as the first robot answer node 341, the second robot answer node 342 and the third robot answer node 343. The configuration information of each node can be directly input or called through the API (Application Programming Interface). For example, when configuring the robot reply "What is your zodiac sign?" of the robot inquiry node 32, you can directly enter the sentence "What is your zodiac sign?" in the robot reply, or you can enter the API call address in the robot reply. For different types of nodes, different types of identifiers can be set for distinction. For the number of each type of node, the configuration personnel can add or delete as needed.

[0077] Step 220: Perform static verification on the dialog configuration information to verify the legitimacy of the dialog configuration information.

[0078] In one embodiment of the present application, static verification is an automatic verification method for verifying the legitimacy of the dialog configuration information, such as whether a field in the dialog configuration information is empty, whether the field length meets the requirements, etc.

[0079] In one embodiment of the present application, if the static check fails, that is, the check result is that the dialogue configuration information is illegal, a test report of the dialogue robot will be directly output, and the test report includes specific matters of illegal dialogue configuration information, such as a field being empty, etc. If the static check passes, that is, the check result is that the dialogue configuration information is legal, then proceed to the next step.

[0080] Exemplary, reference Figure 3 The dialog tree shown. For the robot inquiry node 32, its configuration information includes the robot reply "What's your zodiac sign?". During static verification, it is necessary to verify whether there is this sentence in the robot reply, or whether there is an API call address for this sentence. If no specific robot reply is detected during static verification, a test report is output, indicating the specific test failures, such as no reply in the robot inquiry node 32.

[0081] Step 230: If the verification result shows that the dialog configuration information is legal, all nodes of the dialog tree are traversed to generate multiple test cases; the test cases are composed of multiple dialog combinations; the dialog combination includes dialog content generated according to the configuration information of at least two different types of nodes.

[0082] In one embodiment of the present application, the test case is composed of multiple dialogue combinations, and the dialogue combination is the dialogue content generated according to the configuration information of at least two different types of nodes. The dialogue content of a dialogue combination represents a user input (question) and a robot reply (prompt), the user input is the information provided by the user to the dialogue robot, and the robot reply is the information replied by the dialogue robot to the user. According to the definitions of the various types of nodes mentioned above, the user question node and the robot inquiry node include the user input, and the robot inquiry node and the robot answer node include the robot reply.

[0083] Exemplarily, the dialogue combinations include the following:

[0084] ① The dialogue content is generated based on the configuration information of the user question node and the robot inquiry node.

[0085] This situation refers to the case where the user question configured by the user question node is used as the user input, and the robot reply language configured by the robot inquiry node is used as the robot reply to form a dialogue combination. Figure 3 In the dialogue tree shown, the dialogue combination can be: [I want to check my zodiac personality, what is your zodiac sign?].

[0086] ② The dialogue content is generated based on the configuration information of the user question node and the robot answer node.

[0087] This situation refers to the user question configured in the user question node as the user input, and the final reply language configured in the robot answer node as the robot reply, forming a dialogue combination. For example, the dialogue combination may be: [where is the capital of Guangdong Province, Guangzhou].

[0088] ③ The dialogue content is generated according to the configuration information of the first robot inquiry node, the robot judgment node and the second robot inquiry node.

[0089] This situation means that the parameter of a robot inquiry node is used as the user input, and after passing through one or more robot judgment nodes, the next robot inquiry node is reached, and the robot reply language configured by the next robot inquiry node is used as the robot reply to form a dialogue combination. Among them, the parameter of the robot inquiry node refers to the specific word slot information value extracted from the user's inquiry information after the robot inquiry node asks the user according to the configuration information and obtains the user's inquiry information.

[0090] ④ The dialogue content is generated according to the configuration information of the robot inquiry node, the robot judgment node and the robot answer node.

[0091] This situation refers to the case where the robot query node's parameters are used as user input, and after passing through one or more robot judgment nodes, the robot answer node is reached, and the final reply words configured in the robot answer node are used as the robot reply, forming a dialogue combination. Figure 3 In the dialogue tree shown, the robot asks the user "What's your zodiac sign?" at the robot inquiry node 32, and obtains the user input "Aries", which is the parameter of the robot inquiry node 32. The robot judgment node 33 jumps to the first robot answer node 341 according to the parameter "Aries" of the robot inquiry node 32, and the first robot answer node 341 gives the robot reply "Aries is simple and kind". The dialogue combination can be: [Aries, Aries is simple and kind].

[0092] In one embodiment of the present application, for a dialogue tree, from the starting node to the ending node of the dialogue tree is equivalent to a complete dialogue interaction between the user and the dialogue robot, so the multiple dialogue combinations formed from the starting node to the ending node constitute a test case. Then, after the static verification is passed, when traversing all the nodes of the dialogue tree, starting from the user question node, when traversing to a robot answer node, the multiple dialogue combinations formed from the user question node to the robot answer node constitute a test case. In general, traversing all the nodes of a dialogue tree can generate multiple test cases.

[0093] Exemplary, reference Figure 3 The dialog tree shown. Starting from the user question node 31, the robot inquiry node 32 and the robot judgment node 33 are traversed in sequence. When traversing to the first robot answer node 341, a test case is obtained: {[I want to check the zodiac personality, what zodiac sign are you], [Aries, Aries are simple and kind]}. When traversing to the second robot answer node 342, a test case is obtained: {[I want to check the zodiac personality, what zodiac sign are you], [Libra, Libra pursues beauty]}. When traversing to the third robot answer node 343, a test case is obtained: {[I want to check the zodiac personality, what zodiac sign are you], [Scorpio, Scorpio is independent]}.

[0094] Step 240: Simulate the conversational interaction between the user and the conversational robot based on multiple test cases to obtain multiple test results.

[0095] In one embodiment of the present application, based on the generated test cases, the conversation interaction between the user and the conversation robot can be simulated, and whether the conversation interaction function of the conversation robot is abnormal can be detected by the response effect of the conversation robot in the simulated conversation interaction situation.

[0096] In one embodiment of the present application, the conversation content of the test case also includes user input and robot response as a whole. The user input in the test case is information that simulates the user input, which is called simulated input information; the robot response in the test case is the correct response that the conversation robot should give based on the corresponding user input, which is called the robot expected response. According to the test case, the conversation interaction between the user and the conversation robot is simulated, that is, the simulated input information of the test case is used as the user input information of the conversation robot, the response result fed back by the conversation robot based on the simulated input information is obtained, and the response result is compared with the robot expected response of the test case to obtain the test result. If the response result is consistent with the expected response of the robot, the test result is a test pass; if the response result is inconsistent with the expected response of the robot, the test result is a test failure.

[0097] For example, for the test case: {[I want to check my zodiac sign, what is your zodiac sign], [Aries, Aries is simple and kind]}, the simulated input information is "I want to check my zodiac sign" and "Aries", and the robot's expected response is "Aries is simple and kind". If the simulated input information is "I want to check my zodiac sign" and "Aries", and the dialogue robot's response is "Scorpio is independent", the test result is a test failure.

[0098] In one embodiment of the present application, a conversational interaction between a user and a conversational robot is simulated based on a test case, that is, a full-link verification of the end-to-end call of the conversational robot is performed. For a conversational interaction between a user and a conversational robot, the end where the user inputs information to the conversational robot is the request end of the conversational interaction, and the end where the conversational robot obtains the response result is the result end of the conversational interaction. From user input (request end) to the result return of the conversational robot (result end), this process usually involves a long service link and multiple API calls, such as a service link including access layer services, NLU (Natural Language Understanding) services, skill routing services, conversation management services, etc., and each type of service contains multiple microservices. Any problem in any call link may cause the test to fail. The call verification between various services in the call link from the request end to the result end and the various API call verifications are called end-to-end call full-link verification.

[0099] In one embodiment of the present application, after generating the test case, it also includes: generating similar input information of the simulated input information through a semantic similarity algorithm; generating an extended test case according to the similar input information and the expected response of the robot. That is, similar input information similar to the simulated input information in the test case is generated through a semantic similarity algorithm, and the robot response corresponding to the similar input information is still the same as the expected response of the robot corresponding to the simulated input information, then the extended test case can be generated according to the similar input information and the expected response of the robot. Exemplarily, for the test case: {[I want to check the zodiac personality, what zodiac sign are you], [Aries, Aries personality is simple and kind]}, the simulated input information is "I want to check the zodiac personality". Through the semantic similarity algorithm, similar input information "How to check the zodiac personality", "What is the zodiac personality like", etc. can be generated, thereby generating extended test cases: {[How to check the zodiac personality, what zodiac sign are you], [Aries, Aries personality is simple and kind]}, {[What is the zodiac personality like, what zodiac sign are you], [Aries, Aries personality is simple and kind]}.

[0100] In this way, user input can be generalized, the number of test cases for the conversational robot can be automatically increased, the test set of the conversational robot can be expanded, and the conversational robot can perform more comprehensive detection and improve the detection effect.

[0101] Step 250: Generate a test report based on the multiple test results.

[0102] In one embodiment of the present application, when all test results are tested, it is indicated that the question-and-answer performance of the dialogue robot is normal, and the test report generated at this time indicates that the dialogue robot passes the test. When any one of the multiple test results is a test failure, the generated test report is that the dialogue robot fails the test, and the node type of the test failure, the specific content of the test failure, and the modification suggestions are reflected in the test report. The node type of test failure refers to the node type with an abnormality, such as a robot inquiry node. The specific content of the test failure is the specific abnormal content, such as an API interface exception, a slot failure (such as a questioning failure), etc.

[0103] In one embodiment of the present application, after the test report is generated, it can be stored in a database for reference and use at any time.

[0104] In the technical solution of the embodiment of the present application, a dynamic and static combined dialogue robot detection method is implemented before the dialogue robot goes online. The legality of the dialogue configuration information is initially detected through static verification, and then test cases are automatically generated according to the dialogue configuration information, and dynamic detection of the dialogue robot is realized through test cases. In this way, the detection scope of the dialogue robot is more comprehensive, and more complex intelligent customer service robots with multi-round dialogue configurations can also be automatically detected; test cases do not require manual input, which greatly reduces the time spent on dialogue robot detection, improves detection efficiency, and thus improves the dialogue interaction experience between users and dialogue robots.

[0105] Please refer to Figure 4 . Figure 4 The specific flow chart of step 230 of the method for detecting a conversational robot according to an embodiment of the present application is shown. Step 230 may specifically include steps 2311 to 2312, which are as follows:

[0106] Step 2311, starting from the user question node of the dialogue tree, traverse all nodes of the dialogue tree using the depth-first traversal method.

[0107] In one embodiment of the present application, when traversing all nodes of the dialog tree, a depth-first search (DFS) method is used, and the traversal is performed with the user question node of the dialog tree as the starting node.

[0108] Step 2312: When traversing to a robot answer node, a test case is generated based on multiple dialogue combinations formed from the user question node to the robot answer node.

[0109] In one embodiment of the present application, during the traversal process, a dialog combination consisting of a user input and a robot reply is recorded. When a robot answer node is traversed, multiple dialog combinations from the user question node to the robot answer node are combined to generate a test case. When there are multiple robot answer nodes in a dialog tree, multiple test cases are obtained accordingly.

[0110] Exemplary, reference Figure 3 The dialog tree shown in FIG. 1 is shown. Starting from the user question node 31, the robot inquiry node 32, the robot judgment node 33 and the first robot answer node 341 are traversed in sequence by the depth-first traversal method to obtain a test case. Then, the robot judgment node 33 is returned to be traversed, and the second robot answer node 342 is traversed to obtain a test case. The robot judgment node 33 is returned again, and the third robot answer node 343 is traversed to obtain a test case.

[0111] By traversing all nodes of the dialogue tree through the depth-first traversal method, we can prevent nodes from being missed during the traversal process, making the generated test cases comprehensive and accurate, improving the accuracy of the test set during dialogue robot detection, and thus improving the detection quality.

[0112] Please refer to Figure 5 . Figure 5 A specific flow chart of step 230 of a method for detecting a conversational robot according to another embodiment of the present application is shown. Step 230 may specifically include steps 2321 to 2324, which are as follows:

[0113] Step 2321, starting from the user question node of the dialogue tree, traverse all nodes of the dialogue tree using the depth-first traversal method.

[0114] In one embodiment of the present application, when traversing all nodes of the dialog tree, a depth-first traversal method is adopted, and the traversal is performed with the user question node of the dialog tree as the starting node.

[0115] Step 2322: When traversing to a robot judgment node, traverse the dialogue tree branches derived from the robot judgment node using a breadth-first traversal method.

[0116] In one embodiment of the present application, in the process of traversing the dialogue tree nodes by the depth-first traversal method, when traversing to the robot judgment node, the robot judgment node is usually followed by multiple dialogue tree branches. At this time, the breadth-first traversal (Breadth First Search, BFS) method is used to traverse the dialogue tree branches induced by the robot judgment node.

[0117] Step 2323: For each dialogue tree branch, continue to traverse the remaining nodes using the depth-first traversal method.

[0118] In one embodiment of the present application, for each dialog tree branch, the remaining nodes of the dialog tree branch are traversed continuously by a depth-first traversal method.

[0119] Step 2324: When traversing to a robot answer node, generate a test case based on multiple dialogue combinations formed from the user question node to the robot answer node.

[0120] In one embodiment of the present application, during the traversal of the remaining nodes of the dialogue tree branch, a dialogue combination consisting of a user input and a robot reply is obtained. When a robot answer node is traversed, multiple dialogue combinations from the user question node to the robot answer node are combined to generate a test case. Figure 3The dialog tree shown in FIG. 1 starts from the user question node 31 and traverses the robot inquiry node 32 and the robot judgment node 33 in sequence. When traversing to the robot judgment node 33, the robot judgment node 33 leads to three branches (branch 1, branch 2, and branch 3). At this time, the first robot answer node 341, the second robot answer node 342, and the third robot answer node 343 are traversed using the breadth-first traversal method to obtain three test cases respectively.

[0121] In one embodiment of the present application, during the traversal of the remaining nodes of the dialogue tree branch, if the robot judgment node is traversed again, the dialogue tree sub-branch induced by the robot judgment node is traversed again using the breadth-first traversal method, and then the remaining nodes of the dialogue tree sub-branch are continued to be traversed in depth-first manner.

[0122] In one embodiment of the present application, when traversing to a user question node, the node is the starting node, and user input information can be simulated and generated according to the configuration information of the node; Figure 3 The dialog tree shown can generate user input when traversing the user question node 31. When traversing to the robot inquiry node, the robot reply is obtained according to the configuration information of the node, and the slot information and reference dictionary expected by the node for the user to input can be obtained. The slot information is equivalent to the parameter name. For example, if the user asks questions related to constellations, the slot information can be the name of the constellation; the reference dictionary corresponds to the slot information, which refers to the location or database where the dialog robot processes the information required for the current dialog interaction. For example, the slot information is the name of the constellation, and the reference dictionary is the constellation. When traversing to the robot judgment node, the word slot information value filled in the slot information of the previous robot inquiry node can be obtained according to the configuration information of the node. At this time, it is judged whether the word slot information value is in the reference dictionary to determine whether the word slot information value meets the requirements. When traversing to the robot answer node, the robot's final reply information is obtained according to the configuration information of the node; and when traversing to the robot answer node, it means that a complete dialog interaction is over, and a test case is obtained.

[0123] In the technical solution of the embodiment of the present application, through the depth-first traversal of the dialogue tree combined with the breadth-first traversal of the robot judgment node traversal method, all nodes of the dialogue tree can be traversed quickly. When traversing to the robot judgment node, there is no need to fall back, which speeds up the node traversal speed and the test case generation speed, thereby improving the detection efficiency of the dialogue robot.

[0124] Please refer to Figure 6 . Figure 6 The specific flow chart of step 210 of the method for detecting a conversational robot according to an embodiment of the present application is shown. Step 210 specifically includes steps 211 to 213, which are described in detail as follows:

[0125] Step 211: According to the dialogue robot detection request, the task management server creates a robot detection task.

[0126] In one embodiment of the present application, a conversational robot detection request is initiated by the customer service robot configuration platform and sent to the task management server. After the task management server detects the conversational robot detection request, it creates a robot detection task. Figure 7 As shown, the task management server 72 includes a task scheduling framework, which has two processes: a server process and a worker process. The robot detection task is created by the server process. The worker process is used to manage the status tasks of the robot detection task and provide an API interface for task status query for the customer service robot configuration platform 71. The status tasks are divided into the following categories: Figure 7 The five states shown are: CREATED (creation), STARTED (execution), SUCCESS (successful execution), RETRY (retry), and FAILURE (failed execution).

[0127] In one embodiment of the present application, Figure 7 The following schematically shows an example of a system architecture that uses the technical solution of this application. Figure 7 As shown, the customer service robot configuration platform 71 is the initiator of the dialogue robot detection request. The customer service robot configuration platform 71 sends the dialogue robot detection request to the task management server 72, and the server process of the task management server 72 creates a robot detection task.

[0128] Step 212: Add the robot detection task to the task queue of the message middleware.

[0129] In one embodiment of the present application, after the robot detection task is created, the robot detection task is written into the task queue of the message middleware so that the subsequent process can obtain the robot detection task execution in an orderly manner from the task queue. At the same time, the message middleware will record the task status of the robot detection task, that is, whether the task is completed, so that the customer service robot configuration platform can query the task status. For example, Figure 7 As shown, after the task management server 72 creates the robot detection task, it writes it into the task queue of the message middleware 73.

[0130] Step 213: The task execution server reads the robot detection task from the task queue to obtain the dialogue configuration information of the dialogue robot corresponding to the robot detection task.

[0131] In one embodiment of the present application, the task execution server reads and executes the robot detection task from the task queue, and obtains the corresponding dialogue robot's dialogue configuration information according to the information carried by the robot detection task. Figure 7 As shown, the task execution server 74 reads and executes the robot detection task from the task queue, thereby obtaining the dialogue configuration information and executing the dialogue robot detection method provided in the above embodiment of the present application.

[0132] In one embodiment of the present application, Figure 7 As shown, in the detection method of the dialogue robot executed by the task execution server 74, after obtaining the dialogue configuration information, it can also include steps 741 to 744, which are as follows:

[0133] Step 741: Static verification of the dialog configuration information.

[0134] Specifically, the static verification of the dialogue configuration information is to verify the legality of the dialogue configuration information, such as whether a field in the dialogue configuration information is empty, whether the field length meets the requirements, etc. If the static verification fails, a test report of the dialogue robot will be directly output, and the test report includes specific matters of the illegal dialogue configuration information, such as whether a field is empty, etc. If the static verification passes, the process proceeds to step 742.

[0135] Exemplarily, a test report for a static check failure is shown in Table 1. In Table 1, the answer node refers to the robot answer node, the inquiry node refers to the robot inquiry node, and the judgment node refers to the robot judgment node.

[0136] Table 1

[0137]

[0138] Step 742: Generate test cases based on the configuration.

[0139] Specifically, the dialogue configuration information refers to the dialogue content information when the dialogue robot interacts with the user, which is expressed in the form of a dialogue tree, and the dialogue tree includes multiple types of nodes. When the dialogue configuration information passes the static verification, all nodes of the dialogue tree are traversed to generate test cases. The test case consists of multiple dialogue combinations, and the dialogue combination is the dialogue content generated according to the configuration information of at least two different types of nodes. The dialogue content of a dialogue combination represents a user input (question) and a robot reply (prompt). The user input is the information provided by the user to the dialogue robot, and the robot reply is the information replied by the dialogue robot to the user.

[0140] The dialogue tree includes a user question node, a robot inquiry node, a robot judgment node, and a robot answer node. When traversing the dialogue tree, starting from the user question node, when traversing to the robot answer node, a test case is generated according to the multiple dialogue combinations formed from the user question node to the robot answer node. When a dialogue tree includes multiple robot answer nodes, multiple test cases can be generated. The generated test cases include the simulated input information of the user request and the expected response of the robot.

[0141] Step 743: End-to-end call for full-link verification.

[0142] Specifically, the end-to-end full-link verification is to simulate the conversational interaction between the user and the conversational robot according to the generated test case, and detect whether the response result of the conversational robot in the simulated interaction is consistent with the expected response of the robot in the test case. If the response result of the conversational robot is consistent with the expected response of the robot, it means that the end-to-end full-link verification has passed; if the response result of the conversational robot is inconsistent with the expected response of the robot, it means that an error occurred during the end-to-end full-link verification.

[0143] Step 744: Generate detection results.

[0144] Specifically, the test result is a test report, which is generated based on the verification results of the end-to-end full-link verification of all test cases. If all test cases pass the end-to-end full-link verification, it means that the test is passed and there is no abnormality in the conversational robot; if any test case fails the end-to-end full-link verification, the test result needs to output the specific information of the end-to-end full-link verification error, including error type, error message, modification suggestions, etc. Exemplarily, the test report is shown in Table 2. Among them, the answer node in Table 1 refers to the robot answer node, the inquiry node refers to the robot inquiry node, and the judgment node refers to the robot judgment node.

[0145] Table 2

[0146]

[0147] In one embodiment of the present application, Figure 7 As shown, after the test results are generated, they are also saved in the database 75, so that the test results can be queried and called at any time.

[0148] In one embodiment of the present application, the detection method of the conversational robot can be executed by a system composed of a customer service robot configuration platform, a task management server, a message middleware and a task execution server, or it can be executed by a task execution server alone, and the present application does not impose any restrictions.

[0149] In one embodiment of the present application, the task management server 72, the message middleware 73 and the task execution server 74 will all record the task status, and the test personnel can query the task status through the customer service robot configuration platform 71. When the test personnel initiate a task status query request on the customer service robot configuration platform 71, the task execution server 74, the message middleware 73 and the task management server 72 will report the task status in turn, and finally the task status is fed back to the customer service robot configuration platform 71 through the task status query API interface provided by the Worker process for the customer service robot configuration platform 71.

[0150] It should be noted that although the steps of the method in the present application are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0151] The following introduces an apparatus embodiment of the present application, which can be used to execute the detection method of the conversational robot in the above-mentioned embodiment of the present application. Figure 8 The structure block diagram of the detection device of the dialogue robot provided in the embodiment of the present application is schematically shown. Figure 8 As shown, a detection device for a conversational robot provided by one embodiment of the present application includes:

[0152] An information acquisition module 810 is used to acquire the dialogue configuration information of the dialogue robot, wherein the dialogue configuration information includes one or more dialogue trees, and the dialogue tree includes multiple types of nodes;

[0153] A static verification module 820, configured to perform a static verification on the dialog configuration information to verify the legitimacy of the dialog configuration information;

[0154] The test case generation module 830 is used to traverse all nodes of the dialog tree to generate multiple test cases if the verification result shows that the dialog configuration information is legal; the test cases are composed of multiple dialog combinations; the dialog combinations include dialog content generated according to the configuration information of at least two different types of nodes;

[0155] A testing module 840 is used to simulate a conversation interaction between a user and the conversation robot based on a plurality of the test cases to obtain a plurality of test results;

[0156] The test report generating module 850 is used to generate a test report according to the multiple test results.

[0157] In one embodiment of the present application, the node types of the dialogue tree include at least two of a user question node, a robot inquiry node, a robot judgment node, and a robot answer node; and the dialogue combination includes:

[0158] The dialogue content generated based on the configuration information of the user question node and the robot inquiry node; or

[0159] The dialogue content generated based on the configuration information of the user question node and the robot answer node; or

[0160] The dialogue content is generated according to the configuration information of the first robot inquiry node, the robot judgment node and the second robot inquiry node; or

[0161] The dialogue content is generated based on the configuration information of the robot inquiry node, robot judgment node, and robot answer node.

[0162] In one embodiment of the present application, the test case generation module 830 is specifically used to:

[0163] Starting from the user question node of the dialogue tree, traversing all nodes of the dialogue tree by depth-first traversal method;

[0164] When traversing to a robot answer node, a test case is generated according to multiple dialogue combinations formed from the user question node to the robot answer node.

[0165] In one embodiment of the present application, the test case generation module 830 is further used to:

[0166] When traversing to a robot judgment node, traverse the dialogue tree branches derived from the robot judgment node by using a breadth-first traversal method;

[0167] For each dialogue tree branch, continue to traverse the remaining nodes using the depth-first traversal method.

[0168] In one embodiment of the present application, the dialogue content in the test case includes simulated input information and expected robot responses, and the device further includes:

[0169] A similar input information generating module, used for generating similar input information of the simulated input information by using a semantic similarity algorithm;

[0170] The extended test case generation module is used to generate an extended test case according to the similar input information and the expected response of the robot.

[0171] In one embodiment of the present application, the dialogue content in the test case includes simulated input information and expected responses of the robot; the test module 840 is specifically used to:

[0172] Conducting a dialogue interaction with the dialogue robot based on the simulated input information to obtain a response result fed back by the dialogue robot;

[0173] If the response result is consistent with the expected response of the robot, the test result obtained is that the test passed;

[0174] If the response result is inconsistent with the expected response of the robot, the test result obtained is a test failure.

[0175] In one embodiment of the present application, the information acquisition module 810 is specifically used to:

[0176] According to the dialogue robot detection request, the task management server creates a robot detection task;

[0177] Adding the robot detection task to the task queue of the message middleware;

[0178] The task execution server reads the robot detection task from the task queue to obtain the dialogue configuration information of the dialogue robot corresponding to the robot detection task.

[0179] Please refer to Fig. 9 . Fig. 9 A schematic diagram of the structure of an electronic device provided for one embodiment of the present application. Fig. 9 The electronic device 90 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0180] like Fig. 9 As shown, the electronic device 90 is in the form of a general computing device. The components of the electronic device 90 may include but are not limited to: at least one processing unit 910, at least one storage unit 920, and a bus 930 connecting different system components (including the storage unit 920 and the processing unit 910).

[0181] The storage unit stores program codes, which can be executed by the processing unit 910, so that the processing unit 910 performs the steps according to various exemplary embodiments of the present invention described in the description of the exemplary method described above in this specification. For example, the processing unit 910 can perform the following steps: Figure 1 The steps shown in .

[0182] The storage unit 920 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 9201 and / or a cache storage unit 9202 , and may further include a read-only storage unit (ROM) 9203 .

[0183] The storage unit 920 may also include a program / utility 9204 having a set (at least one) of program modules 9205, such program modules 9205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0184] Bus 930 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0185] The electronic device 90 may also communicate with one or more external devices 10 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 90, and / or communicate with any device that enables the electronic device 90 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 950. The input / output (I / O) interface 950 is connected to a display unit 940. Furthermore, the electronic device 90 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 960. As shown, the network adapter 960 communicates with other modules of the electronic device 90 via a bus 930. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 90, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0186] In an exemplary embodiment of the present application, a computer-readable storage medium is further provided, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer executes the method described in the above method embodiment.

[0187] According to one embodiment of the present application, a program product for implementing the method in the above method embodiment is also provided, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited to this. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0188] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0189] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0190] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0191] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as JAVA, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0192] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0193] In addition, although the steps of the method in the present application are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0194] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation methods of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the implementation methods of the present application.

[0195] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the appended claims.

Claims

1. A method for detecting a conversational robot, characterized in that: The method comprises: Acquire dialogue configuration information of the dialogue robot, wherein the dialogue configuration information includes one or more dialogue trees, wherein the dialogue tree includes multiple types of nodes; the node types of the dialogue tree include at least two of a user question node, a robot inquiry node, a robot judgment node, and a robot answer node; Performing static verification on the dialogue configuration information to verify the legitimacy of the dialogue configuration information; If the verification result shows that the dialogue configuration information is legal, all nodes of the dialogue tree are traversed to generate multiple test cases; the test case is composed of multiple dialogue combinations; the dialogue combination includes dialogue content generated according to the configuration information of at least two different types of nodes; wherein, in the process of traversing all nodes of the dialogue tree: starting from the user question node of the dialogue tree, all nodes of the dialogue tree are traversed by depth-first traversal method; when traversing to a robot judgment node, the dialogue tree branches derived from the robot judgment node are traversed by breadth-first traversal method; for each dialogue tree branch, the remaining nodes are traversed by depth-first traversal method; when traversing to a robot answer node, a test case is generated according to multiple dialogue combinations formed from the user question node to the robot answer node; Simulating a conversation between a user and the conversation robot based on the multiple test cases to obtain multiple test results; A test report is generated according to the plurality of test results.

2. The method for detecting a conversational robot according to claim 1, characterized in that: The dialogue combination includes: The dialogue content generated based on the configuration information of the user question node and the robot inquiry node; or The dialogue content generated based on the configuration information of the user question node and the robot answer node; or The dialogue content is generated according to the configuration information of the first robot inquiry node, the robot judgment node and the second robot inquiry node; or The dialogue content is generated based on the configuration information of the robot inquiry node, robot judgment node, and robot answer node.

3. The method for detecting a conversational robot according to claim 1, characterized in that: The dialogue content in the test case includes simulated input information and expected responses of the robot. After traversing all nodes of the dialogue tree and generating multiple test cases, the method further includes: Generate similar input information of the simulated input information by using a semantic similarity algorithm; An extended test case is generated according to the similar input information and the expected response of the robot.

4. The method for detecting a conversational robot according to claim 1, characterized in that: The conversation content in the test case includes simulated input information and expected responses of the robot; The multiple test results obtained by simulating a user to interact with the conversation robot based on the multiple test cases include: Conducting a dialogue interaction with the dialogue robot based on the simulated input information to obtain a response result fed back by the dialogue robot; If the response result is consistent with the expected response of the robot, the test result obtained is that the test passed; If the response result is inconsistent with the expected response of the robot, the test result obtained is a test failure.

5. The method for detecting a conversational robot according to any one of claims 1 to 4, characterized in that: The obtaining of the dialogue configuration information of the dialogue robot includes: According to the dialogue robot detection request, the task management server creates a robot detection task; Adding the robot detection task to the task queue of the message middleware; The task execution server reads the robot detection task from the task queue to obtain the dialogue configuration information of the dialogue robot corresponding to the robot detection task.

6. A detection device for a conversational robot, characterized in that: include: An information acquisition module, used to acquire dialogue configuration information of the dialogue robot, wherein the dialogue configuration information includes one or more dialogue trees, wherein the dialogue tree includes multiple types of nodes; the node types of the dialogue tree include at least two of a user question node, a robot inquiry node, a robot judgment node, and a robot answer node; A static verification module, used to perform static verification on the dialogue configuration information to verify the legitimacy of the dialogue configuration information; A test case generation module, for traversing all nodes of the dialogue tree and generating multiple test cases if the verification result shows that the dialogue configuration information is legal; the test case is composed of multiple dialogue combinations; the dialogue combination includes dialogue content generated according to the configuration information of at least two different types of nodes; wherein, in the process of traversing all nodes of the dialogue tree: starting from the user question node of the dialogue tree, traversing all nodes of the dialogue tree by depth-first traversal method; when traversing to a robot judgment node, traversing the dialogue tree branches derived from the robot judgment node by breadth-first traversal method; for each dialogue tree branch, continuing to traverse the remaining nodes by depth-first traversal method; when traversing to a robot answer node, generating a test case according to multiple dialogue combinations formed from the user question node to the robot answer node; A testing module, used to simulate a conversation interaction between a user and the conversation robot based on a plurality of the test cases to obtain a plurality of test results; The test report generating module is used to generate a test report according to the multiple test results.

7. An electronic device, characterized in that: include: a memory storing computer-readable instructions; A processor reads computer-readable instructions stored in a memory to execute the detection method for a conversational robot according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and when the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the detection method for a conversational robot according to any one of claims 1 to 5.

9. A computer program product, characterized in that The computer program product includes computer instructions stored in a computer-readable storage medium; The processor of the computer device reads and executes the computer instructions from the computer-readable storage medium, so that the computer device executes the detection method of the conversational robot described in any one of claims 1 to 5.

Images