WO2025107888A1 - 应用调试的方法、信息查询的方法、应用评测的方法 - Google Patents

应用调试的方法、信息查询的方法、应用评测的方法 Download PDF

Info

Publication number
WO2025107888A1
WO2025107888A1 PCT/CN2024/122387 CN2024122387W WO2025107888A1 WO 2025107888 A1 WO2025107888 A1 WO 2025107888A1 CN 2024122387 W CN2024122387 W CN 2024122387W WO 2025107888 A1 WO2025107888 A1 WO 2025107888A1
Authority
WO
WIPO (PCT)
Prior art keywords
query
evaluation
user
target application
information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/CN2024/122387
Other languages
English (en)
French (fr)
Inventor
黄巍
姚璨
黄树炫
马进
蒋世豪
孟德森
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Zitiao Network Technology Co Ltd
Original Assignee
Beijing Zitiao Network Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Zitiao Network Technology Co Ltd filed Critical Beijing Zitiao Network Technology Co Ltd
Priority to JP2025557329A priority Critical patent/JP2026512314A/ja
Publication of WO2025107888A1 publication Critical patent/WO2025107888A1/zh
Anticipated expiration legal-status Critical
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/36Prevention of errors by analysis, debugging or testing of software
    • G06F11/362Debugging of software
    • G06F11/3644Debugging of software by instrumenting at runtime

Definitions

  • Example embodiments of the present disclosure generally relate to the computer field, and in particular, to a method for application debugging, a method for information query, a method for application evaluation, an apparatus, a device, and a computer-readable storage medium.
  • terminal devices can provide various services to people in work and life.
  • Applications that provide services can be deployed in terminal devices.
  • Terminal devices present corresponding content and interact with users through the user interface of the application to meet various needs of users.
  • users may initiate query requests within the application. Therefore, how to improve the quality of query services provided to users is a problem of concern.
  • a method for application debugging includes: executing a test query for a target application; presenting corresponding first operation information of multiple stages of executing the test query, wherein the operation information of the multiple stages at least indicates the output of each of the multiple stages; and generating an operation reference case for the target application based on the corresponding first operation information of the multiple stages of the test query, wherein the operation reference case includes the test query and the operation information of the multiple stages, and the operation reference case is referenced in the operation of the target application to determine the result of the user query.
  • a method for information query includes: in response to receiving a user query for a target application, based on the similarity between the user query and the test query exceeding a threshold similarity, determining an operation reference case including the test query as a target operation reference case for the user query, wherein the operation reference case includes the test query and corresponding operation information of multiple stages of the test query; and executing the user query based on the target operation reference case.
  • a method for application evaluation comprises: generating an evaluation reference case for the target application, the evaluation reference case comprising an evaluation query, a first result generated by the target application for the evaluation query, and evaluation feedback on the first result; selecting a target evaluation sample from the evaluation sample set based on the similarity between the evaluation query and queries in the evaluation sample set, the target evaluation sample comprising a sample query and a second result generated by the target application for the sample query; and The first result, the evaluation feedback and the second result generate an evaluation result for the target application.
  • a device for application debugging includes: a test query execution module, configured to execute a test query for a target application; a first operation information presentation module, configured to present corresponding first operation information of multiple stages of executing the test query, the operation information of the multiple stages at least indicating the output of each of the multiple stages; and an operation reference case generation module, configured to generate an operation reference case for the target application based on the corresponding first operation information of the multiple stages of the test query, the operation reference case including the test query and the operation information of the multiple stages, and the operation reference case is referenced in the operation of the target application to determine the result of the user query.
  • a device for information query includes: an operation reference case determination module, configured to, in response to receiving a user query for a target application, determine an operation reference case including the test query as a target operation reference case for the user query based on the similarity between the user query and the test query exceeding a threshold similarity, wherein the operation reference case includes the test query and corresponding operation information of multiple stages of the test query; and a user query execution module, configured to execute the user query based on the target operation reference case.
  • a device for application evaluation includes: an evaluation reference case generation module, configured to generate an evaluation reference case for the target application, the evaluation reference case including an evaluation query, a first result generated by the target application for the evaluation query, and evaluation feedback on the first result; a target evaluation sample selection module, configured to select a target evaluation sample from the evaluation sample set based on the similarity between the evaluation query and the queries in the evaluation sample set, the target evaluation sample including a sample query and a second result generated by the target application for the sample query; and an evaluation result generation module, configured to generate an evaluation result for the target application based on the first result, the evaluation feedback, and the second result.
  • an electronic device in a seventh aspect of the present disclosure, includes at least one processing unit; and at least one memory, the at least one memory is coupled to the at least one processing unit and stores instructions for execution by the at least one processing unit.
  • the electronic device executes the methods of the first aspect and the second aspect.
  • a computer-readable storage medium wherein a computer program is stored on the medium, and when the computer program is executed by a processor, the method of the first aspect and the method of the second aspect are implemented.
  • FIG1 shows a schematic diagram of an example environment in which embodiments of the present disclosure can be implemented
  • FIG. 2 shows an example of a page of a debugging entry for each function in a target application according to some embodiments of the present disclosure
  • FIG3 shows an example of a page of test query records for a specific function in a target application according to some embodiments of the present disclosure
  • FIG4 shows an example of a page for debugging a target application according to some embodiments of the present disclosure
  • FIG5 shows an example of a page displaying data information according to some embodiments of the present disclosure
  • FIG6A shows an example of a page of a use case library for a specific function in a target application according to some embodiments of the present disclosure
  • FIG6B shows an example of a use case details page according to some embodiments of the present disclosure
  • FIG7A shows a flowchart of a process for application debugging according to some embodiments of the present disclosure
  • FIG7B shows a flowchart of a process of information query according to some embodiments of the present disclosure
  • FIG7C shows a flow chart of a process of application evaluation according to some embodiments of the present disclosure
  • FIG8A shows a schematic structural block diagram of an apparatus for application debugging according to some embodiments of the present disclosure
  • FIG8B shows a schematic structural block diagram of an apparatus for information query according to some embodiments of the present disclosure
  • FIG8C shows a schematic structural block diagram of an apparatus for application evaluation according to some embodiments of the present disclosure.
  • FIG. 9 shows a block diagram of an electronic device in which one or more embodiments of the present disclosure may be implemented.
  • executing a step “in response to A” does not mean executing the step immediately after “A” but may include one or more intermediate steps.
  • the types, scopes of use, usage scenarios, etc. of the information involved in the present disclosure should be informed to relevant users and their authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.
  • the relevant users may include any type of right holders, such as individuals, enterprises, and groups.
  • a prompt message is sent to the relevant user to clearly prompt the relevant user that the operation requested to be performed will require obtaining and using the information of the relevant user, so that the relevant user can independently choose whether to provide information to software or hardware such as an electronic device, application, server or storage medium that executes the operation of the technical solution of the present disclosure based on the prompt message.
  • a prompt message is sent to the relevant user, for example, in the form of a pop-up window, in which the prompt message may be presented in text form.
  • the pop-up window may also carry a selection control for the user to select "agree” or “disagree” to provide information to the electronic device.
  • the data involved in this technical solution shall comply with the requirements of relevant laws, regulations and relevant provisions.
  • the activation of the digital assistant related functions, the acquired data, the processing and storage methods of the data, etc. in the disclosed embodiment shall obtain the advance authorization of the user and other rights holders associated with the user, and shall comply with the relevant laws, regulations and the agreement rules between the rights holders.
  • model can learn the association between the corresponding input and output from the training data, so that after the training is completed, the corresponding output can be generated for a given input.
  • the generation of the model can be based on machine learning technology.
  • Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs by using multi-layer processing units.
  • a neural network model is an example of a model based on deep learning.
  • model may also be referred to as “machine learning model”, “learning model”, “machine learning network” or “learning network”, and these terms are used interchangeably in this article.
  • FIG. 1 shows a schematic diagram of an example environment 100 in which embodiments of the present disclosure can be implemented.
  • Environment 100 involves an application management platform 110, which can support the creation of applications and/or the running of applications.
  • the portion of the application management platform 110 used to support the creation of applications can also be referred to as an application creation portion.
  • the portion of the application management platform 110 used to support the running of applications can also be referred to as an application running portion.
  • the application creation part can provide a user 105 with an environment for creating and publishing applications.
  • the user 105 can be referred to as an application creation user or creator.
  • the application creation part can be a low-code platform that provides a collection of tools for application creation.
  • the application creation part can support visual development of various types of applications, so that developers can skip the manual coding process and speed up the application development cycle and cost.
  • the application creation part can support any appropriate platform for users to develop one or more types of applications, such as application-based Platform as a Service (aPaaS) platforms. Such platforms can support users to develop applications efficiently, and realize operations such as application creation and application function adjustment.
  • aPaaS Application-based Platform as a Service
  • the application creation part can be deployed locally on the terminal device of user 105, and/or can be supported by a server-side device.
  • the terminal device of user 105 can run a client of the application creation part, which can support the interaction between the user and the application creation part provided by the server.
  • the application creation part runs locally on the user's terminal device
  • the user 105 can directly interact with the local application creation part using the terminal device.
  • the server-side device can, based on the communication connection with the terminal device, implement service provision for the client running in the terminal device.
  • the application creation part can present a corresponding page 130 to the user 105 based on the operation of the user 105, so as to output to the user 105 and/or receive from the user 105 information related to application creation.
  • the application creation part can be associated with a corresponding database, which stores data or information required for the application creation process supported by the application creation part.
  • the database can store the code and description information corresponding to each functional module that constitutes the application.
  • the application creation part can also perform operations such as calling, adding, deleting, and updating the functional modules in the database.
  • the database can also store operations that can be performed on different functional blocks. Exemplarily, in a scenario where an application is to be created, the application creation part can call the corresponding functional blocks from the database to build the application.
  • a user 105 can create a target application 120 as needed on the application creation part and publish the target application 120.
  • the target application 120 can be published to any appropriate application running part, as long as the application running part can support the operation of the target application 120.
  • the target application 120 can be used to be operated by one or more users 145.
  • the user 145 can be referred to as the end user of the target application 120.
  • the target application 120 may include or be implemented as a digital assistant 122.
  • the digital assistant 122 can be configured to have intelligent dialogue.
  • the digital assistant 122 can be integrated into the target application 120, as a part of the target application 120 to assist in the task processing within the target application 120.
  • the digital assistant 122 can be configured as an independently running application, such as a web application or other types of applications.
  • the digital assistant 122 and the target application 120 can be regarded as the same application.
  • the digital assistant 122 is provided to assist users in various task processing requirements in different applications and scenarios.
  • the user inputs an interactive message, and the digital assistant 122 provides a reply message in response to the user input.
  • the digital assistant 122 can support users to input questions in natural language, and perform tasks and provide replies based on the understanding of natural language input and logical reasoning ability.
  • digital assistant 122 can interact with user 145 as a contact.
  • digital assistant 122 can be implemented in an instant messaging (IM) application.
  • IM instant messaging
  • Digital assistant 122 can interact with user 145 in a single chat session with user 145.
  • digital assistant 122 can interact with multiple users in a group chat session including multiple users.
  • the client of the application running part can present the target application in the client interface. 120 or an interactive window 142 of the digital assistant 122, such as a conversation window with the digital assistant 122.
  • the user 145 can enter a conversation message in the conversation window, and the target application 120 can determine the reply message of the digital assistant 122 based on the created configuration information and present it to the user in the interactive window 142.
  • the interactive message with the target application 120 may include a message in a multimodal form, such as a text message (e.g., a natural language text), a voice message, an image message, a video message, and the like.
  • the application running part can be deployed locally on the terminal device of each user 145, and/or can be supported by a server-side device.
  • the terminal device of user 145 can run a client with an application running part, which can support the interaction between the user and the application running part provided by the server.
  • the application running part runs locally on the user's terminal device
  • the user 145 can directly interact with the local application running part using the terminal device.
  • the server-side device can realize the service supply to the client running in the terminal device based on the communication connection with the terminal device.
  • the application running part can present the corresponding application page to the user 145 based on the operation of the user 145, so as to output and/or receive information related to the use of the application to the user 145.
  • the implementation of at least part of the functions of the target application 120 and/or the implementation of at least part of the functions of the digital assistant 122 in the target application 120 can be implemented based on the model.
  • one or more models 155 such as the capabilities of the model 155, can be called.
  • the digital assistant 122 can use the model 155 to understand the user input and provide a reply to the user based on the output of the model 155.
  • the application creation platform 110 needs to use the model 155 to test the target application 120 to determine whether the operation result of the target application 120 meets expectations.
  • the application operation part may need to use the model 155 to determine the response result to the user.
  • one or more models 155 may run on the application management platform 110, or other remote servers.
  • the model 155 may be a machine learning model, a deep learning model, a learning model, a neural network, etc.
  • the model may be based on a language model (LM).
  • the language model can have question-answering capabilities by learning from a large amount of corpus.
  • the model 155 may also be based on other appropriate models.
  • the management platform 110 can run on appropriate electronic devices.
  • the electronic devices here can be any type of device with computing capabilities, including terminal devices or server devices.
  • the terminal device can be any type of mobile terminal, fixed terminal or portable terminal, including mobile phones, desktop computers, laptop computers, notebook computers, netbook computers, tablet computers, media computers, multimedia tablets, personal communication systems (PCS) devices, personal navigation devices, personal digital assistants (PDAs), audio/video players, digital cameras/camcorders, positioning devices, television receivers, radio broadcast receivers, e-book devices, gaming devices or any combination of the foregoing, including accessories and peripherals of these devices or any combination thereof.
  • PCS personal communication systems
  • PDAs personal digital assistants
  • audio/video players digital cameras/camcorders
  • positioning devices television receivers, radio broadcast receivers, e-book devices, gaming devices or any combination of the foregoing, including accessories and peripherals of these devices or any combination thereof.
  • the server device may include, for example, a computing system/server, such as a mainframe, an edge computing node, a computing device in a cloud environment, and so on.
  • a computing system/server such as a mainframe, an edge computing node, a computing device in a cloud environment, and so on.
  • the management platform 110 may be implemented based on cloud services.
  • FIG. 1 shows a single user interacting with the application creation portion and a single user interacting with the application execution portion, in reality multiple users may access application management platform 110 to each create a digital assistant, and each digital assistant may be used to interact with multiple users.
  • the application can use the machine learning model to provide query services.
  • the quality of the query service provided by the application may not be high, and multiple rounds of optimization are required to improve the effect of the query service.
  • Traditional optimization requires improving the system at the code level, which is costly, time-consuming, and prone to errors.
  • a solution for application debugging is provided.
  • a test query for a target application is executed.
  • Corresponding first operation information of multiple stages of executing the test query is presented, and the operation information of the multiple stages at least indicates the output of each of the multiple stages.
  • an operation reference case for the target application is generated, the operation reference case includes the test query and the operation information of the multiple stages, and the operation reference case is referenced in the operation of the target application to determine the result of the user query.
  • the application creation user only needs to generate an operation reference case for the target application based on the operation information corresponding to the multiple stages of executing the test query.
  • the operation reference case can be referred to. In this way, the application creation user can improve the quality of service provided by the target application without modifying the code, simplifying the tuning process of the target application.
  • the page presented by the application management platform 110 can be presented via the terminal device of the user 105, and user input can be received via the terminal device of the user 105.
  • FIG2 shows an example of a page 200 of debugging entries for various functions in the target application 120 according to some embodiments of the present disclosure.
  • a page 200 of debugging entries for various functions in the target application 120 is presented on the terminal device of the user 105.
  • Page 200 provides a functional area 210, which can be used for various functions in the target application.
  • the functions in the target application may include, for example, an application data analysis function, an application data update function, a knowledge question and answer function, and the like. However, it should be understood that this is merely an exemplary function and is not intended to be limiting.
  • the target application here may be any suitable application, such as a human resources application, a CRM (Customer Relationship Management) application, a financial application, and the like.
  • the user 105 triggers the debugging control 220 of the data analysis function, for example, the user 105 clicks the debugging control 220. In response to this, a page of test query records may be presented to the user 105.
  • FIG. 3 shows an example of a page 300 of test query records for a specific function in a target application according to some embodiments of the present disclosure.
  • a page 300 of test query records for a specific function in a target application can be presented on a terminal device of a user 105.
  • Page 300 records the test query previously created by the user 105, and the user 105 debugs the corresponding function by entering the test query.
  • the user 105 can re-debug the previously created test query. For example, the user 105 clicks on the control 310 to re-execute the test query "What is the number of commodities with a unit price lower than XX yuan?" to debug the corresponding function in the target application.
  • a debugging page can be presented to the user 105.
  • the user 105 can trigger the execution of a new test query. For example, the user 105 clicks on the newly added control 320 to enter a new test query to debug the corresponding function in the target application. In response to this, the user 105 can enter the debugging page.
  • FIG4 shows an example of a page 400 for debugging a target application according to some embodiments of the present disclosure.
  • the application management platform 110 can execute a test query for the target application.
  • the user 105 enters a test query statement in the text area 410 corresponding to the test query, and clicks the control 420 to execute a test query for the target application.
  • the test query statement entered by the user 105 can be, for example, "What is the quantity of goods with a unit price lower than XX yuan?", "What is the proportion of goods with an origin of XX? Draw a pie chart”.
  • the application management platform 110 can present the corresponding first operation information of multiple stages of executing the test query in the staged debugging area 430, and the operation information of the multiple stages at least indicates the output of each of the multiple stages.
  • the application management platform 110 may present corresponding first operation information of multiple stages of executing the test query in the staged debugging area 430, and the operation information of the multiple stages at least indicates the output of each of the multiple stages.
  • the operation information of five stages is shown in the staged debugging area 430, and the five stages need to execute respective tasks.
  • the corresponding outputs 440, 442, 444, 446 and 448 are generated.
  • the output 444 of the third stage may include the data table required to execute the test query.
  • the operation information of multiple stages may also indicate at least one of the following: the input of each of the multiple stages, the operation time of each of the multiple stages, the operation log of each of the multiple stages, and the operation status of each of the multiple stages (i.e., whether the operation is completed).
  • the operation information of multiple stages since there are limitations on the input information of the machine learning model (e.g., the length of the input information), the operation information of multiple stages may also indicate the input information provided to the machine learning model when the machine learning model is called in at least one of the multiple stages. In this way, user 105 can make the input information provided to the machine learning model meet the requirements of the machine learning model, thereby ensuring the stable operation of the machine learning model.
  • the multiple stages displayed in the staged debugging area 430 may include at least one predefined stage, and the at least one predefined stage includes a stage for generating a database query request.
  • the at least one predefined stage may be some debugging stages pre-set by the application management platform 110 according to the characteristics and/or corresponding functions of the target application.
  • the purpose of the application management platform 110 performing a test query on the target application may be to generate a better
  • the at least one predefined stage may further include a stage of generating a database query request, which may generate an accurate and efficient database query request based on the output of the previous predefined stage.
  • the database query request here may be, for example, a SQL (Structured Query Language) statement.
  • the multiple stages displayed in the staged debugging area 430 may include at least one stage determined based on a test query in addition to at least one predefined stage.
  • at least one stage may also be determined based on the requirements of the test query.
  • the test query is "What is the number of commodities with a unit price lower than XX yuan?" The test query does not have additional data analysis requirements, so the application management platform 110 may only determine one stage, for example, the data viewing stage, and the application management platform 110 may output the number of commodities with a unit price lower than XX yuan in this stage.
  • the test query is "What is the proportion of commodities with a place of origin of XX? Draw a pie chart", and the test query has additional data analysis requirements, so the application management platform 110 may generate multiple stages, for example, a chart generation stage, a data insight stage, etc.
  • the application management platform 110 may output a pie chart of the proportion of commodities with a place of origin of XX to the chart generation stage, and the application management platform 110 may also output the data mining and analysis of the proportion of commodities with a place of origin of XX to the data insight stage. In this way, the application management platform 110 can determine different stages after the predefined stage based on the test query, which can meet different query requirements of users and improve the query experience of users.
  • the test query for the target application executed by the application management platform 110 is a query for the first function in the target application. Since different functions in the target application have great differences, debugging for one function in the target application may not be applicable to another function in the target application, so the test query is a query for a specific function in the target application.
  • the predefined stages may be different for different functions in the target application.
  • at least one predefined stage may be configured for the first function in the target application.
  • the application management platform 110 may configure at least one different predefined stage based on the characteristics of different functions in the target application.
  • the application management platform 110 may configure five predefined stages, namely, the recall and selection of terminology stage, the recall and selection of running reference case stage, the recall and selection of target data set stage, the recall and selection of target data set field stage, and the generation and execution of database query request stage.
  • the recall and selection of terminology stage the interpretation of the terminology used in the test query may be recalled and selected.
  • the running reference case similar to the test query may be recalled and selected.
  • the data set required for executing the test query may be recalled and selected.
  • the recall and selection of target data set field stage the field in the data set required for executing the test query may be recalled and selected.
  • a database query request may be generated and executed based on the output of the previous four stages.
  • phase division can be carried out according to the characteristics of the corresponding functions. In this way, different functions in the target application are debugged separately using test queries, which fully considers the differences between different functions and can effectively improve the service quality provided by different functions.
  • the application management platform 110 can generate an operation reference case for the target application based on the corresponding first operation information of the multiple stages of the test query.
  • the operation reference case includes the test query and the operation information of multiple stages, and the operation reference case is referenced in the operation of the target application to determine the result of the user query.
  • the user 105 can first click the tuned control 450 on the page 400, determine that the test query and the corresponding operation information have been tuned, and then click the save control 460 to generate the operation reference case.
  • the application management platform 110 receives the user query of the user 145, it can refer to the operation reference case to determine the result of the user query.
  • the application management platform 110 can determine the adjustment of the first operation information of at least one of the multiple stages based on the corresponding first operation information of the multiple stages of the test query. Then, the application management platform 110 can re-execute the test query based on the adjusted first operation information of at least one stage. The application management platform 110 can present the corresponding second operation information of the multiple stages of the re-execution of the test query. If a use case generation confirmation is received, the application management platform 110 can generate a reference use case to be run based on the corresponding second operation information.
  • the application management platform 110 can re-execute the test query and present the corresponding second operation information of the multiple stages of the re-execution of the test query in the staged debugging area 430.
  • the application management platform 110 can receive the use case generation confirmation through the triggering of the tuned control 450 and the save control 460 by the user 105 to generate a running reference use case based on the second operation information.
  • the application management platform 110 can receive terminology adjustment information from the user 105 for the first stage of multiple stages, and re-execute the test query based on the terminology adjustment information.
  • the terminology adjustment information indicates adding a term set to the first stage or removing a term set selected in the first stage.
  • the term set includes an explanation of the terms used by the user group of the target application.
  • the terms used by the user group may be used within the scope of the group, but are not common terms, so it is necessary to help the machine learning model understand such terms.
  • the explanation of such a term may include a literal explanation of the term, and may also include a common term corresponding to the term.
  • the first stage of executing the test query may select a term set related to the test query, and the user 105 may add or remove a term set in the first stage.
  • the test query of the user 105 for the target application is "What is the quantity of goods sold through the C2C channel?", where C2C is a special term, and the term set selected in the first stage may include an explanation of C2C, and the user 105 may add other term sets or delete the term set selected in the first stage.
  • the application management platform 110 may determine one or more term sets used in the first stage based on the term adjustment information, and then re-execute the test query based on the one or more term sets.
  • the application management platform 110 may receive a use case for a second phase of the multiple phases. Adjustment information, and re-execute the test query based on the use case adjustment information.
  • the use case adjustment information indicates adding a reference use case to the second stage or removing the reference use case selected in the second stage.
  • the second stage of executing the test query can select a use case similar to the test query, and the user 105 can add a reference use case in the second stage or remove the reference use case.
  • the test query of user 105 for the target application is "What is the number of goods with a unit price of less than 5,000 yuan?"
  • the second stage can select a reference use case similar to the test query based on the similarity calculation.
  • the reference use case selected in the second stage can be "What is the number of goods with a unit price of less than 8,000 yuan?"
  • User 105 can add other reference use cases or move the reference use case selected in the second stage.
  • the application management platform 110 can determine one or more reference use cases used in the second stage based on the use case adjustment information, and then re-execute the test query based on the one or more reference use cases.
  • the application management platform 110 may receive data set adjustment information for the third stage of the multiple stages, and re-execute the test query based on the data set adjustment information.
  • the third stage of executing the test query may select the data set required to execute the test query, and the user 105 may add or remove data sets in the third stage.
  • the test query of the user 105 for the target application is "What is the number of goods with a unit price of less than 5,000 yuan?"
  • the data set selected in the third stage may be a product analysis table, and the user 105 may add other data sets or delete the data set selected in the third stage.
  • the application management platform 110 may also receive data set field adjustment information for the fourth stage among the multiple stages, and re-execute the test query based on the data set field adjustment information.
  • the fourth stage of executing the test query may select the fields of the data set required to execute the test query, and the user 105 may add fields or remove fields in the fourth stage.
  • the test query of the user 105 for the target application is "What is the quantity of goods with a unit price of less than 5,000 yuan?"
  • the fields of the data set selected in the fourth stage are the product number field, unit price field, origin field, and production date field in the product analysis table.
  • the user 105 may add other fields or delete the fields selected in the fourth stage. In this way, the user can adjust the various stages of the test query through a graphical interface, which facilitates the user's understanding of debugging and simplifies the debugging process of the target application.
  • the application management platform 110 may receive a stage operation request after determining the adjustment to the first operation information of at least one stage.
  • the stage operation request indicates that a test query is to be executed from a specified stage among multiple stages, where the specified stage is before at least one stage or is included in at least one stage, and based on the adjusted first operation information of at least one stage, the specified stage among multiple stages and the stage after the specified stage are re-executed.
  • the user 105 has adjusted the second stage, in which case the specified stage may be the first stage before the second stage, or may be the second stage.
  • the application management platform 110 may execute the first stage and the stage after the first stage (i.e., all stages) based on the adjusted first operation information of at least the second stage.
  • a trigger entry for starting to re-execute the stage is provided, such as trigger controls 421, 422, 423, 424, and 425.
  • User 105 can trigger control 420 to execute all stages from the beginning.
  • application management platform 110 can Based on the adjusted first operation information of at least the second stage, the second stage and the stage after the second stage are executed.
  • user 105 can trigger control 422 to execute the second stage and the stage after the second stage. In this way, the user can flexibly run different stages, and by comparing the output results of running different stages, the effectiveness of adjusting different stages can be verified.
  • Figure 4 shows that the corresponding stage is executed by triggering controls 421, 422, 423, 424, and 425, the triggering method is merely schematic.
  • User 105 can also perform corresponding triggering in other ways, for example, by user input representing natural language, to trigger the execution of the corresponding stage.
  • the application management platform 110 may present the operation information associated with the re-execution of the specified stage and the stage after the specified stage in the stage debugging area 430. For example, if the specified stage is the second stage, the operation information associated with the re-execution of the second stage and the stage after the second stage may be presented in the stage debugging area 430.
  • the application management platform 110 may also present data information while presenting the corresponding first operation information of the multiple stages, the data information at least indicating at least one data set used to perform the test query, and generate an operation reference case based on the corresponding first operation information of the multiple stages and the data information.
  • the application management platform 110 may present data information for performing the test query in the data information display area 470.
  • the application management platform 110 may generate an operation reference case based on the data information or an adjustment to the data information.
  • the adjustment of the data information may include updating the at least one data set used to perform the test query.
  • a data set used to perform the test query namely, a commodity analysis table
  • the user 105 may trigger the add control 472 to update the data set.
  • the adjustment of the data information may include updating at least a portion of at least one data set used to perform the test query.
  • a portion of the data set may indicate a field description in the data set.
  • a data set used to perform the test query i.e., a commodity analysis table, is displayed in the data information display area 470.
  • the user 105 may trigger the modification control 474 to update the field description in the data set.
  • the data information also indicates a term set referenced for executing the test query, and the term set includes an explanation of the terms used by the user group of the target application.
  • FIG. 5 shows an example of a page 500 for displaying data information according to some embodiments of the present disclosure.
  • the page 500 includes a term set display area 510, and the user 105 can adjust the content in the term set display area 510.
  • the data information further indicates the conversation record between the user and the digital assistant referenced for executing the test query.
  • the page 500 includes a conversation record display area 520 , and the user 105 can adjust the content in the conversation record display area 520 .
  • the data information also indicates reference prompt words used to perform the test query.
  • page 500 includes a reference prompt word display area 530, and user 105 can adjust the content in reference prompt word display area 530. By adjusting the reference prompt words, better guidance can be provided to the machine learning model to help Machine learning models better complete user tasks.
  • the data information also indicates conditions for determining candidate results for the test query.
  • page 500 includes condition area 540, and user 105 can adjust the similarity and word search quantity in condition area 540.
  • the digital assistant will search for questions similar to the user based on similarity, and provide a predetermined number of search results to user 145 according to the number of single searches.
  • FIG. 6A shows an example of a page 600 of a use case library for a specific function in a target application according to some embodiments of the present disclosure.
  • Page 600 presents a use case library, which contains reference use cases that have been generated.
  • User 105 triggers a detail control 602 to jump to a page of use case details corresponding to the use case, see FIG. 6B .
  • FIG. 6B shows an example of a page 610 of use case details according to some embodiments of the present disclosure.
  • Page 610 may include a test query and corresponding first run information 614 of multiple stages.
  • an operation reference use case set including one or more operation reference use cases can be generated.
  • the operation reference case can be used to provide services to the user 145.
  • the application management platform 110 in response to receiving a user query for the target application, can search for a test query similar to or matching the user query from the operation reference case set. Based on the similarity between the user query and a test query exceeding the threshold similarity, the application management platform 110 can determine the operation reference case as the target operation reference case for the user query, and execute the user query based on the target operation reference case.
  • the application management platform 110 After generating the operation reference case, the application management platform 110 receives the user query for the target application initiated by the user 145, and can determine the similarity between the user query and the test query through similarity calculation, and determine the operation reference case as the target operation reference case for the user query based on the similarity. After determining the target operation reference case, the application management platform 110 can use the target operation reference case to execute the user query. In this way, the result of the user query is determined with reference to the operation reference case, and more accurate query results can be provided to the user, thereby improving the user's satisfaction.
  • a machine learning model can be used to execute user queries.
  • the application management platform 110 can provide the determined target reference use case to the machine learning model as part of the prompt word for the user query. Then, the application management platform 110 can obtain the execution result of the user query from the machine learning model.
  • the target reference use case here is selected from the set of running reference use cases generated during the application creation phase or the application testing phase. Such target reference use cases are generated for debugging purposes, rather than from use cases when users actually use the application.
  • the user using the target application 120 may be a tester, that is, the user query may come from a tester.
  • the tester gives positive feedback on a query result
  • an operation reference case may be generated based on the user query and the corresponding query result.
  • the operation information of multiple stages of the operation process of obtaining the query result may be stored as an operation reference case. In this way, user feedback is used to enrich the source of operation reference cases.
  • the target application includes at least one function, and the test query and the user query are both queries for a first function in the at least one function.
  • the target application may include functions such as application data analysis, application data update, and knowledge question and answer.
  • the test query and the user query need to be for the same function, for example, the test query and the user query are both for the application data analysis function.
  • the target application 120 may be necessary to evaluate the performance of the target application 120. For example, in the case where the target application includes multiple functions, it may be necessary to evaluate one or some of the functions.
  • the application management platform 110 may maintain a test sample set for testing the target application 120, for example, testing one or more functions of the target application 120, such as a data analysis function, a data update function, a knowledge question and answer function, etc.
  • Each test sample in the test sample set may include a sample query and a result (also referred to as a sample query result, a second result) generated by the target application 120 for the sample query.
  • a machine learning model in order to obtain the evaluation result of a function in the target application 120, can be used to obtain feedback on the sample query results in the evaluation sample.
  • the machine learning model can score the sample query results generated by the target application 120 from one or more dimensions.
  • the machine learning model can give feedback in any suitable manner.
  • an automatic evaluation method is described.
  • the data source e.g., knowledge base, data set, etc.
  • the machine learning model extracts queries and corresponding query results therefrom as correct examples.
  • questions and corresponding answers can be extracted.
  • the extracted queries can be executed in the target application 120, and the execution results are generated by the target application 120.
  • the machine learning model can give feedback on the execution results based on the answers and the execution results in various suitable dimensions, such as scoring. Examples of evaluation dimensions may include, but are not limited to, helpfulness to users, language fluency, factual correctness, logic, etc.
  • the feedback given by the machine learning model for each extracted question and execution result can determine the evaluation result of the target application 120. For example, if the evaluation sample set is used for the knowledge question and answer function, the evaluation result of the knowledge question and answer function can be obtained. In this way, the user 105 can debug the target application 120 according to the evaluation result, for example, debug one or more functions.
  • the creator or tester may input some evaluation queries.
  • the target application 120 may generate corresponding query results for the evaluation queries, also referred to as first results or evaluation query results. Then, the creator or tester may give feedback on the evaluation query results, also referred to as evaluation feedback.
  • the evaluation feedback can be used in conjunction with the evaluation sample set to provide a more accurate evaluation result for the target application 120.
  • the operation management platform 110 can generate an evaluation reference case for the target application 120 based on the evaluation query given by the creator or tester.
  • the evaluation reference case includes the evaluation query from the creator or tester, the evaluation query result generated by the target application 120 for the evaluation query, and the evaluation feedback on the evaluation query result.
  • the evaluation feedback can be positive feedback or negative feedback, and can also indicate the user 145's satisfaction with the query result.
  • evaluation results we expect the evaluation results to be consistent or match the evaluation feedback.
  • sample queries similar to the evaluation query can be found from the evaluation sample set, for example, the similarity exceeds the threshold similarity.
  • Evaluation samples including such similar sample queries are also called target evaluation samples, and the target evaluation samples also include sample results generated by the target application 120 for the sample query.
  • the application management platform 110 can generate an evaluation result for the target application 120 based on the evaluation query result, evaluation feedback and sample query result. For example, the score of the sample query result in the target evaluation sample can be matched with the evaluation feedback. It is understandable that the evaluation reference case provides examples of what kind of results the creator or tester is satisfied with or what kind of results the creator or tester is not satisfied with.
  • the evaluation of the target application 120 can be performed using a machine learning model.
  • a prompt word can be generated based on at least a portion of the evaluation reference case (e.g., evaluation feedback and evaluation query results), and the prompt word is provided to the machine learning model.
  • the machine learning model can generate feedback for the sample query results in the target evaluation sample.
  • the machine learning model can score the sample query results.
  • the evaluation results for the target application 120 can be determined.
  • the evaluation reference case gives examples of what kind of results are satisfactory to the creator or tester or what kind of results are unsatisfactory to the creator or tester.
  • the machine learning model can refer to such examples in generating feedback (e.g., scoring) for the sample results.
  • the machine learning model used in the evaluation can be the same as or different from the machine learning model used in the operation of the target application 120.
  • the sample query for evaluation and the evaluation query may be queries for the same function.
  • the evaluation sample set may be a knowledge question and answer function dedicated to evaluating the target application 120.
  • the collected user queries are also for the knowledge question and answer function.
  • FIG7A shows a flow chart of a process 700 of application debugging according to some embodiments of the present disclosure.
  • the process 700 may be implemented at the application management platform 110.
  • the process 700 is described below with reference to FIG7.
  • the application management platform 110 executes a test query for a target application.
  • the application management platform 110 presents corresponding first operation information of multiple stages of executing the test query, where the operation information of the multiple stages at least indicates respective outputs of the multiple stages.
  • the application management platform 110 generates an operation reference case for the target application based on the corresponding first operation information of the multiple stages of the test query.
  • the operation reference case includes the test query and the operation information of the multiple stages.
  • the operation reference case is referenced in the operation of the target application to determine the result of the user query.
  • generating an operation reference case includes: determining an adjustment to first operation information of at least one stage among multiple stages; re-executing a test query based on the adjusted first operation information of at least one stage; presenting corresponding second operation information of multiple stages of the re-executed test query; and in response to receiving a case generation confirmation, generating an operation reference case based on the corresponding second operation information.
  • re-executing the test query includes: receiving a stage execution request, the stage execution request instructing to execute the test query from a specified stage among a plurality of stages, the specified stage being before or included in at least one stage In at least one stage; based on the adjusted first operation information of at least one stage, re-execute a designated stage among the plurality of stages and a stage after the designated stage.
  • presenting corresponding second execution information of the plurality of phases of re-executing the test query includes presenting execution information associated with re-execution of a specified phase and phases subsequent to the specified phase.
  • the operation information of multiple stages also indicates at least one of the following: the input of each of the multiple stages, the operation time of each of the multiple stages, the operation log of each of the multiple stages, the operation status of each of the multiple stages, and the input information provided to the machine learning model when the machine learning model is called in at least one of the multiple stages.
  • determining adjustments to first run information of at least one of multiple stages includes: receiving term adjustment information for a first stage of the multiple stages, the term adjustment information indicating adding a term set to the first stage or removing a term set selected in the first stage, the term set including interpretations of terms used by a user group of a target application, and re-executing the test query includes: determining one or more term sets used in the first stage based on the term adjustment information; and re-executing the test query based on the one or more term sets.
  • determining an adjustment to first operating information of at least one stage among multiple stages includes: receiving use case adjustment information for a second stage among the multiple stages, the use case adjustment information indicating adding a reference case to the second stage or removing a reference case selected in the second stage, and re-executing the test query includes: determining one or more reference use cases used in the second stage based on the use case adjustment information; and re-executing the test query based on the one or more reference use cases.
  • generating an operation reference case for a target application includes: presenting data information while presenting corresponding first operation information of multiple stages, the data information at least indicating at least one data set used to execute the test query; and generating an operation reference case based on the corresponding first operation information of multiple stages and the data information.
  • the data information also indicates at least one of the following: a term set referenced for executing the test query (wherein the term set includes interpretations of terms used by a user group of the target application), a conversation record between a user and a digital assistant referenced for executing the test query, a reference prompt word for executing the test query, and conditions for determining candidate results for the test query.
  • the adjustment of the data information includes at least one of the following: updating at least one data set used to execute the test query, or updating at least a portion of at least one data set used to execute the test query.
  • process 700 also includes: in response to receiving a user query for a target application, determining that a running reference case is a target running reference case for the user query based on a similarity between the user query and the test query exceeding a threshold similarity; and executing the user query based on the target running reference case.
  • the target application includes at least one function, and wherein the test query and the user query are both queries for a first function of the at least one function.
  • the plurality of phases include: at least one predefined phase, the at least one predefined phase including a phase of generating a database query request; and at least one phase determined based on the test query.
  • the test query is a query for a first function in the target application, and wherein the at least one predefined phase is configured for the first function.
  • FIG. 7B shows a flowchart of a process 710 of information query according to some embodiments of the present disclosure.
  • the process 710 may be implemented at the application management platform 110.
  • the process 710 is described below with reference to FIG.
  • the application management platform 110 determines a running reference case including the test query as a target running reference case for the user query based on the similarity between the user query and the test query exceeding a threshold similarity, wherein the running reference case includes the test query and corresponding running information of multiple stages of the test query.
  • the application management platform 110 executes the user query based on the target run reference use case.
  • executing a user query based on a target running reference use case includes: providing the target reference use case as part of a prompt word for the user query to a machine learning model; and obtaining an execution result of the user query from the machine learning model.
  • process 710 if the user query comes from a tester and is implemented in the application testing phase, process 710 also includes: obtaining user feedback on the execution results of the user query; and in response to the user feedback being positive feedback, generating additional running reference use cases based on corresponding execution information for executing the user query in multiple stages.
  • FIG. 7C shows a flow chart of a process 720 of application evaluation according to some embodiments of the present disclosure.
  • the process 720 may be implemented at the application management platform 110.
  • the process 720 is described below with reference to FIG.
  • the application management platform 110 generates an evaluation reference case for the target application, wherein the evaluation reference case includes an evaluation query, a first result generated by the target application for the evaluation query, and evaluation feedback on the first result.
  • the application management platform 110 selects a target evaluation sample from the evaluation sample set based on the similarity between the evaluation query and queries in the evaluation sample set, wherein the target evaluation sample includes a sample query and a second result generated by the target application for the sample query.
  • the application management platform 110 generates an evaluation result for the target application based on the first result, the evaluation feedback, and the second result.
  • generating an evaluation result for the target application includes: providing a prompt word generated based on the first result and the evaluation feedback to a machine learning model; obtaining feedback from the machine learning model for the second result in the target evaluation sample; and determining the evaluation result for the target application based on the feedback for the second result.
  • the target application includes at least one function, and wherein both the sample query and the user query are queries for a second function of the at least one function.
  • FIG. 8A shows a schematic structural block diagram of an apparatus 800 for application debugging according to some embodiments of the present disclosure.
  • the apparatus 800 may be implemented in or included in the application management platform 110.
  • Each module/component in the apparatus 800 may be implemented by hardware, software, firmware or any combination thereof.
  • the device 800 includes a test query execution module 801, which is configured to execute a test query for a target application.
  • the device 800 also includes a first operation information presentation module 802, which is configured to present corresponding first operation information of multiple stages of executing the test query, and the operation information of multiple stages at least indicates the output of each of the multiple stages.
  • the device 800 also includes an operation reference case generation module 803, which is configured to generate an operation reference case for the target application based on the corresponding first operation information of the multiple stages of the test query, and the operation reference case includes the test query and the operation information of multiple stages, and the operation reference case is referenced in the operation of the target application to determine the result of the user query.
  • the reference case generation module 803 includes: a test query re-execution module, configured to determine adjustments to first operation information of at least one of multiple stages; re-execute the test query based on the adjusted first operation information of at least one stage; present corresponding second operation information of multiple stages of the re-executed test query; and in response to receiving a case generation confirmation, generate an operation reference case based on the corresponding second operation information.
  • a test query re-execution module configured to determine adjustments to first operation information of at least one of multiple stages; re-execute the test query based on the adjusted first operation information of at least one stage; present corresponding second operation information of multiple stages of the re-executed test query; and in response to receiving a case generation confirmation, generate an operation reference case based on the corresponding second operation information.
  • the test query re-execution module includes: a stage execution module, configured to receive a stage running request, the stage running request indicating to execute the test query from a specified stage among multiple stages, the specified stage being before at least one stage or included in at least one stage; based on the adjusted first running information of at least one stage, re-execute the specified stage among multiple stages and the stage after the specified stage.
  • the stage execution module includes: a re-execution information presentation module configured to present execution information associated with re-execution of a designated stage and a stage subsequent to the designated stage.
  • the operation information of multiple stages also indicates at least one of the following: the input of each of the multiple stages, the operation time of each of the multiple stages, the operation log of each of the multiple stages, the operation status of each of the multiple stages, and the input information provided to the machine learning model when the machine learning model is called in at least one of the multiple stages.
  • the test query re-execution module includes: a first adjustment information receiving module, configured to receive term adjustment information for a first stage among multiple stages, the term adjustment information indicating adding a term set to the first stage or removing a term set selected in the first stage, the term set including interpretations of terms used by a user group of a target application, and the test query re-execution module is further configured to determine one or more term sets used in the first stage based on the term adjustment information; and re-execute the test query based on the one or more term sets.
  • the test query re-execution module includes: a second adjustment information receiving module configured to receive use case adjustment information for a second stage in the plurality of stages, the use case adjustment information indicating adding a reference use case to the second stage or removing a reference use case selected in the second stage, and the test query re-execution module is further configured to receive a second adjustment information receiving module configured to receive use case adjustment information for a second stage in the plurality of stages, the use case adjustment information indicating adding a reference use case to the second stage or removing a reference use case selected in the second stage
  • the method is configured to determine one or more reference use cases used in the second phase based on the use case adjustment information; and re-execute the test query based on the one or more reference use cases.
  • the device 800 also includes: a data information presentation module, configured to present data information while presenting corresponding first operation information of multiple stages, the data information at least indicating at least one data set used to execute the test query; and generating an operation reference use case based on the corresponding first operation information and data information of multiple stages.
  • a data information presentation module configured to present data information while presenting corresponding first operation information of multiple stages, the data information at least indicating at least one data set used to execute the test query; and generating an operation reference use case based on the corresponding first operation information and data information of multiple stages.
  • the data information also indicates at least one of the following: a terminology set referenced for executing the test query, the terminology set including interpretations of terms used by the user group of the target application, a conversation record between a user and a digital assistant referenced for executing the test query, a reference prompt word for executing the test query, and conditions for determining candidate results for the test query.
  • the data information presenting module includes: a data updating module configured to update at least one data set used to execute a test query, or to update at least a portion of at least one data set used to execute a test query.
  • the device 800 also includes: a user query execution module, configured to, in response to receiving a user query for a target application, determine that the running reference case is a target running reference case for the user query based on the similarity between the user query and the test query exceeding a threshold similarity; and execute the user query based on the target running reference case.
  • a user query execution module configured to, in response to receiving a user query for a target application, determine that the running reference case is a target running reference case for the user query based on the similarity between the user query and the test query exceeding a threshold similarity; and execute the user query based on the target running reference case.
  • the target application includes at least one function, and wherein the test query and the user query are both queries for a first function of the at least one function.
  • the plurality of phases include: at least one predefined phase, the at least one predefined phase including a phase of generating a database query request; and at least one phase determined based on the test query.
  • the test query is a query for a first function in the target application, and wherein the at least one predefined phase is configured for the first function.
  • FIG. 8B shows a schematic structural block diagram of an apparatus 810 for information query according to some embodiments of the present disclosure.
  • the apparatus 810 may be implemented in or included in the application management platform 110.
  • Each module/component in the apparatus 810 may be implemented by hardware, software, firmware or any combination thereof.
  • the device 810 includes a running reference case determination module 811, which is configured to, in response to receiving a user query for a target application, determine a running reference case including a test query as a target running reference case for the user query based on the similarity between the user query and the test query exceeding a threshold similarity, wherein the running reference case includes the test query and corresponding running information of multiple stages of the test query.
  • the device 810 also includes a user query execution module 812, which is configured to execute the user query based on the target running reference case.
  • the user query execution module 812 includes an execution result acquisition module, which is configured to provide the target reference use case as part of the prompt word for the user query to the machine learning model; and obtain the execution result of the user query from the machine learning model.
  • the device 810 also includes an additional operation reference case generation module, which is configured to obtain user feedback on the execution results of the user query; and in response to the user feedback being positive feedback, generate additional operation reference cases based on corresponding execution information for executing the user query in multiple stages.
  • an additional operation reference case generation module configured to obtain user feedback on the execution results of the user query; and in response to the user feedback being positive feedback, generate additional operation reference cases based on corresponding execution information for executing the user query in multiple stages.
  • FIG. 8C shows a schematic structural block diagram of an apparatus 820 for application evaluation according to some embodiments of the present disclosure.
  • the apparatus 820 may be implemented in or included in the application management platform 110.
  • Each module/component in the apparatus 820 may be implemented by hardware, software, firmware or any combination thereof.
  • the device 820 includes an evaluation reference case generation module 821, which is configured to generate an evaluation reference case for the target application, and the evaluation reference case includes an evaluation query, a first result generated by the target application for the evaluation query, and an evaluation feedback on the first result.
  • the device 820 also includes a target evaluation sample selection module 822, which is configured to select a target evaluation sample from the evaluation sample set based on the similarity between the evaluation query and the query in the evaluation sample set, and the target evaluation sample includes a sample query and a second result generated by the target application for the sample query.
  • the device 820 also includes an evaluation result generation module 823, which is configured to generate an evaluation result for the target application based on the first result, the evaluation feedback, and the second result.
  • the evaluation result generation module 823 includes a model feedback acquisition module, which is configured to provide prompt words generated based on the first result and the evaluation feedback to the machine learning model; obtain feedback on the second result in the target evaluation sample from the machine learning model; and determine the evaluation result for the target application based on the feedback on the second result.
  • a model feedback acquisition module configured to provide prompt words generated based on the first result and the evaluation feedback to the machine learning model; obtain feedback on the second result in the target evaluation sample from the machine learning model; and determine the evaluation result for the target application based on the feedback on the second result.
  • the target application includes at least one function, and wherein both the sample query and the user query are queries for a second function of the at least one function.
  • FIG9 shows a block diagram of an electronic device 900 in which one or more embodiments of the present disclosure may be implemented. It should be understood that the electronic device 900 shown in FIG9 is merely exemplary and should not constitute any limitation on the functionality and scope of the embodiments described herein.
  • the electronic device 900 shown in FIG9 may include or be implemented as the application management platform 110 of FIG1 , the device 800 of FIG8A , the device 810 of FIG8B , or the device 820 of FIG8C .
  • the electronic device 900 is in the form of a general electronic device.
  • the components of the electronic device 900 may include, but are not limited to, one or more processors or processing units 910, a memory 920, a storage device 930, one or more communication units 940, one or more input devices 950, and one or more output devices 960.
  • the processing unit 910 may be an actual or virtual processor and is capable of performing various processes according to a program stored in the memory 920. In a multi-processor system, multiple processing units execute computer executable instructions in parallel to improve the parallel processing capability of the electronic device 900.
  • the electronic device 900 typically includes a plurality of computer storage media. Such media may be any accessible media that is accessible to the electronic device 900, including but not limited to volatile and nonvolatile media, removable and non-removable media.
  • the memory 920 may be a volatile memory (e.g., registers, cache, random access memory (RAM)), a nonvolatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof.
  • the storage device 930 may be a removable or non-removable
  • the electronic device 900 may include a machine-readable medium such as a flash drive, a disk, or any other medium that may be used to store information and/or data and may be accessed within the electronic device 900.
  • the electronic device 900 may further include additional removable/non-removable, volatile/non-volatile storage media.
  • a disk drive for reading or writing from a removable, non-volatile disk e.g., a “floppy disk”
  • an optical drive for reading or writing from a removable, non-volatile optical disk may be provided.
  • each drive may be connected to a bus (not shown) by one or more data media interfaces.
  • the memory 920 may include a computer program product 925 having one or more program modules configured to perform various methods or actions of various embodiments of the present disclosure.
  • the communication unit 940 implements communication with other electronic devices through a communication medium. Additionally, the functions of the components of the electronic device 900 can be implemented with a single computing cluster or multiple computing machines that can communicate through a communication connection. Therefore, the electronic device 900 can operate in a networked environment using a logical connection with one or more other servers, a network personal computer (PC), or another network node.
  • PC network personal computer
  • the input device 950 may be one or more input devices, such as a mouse, a keyboard, a tracking ball, etc.
  • the output device 960 may be one or more output devices, such as a display, a speaker, a printer, etc.
  • the electronic device 900 may also communicate with one or more external devices (not shown) through the communication unit 940 as needed, such as a storage device, a display device, etc., communicate with one or more devices that allow a user to interact with the electronic device 900, or communicate with any device that allows the electronic device 900 to communicate with one or more other electronic devices (e.g., a network card, a modem, etc.). Such communication may be performed via an input/output (I/O) interface (not shown).
  • I/O input/output
  • a computer-readable storage medium on which computer-executable instructions are stored, wherein the computer-executable instructions are executed by a processor to implement the method described above.
  • a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, and the computer-executable instructions are executed by a processor to implement the method described above.
  • These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processing unit of the computer or other programmable data processing device, a device that implements the functions/actions specified in one or more boxes in the flowchart and/or block diagram is generated.
  • These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and/or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions/actions specified in one or more boxes in the flowchart and/or block diagram.
  • Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, so that the instructions executed on the computer, other programmable data processing apparatus, or other device implement the functions/actions specified in one or more boxes in the flowchart and/or block diagram.
  • each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of a module, program segment or instruction includes one or more executable instructions for realizing the logical function of the specification.
  • the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved.
  • each square box in the block diagram and/or flow chart, and the combination of the square boxes in the block diagram and/or flow chart can be realized by a special hardware-based system that performs the function or action of the specification, or can be realized by a combination of special hardware and computer instructions.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Debugging And Monitoring (AREA)
  • Stored Programmes (AREA)

Abstract

本公开的实施例提供了应用调试的方法、信息查询的方法、应用评测的方法、装置、设备和存储介质。应用调试的方法包括:执行针对目标应用的测试查询;呈现执行测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出;以及基于测试查询的多个阶段的相应第一运行信息,生成针对目标应用的运行参考用例,运行参考用例包括测试查询和多个阶段的运行信息,运行参考用例在目标应用的运行中被参考用于确定用户查询的结果。以此方式,应用创建用户无需修改代码就可以提升目标应用提供服务的质量,简化了目标应用的调试过程。

Description

应用调试的方法、信息查询的方法、应用评测的方法
本申请要求2023年11月22日递交的、标题为“应用调试的方法、信息查询的方法、应用评测的方法”、申请号为202311570383.1的中国发明专利申请的优先权,该申请的全部内容通过引用结合在本申请中。
技术领域
本公开的示例实施例总体涉及计算机领域,特别地涉及用于应用调试的方法、信息查询的方法、应用评测的方法、装置、设备和计算机可读存储介质。
背景技术
随着信息技术的发展,各种终端设备可以在工作和生活等方面向人们提供各种服务。终端设备中可以部署有提供服务的应用。终端设备通过应用的用户界面呈现相应的内容并实现与用户之间的交互,满足用户的各类需求。在一些情况下,用户可能在应用内发起查询请求。因此,如何提高向用户提供的查询服务的质量是所关注的问题。
发明内容
在本公开的第一方面,提供了一种应用调试的方法。该方法包括:执行针对目标应用的测试查询;呈现执行测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出;以及基于测试查询的多个阶段的相应第一运行信息,生成针对目标应用的运行参考用例,运行参考用例包括测试查询和多个阶段的运行信息,运行参考用例在目标应用的运行中被参考用于确定用户查询的结果。
在本公开的第二方面,提供了一种信息查询的方法。该方法包括:响应于接收到针对目标应用的用户查询,基于用户查询与测试查询之间的相似度超过阈值相似度,将包括测试查询的运行参考用例确定为用户查询的目标运行参考用例,其中运行参考用例包括测试查询和测试查询的多个阶段的相应运行信息;以及基于目标运行参考用例来执行所述用户查询。
在本公开的第三方面,提供了一种应用评测的方法。该方法包括:生成用于所述目标应用的评测参考用例,所述评测参考用例包括评测查询、所述目标应用针对所述评测查询而生成的第一结果和对所述第一结果的评测反馈;基于所述评测查询与评测样本集中的查询之间的相似度,从所述评测样本集中选择目标评测样本,所述目标评测样本包括样本查询和由所述目标应用针对所述样本查询生成的第二结果;以及基于 所述第一结果、所述评测反馈和所述第二结果,生成针对所述目标应用的评测结果。
在本公开的第四方面,提供了一种用于应用调试的装置。该装置包括:测试查询执行模块,被配置为执行针对目标应用的测试查询;第一运行信息呈现模块,被配置为呈现执行测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出;以及运行参考用例生成模块,被配置为基于测试查询的多个阶段的相应第一运行信息,生成针对目标应用的运行参考用例,运行参考用例包括测试查询和多个阶段的运行信息,运行参考用例在目标应用的运行中被参考用于确定用户查询的结果。
在本公开的第五方面,提供了一种用于信息查询的装置。该装置包括:运行参考用例确定模块,被配置为响应于接收到针对目标应用的用户查询,基于所述用户查询与测试查询之间的相似度超过阈值相似度,将包括所述测试查询的运行参考用例确定为所述用户查询的目标运行参考用例,其中所述运行参考用例包括所述测试查询和所述测试查询的多个阶段的相应运行信息;以及用户查询执行模块,被配置为基于所述目标运行参考用例来执行所述用户查询。
在本公开的第六方面,提供了一种用于应用评测的装置。该装置包括:评测参考用例生成模块,被配置为生成用于所述目标应用的评测参考用例,所述评测参考用例包括评测查询、所述目标应用针对所述评测查询而生成的第一结果和对所述第一结果的评测反馈;目标评测样本选择模块,被配置为基于所述评测查询与评测样本集中的查询之间的相似度,从所述评测样本集中选择目标评测样本,所述目标评测样本包括样本查询和由所述目标应用针对所述样本查询生成的第二结果;以及评测结果生成模块,被配置为基于所述第一结果、所述评测反馈和所述第二结果,生成针对所述目标应用的评测结果。
在本公开的第七方面,提供了一种电子设备。该设备包括至少一个处理单元;以及至少一个存储器,至少一个存储器被耦合到至少一个处理单元并且存储用于由至少一个处理单元执行的指令。指令在由至少一个处理单元执行时使电子设备执行第一方面和第二方面的方法。
在本公开的第八方面,提供了一种计算机可读存储介质。介质上存储有计算机程序,计算机程序被处理器执行时实现第一方面和第二方面的方法。
应当理解,该部分中所描述的内容并非旨在限定本公开的实施例的关键特征或重要特征,也不用于限制本公开的范围。本公开的其他特征将通过以下的描述而变得容易理解。
附图说明
结合附图并参考以下详细说明,本公开各实施例的上述和其他特征、优点及方面 将变得更加明显。在附图中,相同或相似的附图标记表示相同或相似的元素,其中:
图1示出了能够在其中实现本公开的实施例的示例环境的示意图;
图2示出了根据本公开的一些实施例的目标应用中各个功能的调试入口的页面的示例;
图3示出了根据本公开的一些实施例的针对目标应用中特定功能的测试查询记录的页面的示例;
图4示出了根据本公开的一些实施例的对目标应用进行调试的页面的示例;
图5示出了根据本公开的一些实施例的展示数据信息的页面的示例;
图6A示出了根据本公开的一些实施例的针对目标应用中特定功能的用例库的页面的示例;
图6B示出了根据本公开的一些实施例的用例详情的页面的示例;
图7A示出了根据本公开的一些实施例的应用调试的过程的流程图;
图7B示出了根据本公开的一些实施例的信息查询的过程的流程图;
图7C示出了根据本公开的一些实施例的应用评测的过程的流程图;
图8A示出了根据本公开的一些实施例的用于应用调试的装置的示意性结构框图;
图8B示出了根据本公开的一些实施例的用于信息查询的装置的示意性结构框图;
图8C示出了根据本公开的一些实施例的用于应用评测的装置的示意性结构框图;以及
图9示出了可以实现本公开的一个或多个实施例的电子设备的框图。
具体实施方式
下面将参照附图更详细地描述本公开的实施例。虽然附图中示出了本公开的某些实施例,然而应当理解的是,本公开可以通过各种形式来实现,而且不应该被解释为限于这里阐述的实施例,相反,提供这些实施例是为了更加透彻和完整地理解本公开。应当理解的是,本公开的附图及实施例仅用于示例性作用,并非用于限制本公开的保护范围。
在本公开的实施例的描述中,术语“包括”及其类似用语应当理解为开放性包含,即“包括但不限于”。术语“基于”应当理解为“至少部分地基于”。术语“一个实施例”或“该实施例”应当理解为“至少一个实施例”。术语“一些实施例”应当理解为“至少一些实施例”。下文还可能包括其他明确的和隐含的定义。
在本文中,除非明确说明,“响应于A”执行一个步骤并不意味着在“A”之后立即执行该步骤,而是可以包括一个或多个中间步骤。
可以理解的是,本技术方案所涉及的数据(包括但不限于数据本身、数据的获得、使用、存储或删除)应当遵循相应法律法规及相关规定的要求。
可以理解的是,在使用本公开各实施例公开的技术方案之前,均应当根据相关法律法规通过适当的方式对本公开所涉及信息的类型、使用范围、使用场景等告知相关用户并获得相关用户的授权,其中,相关用户可以包括任何类型的权利主体,例如个人、企业、团体。
例如,在响应于接收到用户的主动请求时,向相关用户发送提示信息,以明确地提示相关用户,其请求执行的操作将需要获得和使用到相关用户的信息,从而使得相关用户可以根据提示信息来自主地选择是否向执行本公开技术方案的操作的电子设备、应用程序、服务器或存储介质等软件或硬件提供信息。
作为一种可选的但非限制性的实现方式,响应于接收到相关用户的主动请求,向相关用户发送提示信息的方式,例如可以是弹窗的方式,弹窗中可以以文字的方式呈现提示信息。此外,弹窗中还可以承载供用户选择“同意”或“不同意”向电子设备提供信息的选择控件。
可以理解的是,上述通知和获得用户授权过程仅是示意性的,不对本公开的实现方式构成限定,其他满足相关法律法规的方式也可应用于本公开的实现方式中。
可以理解的是,本技术方案所涉及的数据(包括但不限于数据本身、数据的获取或使用)应当遵循相应法律法规及相关规定的要求。本公开实施例数字助手相关功能的启用、获取的数据、数据的处理和存储方式等,均应获得所述用户、以及与所述用户关联的其他权利主体的提前授权,并应符合相关法律法规、权利主体间协议规则的约定。
如本文中所使用的,术语“模型”可以从训练数据中学习到相应的输入与输出之间的关联关系,从而在训练完成后可以针对给定的输入,生成对应的输出。模型的生成可以基于机器学习技术。深度学习是一种机器学习算法,通过使用多层处理单元来处理输入和提供相应输出。神经网络模型是基于深度学习的模型的一个示例。在本文中,“模型”也可以被称为“机器学习模型”、“学习模型”、“机器学习网络”或“学习网络”,这些术语在本文中可互换地使用。
图1示出了本公开的实施例能够在其中实现的示例环境100的示意图。环境100涉及应用管理平台110,其可以支持应用的创建和/或应用的运行。在一些实施例中,应用管理平台110中用于支持应用创建的部分也可以称为应用创建部分。在一些实施例中,应用管理平台110中用于支持应用运行的部分也可以称为应用运行部分。
如图1所示,应用创建部分可以为用户105提供应用的创建和发布环境。用户105可以被称为应用创建用户、创建者。在一些实施例中,应用创建部分可以是低代码平台,其提供应用创建的工具合集。应用创建部分可以支持对各类应用的可视化开发,从而使开发人员可以跳过手工编码的过程,加快应用的开发周期和成本。应用创建部分可以支持用户开发一个或多个类型的应用的任何适当平台,例如可以包括基于应用 平台即服务(aPaaS)的平台。这样的平台能够支持用户对应用进行高效开发,实现应用创建、应用功能调整等操作。
应用创建部分可以被部署在用户105的终端设备本地,和/或可以由服务端设备来支持。例如,用户105的终端设备可以运行有应用创建部分的客户端,该客户端可以支持用户与服务端所提供的应用创建部分的交互。在应用创建部分运行在用户的终端设备本地的情况下,用户105可以直接利用终端设备与本地的应用创建部分进行交互。在应用创建部分运行在服务端设备的情况下,服务端设备可以基于与终端设备之间的通信连接,实现对终端设备中运行的客户端的服务供应。应用创建部分可以基于用户105的操作而向用户105呈现相应的页面130,以向用户105输出和/或从用户105接收与应用创建相关的信息。
在一些实施例中,应用创建部分可以关联到相应的数据库,其中存储应用创建部分所支持的应用创建过程所需的数据或信息。例如,数据库可以存储用于组成应用的各个功能模块对应的代码和描述信息等。应用创建部分还可以对数据库中的功能模块执行调用、增加、删除、更新等操作。数据库还可以存储对不同功能块可执行的操作。示例性的,在要创建应用的场景,应用创建部分可以从数据库中调用相应的功能块来搭建应用。
在本公开的实施例中,用户105可以在应用创建部分上根据需要创建目标应用120,并发布目标应用120。目标应用120可被发布到任一适当的应用运行部分,只要应用运行部分能够支持目标应用120的运行。在发布后,目标应用120可以用于由一个或多个用户145进行操作。用户145可以被称目标应用120的终端用户。在一些实施例中,目标应用120可以包括或者被实现为数字助手122。
数字助手122可以被配置为具有智能对话。在图1所示的示例中,数字助手122可以被集成在目标应用120内,作为目标应用120的一一部分来协助执行目标应用120内的任务处理。在其他示例中,数字助手122可以被配置为独立运行的应用,例如网页应用或其他类型的应用。在这样的示例中,数字助手122与目标应用120可以被视为同一应用。数字助手122被提供用于协助用户在不同应用、场景下的各类任务处理需求。在与数字助手122的交互过程中,用户输入交互消息,并且数字助手122响应于用户输入来提供回复消息。通常,数字助手122能够支持用户以自然语言方式输入提问,并基于对自然语言输入的理解和逻辑推理能力来执行任务和提供回复。
在一些实施例中,数字助手122可以作为用户145的联系人与之交互。例如,数字助手122可以被实现在即时消息(IM)应用中。数字助手122可以在与用户145的单聊会话中与用户145交互。在一些实施例中,数字助手122可以在包括多个用户的群聊会话中与多个用户交互。
对于每个用户145,应用运行部分的客户端可以在客户端界面中呈现目标应用 120或数字助手122的交互窗口142,例如与数字助手122的会话窗口。用户145可以在会话窗口中输入会话消息,目标应用120可以基于所创建的配置信息,来确定数字助手122的回复消息并在交互窗口142中呈现给用户。在一些实施例中,取决于目标应用120的配置,与目标应用120的交互消息可以包括多模态形式的消息,诸如文本消息(例如,自然语言文本)、语音消息、图像消息、视频消息,等等。
与应用创建部分类似,应用运行部分可以被部署在每个用户145的终端设备本地,和/或可以由服务端设备来支持。例如,用户145的终端设备可以运行有应用运行部分的客户端,该客户端可以支持用户与服务端所提供的应用运行部分的交互。在应用运行部分运行在用户的终端设备本地的情况下,用户145可以直接利用终端设备与本地的应用运行部分进行交互。在应用运行部分运行在服务端设备的情况下,服务端设备可以基于与终端设备之间的通信连接,实现对终端设备中运行的客户端的服务供应。应用运行部分可以基于用户145的操作而向用户145呈现相应的应用页面,以向用户145输出和/或从用户145接收与应用使用相关的信息。
在一些实施例中,目标应用120的至少部分功能的实现,和/或目标应用120中的数字助手122的至少部分功能的实现可以基于模型来实现。在目标应用120的创建或运行过程中,可以调用一个或多个模型155,例如模型155的能力。在目标应用120中,数字助手122可以利用模型155来理解用户输入,并基于模型155的输出来提供对用户的回复。
在创建过程中,应用创建平台110对目标应用120的测试需要利用模型155来确定目标应用120的运行结果符合预期。在运行过程中,响应于目标应用120的用户的不同操作请求,应用运行部分可能需要利用模型155来确定对用户的响应结果。
虽然被示出为独立于应用管理平台110,但一个或多个模型155可以运行在应用管理平台110上,或者其他远端服务器。在一些实施例中,模型155可以是机器学习模型、深度学习模型、学习模型、神经网络等。在一些实施例中,模型可以基于语言模型(LM)。语言模型通过从大量语料中学习,能够具备问答能力。模型155也可以基于其他适当的模型。
管理平台110可以运行在适当的电子设备。这里的电子设备可以是任何类型的具有计算能力的设备,包括终端设备或服务端设备。终端设备可以是任何类型的移动终端、固定终端或便携式终端,包括移动手机、台式计算机、膝上型计算机、笔记本计算机、上网本计算机、平板计算机、媒体计算机、多媒体平板、个人通信系统(PCS)设备、个人导航设备、个人数字助理(PDA)、音频/视频播放器、数码相机/摄像机、定位设备、电视接收器、无线电广播接收器、电子书设备、游戏设备或者前述各项的任何组合,包括这些设备的配件和外设或者其任何组合。服务端设备例如可以包括计算系统/服务器,诸如大型机、边缘计算节点、云环境中的计算设备,等等。在一些实 施例中,管理平台110可以基于云服务来实现。
应当理解,仅出于示例性的目的描述环境100的结构和功能,而不暗示对于本公开的范围的任何限制。例如,虽然图1示出了与应用创建部分交互的单个用户和与应用运行部分交互的单个用户,但实际上可以多个用户访问应用管理平台110来各自创建数字助手,并且每个数字助手可以被用于与多个用户进行交互。
用户在应用内发起查询请求时,该应用可以利用机器学习模型来提供查询服务。在一开始该应用提供的查询服务的质量在可能并不高,需要经过多轮优化来提升查询服务的效果。传统的优化需要在代码层面对系统进行改进,存在成本高、耗时长、容易出现错误的问题。
为此,根据本公开的实施例提供了一种应用调试的方案。根据本公开的各种实施例,执行针对目标应用的测试查询。呈现执行测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出。基于测试查询的多个阶段的相应第一运行信息,生成针对目标应用的运行参考用例,运行参考用例包括测试查询和多个阶段的运行信息,运行参考用例在目标应用的运行中被参考用于确定用户查询的结果。
在本公开的各种实施例中,应用创建用户只需要依据执行测试查询的多个阶段相应的运行信息,就可以生成针对目标应用的运行参考用例。在针对终端用户发起的用户查询来生成查询结果时,可以参考该运行参考用例。以此方式,应用创建用户无需修改代码就可以提升目标应用提供服务的质量,简化了目标应用的调优过程。
以下将参考附图的示例来详细描述本公开的一些示例实施例。应当理解,附图中示出的页面仅仅是示例,实际可以存在各种页面设计。页面中的各个图形元素可以具有不同的布置和不同的视觉表示,其中的一个或多个元素可以省略或被替换,并且还可以存在一个或多个其他元素。本公开的实施例在此方面不受限制。
下面继续参考附图来描述本公开的示例实施例。在下文的示例中,出于讨论的需要,从应用管理平台的角度来描述,例如图1所示的应用管理平台110。应用管理平台110所呈现的页面,可以经由用户105的终端设备被呈现,并且可以经由用户105的终端设备来接收用户输入。
图2示出了根据本公开的一些实施例的目标应用120中各个功能的调试入口的页面200的示例。如图2所示,在用户105的终端设备呈现有目标应用120中各个功能的调试入口的页面200。页面200给出了功能区域210,其可以用于目标应用中的各种功能。目标应用中的功能例如可以包括应用数据分析功能、应用数据更新功能、知识问答功能等。但应当理解,这仅是示例性的功能,而无意任何限制。这里的目标应用可以是任何合适的应用,例如人力资源应用、CRM(Customer Relationship Management,客户关系管理)应用、财务应用等。以对目标应用中的应用数据分析功能调试为例, 用户105触发了数据分析功能的调试控件220,例如,用户105点击了调试控件220。响应于此,可以向用户105呈现测试查询记录的页面。
图3示出了根据本公开的一些实施例的针对目标应用中特定功能的测试查询记录的页面300的示例。如图3所示,可以在用户105的终端设备呈现针对目标应用中特定功能的测试查询记录的页面300。页面300记录了用户105之前创建的测试查询,用户105通过输入测试查询来对相应的功能进行调试。在一些实施例中,用户105可以重新调试先前创建的测试查询。例如,用户105点击了控件310即可重新执行测试查询“单价低于XX元的商品的数量是多少?”,以对目标应用中相应的功能进行调试。响应于此,可以向用户105呈现调试页面。在一些实施例中,用户105可以触发执行新的测试查询。例如,用户105点击了新增控件320即可输入新的测试查询,以对目标应用中相应的功能进行调试。响应于此,用户105可以进入调试页面。
图4示出了根据本公开的一些实施例的对目标应用进行调试的页面400的示例。应用管理平台110可以执行针对目标应用的测试查询。参考图4,例如,用户105在测试查询对应的文本区域410输入测试查询语句,并点击控件420,来执行针对目标应用的测试查询。用户105输入的测试查询语句例如可以是“单价低于XX元的商品的数量是多少?”、“产地是XX的商品的占比是多少?画一个饼状图”。应用管理平台110可以在分阶段调试区域430呈现执行测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出。
应用管理平台110在执行针对目标应用的测试查询之后,可以在分阶段调试区域430呈现执行测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出。例如,在分阶段调试区域430中示出了5个阶段的运行信息,这5个阶段需要执行各自的任务。在任务执行完成后,生成对应的输出440、442、444、446和448。例如,第三阶段的输出444可以包括需要执行测试查询所需的数据表。
在一些实施例中,多个阶段的运行信息还可以指示以下至少一项:多个阶段各自的输入,多个阶段各自的运行耗时,多个阶段各自的运行日志,多个阶段各自的运行状态(即,是否运行完成)。在一些实施例中,由于机器学习模型的输入信息(例如,输入信息的长度)存在限制,因此,多个阶段的运行信息还可以指示多个阶段中的至少一个阶段中调用机器学习模型时,被提供给所述机器学习模型的输入信息。以此方式,用户105可以使提供给机器学习模型的输入信息满足机器学习模型的要求,从而保证机器学习模型稳定运行。
在一些实施例中,分阶段调试区域430中展示的多个阶段可以包括至少一个预定义阶段,至少一个预定义阶段包括生成数据库查询请求的阶段。该至少一个预定义阶段可以是应用管理平台110根据目标应用的特性和/或对应的功能预先设置的一些调试阶段。应用管理平台110执行针对目标应用的测试查询的目的可以是生成一个更好的 数据库查询请求,以向用户返回更加精确的查询结果。因此,该至少一个预定义阶段还可以包括生成数据库查询请求的阶段,该阶段可以基于之前的预定义阶段的输出生成准确和高效的数据库查询请求。这里的数据库查询请求例如可以是SQL(Structured Query Language,结构化查询语言)语句。
在一些实施例中,分阶段调试区域430中展示的多个阶段除了至少一个预定义阶段,还可以包括基于测试查询确定的至少一个阶段。在应用管理平台110运行数据库查询请求获取到对应的数据之后,还可以基于测试查询的要求确定至少一个阶段。在一个示例中,测试查询是“单价低于XX元的商品的数量是多少?”,该测试查询并没有附加的数据分析要求,因此应用管理平台110可以只确定一个阶段,例如,数据查看阶段,应用管理平台110可以将单价低于XX元的商品的数量在该阶段输出。在另一示例中,测试查询是“产地是XX的商品的占比是多少?画一个饼状图”,该测试查询有附加的数据分析要求,因此应用管理平台110可以生成多个阶段,例如,生成图表阶段、数据洞察阶段等,应用管理平台110可以将产地是XX的商品的占比是多少的饼图输出到生成图表阶段,应用管理平台110还可以将产地是XX的商品的占比的数据的挖掘和分析输出到数据洞察阶段。以此方式,应用管理平台110可以基于测试查询在预定义阶段之后确定不同的阶段,可以满足用户不同的查询需求,提升用户的查询体验。
在一些实施例中,应用管理平台110执行的针对目标应用的测试查询是针对目标应用中的第一功能的查询。由于目标应用中不同的功能具有较大的差异性,针对目标应用中一个功能的调试可能对目标应用中另一个功能并不适用,因此测试查询是针对目标应用中一个特定功能的查询。
在一些实施例中,对于目标应用中的不同功能而言,预定义阶段可以是不同的。例如,以第一功能为例,至少一个预定义阶段可以是针对目标应用中的第一功能来配置的。应用管理平台110可以基于目标应用中不同功能的特性,来配置不同的至少一个预定义阶段。示例性地,针对目标应用中的应用数据分析功能,应用管理平台110可以配置5个预定义阶段,分别是召回并选择术语阶段、召回并选择运行参考用例阶段、召回并选择目标数据集阶段、召回并选择目标数据集所属字段阶段、生成并执行数据库查询请求阶段。在召回并选择术语阶段可以召回并选择测试查询中使用的术语的解释。在召回并选择运行参考用例阶段可以召回并选择与测试查询相似的运行参考用例。在召回并选择目标数据集阶段可以召回并选择执行测试查询所需的数据集。在召回并选择目标数据集所属字段阶段可以召回并选择执行测试查询所需的数据集中的字段。在生成并执行数据库查询请求阶段可以基于前面4个阶段的输出,生成并执行数据库查询请求。
上文以数据分析功能作为示例描述了阶段的划分,但这种划分仅是示例性的。此 外,针对其他的功能(例如,数据更新、知识问答等)可以根据相应功能的特性进行合适的阶段划分。以此方式,使用测试查询对目标应用中的不同功能分别进行调试,充分考虑到了不同功能之间的差异性,可以有效地提升不同功能提供的服务质量。
在呈现了执行测试查询的多个阶段的相应第一运行信息之后,应用管理平台110可以基于测试查询的多个阶段的相应第一运行信息,生成针对目标应用的运行参考用例。运行参考用例包括测试查询和多个阶段的运行信息,运行参考用例在目标应用的运行中被参考用于确定用户查询的结果。例如,用户105可以在页面400中先将点击已调优控件450,将测试查询和对应的运行信息确定已经完成了调优,再点击保存控件460以生成运行参考用例。在生成了运行参考用例之后,应用管理平台110在接收到用户145的用户查询时,可以参考运行参考用例来确定用户查询的结果。
在一些实施例中,应用管理平台110基于测试查询的多个阶段的相应第一运行信息,可以确定对多个阶段中的至少一个阶段的第一运行信息的调整。而后,应用管理平台110可以基于经调整的至少一个阶段的第一运行信息,重新执行测试查询。应用管理平台110可以呈现重新执行测试查询的多个阶段的相应第二运行信息。如果接收到用例生成确认,应用管理平台110可以基于相应第二运行信息,生成所运行参考用例。在至少一个阶段的第一运行信息调整之后,应用管理平台110可以重新执行测试查询,并在分阶段调试区域430中呈现重新执行测试查询的多个阶段的相应第二运行信息。参考图4,应用管理平台110可以通过用户105对已调优控件450和保存控件460的触发接收到用例生成确认,以基于第二运行信息生成运行参考用例。
在一些实施例中,应用管理平台110可以接收到用户105针对多个阶段中的第一阶段的术语调整信息,并基于术语调整信息重新执行测试查询。该术语调整信息指示向第一阶段添加术语集或移除在所述第一阶段中选择的术语集。术语集包括对目标应用的用户群体所使用的术语的解释。用户群体所使用的术语可能是在该群体的范围内被使用的,但并不是一种通用术语,因此需要帮助机器学习模型理解这样的术语。对这种术语的解释可以包括对该术语的文字性解释,也可以包括与该术语相对应的通用术语。
示例性地,执行测试查询的第一阶段可以选择与测试查询相关的术语集,用户105可以在第一阶段添加术语集,也可以移除术语集。例如,用户105针对目标应用的测试查询为“通过C2C渠道销售的商品的数量是多少?”,其中C2C是专用术语,第一阶段选择的术语集可以包括对C2C的解释,用户105可以添加其他术语集或删除第一阶段选择的术语集。在调整完术语集后,应用管理平台110可以基于术语调整信息,确定在第一阶段中使用的一个或多个术语集,然后基于该一个或多个术语集重新执行测试查询。
在一些实施例中,应用管理平台110可以接收针对多个阶段中的第二阶段的用例 调整信息,并基于用例调整信息重新执行测试查询。用例调整信息指示向第二阶段添加参考用例或移除所述第二阶段中所选择的参考用例。示例性地,执行测试查询的第二阶段可以选择与该测试查询相似的用例,用户105可以在第二阶段添加参考用例,也可以移除参考用例。例如,用户105针对目标应用的测试查询为“单价小于5000元的商品的数量是多少?”,第二阶段可以基于相似度计算选择与该测试查询相似的参考用例,第二阶段选择的参考用例可以是“单价小于8000元的商品的数量是多少?”,用户105可以添加其他参考用例或移动第二阶段选择的参考用例。在调整完参考用例之后,应用管理平台110可以基于用例调整信息,确定在第二阶段中使用的一个或多个参考用例,然后基于该一个或多个参考用例,重新执行测试查询。
在一些实施例中,应用管理平台110可以接收针对多个阶段中的第三阶段的数据集调整信息,并基于数据集调整信息重新执行测试查询。示例性地,执行测试查询的第三阶段可以选择执行测试查询所需的数据集,用户105可以在第三阶段添加数据集,也可以移除数据集。例如,用户105针对目标应用的测试查询为“单价小于5000元的商品的数量是多少?”,第三阶段选择的数据集可以是商品分析表,用户105可以添加其他数据集,也可以删除第三阶段选择的数据集。
在一些实施例中,应用管理平台110还可以接收针对多个阶段中的第四阶段的数据集字段调整信息,并基于数据集字段调整信息重新执行测试查询。示例性地,执行测试查询的第四阶段可以选择执行测试查询所需的数据集的字段,用户105可以在第四阶段添加字段,也可以移除字段。例如,用户105针对目标应用的测试查询为“单价小于5000元的商品的数量是多少?”,第四阶段选择的数据集的字段是商品分析表中的商品号字段、单价字段、产地字段和生产日期字段,用户105可以添加其他字段,也可以删除第四阶段选择的字段。以此方式,用户可以通过图形界面的方式对测试查询的各个阶段进行调整,方便用户对于调试的理解,简化了目标应用的调试过程。
在一些实施例中,应用管理平台110在确定对至少一个阶段的第一运行信息的调整之后,可以接收阶段运行请求。阶段运行请求指示从多个阶段中的指定阶段起执行测试查询,指定阶段在至少一个阶段之前或被包括在至少一个阶段中,并且基于经调整的至少一个阶段的第一运行信息,重新执行多个阶段中的指定阶段以及指定阶段之后的阶段。示例性地,用户105对第二阶段进行了调整,此时指定阶段可以是第二阶段之前的第一阶段,也可以是第二阶段。在指定阶段是第一阶段的情况下,应用管理平台110可以基于经调整的至少第二阶段的第一运行信息,执行第一阶段和第一阶段之后的阶段(即,全部的阶段)。
例如,参考图4,针对这五个阶段中的每个阶段都提供了用于该阶段开始重新执行的触发入口,例如触发控件421、422、423、424、425。用户105可以触发控件420来从头开始执行全部的阶段。在指定阶段是第二阶段的情况下,应用管理平台110可 以基于经调整的至少第二阶段的第一运行信息,执行第二阶段和第二阶段之后的阶段。例如,继续参考图4,用户105可以触发控件422来执行第二阶段和第二阶段之后的阶段。以此方式,用户可以灵活地运行不同的阶段,通过对比运行不同阶段的输出结果,可以验证对不同阶段进行调整的有效性。可以理解的是,尽管图4示出了通过控件421、422、423、424、425的触发来执行相应的阶段,但该触发方式仅仅是示意性的。用户105还可以通过其他的方式来进行相应的触发,例如,通过表征自然语言的用户输入,来触发相应阶段的执行。
在一些实施例中,在重新执行测试查询之后,应用管理平台110可以在分阶段调试区域430呈现与指定阶段以及在指定阶段之后的阶段的重新运行相关联的运行信息。例如,指定阶段是第二阶段,则可以在分阶段调试区域430呈现与第二阶段和第二阶段之后的阶段重新运行相关联的运行信息。
在一些实施例中,应用管理平台110还可以在呈现多个阶段的相应第一运行信息的同时呈现数据信息,数据信息至少指示用于执行测试查询的至少一个数据集,并且基于多个阶段的相应第一运行信息和数据信息,生成运行参考用例。参考图4,应用管理平台110可以在数据信息展示区域470用于执行测试查询的数据信息。此外,应用管理平台110可以基于该数据信息或对该数据信息的调整来生成运行参考用例。
在一些实施例中,对数据信息的调整可以包括对用于执行所述测试查询的所述至少一个数据集的更新。例如,在数据信息展示区域470展示了用于执行测试查询的一个数据集,即商品分析表。参考图4,用户105可以触发添加控件472以对该数据集进行更新。
在一些实施例中,对数据信息的调整可以包括对用于执行测试查询的至少一个数据集中的至少一部分的更新。在这里数据集中的一部分可以指示数据集中的字段描述。例如,在数据信息展示区域470展示了用于执行测试查询的一个数据集,即商品分析表。参考图4,用户105可以触发修改控件474以对该数据集中的字段描述进行更新。
在一些实施例中,数据信息还指示用于执行测试查询所参考的术语集,术语集包括对目标应用的用户群体所使用的术语的解释。参考图5,图5示出了根据本公开的一些实施例的展示数据信息的页面500的示例。页面500包括术语集展示区域510,用户105可以调整术语集展示区域510中的内容。
在一些实施例中,数据信息还指示用于执行测试查询所参考的用户与数字助手之间的对话记录。继续参考图5,页面500包括对话记录展示区域520,用户105可以调整对话记录展示区域520中的内容。
在一些实施例中,数据信息还指示用于执行所述测试查询的参考提示词。继续参考图5,页面500包括参考提示词展示区域530,用户105可以调整参考提示词展示区域530中的内容。通过调整参考提示词,可以给机器学习模型提供更好的指引,帮助 机器学习模型更好地完成用户的任务。
在一些实施例中,数据信息还指示用于确定测试查询的候选结果的条件。继续参考图5,页面500包括条件区域540,用户105可以调整条件区域540中的相似度和单词搜索数量。例如,在知识问答场景下,数字助手会基于相似度搜索与用户相似的提问,并根据单次搜索数量向用户145提供预定数量的搜索结果。
参考图6A,图6A示出了根据本公开的一些实施例的针对目标应用中特定功能用例库的页面600的示例。页面600呈现了用例库,用例库中包含了已经生成的参考用例。用户105触发详情控件602可以跳转至该用例对应的用例详情的页面,参考图6B。图6B示出了根据本公开的一些实施例的用例详情的页面610的示例。页面610可以包括测试查询以及多个阶段的相应第一运行信息614。
可以理解的是,根据上文所描述的实施例,可以生成包括一个或多个运行参考用例的运行参考用例集合。
在一些实施例中,在目标应用120被发布之后,在目标应用120的运行中,可以利用运行参考用例来向用户145提供服务。示例性的,响应于接收到针对目标应用的用户查询,应用管理平台110可以从运行参考用例集合中查找与用户查询相似或匹配的测试查询。基于用户查询与某一测试查询之间的相似度超过阈值相似度,应用管理平台110可以确定运行参考用例为用户查询的目标运行参考用例,并基于目标运行参考用例来执行用户查询。在生成运行参考用例之后,应用管理平台110接收到用户145发起的针对目标应用的用户查询,可以通过相似度计算确定用户查询与测试查询之间的相似度,并基于该相似度确定运行参考用例为用户查询的目标运行参考用例,在确定出目标运行参考用例之后,应用管理平台110可以使用该目标运行参考用例来执行用户查询。以此方式,参考运行参考用例来确定用户查询的结果,可以提供给用户更加精确的查询结果,提高用户的满意度。
在一些实施例中,可以利用机器学习模型来执行用户查询。例如,应用管理平台110可以将所确定的目标参考用例作为针对用户查询的提示词的一部分提供给机器学习模型。而后,应用管理平台110可以从机器学习模型获取用户查询的执行结果。需要注意的是,这里的目标参考用例是从应用创建阶段或者应用测试阶段所生成的运行参考用例集合中所选择的。这样的目标参考用例是出于调试的目的而生成的,而不是来自用户真实使用该应用时的用例。
在一些实施例中,使用目标应用120的用户可以是测试人员,也即用户查询可以来自测试人员。在这种情况下,如果测试人员针对某个查询结果给出了正面反馈,可以基于该用户查询和对应的查询结果生成运行参考用例。例如,可以将得出该查询结果的运行过程的多个阶段的运行信息存储作为运行参考用例。以此方式,利用用户反馈丰富了运行参考用例的来源。
在一些实施例中,目标应用包括至少一个功能,并且测试查询和用户查询均为针对至少一个功能中的第一功能的查询。目标应用可以包括应用数据分析、应用数据更新、知识问答等功能。测试查询和用户查询需要针对同一功能,例如,测试查询和用户查询均是针对应用数据分析功能。
在目标应用120的创建或测试期间,可能需要评测目标应用120的性能。例如,在目标应用包括多个功能的情况下,可能需要评测这些功能中的某个功能或某些功能。
为此,在一些实施例中,应用管理平台110可以维护评测样本集,以用于评测目标应用120,例如评测目标应用120的一个或多个功能,诸如数据分析功能、数据更新功能、知识问答功能等。评测样本集中的每个评测样本可以包括样本查询和目标应用120针对样本查询生成的结果(也称为样本查询结果、第二结果)。
在一些实施例中,为了获得对目标应用120中的某个功能的评测结果,可以利用机器学习模型获取针对评测样本中的样本查询结果的反馈。例如,机器学习模型可以从一个或多个维度对目标应用120生成的样本查询结果打分。
机器学习模型可以采用任何合适的方式给出反馈。作为示例,描述一种自动评测方式。可以将目标应用120提供服务所使用的数据源(例如,知识库、数据集等)提供给机器学习模型,使得机器学习模型从其中提取查询和对应的查询结果作为正确示例。例如,对于知识问答功能,可以提取问题和对应的答案。而后,可以在目标应用120中执行所提取的查询,并且由目标应用120生成执行结果。进而,机器学习模型可以基于答案和执行结果在各种合适的维度来对执行结果给出反馈,例如打分。评测维度的示例可以包括但不限于对用户的帮助性、语言流畅度、事实正确度、逻辑性等。
由此,通过机器学习模型针对所提取的各个问题和执行结果给出的反馈,可以确定对目标应用120的评测结果。例如,如果评测样本集用于知识问答功能,可以获得对知识问答功能的评测结果。这样,用户105可以根据评测结果来调试目标应用120,例如调试一个或多个功能。
在目标应用120的创建或测试期间,创建人员或测试人员可能输入一些评测查询。目标应用120可以针对评测查询生成对应的查询结果,也称为第一结果或评测查询结果。而后,创建人员或测试人员可以对评测查询结果给出反馈,也称为评测反馈。
在一些实施例中,评测反馈可以与评测样本集结合使用,以针对目标应用120给出更准确的评测结果。示例性的,运行管理平台110可以基于创建人员或测试人员给出的评测查询,生成用于目标应用120的评测参考用例。评测参考用例包括来自创建人员或测试人员的评测查询、目标应用120针对该评测查询而生成的评测查询结果和对该评测查询结果的评测反馈。评测反馈可以是正面反馈或负面反馈,也可以指示用户145对查询结果的满意程度。
通常而言,针对相似的查询,期望评测结果与评测反馈是一致的或匹配的。有鉴 于此,可以从评测样本集中找出与评测查询相似的样本查询,例如相似度超过阈值相似度。包括这种相似的样本查询的评测样本也称为目标评测样本,目标评测样本还包括目标应用120针对该样本查询而生成的样本结果。
应用管理平台110可以基于评测查询结果、评测反馈和样本查询结果,生成针对目标应用120的评测结果。例如,可以使对目标评测样本中的样本查询结果的打分与评测反馈是匹配的。可以理解的是,评测参考用例给出了什么样的结果是创建人员或测试人员满足的或什么样的结果是创建人员或测试人员不满意的示例。
在一些实施例中,如上文所描述的,可以利用机器学习模型执行对目标应用120的评测。相应地,可以基于评测参考用例的至少一部分(例如,评测反馈和评测查询结果)生成提示词,并且将提示词提供给机器学习模型。机器学习模型可以生成针对目标评测样本中的样本查询结果的反馈。例如,机器学习模型可以对样本查询结果进行打分。而后可以基于该反馈,确定针对目标应用120的评测结果。可以理解的是,评测参考用例给出了什么样的结果是创建人员或测试人员满足的或什么样的结果是创建人员或测试人员不满意的示例。这样,机器学习模型在生成针对样本结果的反馈(例如,打分)中可以参考这样的示例。注意,在评测中所利用的机器学习模型与目标应用120的运行中所利用的机器学习模型可以相同,也可以不同。
在一些实施例中,如果目标应用120包括至少一个功能,上述用于评测的样本查询和评测查询可以是针对同一功能的查询。例如,评测样本集可以是专用于评测目标应用120的知识问答功能的。在这种情况下,所收集的用户查询也是针对知识问答功能的。
图7A示出了根据本公开的一些实施例的应用调试的过程700的流程图。过程700可以被实现在应用管理平台110处。下面参考图7描述过程700。
在框701,应用管理平台110执行针对目标应用的测试查询。
在框702,应用管理平台110呈现执行所述测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出。
在框703,应用管理平台110基于所述测试查询的多个阶段的相应第一运行信息,生成针对目标应用的运行参考用例,运行参考用例包括测试查询和多个阶段的运行信息,运行参考用例在目标应用的运行中被参考用于确定用户查询的结果。
在一些实施例中,生成运行参考用例包括:确定对多个阶段中的至少一个阶段的第一运行信息的调整;基于经调整的至少一个阶段的第一运行信息,重新执行测试查询;呈现重新执行测试查询的多个阶段的相应第二运行信息;以及响应于接收到用例生成确认,基于相应第二运行信息,生成运行参考用例。
在一些实施例中,重新执行测试查询包括:接收阶段运行请求,阶段运行请求指示从多个阶段中的指定阶段起执行测试查询,指定阶段在至少一个阶段之前或被包括 在至少一个阶段中;基于经调整的至少一个阶段的第一运行信息,重新执行多个阶段中的指定阶段以及所述指定阶段之后的阶段。
在一些实施例中,呈现重新执行测试查询的所述多个阶段的相应第二运行信息包括:呈现与指定阶段以及在所述指定阶段之后的阶段的重新运行相关联的运行信息。
在一些实施例中,多个阶段的运行信息还指示以下至少一项:多个阶段各自的输入,多个阶段各自的运行耗时,多个阶段各自的运行日志,多个阶段各自的运行状态,在多个阶段中的至少一个阶段中调用机器学习模型时,被提供给机器学习模型的输入信息。
在一些实施例中,确定对多个阶段中的至少一个阶段的第一运行信息的调整包括:接收针对多个阶段中的第一阶段的术语调整信息,术语调整信息指示向第一阶段添加术语集或移除在第一阶段中选择的术语集,术语集包括对目标应用的用户群体所使用的术语的解释,并且重新执行测试查询包括:基于术语调整信息,确定在第一阶段中使用的一个或多个术语集;以及基于一个或多个术语集,重新执行测试查询。
在一些实施例中,确定对多个阶段中的至少一个阶段的第一运行信息的调整包括:接收针对多个阶段中的第二阶段的用例调整信息,用例调整信息指示向第二阶段添加参考用例或移除第二阶段中所选择的参考用例,并且重新执行测试查询包括:基于用例调整信息,确定在第二阶段中使用的一个或多个参考用例;以及基于一个或多个参考用例,重新执行测试查询。
在一些实施例中,生成针对目标应用的运行参考用例包括:在呈现多个阶段的相应第一运行信息的同时呈现数据信息,数据信息至少指示用于执行所述测试查询的至少一个数据集;以及基于多个阶段的相应第一运行信息和数据信息,生成运行参考用例。
在一些实施例中,数据信息还指示以下至少一项:用于执行测试查询所参考的术语集(其中,术语集包括对目标应用的用户群体所使用的术语的解释),用于执行测试查询所参考的用户与数字助手之间的对话记录,用于执行测试查询的参考提示词,用于确定测试查询的候选结果的条件。
在一些实施例中,对数据信息的调整包括以下至少一项:对用于执行测试查询的至少一个数据集的更新,或对用于执行测试查询的至少一个数据集中的至少一部分的更新。
在一些实施例中,过程700还包括:响应于接收到针对目标应用的用户查询,基于用户查询与测试查询之间的相似度超过阈值相似度,确定运行参考用例为用户查询的目标运行参考用例;以及基于目标运行参考用例来执行用户查询。
在一些实施例中,目标应用包括至少一个功能,并且其中测试查询和用户查询均为针对至少一个功能中的第一功能的查询。
在一些实施例中,多个阶段包括:至少一个预定义阶段,至少一个预定义阶段包括生成数据库查询请求的阶段;以及基于所述测试查询确定的至少一个阶段。
在一些实施例中,测试查询是针对目标应用中的第一功能的查询,并且其中至少一个预定义阶段针对第一功能来配置。
图7B示出了根据本公开的一些实施例的信息查询的过程710的流程图。过程710可以被实现在应用管理平台110处。下面参考图7描述过程710。
在框711,应用管理平台110响应于接收到针对目标应用的用户查询,基于用户查询与测试查询之间的相似度超过阈值相似度,将包括测试查询的运行参考用例确定为用户查询的目标运行参考用例,其中运行参考用例包括测试查询和测试查询的多个阶段的相应运行信息。
在框712,应用管理平台110基于目标运行参考用例来执行用户查询。
在一些实施例中,基于目标运行参考用例来执行用户查询包括:将目标参考用例作为针对用户查询的提示词的一部分提供给机器学习模型;以及从机器学习模型获取用户查询的执行结果。
在一些实施例中,如果用户查询是来自测试人员并且在应用测试阶段实施,那么过程710还包括:获取对用户查询的执行结果的用户反馈;以及响应于用户反馈是正向反馈,基于多个阶段的用于执行用户查询的相应执行信息,生成另外的运行参考用例。
图7C示出了根据本公开的一些实施例的应用评测的过程720的流程图。过程720可以被实现在应用管理平台110处。下面参考图7描述过程720。
在框721,应用管理平台110生成用于所述目标应用的评测参考用例,所述评测参考用例包括评测查询、所述目标应用针对所述评测查询而生成的第一结果和对所述第一结果的评测反馈。
在框722,应用管理平台110基于所述评测查询与评测样本集中的查询之间的相似度,从所述评测样本集中选择目标评测样本,所述目标评测样本包括样本查询和由所述目标应用针对所述样本查询生成的第二结果。
在框723,应用管理平台110基于所述第一结果、所述评测反馈和所述第二结果,生成针对所述目标应用的评测结果。
在一些实施例中,生成针对所述目标应用的评测结果包括:将基于所述第一结果和所述评测反馈而生成的提示词提供给机器学习模型;从所述机器学习模型获得针对所述目标评测样本中的所述第二结果的反馈;以及基于针对所述第二结果的反馈,确定针对所述目标应用的评测结果。
在一些实施例中,目标应用包括至少一个功能,并且其中样本查询和用户查询均为针对至少一个功能中的第二功能的查询。
图8A示出了根据本公开的一些实施例的用于应用调试的装置800的示意性结构框图。装置800例如可以被实现在或被包括在应用管理平台110中。装置800中的各个模块/组件可以由硬件、软件、固件或者它们的任何组合来实现。
如图所示,装置800包括测试查询执行模块801,被配置为执行针对目标应用的测试查询。装置800还包括第一运行信息呈现模块802,被配置为呈现执行所述测试查询的多个阶段的相应第一运行信息,多个阶段的运行信息至少指示多个阶段各自的输出。装置800还包括运行参考用例生成模块803,被配置为基于所述测试查询的多个阶段的相应第一运行信息,生成针对目标应用的运行参考用例,运行参考用例包括测试查询和多个阶段的运行信息,运行参考用例在目标应用的运行中被参考用于确定用户查询的结果。
在一些实施例中,参考用例生成模块803包括:测试查询重新执行模块,被配置为确定对多个阶段中的至少一个阶段的第一运行信息的调整;基于经调整的至少一个阶段的第一运行信息,重新执行测试查询;呈现重新执行测试查询的多个阶段的相应第二运行信息;以及响应于接收到用例生成确认,基于相应第二运行信息,生成运行参考用例。
在一些实施例中,测试查询重新执行模块包括:阶段执行模块,被配置为接收阶段运行请求,阶段运行请求指示从多个阶段中的指定阶段起执行所述测试查询,指定阶段在至少一个阶段之前或被包括在至少一个阶段中;基于经调整的至少一个阶段的第一运行信息,重新执行多个阶段中的指定阶段以及指定阶段之后的阶段。
在一些实施例中,阶段执行模块包括:重新执行信息呈现模块,被配置为呈现与指定阶段以及在所述指定阶段之后的阶段的重新运行相关联的运行信息。
在一些实施例中,多个阶段的运行信息还指示以下至少一项:多个阶段各自的输入,多个阶段各自的运行耗时,多个阶段各自的运行日志,多个阶段各自的运行状态,在多个阶段中的至少一个阶段中调用机器学习模型时,被提供给机器学习模型的输入信息。
在一些实施例中,测试查询重新执行模块包括:第一调整信息接收模块,被配置为接收针对多个阶段中的第一阶段的术语调整信息,术语调整信息指示向第一阶段添加术语集或移除在第一阶段中选择的术语集,术语集包括对目标应用的用户群体所使用的术语的解释,并且测试查询重新执行模块进一步被配置为基于术语调整信息,确定在第一阶段中使用的一个或多个术语集;以及基于一个或多个术语集,重新执行测试查询。
在一些实施例中,测试查询重新执行模块包括:第二调整信息接收模块,被配置接收针对多个阶段中的第二阶段的用例调整信息,用例调整信息指示向第二阶段添加参考用例或移除第二阶段中所选择的参考用例,并且测试查询重新执行模块进一步被 配置为基于用例调整信息,确定在第二阶段中使用的一个或多个参考用例;以及基于一个或多个参考用例,重新执行测试查询。
在一些实施例中,装置800还包括:数据信息呈现模块,被配置为在呈现多个阶段的相应第一运行信息的同时呈现数据信息,数据信息至少指示用于执行所述测试查询的至少一个数据集;以及基于多个阶段的相应第一运行信息和数据信息,生成运行参考用例。
在一些实施例中,数据信息还指示以下至少一项:用于执行测试查询所参考的术语集,术语集包括对目标应用的用户群体所使用的术语的解释,用于执行测试查询所参考的用户与数字助手之间的对话记录,用于执行测试查询的参考提示词,用于确定测试查询的候选结果的条件。
在一些实施例中,数据信息呈现模块包括:数据更新模块,被配置为对用于执行测试查询的至少一个数据集的更新,或对用于执行测试查询的至少一个数据集中的至少一部分的更新。
在一些实施例中,装置800还包括:用户查询执行模块,被配置为响应于接收到针对目标应用的用户查询,基于用户查询与测试查询之间的相似度超过阈值相似度,确定运行参考用例为用户查询的目标运行参考用例;以及基于目标运行参考用例来执行用户查询。
在一些实施例中,目标应用包括至少一个功能,并且其中测试查询和用户查询均为针对至少一个功能中的第一功能的查询。
在一些实施例中,多个阶段包括:至少一个预定义阶段,至少一个预定义阶段包括生成数据库查询请求的阶段;以及基于所述测试查询确定的至少一个阶段。
在一些实施例中,测试查询是针对目标应用中的第一功能的查询,并且其中至少一个预定义阶段针对第一功能来配置。
图8B示出了根据本公开的一些实施例的用于信息查询的装置810的示意性结构框图。装置810例如可以被实现在或被包括在应用管理平台110中。装置810中的各个模块/组件可以由硬件、软件、固件或者它们的任何组合来实现。
如图所示,装置810包括运行参考用例确定模块811,被配置为响应于接收到针对目标应用的用户查询,基于用户查询与测试查询之间的相似度超过阈值相似度,将包括测试查询的运行参考用例确定为用户查询的目标运行参考用例,其中运行参考用例包括测试查询和测试查询的多个阶段的相应运行信息。装置810还包括用户查询执行模块812,被配置为基于目标运行参考用例来执行用户查询。
在一些实施例中,用户查询执行模块812包括执行结果获取模块,被配置为将目标参考用例作为针对用户查询的提示词的一部分提供给机器学习模型;以及从机器学习模型获取用户查询的执行结果。
在一些实施例中,装置810还包括附加运行参考用例生成模块,被配置为获取对用户查询的执行结果的用户反馈;以及响应于用户反馈是正向反馈,基于多个阶段的用于执行用户查询的相应执行信息,生成另外的运行参考用例。
图8C示出了根据本公开的一些实施例的用于应用评测的装置820的示意性结构框图。装置820例如可以被实现在或被包括在应用管理平台110中。装置820中的各个模块/组件可以由硬件、软件、固件或者它们的任何组合来实现。
如图所示,装置820包括评测参考用例生成模块821,被配置为生成用于所述目标应用的评测参考用例,所述评测参考用例包括评测查询、所述目标应用针对评测查询而生成的第一结果和对第一结果的评测反馈。装置820还包括目标评测样本选择模块822,被配置为基于评测查询与评测样本集中的查询之间的相似度,从评测样本集中选择目标评测样本,目标评测样本包括样本查询和由目标应用针对样本查询生成的第二结果。装置820还包括评测结果生成模块823,被配置为基于第一结果、评测反馈和第二结果,生成针对目标应用的评测结果。
在一些实施例中,评测结果生成模块823包括模型反馈获得模块,被配置为将基于第一结果和评测反馈而生成的提示词提供给机器学习模型;从机器学习模型获得针对目标评测样本中的第二结果的反馈;以及基于针对第二结果的反馈,确定针对目标应用的评测结果。
在一些实施例中,目标应用包括至少一个功能,并且其中样本查询和用户查询均为针对至少一个功能中的第二功能的查询。
图9示出了其中可以实现本公开的一个或多个实施例的电子设备900的框图。应当理解,图9所示出的电子设备900仅仅是示例性的,而不应当构成对本文所描述的实施例的功能和范围的任何限制。图9所示出的电子设备900可以包括或被实现为图1的应用管理平台110、图8A的装置800、图8B的装置810,或图8C的装置820。
如图9所示,电子设备900是通用电子设备的形式。电子设备900的组件可以包括但不限于一个或多个处理器或处理单元910、存储器920、存储设备930、一个或多个通信单元940、一个或多个输入设备950以及一个或多个输出设备960。处理单元910可以是实际或虚拟处理器并且能够根据存储器920中存储的程序来执行各种处理。在多处理器系统中,多个处理单元并行执行计算机可执行指令,以提高电子设备900的并行处理能力。
电子设备900通常包括多个计算机存储介质。这样的介质可以是电子设备900可访问的任何可以获取的介质,包括但不限于易失性和非易失性介质、可拆卸和不可拆卸介质。存储器920可以是易失性存储器(例如寄存器、高速缓存、随机访问存储器(RAM))、非易失性存储器(例如,只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、闪存)或它们的某种组合。存储设备930可以是可拆卸或不可拆 卸的介质,并且可以包括机器可读介质,诸如闪存驱动、磁盘或者任何其他介质,其可以能够用于存储信息和/或数据并且可以在电子设备900内被访问。
电子设备900可以进一步包括另外的可拆卸/不可拆卸、易失性/非易失性存储介质。尽管未在图9中示出,可以提供用于从可拆卸、非易失性磁盘(例如“软盘”)进行读取或写入的磁盘驱动和用于从可拆卸、非易失性光盘进行读取或写入的光盘驱动。在这些情况中,每个驱动可以由一个或多个数据介质接口被连接至总线(未示出)。存储器920可以包括计算机程序产品925,其具有一个或多个程序模块,这些程序模块被配置为执行本公开的各种实施例的各种方法或动作。
通信单元940实现通过通信介质与其他电子设备进行通信。附加地,电子设备900的组件的功能可以以单个计算集群或多个计算机器来实现,这些计算机器能够通过通信连接进行通信。因此,电子设备900可以使用与一个或多个其他服务器、网络个人计算机(PC)或者另一个网络节点的逻辑连接来在联网环境中进行操作。
输入设备950可以是一个或多个输入设备,例如鼠标、键盘、追踪球等。输出设备960可以是一个或多个输出设备,例如显示器、扬声器、打印机等。电子设备900还可以根据需要通过通信单元940与一个或多个外部设备(未示出)进行通信,外部设备诸如存储设备、显示设备等,与一个或多个使得用户与电子设备900交互的设备进行通信,或者与使得电子设备900与一个或多个其他电子设备通信的任何设备(例如,网卡、调制解调器等)进行通信。这样的通信可以经由输入/输出(I/O)接口(未示出)来执行。
根据本公开的示例性实现方式,提供了一种计算机可读存储介质,其上存储有计算机可执行指令,其中计算机可执行指令被处理器执行以实现上文描述的方法。根据本公开的示例性实现方式,还提供了一种计算机程序产品,计算机程序产品被有形地存储在非瞬态计算机可读介质上并且包括计算机可执行指令,而计算机可执行指令被处理器执行以实现上文描述的方法。
这里参照根据本公开实现的方法、装置、设备和计算机程序产品的流程图和/或框图描述了本公开的各个方面。应当理解,流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合,都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其他可编程数据处理装置的处理单元,从而生产出一种机器,使得这些指令在通过计算机或其他可编程数据处理装置的处理单元执行时,产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中,这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作,从而,存储有指令的计算机可读介质则包括一个制造品,其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
可以把计算机可读程序指令加载到计算机、其他可编程数据处理装置、或其他设备上,使得在计算机、其他可编程数据处理装置或其他设备上执行一系列操作步骤,以产生计算机实现的过程,从而使得在计算机、其他可编程数据处理装置、或其他设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本公开的多个实现的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上,流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分,模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中,方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如,两个连续的方框实际上可以基本并行地执行,它们有时也可以按相反的顺序执行,这依所涉及的功能而定。也要注意的是,框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合,可以用执行规定的功能或动作的专用的基于硬件的系统来实现,或者可以用专用硬件与计算机指令的组合来实现。
以上已经描述了本公开的各实现,上述说明是示例性的,并非穷尽性的,并且也不限于所公开的各实现。在不偏离所说明的各实现的范围和精神的情况下,对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择,旨在最好地解释各实现的原理、实际应用或对市场中的技术的改进,或者使本技术领域的其他普通技术人员能理解本文公开的各个实现方式。

Claims (24)

  1. 一种应用调试的方法,包括:
    执行针对目标应用的测试查询;
    呈现执行所述测试查询的多个阶段的相应第一运行信息,所述多个阶段的运行信息至少指示所述多个阶段各自的输出;以及
    基于所述测试查询的多个阶段的相应第一运行信息,生成针对所述目标应用的运行参考用例,所述运行参考用例包括所述测试查询和所述多个阶段的运行信息,所述运行参考用例在所述目标应用的运行中被参考用于确定用户查询的结果。
  2. 根据权利要求1所述的方法,其中生成所述运行参考用例包括:
    确定对所述多个阶段中的至少一个阶段的第一运行信息的调整;
    基于经调整的所述至少一个阶段的第一运行信息,重新执行所述测试查询;
    呈现重新执行所述测试查询的所述多个阶段的相应第二运行信息;以及
    响应于接收到用例生成确认,基于所述相应第二运行信息,生成所述运行参考用例。
  3. 根据权利要求2所述的方法,其中重新执行所述测试查询包括:
    接收阶段运行请求,所述阶段运行请求指示从所述多个阶段中的指定阶段起执行所述测试查询,所述指定阶段在所述至少一个阶段之前或被包括在所述至少一个阶段中;以及
    基于经调整的所述至少一个阶段的第一运行信息,重新执行所述多个阶段中的所述指定阶段以及所述指定阶段之后的阶段。
  4. 根据权利要求3所述的方法,其中呈现重新执行所述测试查询的所述多个阶段的相应第二运行信息包括:
    呈现与所述指定阶段以及在所述指定阶段之后的阶段的重新运行相关联的运行信息。
  5. 根据权利要求1所述的方法,其中所述多个阶段的运行信息还指示以下至少一项:
    所述多个阶段各自的输入,
    所述多个阶段各自的运行耗时,
    所述多个阶段各自的运行日志,
    所述多个阶段各自的运行状态,
    在所述多个阶段中的至少一个阶段中调用机器学习模型时,被提供给所述机器学习模型的输入信息。
  6. 根据权利要求2所述的方法,其中确定对所述多个阶段中的至少一个阶段的 第一运行信息的调整包括:
    接收针对所述多个阶段中的第一阶段的术语调整信息,所述术语调整信息指示向所述第一阶段添加术语集或移除在所述第一阶段中选择的术语集,术语集包括对所述目标应用的用户群体所使用的术语的解释,并且
    其中重新执行所述测试查询包括:
    基于所述术语调整信息,确定在所述第一阶段中使用的一个或多个术语集;以及
    基于所述一个或多个术语集,重新执行所述测试查询。
  7. 根据权利要求2所述的方法,其中确定对所述多个阶段中的至少一个阶段的第一运行信息的调整包括:
    接收针对所述多个阶段中的第二阶段的用例调整信息,所述用例调整信息指示向所述第二阶段添加参考用例或移除所述第二阶段中所选择的参考用例,并且
    其中重新执行所述测试查询包括:
    基于所述用例调整信息,确定在所述第二阶段中使用的一个或多个参考用例;以及
    基于所述一个或多个参考用例,重新执行所述测试查询。
  8. 根据权利要求1所述的方法,其中生成针对所述目标应用的运行参考用例包括:
    在呈现所述多个阶段的相应第一运行信息的同时呈现数据信息,所述数据信息至少指示用于执行所述测试查询的至少一个数据集;以及
    基于所述多个阶段的相应第一运行信息和所述数据信息,生成所述运行参考用例。
  9. 根据权利要求8所述的方法,其中所述数据信息还指示以下至少一项:
    用于执行所述测试查询所参考的术语集,所述术语集包括对所述目标应用的用户群体所使用的术语的解释,
    用于执行所述测试查询所参考的用户与数字助手之间的对话记录,
    用于执行所述测试查询的参考提示词,
    用于确定所述测试查询的候选结果的条件。
  10. 根据权利要求8所述的方法,其中对所述数据信息的调整包括以下至少一项:
    对用于执行所述测试查询的所述至少一个数据集的更新,或
    对用于执行所述测试查询的所述至少一个数据集中的至少一部分的更新。
  11. 根据权利要求1所述的方法,还包括:
    响应于接收到针对所述目标应用的用户查询,基于所述用户查询与所述测试查询之间的相似度超过阈值相似度,确定所述运行参考用例为所述用户查询的目标运行参考用例;以及
    基于所述目标运行参考用例来执行所述用户查询。
  12. 根据权利要求11所述的方法,其中所述目标应用包括至少一个功能,并且其中所述测试查询和所述用户查询均为针对所述至少一个功能中的第一功能的查询。
  13. 根据权利要求1所述的方法,其中所述多个阶段包括:
    至少一个预定义阶段,所述至少一个预定义阶段包括生成数据库查询请求的阶段;以及
    基于所述测试查询确定的至少一个阶段。
  14. 根据权利要求13所述的方法,其中所述测试查询是针对所述目标应用中的第一功能的查询,并且其中所述至少一个预定义阶段针对所述第一功能来配置。
  15. 一种信息查询的方法,包括:
    响应于接收到针对目标应用的用户查询,基于所述用户查询与测试查询之间的相似度超过阈值相似度,将包括所述测试查询的运行参考用例确定为所述用户查询的目标运行参考用例,其中所述运行参考用例包括所述测试查询和所述测试查询的多个阶段的相应运行信息;以及
    基于所述目标运行参考用例来执行所述用户查询。
  16. 根据权利要求15所述的方法,其中基于所述目标运行参考用例来执行所述用户查询包括:
    将所述目标参考用例作为针对所述用户查询的提示词的一部分提供给机器学习模型;以及
    从所述机器学习模型获取所述用户查询的执行结果。
  17. 一种应用评测的方法,包括:
    生成用于所述目标应用的评测参考用例,所述评测参考用例包括评测查询、所述目标应用针对所述评测查询而生成的第一结果和对所述第一结果的评测反馈;
    基于所述评测查询与评测样本集中的查询之间的相似度,从所述评测样本集中选择目标评测样本,所述目标评测样本包括样本查询和由所述目标应用针对所述样本查询生成的第二结果;以及
    基于所述第一结果、所述评测反馈和所述第二结果,生成针对所述目标应用的评测结果。
  18. 根据权利要求17所述的方法,其中生成针对所述目标应用的评测结果包括:
    将基于所述第一结果和所述评测反馈而生成的提示词提供给机器学习模型;
    从所述机器学习模型获得针对所述目标评测样本中的所述第二结果的反馈;以及
    基于针对所述第二结果的反馈,确定针对所述目标应用的评测结果。
  19. 根据权利要求17所述的方法,其中所述目标应用包括至少一个功能,并且其中所述样本查询和所述用户查询均为针对所述至少一个功能中的第二功能的查询。
  20. 一种用于应用调试的装置,包括:
    测试查询执行模块,被配置为执行针对目标应用的测试查询;
    第一运行信息呈现模块,被配置为呈现执行所述测试查询的多个阶段的相应第一运行信息,所述多个阶段的运行信息至少指示所述多个阶段各自的输出;以及
    运行参考用例生成模块,被配置为基于所述测试查询的多个阶段的相应第一运行信息,生成针对所述目标应用的运行参考用例,所述运行参考用例包括所述测试查询和所述多个阶段的运行信息,所述运行参考用例在所述目标应用的运行中被参考用于确定用户查询的结果。
  21. 一种用于信息查询的装置,包括:
    运行参考用例确定模块,被配置为响应于接收到针对目标应用的用户查询,基于所述用户查询与测试查询之间的相似度超过阈值相似度,将包括所述测试查询的运行参考用例确定为所述用户查询的目标运行参考用例,其中所述运行参考用例包括所述测试查询和所述测试查询的多个阶段的相应运行信息;以及
    用户查询执行模块,被配置为基于所述目标运行参考用例来执行所述用户查询。
  22. 一种用于应用评测的装置,包括:
    评测参考用例生成模块,被配置为生成用于所述目标应用的评测参考用例,所述评测参考用例包括评测查询、所述目标应用针对所述评测查询而生成的第一结果和对所述第一结果的评测反馈;
    目标评测样本选择模块,被配置为基于所述评测查询与评测样本集中的查询之间的相似度,从所述评测样本集中选择目标评测样本,所述目标评测样本包括样本查询和由所述目标应用针对所述样本查询生成的第二结果;以及
    评测结果生成模块,被配置为基于所述第一结果、所述评测反馈和所述第二结果,生成针对所述目标应用的评测结果。
  23. 一种电子设备,包括:
    至少一个处理单元;以及
    至少一个存储器,所述至少一个存储器被耦合到所述至少一个处理单元并且存储用于由所述至少一个处理单元执行的指令,所述指令在由所述至少一个处理单元执行时使所述电子设备执行根据权利要求1至19中任一项所述的方法。
  24. 一种计算机可读存储介质,其上存储有计算机程序,所述计算机程序可由处理器执行以实现根据权利要求1至19中任一项所述的方法。
PCT/CN2024/122387 2023-11-22 2024-09-29 应用调试的方法、信息查询的方法、应用评测的方法 Pending WO2025107888A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2025557329A JP2026512314A (ja) 2023-11-22 2024-09-29 アプリケーションデバッグの方法、情報クエリの方法、及びアプリケーションアセスメントの方法

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202311570383.1A CN119003326B (zh) 2023-11-22 2023-11-22 应用调试的方法、信息查询的方法、应用评测的方法
CN202311570383.1 2023-11-22

Publications (1)

Publication Number Publication Date
WO2025107888A1 true WO2025107888A1 (zh) 2025-05-30

Family

ID=93478975

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2024/122387 Pending WO2025107888A1 (zh) 2023-11-22 2024-09-29 应用调试的方法、信息查询的方法、应用评测的方法

Country Status (3)

Country Link
JP (1) JP2026512314A (zh)
CN (1) CN119003326B (zh)
WO (1) WO2025107888A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103164334A (zh) * 2011-12-19 2013-06-19 国际商业机器公司 检测web应用自动测试用例中的断裂点的系统和方法
CN105740312A (zh) * 2014-12-27 2016-07-06 达索系统公司 使数据库查询形成聚类以用于运行时间预测
US20230251960A1 (en) * 2022-02-04 2023-08-10 Optum, Inc. Machine learning techniques for automated software testing configuration management

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103207871A (zh) * 2012-01-17 2013-07-17 深圳市腾讯计算机系统有限公司 对搜索系统的查询串改写效果进行评测的方法和装置
US11416229B2 (en) * 2019-03-13 2022-08-16 Google Llc Debugging applications for delivery via an application delivery server
CN111382256B (zh) * 2020-03-20 2024-04-09 北京百度网讯科技有限公司 信息推荐方法及装置
CN113515440A (zh) * 2020-12-22 2021-10-19 腾讯科技(深圳)有限公司 测试用例分配方法、装置、计算机设备及存储介质

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103164334A (zh) * 2011-12-19 2013-06-19 国际商业机器公司 检测web应用自动测试用例中的断裂点的系统和方法
CN105740312A (zh) * 2014-12-27 2016-07-06 达索系统公司 使数据库查询形成聚类以用于运行时间预测
US20230251960A1 (en) * 2022-02-04 2023-08-10 Optum, Inc. Machine learning techniques for automated software testing configuration management

Also Published As

Publication number Publication date
CN119003326A (zh) 2024-11-22
JP2026512314A (ja) 2026-04-15
CN119003326B (zh) 2025-08-22

Similar Documents

Publication Publication Date Title
US20250021842A1 (en) Monitoring generative model quality
CN118981527B (zh) 基于大模型的问答方法、装置、电子设备、存储介质、智能体和程序产品
US11907863B2 (en) Natural language enrichment using action explanations
US12411662B2 (en) Digital assistant creation
WO2024234614A1 (zh) 信息处理方法、装置、设备、计算机可读存储介质及计算机程序产品
EP4575760A1 (en) Method and apparatus for managing workflow, and device and medium
CN110647314B (zh) 技能生成方法、装置及电子设备
WO2025223028A1 (zh) 信息处理方法、装置、设备和存储介质
CN120316207A (zh) 一种基于检索-生成模型的智能问答方法及终端
CN119005327A (zh) 任务处理方法、装置、设备和计算机可读存储介质
WO2026040472A1 (zh) 应用配置方法、信息处理方法、装置、设备和存储介质
CN120596356A (zh) Ai应用程序的调优方法、装置和计算设备
WO2025107890A1 (zh) 用于应用创建的方法、装置、设备和存储介质
CN118839769A (zh) 任务处理方法、装置、设备和计算机可读存储介质
CN118838519A (zh) 信息处理方法、装置、设备和存储介质
WO2025107888A1 (zh) 应用调试的方法、信息查询的方法、应用评测的方法
CN119474362A (zh) 用于文档生成的方法、装置、设备、存储介质和产品
CN119003558A (zh) 信息处理方法、装置、设备和存储介质
CN117271514A (zh) 数据处理方法、装置、设备、存储介质和程序产品
CN114237588A (zh) 一种代码仓库选择方法、装置、设备及存储介质
US20260093702A1 (en) Method for data processing, apparatus, device, storage medium and program product
US20260050800A1 (en) Digital assistant evaluation
US12314743B2 (en) Digital assistant interaction
US12242742B1 (en) Storing data in a digital assistant
US20260093741A1 (en) Performance optimization with reflection tokens in a self-reflective retrieval-augmented generation framework

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24893047

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2025557329

Country of ref document: JP

Kind code of ref document: A

WWE Wipo information: entry into national phase

Ref document number: 2025557329

Country of ref document: JP