CN114519005A - Method, device and equipment for testing response text and storage medium - Google Patents

Abstract

The invention relates to an artificial intelligence technology, and discloses a method for testing a response text, which comprises the following steps: acquiring test event information set based on an actual service scene; based on a full combination algorithm, carrying out permutation and combination on the intention dictionary corresponding to the dialogue model, and splicing the test event information and the result of the permutation and combination to obtain a test intention scene table; selecting the test intention data in the test intention scene table one by one as a current test scene according to the sequence, and calling a dialogue model through an interface to generate a response text and configuration parameters; and carrying out verification analysis on the response text and the configuration parameters based on a text similarity algorithm to obtain a test result. In addition, the invention also relates to a block chain technology, and the test scene diagram can be stored in the node of the block chain. The invention also provides a test device, electronic equipment and a storage medium for the answer operation. The invention can solve the problems of low test efficiency and incomplete test scene of the response text.

Description

Test method, device, equipment and storage medium for response text

technical field

The present invention relates to the technical field of artificial intelligence, and in particular, to a test method, device, electronic device and computer-readable storage medium for answering text.

Background technique

With the continuous development of artificial intelligence technology, intelligent dialogue systems have been widely used, and more and more scenarios will use anthropomorphic AI customer service instead of human customer service to conduct dialogues to collect customer information and answer customer questions.

However, for the test of response text in the current dialogue system, manual testing is often used. Due to the multiple rounds of dialogue, there are many branch nodes, and the differences in the length of sentence segmentation and silence under different parent nodes lead to a large amount of testing workload. , the test time is long, and the test scene is prone to be incomplete. Therefore, a test method that improves the efficiency and ensures the comprehensiveness of the test scene is required.

SUMMARY OF THE INVENTION

The present invention provides a method, device and computer-readable storage medium for testing response text, the main purpose of which is to solve the problems of low testing efficiency and incomplete testing scenarios of response text.

To achieve the above object, a kind of test method of response text provided by the present invention comprises:

Obtain test event information set based on an actual business scenario, wherein the test event information includes a dialog model name;

Arrange and combine the intent dictionary corresponding to the dialog model name based on the full combination algorithm, and splicing the test event information with the result of the arrangement and combination to obtain a test intent scene table;

Select the test intent data in the test intent scene table one by one as the current test scene in sequence, and call the dialog model corresponding to the dialog model name through the interface to generate the response text and configuration parameters of the current test scene;

The response text and configuration parameters are verified and analyzed based on a text similarity algorithm to obtain a test result.

Optionally, the acquiring test event information set based on an actual business scenario includes:

Obtain the set test event table preset according to the actual business scenario;

Set test data according to the business function of the dialogue system and fill in the test event table to obtain test event information.

Optionally, the intent dictionary corresponding to the dialog model name is arranged and combined based on the full combination algorithm, and the test event information is spliced with the result of the arrangement and combination to obtain a test intent scene table, including:

Obtain a corresponding intent dictionary according to the dialog model name;

Using a full combination algorithm to arrange and combine the intents of the intent dictionary according to the number of dialogue polls in the test event information;

Using a data manipulation tool, each combination in the result of the arrangement and combination is spliced with the test event information to obtain a test intent scene table.

Optionally, the use of a full combination algorithm to arrange and combine the total number of intents of the intent dictionary according to the number of dialogue polls in the test event information, including:

Obtain the number of dialog polls in the test event information and the total number of intents in the intent dictionary;

Use all intents of the intent dictionary as candidate intents for each dialogue;

Selecting the same number of intent combinations as the number of dialogue polling times each time, wherein, in each combination, one intent is selected from the candidate intents of each round in turn, and recursively arranging and combining according to the number of dialogue polling times , get the result of permutation and combination.

Optionally, generating the response text and configuration parameters of the current test scenario by connecting the dialogue model through an interface, including:

Generate a dialog request corresponding to the current test scenario based on the request;

Connect the dialogue system through the interface according to the dialogue request, and call the dialogue model with the dialogue model name in the dialogue system;

According to the dialog request, the test data in the current test scene is transferred to the dialog model and the response text generated by the dialog model is obtained;

Acquire the current configuration parameters of the dialogue system, and return the configuration parameters and the response text to the tester through the interface.

Optionally, the response text and configuration parameters are verified and analyzed based on the text similarity algorithm to obtain a test result, including:

Compare the response text and configuration parameters with the data of the corresponding field in the test intent scene table of the current test scene;

If the comparison is inconsistent, it means that the verification fails, and the inconsistent content is written into the test intent scene table by using the data operation;

Connecting to the database to obtain a preset drop-off record, and looking up the actual drop-off intention corresponding to the current test scene in the drop-off record;

Detecting the consistency between the actual drop-off intention and the preset intention of the current test scene based on a text similarity algorithm, and if they are inconsistent, it means that the current test scene is abnormal;

After splicing the verification failure and the abnormality as test results in the test scene table corresponding to the test scenes in the test scene table, and recording the test results of the remaining test scenes as normal, the test scene is obtained Table of full test results.

Optionally, the detection of the consistency between the actual deposit intention and the preset intention of the current test scene based on the text similarity algorithm includes:

Using a tokenizer to segment the actual storage intention and the preset intention, respectively, to obtain the actual segmentation sequence and the preset segmentation sequence;

Calculate the string similarity according to the actual word segmentation sequence and the preset word segmentation sequence;

When the string similarity is greater than or equal to a preset threshold, judging that the actual storage intention is consistent with the preset intention of the current test scene;

When the string similarity is less than a preset threshold, it is determined that the actual drop-in intention is inconsistent with the preset intention of the current test scene.

In order to solve the above problems, the present invention also provides a test device for answering text, the device comprising:

an event information acquisition module, configured to acquire test event information set based on an actual business scenario, wherein the test event information includes a dialog model name;

A test scene acquisition module, used for arranging and combining the intent dictionary corresponding to the dialogue model name based on a full combination algorithm, and splicing the test event information and the result of the permutation and combination to obtain a test intent scene table;

The response generation module is used to select the test intent data in the test intent scene table as the current test scene one by one in sequence, and generate the response text and configuration parameters;

The verification and analysis module is configured to perform verification and analysis on the response text and configuration parameters based on a text similarity algorithm to obtain a test result.

In order to solve the above problems, the present invention also provides an electronic device, the electronic device includes:

at least one processor; and,

a memory communicatively coupled to the at least one processor; wherein,

The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the above-mentioned method for testing response texts .

In order to solve the above problems, the present invention also provides a computer-readable storage medium, where at least one computer program is stored in the computer-readable storage medium, and the at least one computer program is executed by a processor in an electronic device to realize the above-mentioned The test method for the response text described above.

The embodiment of the present invention is based on arranging and combining the intent dictionaries corresponding to the dialogue model names based on the full combination algorithm, and splicing the test event information and the result of the permutation and combination to obtain a test intent scene table, where the intent scene table is The full combination is obtained, which ensures the comprehensiveness of the test scene; at the same time, the test scene is connected to the dialogue system by calling the interface, and the test scene is simulated as the input user question and answer text to obtain the response of the dialogue system and related configuration parameter information, and the response is verified by verifying the response. The correctness of the text and related configuration parameters is used to verify whether the node flow of the dialogue system is correct, and the verification of the intent output result of the node model is added. The accuracy of the results, and automated testing of the dialogue system to improve efficiency. Therefore, the test method, device, electronic device and computer-readable storage medium of the response text proposed by the present invention can solve the problems of low test efficiency and incomplete test scenarios of the response text.

Description of drawings

1 is a schematic flowchart of a method for testing a response text provided by an embodiment of the present invention;

2 is a schematic flowchart of a test scenario generation provided by an embodiment of the present invention;

3 is a functional block diagram of a device for testing response text provided by an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of an electronic device implementing the method for testing the response text according to an embodiment of the present invention.

The realization, functional characteristics and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments.

Detailed ways

It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

The embodiment of the present application provides a method for testing response text. The execution subject of the response text testing method includes, but is not limited to, at least one of electronic devices that can be configured to execute the method provided by the embodiments of the present application, such as a server and a terminal. In other words, the test method of the response text can be executed by software or hardware installed in the terminal device or the server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, and the like. The server can be an independent server, or can provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery network (Content Delivery Network) , CDN), and cloud servers for basic cloud computing services such as big data and artificial intelligence platforms.

Referring to FIG. 1 , it is a schematic flowchart of a method for testing response text according to an embodiment of the present invention. In this embodiment, the test method of the response text includes:

S1. Acquire test event information set based on an actual business scenario, where the test event information includes a dialog model name.

The test event information in this embodiment of the present invention is data information that is set according to actual business scenarios and can cause the dialogue system to respond to dialogue, including but not limited to driving events, vocabulary text, dialogue polling times, dialogue model name, sentence segmentation duration, Silence time.

In detail, the acquisition of test event information set based on actual business scenarios includes:

Obtain the set test event table preset according to the actual business scenario;

Set test data according to the business function of the dialogue system and fill in the test event table to obtain test event information.

Wherein, the test event table is the event information of the driving dialogue system and the corresponding configuration information, and the fields contained in the event information include the driving event, the event language text, the number of dialogue polling times and the dialogue model name; the configuration information includes The fields include segment duration and silence duration.

When the AI dialogue system is in use, when the response text of the dialogue model is obtained, it is also necessary to inform the voice system of the AI phrases, hang-up signs, manual transfer signs, and the configuration of the segment duration and silence duration to be broadcast, so that the voice can be heard. The system converts the corresponding response text into voice, and then transmits it to the user, which can improve the user experience.

S2. Arrange and combine the intent dictionaries corresponding to the dialog model names based on the full combination algorithm, and splicing the test event information and the result of the arrangement and combination to obtain a test intent scene table.

Since there may be multiple rounds of dialogue in a test scenario, and the user intentions of each round may have multiple possibilities, in order to ensure the integrity of the test results, it is necessary to fully combine the dialogue rounds with the multiple intentions of the corresponding model to obtain A full coverage test intent scenario table.

In detail, as shown in FIG. 2 , the S2 includes:

S21, obtaining a corresponding intent dictionary according to the dialog model name;

S22, using a full combination algorithm to arrange and combine the intents of the intent dictionary according to the number of dialogue polls in the test event information;

S23. Use a data manipulation tool to splicing each combination in the result of the arrangement and combination with the test event information to obtain a test intent scene table.

The intent dictionary is an intent set that combines intents that can be recognized by the recognition model of the dialogue system with corresponding corpus.

The data manipulation tool described in the embodiment of the present invention is a data tool that can efficiently manipulate a large data set, such as pandas. In the full coverage test mode of the multi-round dialogue system, the data scale will also increase sharply and become particularly large. Therefore, special data tools are required to operate the data set, such as read and write operations, which can effectively improve work efficiency.

Optionally, in order to further improve the storage scale and processing efficiency of the test intent scenario table, the test intent scenario table may also be stored in a node of a blockchain.

Further, the full combination algorithm is used to arrange and combine the total number of intents of the intent dictionary according to the number of dialogue polls in the test event information, including:

Obtain the number of dialog polls in the test event information and the total number of intents in the intent dictionary;

Use all intents of the intent dictionary as candidate intents for each round of dialogue;

Selecting the same number of intent combinations as the number of dialogue polling times each time, wherein, in each combination, one intent is selected from the candidate intents of each round in turn, and recursively arranging and combining according to the number of dialogue polling times , get the result of permutation and combination.

If the number of rounds is m and the total number of intentions is n, the combined result is:

1st round: answer combination {intent 1, intent 1, ..., intent 1 (mth round of answer intent)}

2nd round: answer combination {intent 1, intent 1, ..., intent 2 (the mth round of answering intent)}

…

nth round: answer combinations {intent 1, intent 1, ..., intent n (mth round answer intent)}

n+1th round: answer combination {intent 2, intent 1, ..., intent 1 (mth round of answering intent)}

…

2nth round: answer combination {intent 2, intent 1, ..., intent n (mth round answer intent)}

…

mnth round: answer combination {intent n, intent n, ..., intent n (mth round answer intent)}

S3. Select the test intent data in the test intent scene table as the current test scene one by one in sequence, and invoke the dialog model corresponding to the dialog model name through the interface to generate the response text and configuration parameters of the current test scene.

Specifically, invoking the dialogue model corresponding to the dialogue model name through the interface to generate the response text and configuration parameters of the current test scenario includes:

Generate a dialog request corresponding to the current test scenario based on the request;

Connect the dialogue system through the interface according to the dialogue request, and call the dialogue model with the dialogue model name in the dialogue system;

According to the dialog request, the test data in the current test scene is transferred to the dialog model and the response text generated by the dialog model is obtained;

Acquire the current configuration parameters of the dialogue system, and return the configuration parameters and the response text to the tester through the interface.

In this embodiment of the present invention, the data in the test intent scene table is used as the current test scene, and is connected to the dialog system through the calling interface, so that the current test scene is simulated as the user inputting answer text, so that the dialog system calls the dialog model to generate the corresponding reply, and uses the request (ie request request) to implement the dialog system interface call and obtain the return result of the interface, that is, the response text and configuration parameters.

S4. Verify and analyze the response text and configuration parameters based on a text similarity algorithm to obtain a test result.

This embodiment of the present invention tests whether the response of the dialogue system is correct by comparing the response text and configuration parameters with the expected response of the dialogue system (that is, the response text and configuration parameters), and compares the response text with the pre-stored dialogue system The intention output of the model is checked for consistency, and it is further analyzed whether it is the problem of the dialogue system or the problem of the dialogue model.

The test result refers to a result of judging whether the preset reply of each test scene in the test scene table is consistent with the reply text and configuration parameters based on the verification analysis.

In detail, the text-based similarity algorithm performs verification and analysis on the response text and configuration parameters to obtain test results, including:

Compare the response text and configuration parameters with the data of the corresponding field in the test intent scene table of the current test scene;

If the comparison is inconsistent, it means that the verification fails, and the inconsistent content is written into the test intent scene table by using the data operation;

Connecting to the database to obtain a preset drop-off record, and looking up the actual drop-off intention corresponding to the current test scene in the drop-off record;

Detecting the consistency between the actual drop-off intention and the preset intention of the current test scene based on a text similarity algorithm, and if they are inconsistent, it means that the current test scene is abnormal;

After splicing the verification failure and the abnormality as test results in the test scene table corresponding to the test scenes in the test scene table, and recording the test results of the remaining test scenes as normal, the test scene is obtained Table of full test results.

Wherein, the drop-list record refers to identifying the corresponding intention in the dialogue system through the actual corpus. The preset intent of the current test scene is the corresponding intent of the test corpus, that is, the intent in the intent dictionary.

Further, the detection of the consistency of the actual deposit intention and the preset intention of the current test scene based on the text similarity algorithm includes:

Using a tokenizer to segment the actual storage intention and the preset intention, respectively, to obtain the actual segmentation sequence and the preset segmentation sequence;

Calculate the string similarity according to the actual word segmentation sequence and the preset word segmentation sequence;

When the string similarity is greater than or equal to a preset threshold, judging that the actual storage intention is consistent with the preset intention of the current test scene;

When the string similarity is less than a preset threshold, it is determined that the actual drop-in intention is inconsistent with the preset intention of the current test scene.

This embodiment of the present invention verifies whether the node flow of the dialogue system is correct by verifying the correctness of the response text and related configuration parameters. At the same time, model result verification is added. If there is an abnormality, it can be intuitively seen whether it is a dialogue system problem or a dialogue model For example, when the test result is a verification failure, it means that there is a problem with the dialogue system, and when the test result is abnormal, it means that there is a problem with the intention output of the dialogue model.

The embodiment of the present invention is based on arranging and combining the intent dictionaries corresponding to the dialogue model names based on the full combination algorithm, and splicing the test event information and the result of the permutation and combination to obtain a test intent scene table, where the intent scene table is The full combination is obtained to ensure the comprehensiveness of the test scene; at the same time, the test scene is connected to the dialogue system by calling the interface, and the user's question and answer text is simulated by the test scene to obtain the dialogue system's response and related configuration parameter information, and the response is verified by verifying the response. The correctness of the text and related configuration parameters is used to verify whether the node flow of the dialogue system is correct, and the verification of the intent output result of the node model is added. The accuracy of the results, and automated testing of the dialogue system to improve efficiency. Therefore, the test method, device, electronic device and computer-readable storage medium of the response text proposed by the present invention can solve the problems of low test efficiency and incomplete test scenarios of the response text.

As shown in FIG. 3 , it is a functional block diagram of an apparatus for testing response text provided by an embodiment of the present invention.

The test device 100 for answering text according to the present invention can be installed in an electronic device. According to the implemented functions, the response text testing apparatus 100 may include an event information acquisition module 101 , a test scene acquisition module 102 , a response generation module 103 and a verification analysis module 104 . The modules in the present invention can also be called units, which refer to a series of computer program segments that can be executed by the electronic device processor and can perform fixed functions, and are stored in the memory of the electronic device.

In this embodiment, the functions of each module/unit are as follows:

The event information acquisition module 101 is configured to acquire test event information set based on an actual business scenario, wherein the test event information includes a dialog model name;

The test scene acquisition module 102 is configured to arrange and combine the intent dictionaries corresponding to the dialogue model names based on a full combination algorithm, and splicing the test event information and the result of the arrangement and combination to obtain a test intent scene table;

The response generation module 103 is configured to select the test intent data in the test intent scene table as the current test scene one by one in sequence, and generate the current test scene by invoking the dialog model corresponding to the dialog model name through the interface. Response text and configuration parameters;

The verification and analysis module 104 is configured to perform verification and analysis on the response text and configuration parameters based on a text similarity algorithm to obtain a test result.

In detail, each module described in the device 100 for testing the response text in the embodiment of the present invention adopts the same technical means as the testing method for the response text described in FIG. 1 to FIG. 2, and can generate The same technical effect will not be repeated here.

As shown in FIG. 4 , it is a schematic structural diagram of an electronic device implementing a method for testing response text provided by an embodiment of the present invention.

The electronic device 1 may include a processor 10, a memory 11, a communication bus 12 and a communication interface 13, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as an answer text test program.

The processor 10 may be composed of integrated circuits in some embodiments, for example, may be composed of a single packaged integrated circuit, or may be composed of multiple integrated circuits packaged with the same function or different functions, including one or A combination of multiple central processing units (Central Processing Units, CPUs), microprocessors, digital processing chips, graphics processors, and various control chips, etc. The processor 10 is the control core (ControlUnit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and by running or executing the program or module stored in the memory 11 (for example, executing the response) Text test program, etc.), and call the data stored in the memory 11 to execute various functions of the electronic device and process data.

The storage 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, removable hard disk, multimedia card, card-type storage (for example: SD or DX storage, etc.), magnetic storage, magnetic disk, optical disk, etc. . In some embodiments, the memory 11 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 11 may also be an external storage device of the electronic device, for example, a pluggable mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) equipped on the electronic device. ) card, flash memory card (Flash Card) and so on. Further, the memory 11 may also include both an internal storage unit of an electronic device and an external storage device. The memory 11 can not only be used to store application software installed in the electronic device and various types of data, such as the code of the test program in response to the text, etc., but also can be used to temporarily store the data that has been output or will be output.

The communication bus 12 may be a peripheral component interconnect (PCI for short) bus or an extended industry standard architecture (EISA for short) bus or the like. The bus can be divided into address bus, data bus, control bus and so on. The bus is configured to enable connection communication between the memory 11 and at least one processor 10 and the like.

The communication interface 13 is used for communication between the above electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and/or a wireless interface (eg, a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, and the like. The display may also be appropriately called a display screen or a display unit, which is used for displaying information processed in the electronic device and for displaying a visual user interface.

FIG. 4 only shows an electronic device with components. Those skilled in the art can understand that the structure shown in FIG. 4 does not constitute a limitation on the electronic device 1, and may include fewer or more components than those shown in the drawings. components, or a combination of certain components, or a different arrangement of components.

For example, although not shown, the electronic device may also include a power source (such as a battery) for powering the various components, preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power source can be logically connected through the power management device. Implement functions such as charge management, discharge management, and power management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and any other components. The electronic device may further include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

It should be understood that the embodiments are only used for illustration, and are not limited by this structure in the scope of the patent application.

The test program of the response text stored in the memory 11 in the electronic device 1 is a combination of multiple instructions, and when running in the processor 10, it can realize:

Obtain test event information set based on an actual business scenario, wherein the test event information includes a dialog model name;

Arrange and combine the intent dictionary corresponding to the dialog model name based on the full combination algorithm, and splicing the test event information with the result of the arrangement and combination to obtain a test intent scene table;

Select the test intent data in the test intent scene table one by one as the current test scene in sequence, and call the dialog model corresponding to the dialog model name through the interface to generate the response text and configuration parameters of the current test scene;

The response text and configuration parameters are verified and analyzed based on a text similarity algorithm to obtain a test result.

Specifically, for the specific implementation method of the above-mentioned instruction by the processor 10, reference may be made to the description of the relevant steps in the embodiment corresponding to the accompanying drawings, which will not be repeated here.

Further, if the modules/units integrated in the electronic device 1 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. The computer-readable storage medium may be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a removable hard disk, a magnetic disk, an optical disc, a computer memory, a read-only memory (ROM, Read-Only). Memory).

The present invention also provides a computer-readable storage medium, where the readable storage medium stores a computer program, and when executed by a processor of an electronic device, the computer program can realize:

Obtain test event information set based on an actual business scenario, wherein the test event information includes a dialog model name;

Arrange and combine the intent dictionary corresponding to the dialog model name based on the full combination algorithm, and splicing the test event information with the result of the arrangement and combination to obtain a test intent scene table;

Select the test intent data in the test intent scene table one by one as the current test scene in sequence, and call the dialog model corresponding to the dialog model name through the interface to generate the response text and configuration parameters of the current test scene;

The response text and configuration parameters are verified and analyzed based on a text similarity algorithm to obtain a test result.

In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus, apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are only illustrative. For example, the division of the modules is only a logical function division, and there may be other division manners in actual implementation.

The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware, or can be implemented in the form of hardware plus software function modules.

It will be apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, but that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics of the invention.

Therefore, the embodiments are to be regarded in all respects as illustrative and not restrictive, and the scope of the invention is to be defined by the appended claims rather than the foregoing description, which are therefore intended to fall within the scope of the claims. All changes within the meaning and range of the equivalents of , are included in the present invention. Any reference signs in the claims shall not be construed as limiting the involved claim.

The blockchain referred to in the present invention is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain, essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information to verify its Validity of information (anti-counterfeiting) and generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

The embodiments of the present application may acquire and process related data based on artificial intelligence technology. Among them, 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. .

Furthermore, it is clear that the word "comprising" does not exclude other units or steps and the singular does not exclude the plural. Several units or means recited in the system claims can also be realized by one unit or means by means of software or hardware. The words first, second, etc. are used to denote names and do not denote any particular order.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent substitutions can be made without departing from the spirit and scope of the technical solutions of the present invention.

Claims (10)

1. A method for testing response text, the method comprising:

Acquiring test event information set based on an actual service scene, wherein the test event information comprises a conversation model name;

based on a full combination algorithm, carrying out permutation and combination on the intention dictionary corresponding to the conversation model name, and splicing the test event information and the result of the permutation and combination to obtain a test intention scene table;

selecting the test intention data in the test intention scene table as a current test scene one by one according to the sequence, and calling a dialogue model corresponding to the dialogue model name through an interface to generate a response text and configuration parameters of the current test scene;

and carrying out verification analysis on the response text and the configuration parameters based on a text similarity algorithm to obtain a test result.

2. The method for testing response text according to claim 1, wherein the obtaining of the test event information set based on the actual service scenario comprises:

acquiring a test event table preset according to an actual service scene;

and setting test data according to the service function of the dialog system and filling the test data into the test event table to obtain test event information.

3. The method for testing response text according to claim 1, wherein the ranking and combining the intention dictionaries corresponding to the dialogue model names based on a full-combination algorithm, and the splicing the test event information and the results of the ranking and combining to obtain a test intention scenario table comprises:

Acquiring a corresponding intention dictionary according to the conversation model name;

utilizing a full combination algorithm to arrange and combine the intentions of the intention dictionary according to the number of dialog polling times in the test event information;

and splicing each combination in the results of the permutation and combination with the test event information by using a data operation tool to obtain a test intention scene table.

4. The method for testing answer text according to claim 3, wherein the ranking and combining the total number of intentions of the intention dictionary according to the number of dialog polls in the test event information by using a full combination algorithm comprises:

acquiring the number of dialog polling times in the test event information and the total number of intentions of the intention dictionary;

all intentions of the intention dictionary are used as candidate intentions of each dialog;

and selecting the intention combinations with the same number as the number of times of the dialogue polling each time, wherein one intention is selected from the candidate intentions of each turn in each combination in sequence, and carrying out permutation and combination according to the number of times of the dialogue polling in a recursion manner to obtain a permutation and combination result.

5. The method for testing the response text according to claim 1, wherein the interfacing the dialog model to generate the response text and the configuration parameters of the current test scenario comprises:

Generating a dialogue request corresponding to the current test scene based on the request;

connecting a dialog system through an interface according to the dialog request, and calling a dialog model with the name of the dialog model in the dialog system;

transmitting test data in the current test scene into the dialogue model according to the dialogue request and acquiring a response text generated by the dialogue model;

and acquiring the current configuration parameters of the dialog system, and returning the configuration parameters and the response text to a tester through the interface.

6. The method for testing response text according to any one of claims 1 to 5, wherein the verification analysis of the response text and the configuration parameters based on the text similarity algorithm to obtain the test result comprises:

comparing the response text and the configuration parameters with the data of the corresponding field of the current test scene in the test intention scene table;

if the comparison is inconsistent, the verification is failed, and inconsistent contents are written into the test intention scene table by utilizing the data operation;

connecting a database to obtain a preset drop record, and searching an actual drop intention corresponding to the current test scene in the drop record;

Detecting the consistency of the actual database falling intention and the preset intention of the current test scene based on a text similarity algorithm, and if the actual database falling intention and the preset intention of the current test scene are inconsistent, indicating that the current test scene is abnormal;

and splicing the verification failure and the abnormity serving as test results in the test scene table in a corresponding test scene, and recording the test results of the other test scenes as normal to obtain the complete test result of the test scene table.

7. The method for testing the answer text according to claim 6, wherein the detecting the consistency of the actual landing intent and the preset intent of the current test scenario based on the text similarity algorithm comprises:

respectively segmenting the actual database falling intention and the preset intention by using a word segmentation device to obtain an actual word segmentation sequence and a preset word segmentation sequence;

calculating the similarity of character strings according to the actual word segmentation sequence and the preset word segmentation sequence;

when the similarity of the character strings is larger than or equal to a preset threshold value, judging that the actual library falling intention is consistent with the preset intention of the current test scene;

and when the similarity of the character strings is smaller than a preset threshold value, judging that the actual library falling intention is inconsistent with the preset intention of the current test scene.

8. A test apparatus for response text, the apparatus comprising:

the system comprises an event information acquisition module, a conversion module and a conversion module, wherein the event information acquisition module is used for acquiring test event information set based on an actual service scene, and the test event information comprises a conversation model name;

the test scenario acquisition module is used for carrying out permutation and combination on the intention dictionary corresponding to the conversation model name based on a full combination algorithm, and splicing the test event information and the result of the permutation and combination to obtain a test intention scenario table;

the response generation module is used for selecting the test intention data in the test intention scene table as the current test scene one by one according to the sequence, and calling the dialogue model corresponding to the dialogue model name through an interface to generate a response text and configuration parameters of the current test scene;

and the verification analysis module is used for verifying and analyzing the response text and the configuration parameters based on a text similarity algorithm to obtain a test result.

9. An electronic device, characterized in that the electronic device comprises:

at least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein,

the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform a method of testing a response text according to any one of claims 1 to 7.

10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out a method for testing a response document according to any one of claims 1 to 7.

Images