Interaction method, platform and equipment based on AI gateway and AIAgent, and storage medium

Abstract

The invention relates to an interaction method, platform and device based on an AI gateway and an AIAgent and a storage medium, and belongs to the technical field of interactive communication of an intelligent terminal and a server end, the method comprises the steps of obtaining a service request sent by a client, and analyzing the service request to obtain a service demand; matching an API interface for the business demand, and calling the API interface to process the service request; and feeding back a processing result of the corresponding service request to the client. The method has the advantages that self-adaptive definition and flexible adjustment of the API interface function are achieved, and diversified interaction requirements proposed by the client side can be met.

Description

Technical Field

[0001] This application relates to the technical field of interactive communication between intelligent terminals and server - sides, and particularly relates to an interaction method, platform, device, and storage medium based on an AI gateway and an AI Agent. Background Art

[0002] With the rapid development of Internet of Things (IoT) technology, more and more intelligent terminal devices are connected to cloud servers through networks to achieve functions such as data upload / download and remote control. The realization of these functions depends on preset standard protocols or API interfaces to complete specific service requests. However, in actual applications, due to the rapidly growing personalized needs of intelligent terminal users and complex and changeable business scenarios, since traditional API designs are often fixed, it is not easy to meet the personalized customization needs of users, thus restricting the improvement of the intelligent level, so it needs to be improved. Summary of the Invention

[0003] To solve the technical problem that traditional API interfaces are inflexible and difficult to adapt to diverse requirements, this application provides an interaction method, platform, device, and storage medium based on an AI gateway and an AI Agent.

[0004] In a first aspect, this application provides an interaction method based on an AI gateway and an AI Agent, including: Obtain a service request sent by a client, and parse the service request to obtain a business requirement; Match an API interface for the business requirement, and call the API interface to process the service request; Feedback the processing result of the corresponding service request to the client.

[0005] By adopting the above - mentioned technical solution, by deeply parsing the service request proposed by the client and analyzing to obtain the business requirement, then matching an API interface for this business requirement. Here, the matching can either refer to associating an existing API interface for this business requirement or custom - generating a new API interface for this business requirement, and then calling the API interface to process the service request to meet the customer's business requirement. Finally, the processing result is fed back to the client, realizing the adaptive adjustment of the API interface to enable it to adapt to diverse requirements.

[0006] Optionally, the feedback of the processing result of the corresponding service request to the client includes: After performing desensitization processing on the processing result of the service request, feedback the processing result of the corresponding service request to the client.

[0007] By adopting the above technical solution, to ensure data security and customer privacy security, before the processing result is fed back, the processing result will be desensitized to strictly comply with the regulatory requirements of data protection. The desensitization process here can be considered as identifying and hiding sensitive information to effectively avoid the risk of leakage of sensitive information.

[0008] Optionally, matching an API interface for the service requirement includes: Matching an API interface for the service requirement and defining the configuration information of the API interface based on the service requirement, where the configuration information at least includes interface bandwidth; The method further includes: Predicting the change trend of the number of times each API interface is called over time, and determining the interface bandwidth of each API interface at different times based on the change trend; Adjusting the interface bandwidth included in the configuration information of the corresponding API interface in real time according to the interface bandwidth of each API interface.

[0009] By adopting the above technical solution, a custom function for API interface configuration is provided, thereby improving the matching degree between the API interface and the service requirement. At the same time, the interface bandwidth required by the API interface is predicted based on the number of times the API interface is called to ensure that the processing operation of the service request can be successfully completed with the help of the API interface.

[0010] Optionally, the method further includes: Whenever an API interface is matched for a service requirement, a test plan is periodically executed to generate a virtual service request corresponding to the service requirement, parse the virtual service request to match the API interface, call the API interface to process the virtual service request, verify the processing result of the virtual service request, and output the verification result; where the test plan at least includes a virtual service request corresponding to the service requirement.

[0011] By adopting the above technical solution, after the service requirement and the API interface are matched, a virtual service request can be periodically triggered to obtain the processing result corresponding to the virtual service request, and then the processing result is verified (such as comparing the processing result with the processing result corresponding to the service request proposed by the client for consistency), so that the tester can use the verification result to analyze the matching degree and stability between the service requirement and the API interface it matches.

[0012] Optionally, each of the API interfaces corresponds to at least one functional attribute; The calling the API interface to process the service request includes: Generate a call instruction based on a service request, where the call instruction at least includes the information of the first API interface to be called and the first function attribute to be enabled for the first API interface; Enable the function attribute in the first API interface that is consistent with the first function attribute, and process the service request based on the function attribute enabled for the first API interface.

[0013] By adopting the above technical solution, the function attribute can be considered as the specific business requirements processed by means of the API interface. Considering the components of the API, an API consists of a request address, a request method, etc. Then, when an API interface supports multiple request methods (such as GET, POST, PUT, DELETE, etc.), it correspondingly has multiple function attributes (such as GET, POST, PUT, DELETE, etc.). At this time, the enabling conditions of the request methods can be dynamically or statically limited by defining the call instruction, that is, the enabling conditions of the function attributes included in the API interface are limited, and the applicable range of the usage methods included in the API interface is flexibly controlled, realizing the dynamic switching of the function attributes of a single API interface.

[0014] Optionally, the first API interface includes a second API interface and / or a third API interface; Matching the API interface for the business requirements further includes: Determine the function range corresponding to the business requirements, determine the second API interface from the existing API interfaces according to the function attributes corresponding to the existing API interfaces, and determine the function attributes to be enabled for each of the second API interfaces; wherein, the union of the function attributes to be enabled for all the second API interfaces is included in the function range; If the union range of the function attributes to be enabled for the second API interfaces is less than the function range, determine the newly added function attributes, and generate a third API interface for the newly added function attributes so that the third API interface corresponds to the newly added function attributes; wherein, the newly added function attributes are the complement of the union of the function attributes to be enabled for the second API interfaces.

[0015] By adopting the above technical solution, when matching API interfaces for business needs, the functional attributes of existing API interfaces are analyzed first to determine whether there is an API interface (i.e., a second API interface) whose functional attributes overlap with or completely cover the business needs. If so, all available second API interfaces are collected. When the union range of the functional attributes required to be enabled for all second API interfaces cannot completely cover the functional range, a third API interface is generated so that the second API interface and the third API interface are combined to form a first API interface to process the corresponding service request. The main purpose of the above method is to avoid redundancy in the corresponding matched generated API interface as the service requests made by the client increase, and to achieve the combined use of multiple API interfaces by limiting the activation or not of the API interface, thereby increasing the probability of reuse of the API interface.

[0016] Optionally, the determining the functional scope corresponding to the business requirement further includes: Determine a guidance set based on business requirements, wherein the guidance set includes at least a guidance question; Determine the functional scope corresponding to the business needs, send the guidance questions contained in the guidance set to the client one by one in sequence, and determine whether there are new needs based on the client's feedback results on the target guidance questions. If so, add the functional attributes corresponding to the new needs to the functional scope.

[0017] By adopting the above technical solution, for the business needs proposed by the client, after preliminarily determining its functional scope, a guidance set consisting of guidance questions is further formulated based on the business needs, aiming to solve the problem that other business needs may be generated after the user proposes the current business needs. Therefore, this application proposes to query the user by sending guidance questions to the client, so as to guide and obtain the business needs that the user may propose in the future at one time, and then finally determine the functional scope of the matched API interface based on the guidance results (i.e. feedback results), so as to improve the efficiency of responding to user needs.

[0018] In the second aspect, the present application provides an interactive platform based on AI gateway and AIAgent, including: A demand analysis module is used to obtain a service request from a client and analyze the service request to obtain a business demand; An interface matching module is used to match an API interface to the business requirement and call the API interface to process the service request; The demand response module is used to feed back the processing result of the corresponding service request to the client.

[0019] In a third aspect, the present application provides an interaction device based on an AI gateway and an AI Agent, including a memory and a processor. A computer program capable of being loaded and executed by the processor to perform any of the methods described in the first aspect is stored on the memory.

[0020] In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program capable of being loaded and executed by the processor to perform any of the methods described in the first aspect.

[0021] In summary, the present application includes at least one of the following beneficial technical effects: 1. Since the function of existing AI gateway products is only to realize the switching and calling of various AI large models, that is, limited to the management and agency of background large models, while in the present application, the AI Agent is used to automatically generate or redefine API interfaces to flexibly adjust the functions supported by the API interfaces so as to adapt to the diverse requirements proposed by the client. 2. Further, the present application realizes plug-in expansion based on LLMs and adds functions such as data desensitization processing and malicious behavior detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0023] Figure 1 is a schematic flowchart of an interaction method based on an AI gateway and an AI Agent disclosed in an embodiment of the present application.

[0024] Figure 2 is a schematic structural diagram of an interaction platform based on an AI gateway and an AI Agent disclosed in an embodiment of the present application.

[0025] Figure 3 is a block diagram of modules of an interaction platform based on an AI gateway and an AI Agent disclosed in an embodiment of the present application.

[0026] Description of reference numerals: 201, requirement analysis module; 202, interface matching module; 203, requirement response module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The following will further elaborate on the present application in conjunction with the attached Figures 1-3 for a more detailed description.

[0028] An embodiment of the present application discloses an interaction method based on an AI gateway and an AI Agent (hereinafter simply referred to as the interaction method), which aims to utilize the business functions of AI to analyze personalized and diverse business requirements proposed by a client, customize and match API interfaces for the business requirements, and the API interfaces can serve as an interaction bridge between the client and the backend server to help solve the business requirements proposed by the client and feedback the processing results; ultimately, an adaptive adjustment of the API interfaces is realized so that they can adapt to the diverse needs of users. In addition, the execution entity of the interaction method of the present application is an interaction platform based on an AI gateway and an AI Agent (hereinafter simply referred to as the interaction platform), and the execution process of the interaction method by the interaction platform will be specifically described below in conjunction with the attached Figures 1-2 Specifically describe the execution process of the interaction method by the interaction platform.

[0029] S101, Obtain the service request sent by the client, and parse the service request to obtain the business requirement.

[0030] In implementation, the user can access the interaction platform in the form of a web page or an APP through a smart terminal device (i.e., the client) and trigger a service request. The service request includes at least the request content, and the request content can be embodied in the form of text, such as "Upload document A to server B".

[0031] The interaction platform includes an AI gateway and an AI Agent. The AI gateway is responsible for receiving the service request proposed by the client. The AI Agent is integrated with a variety of pre-trained AI models (such as various large language models (LLMs)) to parse the service request to obtain the business requirement. The business request can be considered as the content predefined and stored in the interaction platform, corresponding to the service request, and used to describe the actual needs of the user.

[0032] S102, Match an API interface for the business requirement, and call the API interface to process the service request.

[0033] S103, Feedback the processing result of the corresponding service request to the client.

[0034] Among them, S102 specifically includes the following sub-steps: S1021, Determine the function scope corresponding to the business requirement. According to the function attributes corresponding to the existing API interfaces, determine the second API interfaces from the existing API interfaces, and determine the function attributes to be enabled for each second API interface; among them, the union of the function attributes to be enabled for all second API interfaces is included in the function scope; S1022. If the union range of the function attributes required to enable the second API interface is smaller than the function range, determine the new function attributes, and generate a third API interface for the new function attributes, so that the third API interface corresponds to the new function attributes; wherein, the new function attributes are the complement of the union of the function attributes required to enable the second API interface. S1023. Generate a call instruction based on the service request. The call instruction at least includes the information of the first API interface to be called and the first function attribute required to enable the first API interface. S1024. Enable the function attributes in the first API interface that are consistent with the first function attribute, and process the service request based on the function attributes enabled by the first API interface.

[0035] In implementation, for business requirements, match an API interface (i.e., the first API interface) for it. The specific process for finding the first API interface is as follows: First, determine the function range of the business requirements, define the function attributes according to the business requirements. When the business requirements are relatively complex, a business requirement can also include multiple function attributes, and the set of all function attributes is the function range of the business requirements. Among them, the function attributes can be described in combination with the composition structure of the API interface (URL and request method), such as described as "POST the A document to URL1". Here, URL1 can be considered as the URL on the B server, and the function of this function attribute is to upload the A document to the B server. Therefore, by default, a function attribute refers to including a request address (URL) and a request method. The function attributes are used to describe the functions of the API interface. A business requirement corresponds to an API interface. When the function range corresponding to the business requirement contains multiple function attributes, the composition structure of the corresponding API interface includes several request addresses or several request methods, that is, the API interface supports one or more request methods / URLs.

[0036] Next, determine whether there is a second API interface among the existing API interfaces. The second API interface refers to an API interface that meets any of the following conditions: the API interface corresponding to the function attribute and the function attributes included in the function range (hereinafter referred to as the second API interface). 1. The second API interface is unique, and the function attributes corresponding to the second API interface are exactly the same as the function attributes included in the function range. 2. There are several second API interfaces, and the set formed by the function attributes corresponding to each second API interface has an intersection with the function range.

[0037] For case 1, the interaction platform is used to use the second API interface as the first API interface to establish a corresponding relationship with the business requirements and complete the matching operation. For Case 2, the interaction platform is used to further determine whether all the functional attributes included in all the second API interfaces cover all the functional attributes included in the functional scope after taking the union. If so, the set composed of all the second API interfaces is used as the first API interface, and the corresponding relationship between the first API interface and the business requirement is established. That is, the first API interface at this time includes several second API interfaces.

[0038] If not, that is, after taking the union of the functional attributes included in all the second API interfaces, there are functional attributes (i.e., newly added functional attributes) that are included in the functional scope but not included in the aforementioned union. At this time, an API interface (i.e., the third API interface) is generated based on the newly added functional attributes so that the third API interface has the functions corresponding to the newly added functional attributes. The set of functional attributes included in the third API interface and the union of the functional attributes included in all the second API interfaces are complementary to each other. At this time, the interaction platform is used to take the set composed of all the second API interfaces and the third API interface as the first API interface, and establish the corresponding relationship between the first API interface and the business requirement. That is, the first API interface at this time includes the third API interface and several second API interfaces.

[0039] Furthermore, it should be noted that since an API interface may include multiple functional attributes, for Case 2, even if the union of the functional attributes corresponding to all the second API interfaces can completely cover the functional scope or cover part of the functional scope, there are likely to be functional attributes that do not need to be enabled (hereinafter simply referred to as disabled functional attributes) in the second API interfaces. Disabled functional attributes refer to the functional attributes that are included in the aforementioned union but not included in the functional scope. Exemplarily, if the functional scope is (X, Y, Z); at this time, there are a second API interface 1, a second API interface 2, and a third API interface, and the functional attributes included in the second API interface 1 are (X, M), the functional attributes included in the second API interface 2 are (Y, N), and the functional attributes corresponding to the third API interface are Z. Then M and N are disabled functional attributes, and only X included in the second API interface 1 and Y included in the second API interface 2 are the functional attributes that need to be enabled. Therefore, this application proposes that when determining the first API interface, it is also necessary to determine the functional attributes (i.e., the first functional attributes) that each second API interface included in the first API interface needs to be enabled. The determination process can be as shown in the above example, by taking the intersection of the set of functional attributes included in each second API interface and the functional scope.

[0040] If the second API interface does not exist in the existing API interfaces, a third API interface is generated, and according to all the functional attributes included in the functional scope, the configuration file and composition structure of the API interface are defined, so that the URLs and request methods included in the API interface can correspond to all the functional attributes included in the functional scope; then the third API interface is used as the first API interface, and the corresponding relationship between the first API interface and the business requirements is established. In summary, the first API interface includes the second API interface and / or the third API interface.

[0041] Then a call instruction is generated, which includes the first API interface information and the first functional attribute. Here, the information at least includes the interface name. Then the interaction platform calls the first API interface according to the call instruction, and only enables the functional attributes in the first API interface that are consistent with the first functional attribute. Exemplarily, the specific implementation manner of the enabling operation here can be: whenever an API interface is generated and the composition structure of the API interface is defined according to the functional attributes, the configuration file of the API interface can also include the enabling conditions of each functional attribute, specifying that only when the enabling conditions are met, the API interface supports the corresponding functional attribute. Correspondingly, the interaction platform can store each functional attribute and its corresponding enabling order for meeting the enabling conditions. Therefore, in the process of generating the call instruction and calling the first API interface, the interaction platform will query the corresponding enabling order according to the first functional attribute included in the call instruction, and use the enabling order to enable the first functional attribute. In addition, the status of each functional attribute, such as the enabled status or disabled status, can be updated in real time in the configuration file of the API interface, and the initial status of all functional attributes included in all API interfaces is defaulted to the disabled status when they are called, so that when processing service requests using the first API interface, the first API interface can only support the functions corresponding to the first functional attribute. In the embodiments of the present application, disabling or enabling a functional attribute specifically refers to disabling or enabling any of the multiple request methods included in the interface. The above enabling order and its corresponding enabling conditions can specifically be client identity information or time information, so as to restrict the access rights of different clients or different time periods to different request methods included in the interface. In addition, the request methods to be enabled can also be dynamically controlled through the configuration file, database or configuration center without modifying the code (such as directly disabling the request methods to be disabled through the AWS console or Terraform configuration); or filtering the disabled request methods through the API gateway / firewall layer / reverse proxy (such as Nginx)... Finally, the server - side corresponding to the service request is found by calling the API interface, so that the server - side processes the service request and returns a processing result, and finally the processing result is fed back to the client. The feedback result here can specifically be the feedback content on whether the service request is processed (such as task failure / upload success, etc.) or specific data. For example, if the service request is to download a C document, then the corresponding feedback result can be the downloaded C document.

[0042] Optionally, "determining the function scope corresponding to the business requirement" in S1021 specifically further includes the following sub - steps: Determine a guidance set based on the business requirement, where the guidance set includes at least guidance questions; Determine the function scope corresponding to the business requirement, sequentially send each guidance question included in the guidance set to the client one by one, and based on the feedback result of the client for the target guidance question, determine whether there is a new requirement. If so, add the function attributes corresponding to the new requirement to the function scope.

[0043] In implementation, the guidance questions are pre - defined and stored in the interaction platform, and the guidance questions are specifically divided into general questions and associated questions. The specific content of the general question can be: "Are there any other requirements?"; The associated question is a question summarized after the interaction platform regularly analyzes the associations between different business requirements proposed by the same client within a specified duration in a historical period. Specifically, the interaction platform will regularly analyze all service requests received within a specified period and their corresponding business requirements to learn and analyze whether there are associated business requirements. For example, based on all business requirements within the aforementioned specified period, it is summarized that whenever the client submits a business request with the content "download Q - type documents from server D", the corresponding client will further submit a business request with the content "download P - type documents from server D". Therefore, the interaction platform can generate an associated question with the content "Do you need to download other documents from this server?" and establish an association relationship between the associated question and the business requirement.

[0044] Next, the interaction platform is used to sequentially send each guidance question included in the guidance set to the client and receive the feedback result from the client regarding each of the aforementioned guidance questions. Use the AI Agent to parse and determine whether the client has issued a new business requirement (i.e., a new requirement) in combination with the context (i.e., the guidance question and its corresponding feedback result). If so, define the function attributes based on the new requirement and add the function attributes to the function scope; until the parsing of the feedback results corresponding to all the guidance questions in the set is completed, finally obtain the function scope, and then determine the first API interface based on the function scope, and this function scope is all the functions that the first API interface needs to support.

[0045] Optionally, S103 specifically includes the following sub-steps: After performing desensitization processing on the processing result of the service request, feedback the processing result of the corresponding service request to the client.

[0046] In implementation, the interaction platform is further configured to identify whether there is sensitive information (such as identity information) in the feedback result through a preset machine learning algorithm, and perform desensitization processing on the identified sensitive information, such as replacing the sensitive information with randomly generated content. The desensitization processing operation can be in the form of a plugin, such as a preset security plugin, for performing the above desensitization operation; in other embodiments, the security plugin can also have a malicious behavior detection function, which is used to identify malicious code samples by extracting features during the parsing process of the service request.

[0047] Optionally, "matching an API interface for the service requirement" in S102 further includes: Match an API interface for the service requirement, and define the configuration information of the API interface based on the service requirement, where the configuration information includes at least the interface bandwidth; The interaction method further includes the following steps: Predict the change trend of the number of times each API interface is called over time, and determine the interface bandwidth of each API interface at different times based on the change trend; Adjust the interface bandwidth included in the configuration information of the corresponding API interface in real time according to the interface bandwidth of each API interface.

[0048] In implementation, whenever the API interface corresponding to the service requirement is determined (i.e., the first API interface), in addition to the above-mentioned: enabling the first functional attribute included in the API interface, other configuration information of the API interface can also be defined, such as the interface name, interface bandwidth, etc.; the interface bandwidth here is a dynamically adjustable value, specifically by the interaction platform regularly determining the curve of the number of times the corresponding API interface is called over time, and then according to the predefined relationship between the number of calls and the interface bandwidth, plotting the curve of the interface bandwidth over time, and using this curve to update the interface bandwidth in the configuration information of the corresponding API interface in real time for real-time adjustment of the API interface bandwidth. In other embodiments, the configuration information of the API interface can also include the applicable scope, and the applicable scope includes client identity information, that is, it is limited that only the specified client can use this API interface, so as to ensure the smoothness of the interaction between the client and the server by adjusting the API interface bandwidth during the high concurrency period.

[0049] Optionally, the interaction method further includes: After matching the API interface for business requirements, the test plan is executed regularly to generate virtual service requests corresponding to the business requirements, parse the virtual service requests, match the API interfaces, call the API interfaces to process the virtual service requests, verify the processing results of the virtual service requests, and output the verification results. Among them, the test plan at least includes virtual service requests corresponding to the business requirements.

[0050] In implementation, the interaction platform uses the preset API documentation and test documentation (i.e., the test plan) to conduct tests on the API interfaces and the front-end pages (i.e., the access pages when the client accesses the interaction platform and triggers service requests) by means of AI simulation technology to simulate the operations of users. That is, it simulates the operation of the user triggering a service request to generate a virtual service request, then calls the corresponding API interface to process the virtual service request and obtains the corresponding processing result, and displays the processing result (hereinafter referred to as the virtual processing result) on the front-end page as the verification result. In addition, the interaction platform can also display the processing results corresponding to the service requests processed by the API interface in the historical period (here, the processing result refers to the processing result obtained by the API interface for the service requests made by real users in the historical period, hereinafter referred to as the actual processing result) on the front-end page, and can also be used to verify the similarity between the virtual processing result and the actual processing result, and display the similarity on the front-end page as the verification result to achieve AI automated testing for the testers to perform subsequent maintenance operations based on the test results.

[0051] The embodiment of the present application also discloses an interaction platform based on an AI gateway and an AI Agent. Refer to Figure 3 , including: A requirement analysis module 201, configured to obtain the service request sent by the client and analyze the service request to obtain the business requirements; An interface matching module 202, configured to match the API interface for the business requirements and call the API interface to process the service request; A requirement response module 203, configured to feedback the processing result of the corresponding service request to the client.

[0052] Optionally, the requirement response module is further configured to desensitize the processing result of the service request and then feedback the processing result of the corresponding service request to the client.

[0053] Optionally, the interface matching module 202 is further configured to match the API interface for the business requirements and define the configuration information of the API interface based on the business requirements, where the configuration information at least includes the interface bandwidth; It further includes a bandwidth adjustment module for predicting the change trend of the number of times each API interface is called over time, determining the interface bandwidth of each API interface at different times based on the change trend, and adjusting the interface bandwidth included in the configuration information of the corresponding API interface in real time according to the interface bandwidth of each API interface.

[0054] Optionally, it further includes an AI test module for regularly executing a test plan whenever an API interface is matched for a business requirement, so as to generate a virtual service request corresponding to the business requirement, parsing the virtual service request to match the API interface, calling the API interface to process the virtual service request, verifying the processing result of the virtual service request, and outputting the verification result; wherein, the test plan at least includes a virtual service request corresponding to the business requirement.

[0055] Optionally, the interface matching module 202 is further configured to generate a call instruction based on the service request, where the call instruction at least includes the first API interface information to be called and the first functional attribute to be enabled by the first API interface; and is further configured to enable the functional attribute in the first API interface that is consistent with the first functional attribute, and process the service request based on the functional attribute enabled by the first API interface.

[0056] Optionally, the interface matching module 202 is further configured to determine the functional range corresponding to the business requirement, determine the second API interface from the existing API interfaces according to the functional attributes corresponding to the existing API interfaces, and determine the functional attributes to be enabled for each second API interface; wherein, the union of the functional attributes to be enabled for all second API interfaces is included in the functional range; and is further configured to determine a new functional attribute if the union range of the functional attributes to be enabled for the second API interface is less than the functional range, and generate a third API interface for the new functional attribute so that the third API interface corresponds to the new functional attribute; wherein, the new functional attribute is the complement of the union of the functional attributes to be enabled for the second API interface.

[0057] Optionally, the interface matching module 202 is further configured to determine a guidance set based on the business requirement, where the guidance set at least includes guidance questions; determine the functional range corresponding to the business requirement, sequentially send the guidance questions included in the guidance set to the client one by one, and determine whether there is a new requirement according to the feedback result of the client on the target guidance question. If so, add the functional attributes corresponding to the new requirement to the functional range.

[0058] An embodiment of the present application also discloses an interaction device based on an AI gateway and an AI Agent. The interaction device based on the AI gateway and the AI Agent includes a memory and a processor. The memory stores a computer program that can be loaded and executed by the processor, and the computer program is the above-mentioned interaction method based on the AI gateway and the AI Agent.

[0059] An embodiment of the present application also discloses a computer-readable storage medium, which stores a computer program that can be loaded and executed by the processor, and the computer program is the above-mentioned interaction method based on the AI gateway and the AI Agent. The computer-readable storage medium includes, for example: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs.

[0060] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0061] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit the protection scope of the application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all embodiments. Based on these embodiments, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope to be protected by the present application.

Claims

1. An interaction method based on an AI gateway and an AI Agent, characterized in that, Including: Obtain a service request sent by a client, and parse the service request to obtain a business requirement; Match an API interface for the business requirement, and call the API interface to process the service request; Feedback the processing result of the corresponding service request to the client.

2. The interactive method based on the AI gateway and the AI Agent according to claim 1, wherein The feedback of the processing result of the corresponding service request to the client includes: After performing desensitization processing on the processing result of the service request, feedback the processing result of the corresponding service request to the client.

3. The interactive method based on the AI gateway and the AI Agent according to claim 1, wherein The matching of the API interface for the business requirement includes: Match an API interface for the business requirement, and define the configuration information of the API interface based on the business requirement, where the configuration information includes at least the interface bandwidth; The method further includes: Predict the change trend of the number of times each API interface is called over time, and determine the interface bandwidth of each API interface at different times based on the change trend; Adjust the interface bandwidth included in the configuration information of the corresponding API interface in real time according to the interface bandwidth of each API interface.

4. The interactive method based on the AI gateway and the AI Agent according to claim 3, wherein The method further includes: Whenever an API interface is matched for a business requirement, regularly execute a test plan to generate a virtual service request corresponding to the business requirement, parse the virtual service request, match an API interface, call the API interface to process the virtual service request, verify the processing result of the virtual service request, and output a verification result; where the test plan includes at least a virtual service request corresponding to the business requirement.

5. The interactive method based on an AI gateway and an AI Agent according to claim 1, wherein, Each of the API interfaces corresponds to at least one functional attribute; The calling of the API interface to process the service request includes: Generate a call instruction based on the service request, where the call instruction includes at least the first API interface information to be called and the first functional attribute to be enabled by the first API interface; Enable the functional attribute in the first API interface that is consistent with the first functional attribute, and process the service request based on the functional attribute enabled by the first API interface.

6. The interactive method based on an AI gateway and an AI Agent according to claim 5, wherein The first API interface includes a second API interface and / or a third API interface; The matching of the API interface for the business requirement further includes: Determine the functional scope corresponding to the business requirement, determine the second API interface from the existing API interfaces according to the functional attributes corresponding to the existing API interfaces, and determine the functional attributes to be enabled for each of the second API interfaces; where the union of the functional attributes to be enabled for all the second API interfaces is included in the functional scope; If the union range of the functional attributes to be enabled for the second API interface is less than the functional scope, determine a new functional attribute, and generate a third API interface for the new functional attribute so that the third API interface corresponds to the new functional attribute; where the new functional attribute is the complement of the union of the functional attributes to be enabled for the second API interface.

7. The interactive method based on the AI gateway and the AI Agent according to claim 5, wherein The determination of the functional scope corresponding to the business requirement further includes: Determine a guidance set based on the business requirement, where the guidance set includes at least guidance questions; Determine the functional scope corresponding to the business requirements, sequentially send each guiding problem included in the guiding set to the client one by one, and determine whether there are new requirements according to the feedback result of the client on the target guiding problem. If so, add the functional attributes corresponding to the new requirements to the functional scope.

8. An interaction platform based on an AI gateway and an AI Agent, characterized in that, Including, A requirement analysis module (201) for obtaining a service request sent by the client and parsing the service request to obtain business requirements; An interface matching module (202) for matching an API interface for the business requirements and calling the API interface to process the service request; A requirement response module (203) for feeding back the processing result of the corresponding service request to the client.

9. An interactive device based on an AI gateway and an AI Agent, characterized in that, Including a memory and a processor, and a computer program capable of being loaded and executed by the processor is stored on the memory, and the computer program is the same as any one of the methods described in claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Stored with a computer program capable of being loaded and executed by the processor, and the computer program is the same as any one of the methods described in claims 1 to 7.

Images