An agent parsing method, device and program product

Abstract

The application discloses an agent analysis method and device and a program product. The method comprises the following steps: receiving a first query request for a domain name system, wherein the first query request comprises a first query domain name, the first query domain name comprises multiple fields, and the multiple fields are respectively used for describing information of different dimensions of an agent to be queried; determining at least one target agent matched with the first query domain name in response to the first query request; and feeding back an analysis result, wherein the analysis result comprises resource records of the at least one target agent. In this way, the private identification or the special interface of each agent platform does not need to be relied on, and the cross-platform adaptation cost is significantly reduced. The application of the domain name system can improve the standardization degree and the system compatibility of the analysis process, and an additional special communication protocol or analysis mechanism does not need to be established. The communication overhead and processing delay caused by multiple invocations of different platforms are avoided, and the analysis efficiency is improved. The returned analysis result can be more consistent with the query demand, and the agent matching accuracy is effectively improved.

Description

Technical Field

[0001] This application relates to the field of computers, and in particular to a method, apparatus and program product for intelligent agent analysis. Background Technology

[0002] With the development of artificial intelligence technology, intelligent agents can perform operations such as information acquisition, content generation, data processing, or device control according to task requirements. In scenarios where multiple intelligent agents collaborate on tasks, the requesting party needs to identify the target intelligent agent with the corresponding capabilities from among multiple agents and obtain information such as the target intelligent agent's communication address, capability description, and operating status to achieve discovery, selection, and communication between intelligent agents. Therefore, agent resolution is a crucial step in realizing agent collaboration.

[0003] In current practical applications, agents are typically managed by corresponding agent platforms. These platforms can distinguish different agents using universally unique identifiers, random strings, custom serial numbers, or other private identifiers, and associate these private identifiers with the agent's capability descriptions and communication addresses. The requesting party needs to submit query conditions through the interface provided by the agent platform, which then searches its managed agent data for agents that match the query conditions.

[0004] However, different agent platforms typically use different identifier formats, information description formats, and query interfaces. Private identifiers can generally only be used within their corresponding agent platform, making it difficult to support cross-platform agent parsing. This can also lead to complex parsing processes, low matching efficiency, and insufficient accuracy of parsing results. Summary of the Invention

[0005] This application provides a method for intelligent agent parsing, enabling intelligent agents to perform parsing across platforms and improving parsing accuracy. Furthermore, this application also provides corresponding apparatus, computing devices, computer-readable storage media, and computer program products.

[0006] In a first aspect, embodiments of this application provide an agent resolution method, comprising: receiving a first query request for a domain name system, the first query request including a first query domain name, the first query domain name including multiple fields, the multiple fields being used to describe information of different dimensions of the agent to be queried; responding to the first query request, determining at least one target agent that matches the first query domain name; and providing a resolution result, the resolution result including resource records of at least one target agent.

[0007] In one possible implementation, determining at least one target agent that matches the first query domain name includes: determining at least one registered domain name that matches the first query domain name from a plurality of registered domain names, wherein the plurality of registered domain names correspond one-to-one with a plurality of agents, and at least one registered domain name corresponds to at least one target agent, and the registered domain name includes a plurality of fields; or, determining at least one resource record that matches the first query domain name from a plurality of resource records, wherein at least one resource record corresponds to at least one target agent.

[0008] In one possible implementation, before receiving a first query request for the Domain Name System, the method further includes: receiving a first registration request, the first registration request including a first registered domain name and information for describing a first intelligent agent; establishing a correspondence between the first registered domain name and a first resource record, the first resource record including information for describing the first intelligent agent.

[0009] In one possible implementation, the multiple fields include one or more of a function field, a scenario field, an organization field, and a management domain field; the function field is used to indicate the functions provided by the agent, the scenario field is used to indicate the scenarios to which the agent is applicable, the organization field is used to indicate the organization to which the agent belongs, and the management domain field is used to indicate the domain name service for managing the agent.

[0010] In one possible implementation, before establishing the correspondence between the first registered domain name and the first resource record, the method further includes: determining whether the first registered domain name conforms to the domain name naming rules; establishing the correspondence between the first registered domain name and the first resource record includes: if the first registered domain name conforms to the domain name naming rules, establishing the correspondence between the first registered domain name and the first resource record.

[0011] In one possible implementation, the number of at least one target agent is greater than 1, and in the parsing result, the resource records of at least one target agent are sorted based on one or more of the following: the degree of matching between at least one target agent and the first query domain name, the running status of at least one target agent, and the response speed of at least one target agent.

[0012] In one possible implementation, the resource record includes one or more of the following: agent version information, scene information, capability information, running status information, communication information for accessing the agent, input data format, and output data format.

[0013] In one possible implementation, the first query request is sent by a recursive domain name system server, or the first query request is sent by a user terminal device.

[0014] Secondly, embodiments of this application provide an intelligent agent parsing device, comprising: a communication module, configured to receive a first query request for a domain name system, the first query request including a first query domain name, the first query domain name including multiple fields, the multiple fields being used to describe information of different dimensions of the intelligent agent to be queried; a determination module, configured to, in response to the first query request, determine at least one target intelligent agent matching the first query domain name; the communication module is further configured to provide a parsing result, the parsing result including resource records of at least one target intelligent agent.

[0015] In one possible implementation, the determining module is specifically configured to determine at least one registered domain name that matches the first query domain name from a plurality of registered domain names, wherein the plurality of registered domain names correspond one-to-one with a plurality of intelligent agents, and at least one registered domain name corresponds to at least one target intelligent agent, and the registered domain name includes a plurality of fields; or, specifically configured to determine at least one resource record that matches the first query domain name from a plurality of resource records, wherein at least one resource record corresponds to at least one target intelligent agent.

[0016] In one possible implementation, the communication module is further configured to receive a first registration request before receiving a first query request for the Domain Name System, the first registration request including a first registered domain name and information for describing a first intelligent agent; the apparatus further includes an establishment module configured to establish a correspondence between the first registered domain name and a first resource record, the first resource record including information for describing the first intelligent agent.

[0017] In one possible implementation, the multiple fields include one or more of a function field, a scenario field, an organization field, and a management domain field; the function field is used to indicate the functions provided by the agent, the scenario field is used to indicate the scenarios to which the agent is applicable, the organization field is used to indicate the organization to which the agent belongs, and the management domain field is used to indicate the domain name service for managing the agent.

[0018] In one possible implementation, the determining module is further configured to determine whether the first registered domain name conforms to the domain name naming rules before establishing the correspondence between the first registered domain name and the first resource record; the establishing module is specifically configured to establish the correspondence between the first registered domain name and the first resource record if the first registered domain name conforms to the domain name naming rules.

[0019] In one possible implementation, the number of at least one target agent is greater than 1, and in the parsing result, the resource records of at least one target agent are sorted based on one or more of the following: the degree of matching between at least one target agent and the first query domain name, the running status of at least one target agent, and the response speed of at least one target agent.

[0020] In one possible implementation, the resource record includes one or more of the following: agent version information, scene information, capability information, running status information, communication information for accessing the agent, input data format, and output data format.

[0021] In one possible implementation, the first query request is sent by a recursive domain name system server, or the first query request is sent by a user terminal device.

[0022] Thirdly, embodiments of this application also provide a computing device, which may include a processor and a memory: the memory is used to store a computer program; the processor is used to execute the method described in the first aspect and any one of the embodiments in the first aspect according to the computer program.

[0023] Fourthly, embodiments of this application also provide a computer-readable storage medium for storing a computer program for performing the methods described in the first aspect and any one of the embodiments of the first aspect.

[0024] Fifthly, embodiments of this application also provide a computer program product including instructions that, when run on a computing device, cause the computing device to perform the methods described in the first aspect and any one of the embodiments of the first aspect.

[0025] In the above implementation of this application embodiment, the first query domain name can, for example, carry information about the intelligent agent to be queried in different dimensions such as function, scenario, organization, and management domain name through multiple fields. This allows for direct intelligent agent matching based on the first query domain name, without relying on the private identifiers or dedicated interfaces of each intelligent agent platform, thereby achieving a unified query entry point and significantly reducing cross-platform adaptation costs. Simultaneously, applying the intelligent agent resolution process to the domain name system can fully utilize the existing hierarchical structure and distributed resolution mechanism of the domain name system, enabling query requests to be automatically routed to the corresponding server according to the existing recursive resolution process. This improves the standardization of the resolution process and system compatibility, and eliminates the need to establish additional dedicated communication protocols or resolution mechanisms, thus simplifying the system architecture. Furthermore, by having the domain name system return resource records of at least one target intelligent agent, the communication overhead and processing delays caused by multiple calls to different platform interfaces can be avoided, reducing redundant matching and invalid queries, and improving overall resolution efficiency. Moreover, since the first query domain name contains information about the intelligent agent to be queried in different dimensions, the returned resolution results can better match the query requirements, thereby further improving the accuracy of intelligent agent matching. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0027] Figure 1 This is a schematic diagram of the architecture of a data processing system according to an embodiment of this application;

[0028] Figure 2 This is a flowchart illustrating an intelligent agent parsing method in an embodiment of this application;

[0029] Figure 3 This is a schematic diagram of the structure of an intelligent agent analysis device according to an embodiment of this application;

[0030] Figure 4 This is a schematic diagram of the hardware structure of a computing device according to an embodiment of this application. Detailed Implementation

[0031] See Figure 1 This is a schematic diagram of a data processing system provided in this application. Figure 1 As shown, the data processing system 10 includes a processor and a memory. Figure 1 Taking the processor 100 and memory 201 as an example, in actual application scenarios, the data processing system 10 may include a greater number of processors or memories.

[0032] The processor 100 and the memory 201 can be connected via a bus, such as a peripheral component interconnect express (PCIe) bus. Alternatively, the processor 100 and the memory 201 can also be connected via a network, such as a local area network (LAN) or a wide area network (WAN), and in terms of connection method, such as a wired network or a wireless network.

[0033] The processor 100 refers to a device with data processing and control capabilities. For example, the processor 100 can be a graphics processing unit (GPU), a central processing unit (CPU), etc., and there is no limitation on this.

[0034] Memory 201 refers to a device with file storage capabilities, such as semiconductor memory, magnetic memory, optical memory, etc., without limitation. Additionally, memory 201 may store a database for data management and retrieval, without limitation.

[0035] This application can be applied to the aforementioned data processing system 10. The data processing system 10 can be a domain name server, specifically a recursive domain name system (DNS) server or an authoritative DNS server. Alternatively, the data processing system 10 can be a local resolution component in a terminal device, gateway device, browser, local proxy program, or other device or module, used to provide domain name resolution services.

[0036] The following section introduces the relevant technologies for agent parsing.

[0037] An intelligent agent can be a data processing entity built based on an artificial intelligence (AI) model, used to perform tasks such as information acquisition, content generation, data analysis, business processing, or device control. In practical applications, multiple intelligent agents can each possess different functions and be created, deployed, and managed by different intelligent agent platforms. When an intelligent agent needs to collaborate with other intelligent agents to perform a task, it first needs to identify a target intelligent agent with the desired capabilities from multiple candidate intelligent agents and obtain information such as the target intelligent agent's communication address, capability description, and operational status. The process of finding the target intelligent agent and obtaining its relevant information based on task requirements can be called intelligent agent resolution.

[0038] In current practical applications, agent resolution typically relies on private identifiers and private query interfaces set up within the agent platform. When creating an agent, the agent platform can assign it a universally unique identifier (UUID), a random string, a custom serial number, or other forms of private identifier. This private identifier is used to uniquely distinguish different agents within the corresponding agent platform. The agent platform can also establish a relationship between the private identifier and the agent's description information, and store this relationship in a database maintained by the agent platform.

[0039] The agent description information may include, for example, information used to access the agent. This information typically uses a data format defined by the agent platform and can be stored as database fields, configuration files, or capability description documents. The data structure for agent description information may differ between different agent platforms.

[0040] When the requesting party knows the private identifier of the target agent, it can send a query request carrying that private identifier to the agent platform. The agent platform queries pre-stored relationships based on the private identifier to determine the agent description information corresponding to that private identifier, and then returns the agent description information to the requesting party. The requesting party can then obtain relevant information for accessing the agent from the agent description information and send a communication request based on this.

[0041] However, the aforementioned agent parsing methods based on private identifiers are typically only applicable within the corresponding agent platform. Different agent platforms usually employ different private identifier generation rules, and the same private identifier may correspond to different agents on different platforms. Furthermore, different forms of private identifiers are difficult to process using a unified parsing method. When a requester needs to find agents managed by other agent platforms, it is necessary to adapt the query interfaces, authentication methods, and data formats of different agent platforms, making the implementation process of cross-platform agent parsing quite complex.

[0042] Furthermore, private identifiers are typically composed of random characters, numbers, or hash values, and cannot directly represent semantic information such as the agent's function, applicable scenarios, or affiliated organization. When the requesting party is unaware of the target agent's private identifier, it often cannot directly determine the target agent based on the private identifier alone.

[0043] Furthermore, agent platforms typically only maintain information about agents managed by that platform, making it difficult for the requester to obtain candidate agents from different platforms in a single query. The requester may need to send query requests to multiple agent platforms separately and then merge and filter the results returned by each platform. As the number of agent platforms and agents increases, the number of query requests, interface adaptations, and result processing also increases, potentially leading to reduced agent parsing efficiency. Therefore, current agent parsing methods struggle to balance cross-platform applicability, parsing efficiency, and semantic matching capabilities.

[0044] Based on this, in this application, the processor 100 can receive a first query request for the Domain Name System, the first query request including a first query domain name, the first query domain name including multiple fields, the multiple fields being used to describe information of different dimensions of the intelligent agent to be queried; then, in response to the first query request, the processor 100 determines at least one target intelligent agent that matches the first query domain name; subsequently, the processor 100 feeds back the parsing result, the parsing result including the resource records of at least one target intelligent agent.

[0045] In this way, the first query domain name can carry information about the target agent in different dimensions, such as function, scenario, organization, and management domain name, through multiple fields. This allows the processor 100 to directly match agents based on the first query domain name without relying on the private identifiers or dedicated interfaces of each agent platform, thus achieving a unified query entry point and significantly reducing cross-platform adaptation costs. Simultaneously, applying the agent resolution process to the Domain Name System (DNS) fully utilizes the existing hierarchical structure and distributed resolution mechanism of DNS, enabling query requests to be automatically routed to the corresponding server according to the existing recursive resolution process. This improves the standardization of the resolution process and system compatibility, and eliminates the need to establish additional dedicated communication protocols or resolution mechanisms, thereby simplifying the system architecture. Furthermore, by providing at least one resource record of the target agent through DNS feedback, the communication overhead and processing latency caused by multiple calls to different platform interfaces can be avoided, reducing redundant matching and invalid queries, and improving overall resolution efficiency. Moreover, since the first query domain name contains information about the target agent in different dimensions, the returned resolution results are more consistent with the query requirements, further improving the accuracy of agent matching.

[0046] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, various non-limiting embodiments of the present application will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0047] See Figure 2 , Figure 2 This paper illustrates a flowchart of an agent parsing method according to an embodiment of this application. This method can be applied to... Figure 1 The data processing system 10 shown can be applied to other applicable data processing systems, as described below. Figure 1 The data processing system 10 shown is illustrated using agent parsing as an example. Figure 2 As shown, the method may specifically include the following steps.

[0048] S201: Processor 100 receives a first query request for the Domain Name System. The first query request includes a first query domain name, which includes multiple fields, each of which describes information about different dimensions of the agent to be queried.

[0049] In this embodiment of the application, the processor 100 in the data processing system 10 can be used to manage the domain names and corresponding resolution records of multiple smart agents, and can determine the smart agent that matches the smart agent to be queried based on the received first query request.

[0050] Understandably, DNS can be used to resolve data between domain names. Generally, a domain name can include multiple fields in a hierarchical structure, with adjacent fields separated by periods. After receiving a query request containing a domain name, the DNS server can query the resource record corresponding to that domain name based on the query type in the query request and return the query result to the sender of the query request. This application embodiment utilizes the existing hierarchical domain name structure of DNS to carry relevant information about the queried agent, allowing different fields in the first query domain name to describe different dimensions of information about the queried agent.

[0051] Optionally, the first query request may be sent by a recursive DNS server, for example. Specifically, the user device may generate a first query domain name based on the task to be performed and send a DNS query request including the first query domain name to the recursive DNS server. If the recursive DNS server cannot perform the query process for the first query domain name, the recursive DNS server may perform a recursive query or an iterative query and send the first query request to the processor 100. The user device may be a terminal device, a server, an edge computing device, or other device with data processing and communication capabilities. An intelligent agent may be running on the user device, and the first query request may be generated by the intelligent agent based on the task to be performed, or it may be generated by the intelligent agent management program, the DNS client, or other applications in the user device.

[0052] Alternatively, the first query request can also be sent directly from the aforementioned user terminal device to the device where the processor 100 is located, without any limitation.

[0053] The first query request can adopt the query message format specified by the DNS protocol. The first query request may include message identifier, query flag, query quantity, first query domain name, query type, and query category. The query type can be the query type corresponding to text (TXT) resource records. It is understood that in other implementations, the first query request can also adopt other query types, and this is not limited.

[0054] The first query domain name is described below as an example.

[0055] The first query domain name includes multiple fields, each describing different dimensions of information about the agent being queried. For example, these fields may include one or more of the following: functional fields, scenario fields, organizational fields, and management domain name fields. The fields in the first query domain name can be arranged in a preset order and form structured query information through the hierarchical relationship of the domain names.

[0056] For example, the first query domain name can be in the form of "functional field.scenario field.organization field.management domain name field". The first query domain name can also omit one or more fields, for example, it can be in the form of "functional field.scenario field.management domain name field", "functional field.organization field.management domain name field" or "functional field.management domain name field".

[0057] The function field indicates the functionality provided by the agent being queried. The function field can include one or more function keywords describing the agent's core functions. For example, the function field can include characters such as translator, image-recognition, data-analysis, audit, or compliance to represent the corresponding function. Function fields can consist of lowercase English letters, numbers, or hyphens. For functions that need to be described by multiple keywords, these keywords can be connected by hyphens or set separately in different subfields of the first query domain. For example, audit-compliance can be used to describe the compliance audit function.

[0058] The scenario field indicates the applicable scenario for the queried agent. This field can include characters corresponding to business scenarios such as finance, office, IoT, healthcare, education, or transportation. For example, "fintech" can indicate a financial technology scenario, "office" can indicate an office scenario, and "iot" can indicate an IoT scenario. For agents applicable to multiple scenarios, the scenario field can include characters corresponding to multiple scenarios, which can be connected by hyphens. For example, "office-iot" can indicate that the queried agent is applicable to both office and IoT scenarios.

[0059] The organization field indicates the organization to which the queried agent belongs. The organization field can include the organization's name, its English name, its standardized abbreviation, or a unique identifier for the individual developer. When a client device needs to query an agent provided by a specific organization, it can set the corresponding organization field in the first query domain. When the client device does not specify the organization to which the queried agent belongs, the first query domain may not include the organization field, or the organization field can be set to a preset character used to represent any organization.

[0060] The management domain field is used to indicate the domain name service for the management agent. The management domain field can be a domain name provided by a domain name registry, domain name registrar, agent service provider, or other domain name management entity. In this application, the data processing system 10 can correspond to an authoritative DNS server, which can provide authoritative resolution services for the domain name corresponding to the management domain field and manage multiple agent domain names and their resolution records under the management domain field. By setting the management domain field in the first query domain name, the recursive DNS server can send the first query request to the authoritative DNS server corresponding to the management domain field according to the DNS resolution process.

[0061] For example, the management domain field could be dnsagent.cn, and the first query domain could be audit-compliance.fintech.globaltrust.dnsagent.cn. Here, audit-compliance is a function field indicating that the agent being queried provides compliance auditing functions; fintech is a scenario field indicating that the agent being queried is suitable for fintech scenarios; globaltrust is an organization field indicating that the agent being queried belongs to the corresponding organization; and dnsagent.cn is the management domain field indicating the domain name service that manages this agent.

[0062] It should be noted that the multiple fields in the first query domain name can represent precise query conditions or retrieval conditions used for semantic matching. This will be explained in detail in step S202 and will not be repeated here.

[0063] It should be noted that the first query domain name may also include label fields, status fields, protocol fields, or other information fields used to describe the intelligent agent to be queried, without any restrictions.

[0064] Optionally, the first query request may also include requester information, request time, query identifier, or other information related to the query process. The requester information may be used to indicate the user terminal device, agent, or recursive DNS server sending the first query request. The query identifier may be used to distinguish different first query requests and to associate the first query request with subsequent resolution responses. After receiving the first query request, the processor 100 may parse the message format of the first query request, extract the first query domain name from the first query request, and obtain multiple fields included in the first query domain name for subsequent matching based on the first query domain name.

[0065] S202: In response to the first query request, the processor 100 identifies at least one target agent that matches the first query domain name.

[0066] In this embodiment, after receiving a first query request, the processor 100 can extract a first query domain name from the first query request and, based on multiple fields included in the first query domain name, determine at least one target intelligent agent that matches the first query domain name from among multiple registered intelligent agents. The target intelligent agent may be an intelligent agent capable of providing the function indicated by the first query domain name, applicable to the scenario indicated by the first query domain name, belonging to the organization indicated by the first query domain name, or managed by the domain name service indicated by the first query domain name.

[0067] For example, in the data processing system 10, the memory 201 can store the correspondence between multiple agents, multiple registered domain names, and multiple resource records. Specifically, multiple agents can correspond one-to-one with multiple registered domain names, and / or, multiple registered domain names can also correspond one-to-one with multiple resource records. The registered domain name corresponding to an agent is used to identify the agent in the DNS, and the resource record corresponding to a registered domain name is used to carry information describing the agent. Therefore, the processor 100 can determine the corresponding agent by the registered domain name or by the resource record.

[0068] The following describes several different implementations of how processor 100 determines at least one target agent that matches the first query domain name.

[0069] In one implementation, the processor 100 can determine at least one registered domain name that matches the first queried domain name from a plurality of registered domain names. Each of the plurality of registered domain names corresponds one-to-one with a plurality of intelligent agents, and the aforementioned at least one registered domain name corresponds to at least one target intelligent agent.

[0070] The registered domain name includes multiple fields, which may include one or more of the following: function field, scenario field, organization field, and management domain name field. The meaning of each field in the registered domain name can be the same as the meaning of the corresponding field in the first queried domain name. For details, please refer to the relevant introduction in step S201 above, which will not be repeated here.

[0071] The processor 100 can compare the first query domain name with multiple registered domain names stored in the memory 201. Matching the first query domain name with a registered domain name can be either an exact match or a match between one or more fields in the first query domain name and corresponding fields in the registered domain name. The specific matching method can be determined, for example, based on the fields included in the first query domain name and the query method indicated by the first query request.

[0072] If the first query domain name is exactly the same as the first registered domain name among multiple registered domain names, the processor 100 can identify the agent corresponding to the first registered domain name as the target agent. In this case, the first query domain name can be the complete registered domain name corresponding to an agent that has already completed registration. The sender of the first query request can know the complete registered domain name of the target agent in advance and obtain the resource record corresponding to the target agent by querying the complete registered domain name.

[0073] If the first query domain name is not completely identical to any registered domain name, the processor 100 can parse one or more fields included in the first query domain name and match the parsed fields with corresponding fields in multiple registered domain names. The first query domain name may only include some information that the queried agent needs to satisfy, but not all information used to uniquely identify an agent. Accordingly, the processor 100 can determine one or more registered domain names whose fields successfully match as the registered domain names that match the first query domain name.

[0074] For example, the first query domain name includes a function field and a scenario field, but does not specify an organization field. The processor 100 can match the function field in the first query domain name with the function fields in multiple registered domain names, and match the scenario field in the first query domain name with the scenario fields in multiple registered domain names. When the function field of a registered domain name matches the function field in the first query domain name, and the scenario field of that registered domain name also matches the scenario field in the first query domain name, the processor 100 can identify that registered domain name as the registered domain name that matches the first query domain name.

[0075] Matching between fields can be based on identical characters, one field including another, or a pre-defined correspondence between the two fields. For example, if the function field in the first query domain is "audit" and the function field in a certain registered domain is "audit-compliance," the processor 100 can determine that these two function fields match. As another example, if the function field in the first query domain is "translation" and the function field in a certain registered domain is "translator," the processor 100 can determine that these two function fields match based on a pre-configured word correspondence.

[0076] For intelligent agents applicable to multiple scenarios, the scenario field in the registered domain name can include characters corresponding to multiple scenarios. For example, the scenario field in the registered domain name can be office-iot, indicating that the corresponding intelligent agent is applicable to both office scenarios and IoT scenarios. When the scenario field in the first query domain name is office or iot, the processor 100 can determine that the scenario field matches office-iot.

[0077] Processor 100 can determine whether a registered domain name matches a first query domain name based on the matching results of each field. For example, processor 100 can determine a registered domain name whose fields in the first query domain name all match successfully as a registered domain name that matches the first query domain name. Optionally, processor 100 can also calculate the degree of matching between the registered domain name and the first query domain name based on the matching priority or field weight corresponding to different fields, and determine the registered domain name whose matching degree meets preset conditions as a registered domain name that matches the first query domain name.

[0078] For example, the matching priority of a function field can be higher than that of a scenario field, and the matching priority of a scenario field can be higher than that of an organization field. The processor 100 can first determine the registered domain names that match the function field, and then determine the registered domain names that match the scenario field and the organization field from among the registered domain names that match the function field. For fields not included in the first query domain names, the processor 100 may choose not to match that field.

[0079] As another implementation, the processor 100 can determine at least one resource record from a plurality of resource records that matches the first query domain name, wherein the at least one resource record corresponds to at least one target agent.

[0080] The resource record can be a text (TXT) resource record. The TXT record can store information describing the intelligent agent. The information describing the intelligent agent can adopt a preset structured data format so that the processor 100 can extract various information from the resource record.

[0081] Optionally, the resource record may include one or more of the following: agent version information, scene information, capability information, running status information, communication information for accessing the agent, input data format, and output data format.

[0082] Version information can be used to indicate the version of the data structure used in the resource record, or to indicate the updated version of the agent description information.

[0083] Scenario information can be used to indicate the applicable business scenarios for intelligent agents.

[0084] Capability information can be used to describe the tasks that an intelligent agent can perform, the services it can provide, or the data processing capabilities it possesses.

[0085] Operational status information can be used to indicate whether an agent is currently online, under maintenance, offline, or in an abnormal state.

[0086] Communication information used to access an intelligent agent may include one or more of the following: a communication address, a communication port, and a communication protocol. The communication address may be a Uniform Resource Locator (URL), an Internet Protocol (IP) address, or any other address that can be used to access the intelligent agent.

[0087] The input data format can be used to indicate the type or structure of input data that an agent can receive.

[0088] The output data format can be used to indicate the type or structure of the output data that the agent can generate.

[0089] In addition to the above, resource records may also include one or more of the following: agent name, identifier, organization affiliation, tag information, functional description, update time, and other extended information; there are no restrictions on this. The information in a resource record can be stored according to preset field names and field types. For example, resource records can be encoded using JavaScript object notation (JSON), allowing different resource records to use the same or compatible data structure.

[0090] Processor 100 can extract functional fields, scenario fields, organization fields, or other fields from the first query domain name, and match the extracted fields with the corresponding information in multiple resource records. For example, processor 100 can match the functional fields in the first query domain name with the capability information in the resource records, match the scenario fields in the first query domain name with the scenario information in the resource records, and match the organization fields in the first query domain name with the organization to which the resource records belong.

[0091] The capability information in the resource record may include structured capability names or function descriptions in natural language. The processor 100 can perform character matching, keyword matching, or semantic matching between the function fields in the first query domain and the capability information in the resource record. Semantic matching can determine the degree of semantic relevance between the function fields and the capability information, and identify resource records whose semantic relevance meets preset conditions as matching the first query domain.

[0092] For example, the functional field in the first query domain name could be "audit," and the capability information in a resource record could include "cross-border transaction compliance audit." The processor 100 can determine that the resource record matches the first query domain name based on the semantic correspondence between "audit" and "cross-border transaction compliance audit." As another example, the functional field in the first query domain name could be "translator," and the capability information in a resource record could include "supports text translation between Chinese and English." The processor 100 can then determine that the resource record matches the first query domain name.

[0093] The processor 100 can also filter the matching results by combining the runtime status information in the resource record. For example, when the capability information and scenario information in the resource record match the first query domain name, but the runtime status information indicates that the corresponding agent is offline or in an abnormal state, the processor 100 may not identify the agent corresponding to that resource record as the target agent. Optionally, the processor 100 may also retain the agent and feed back the corresponding runtime status information to the sender of the first query request in subsequent steps.

[0094] If multiple resource records are found to match the first query domain name, the processor 100 can identify the multiple agents corresponding to each resource record as multiple target agents. The processor 100 can also sort the multiple resource records according to the degree of matching between each resource record and the first query domain name. The degree of matching can be determined based on one or more of the following: functional matching degree, scenario matching degree, organizational matching degree, and tag matching degree.

[0095] It should be noted that matching by registered domain name and matching by resource record can be performed separately or in combination. For example, processor 100 can first match the first query domain name with multiple registered domain names. When a registered domain name exists that is exactly the same as the first query domain name, processor 100 can directly identify the agent corresponding to that registered domain name as the target agent. When no registered domain name exists that is exactly the same as the first query domain name, processor 100 can further extract multiple fields from the first query domain name and perform matching on multiple resource records based on these fields.

[0096] Processor 100 can also first filter out a portion of registered domain names based on the fields in the first query domain name, and then match the resource records corresponding to those registered domain names. For example, processor 100 can filter out multiple registered domain names belonging to a specific organization and managed by a specific domain name service based on the organization field and management domain name field in the first query domain name, and then perform semantic matching on the resource records corresponding to the multiple filtered registered domain names based on the function field and scenario field in the first query domain name.

[0097] In addition, there may be other ways to determine at least one target agent, and no limitation is imposed on this.

[0098] Additionally, the memory 201 can store the direct correspondence between agents, registered domain names, and resource records, or it can store a first correspondence between agents and registered domain names, and a second correspondence between registered domain names and resource records, respectively. The processor 100 can determine the agent corresponding to the resource record based on the first and second correspondences. Optionally, the resource record can also directly include the identifier of the corresponding agent, thereby enabling the processor 100 to determine the corresponding agent based on the identifier included in the resource record.

[0099] The registration process of the first intelligent agent will be introduced below, taking the first intelligent agent as an example.

[0100] Before receiving the first DNS query request, the processor 100 may receive a first registration request. The first registration request includes a first registration domain name and information describing the first intelligent agent. The first registration request may be sent by a user terminal device used by the developer of the first intelligent agent, or by an intelligent agent platform, server, or other device used to manage the first intelligent agent.

[0101] The first registered domain name can be generated by the developer of the first intelligent agent based on the function, applicable scenario, affiliated organization, and domain name service managing the first intelligent agent. The first registered domain name can include one or more of the following fields: function field, scenario field, organization field, and management domain name field. Multiple fields in the first registered domain name can be arranged in a preset order and separated by periods.

[0102] Information describing the first intelligent agent may include one or more of the following: version information, scenario information, capability information, operational status information, communication information for accessing the first intelligent agent, input data format, and output data format. Information describing the first intelligent agent may also include its name, organization, tag information, update time, or other extended information. This information may be directly carried in the first registration request or obtained through file addresses or data addresses included in the first registration request.

[0103] After receiving the first registration request, the processor 100 can generate a first resource record based on the information used to describe the first intelligent agent, and establish a correspondence between the first registered domain name and the first resource record. The first resource record includes information used to describe the first intelligent agent. Specifically, the first resource record may be a TXT record corresponding to the first registered domain name. The processor 100 can also establish a correspondence between the first intelligent agent, the first registered domain name, and the first resource record, and store this correspondence in the memory 201.

[0104] Before establishing the correspondence between the first registered domain name and the first resource record, the processor 100 can also determine whether the first registered domain name conforms to the domain name naming rules. The domain name naming rules can specify one or more of the following: the fields included in the first registered domain name, the order of the fields, the characters allowed in each field, the length of each field, and the value range of the domain name fields.

[0105] For example, a function field can consist of one or more of lowercase English letters, numbers, and hyphens. The length of a function field can be within a first preset length range. A scenario field can consist of characters representing one or more application scenarios. An organization field can be the standardized name or abbreviation of the organization to which the first intelligent agent belongs. A management domain field can be a domain name that the data processing system 10 has management authority over.

[0106] Optionally, the processor 100 can also determine whether the first registered domain name is a duplicate of an existing registered domain name. Specifically, the processor 100 can compare the first registered domain name with multiple registered domain names stored in the memory 201. If the first registered domain name is different from any existing registered domain name, the processor 100 can determine that the first registered domain name does not conflict. If the first registered domain name is the same as an existing registered domain name, the processor 100 may not establish a correspondence between the first registered domain name and the first resource record, or may send information indicating a domain name conflict to the sender of the first registration request.

[0107] Optionally, the processor 100 can also determine whether the information used to describe the first intelligent agent conforms to a preset data format. For example, the processor 100 can determine whether the information used to describe the first intelligent agent includes required fields, whether the data types of each field are correct, whether the communication information conforms to a preset address format, and whether the version information is a version supported by the processor 100. The processor 100 can also determine whether the capability information, scenario information, and affiliated organization are consistent with the corresponding fields in the first registered domain name.

[0108] Furthermore, if the first registered domain name conforms to the domain name naming rules, the processor 100 can generate a first resource record and establish a correspondence between the first registered domain name and the first resource record. Then, the processor 100 adds the first registered domain name to the aforementioned plurality of registered domain names and adds the first resource record to the aforementioned plurality of resource records, so that when the processor 100 receives a subsequent query request, it can match the first intelligent agent based on the first registered domain name or the first resource record.

[0109] When the capabilities, applicable scenarios, operating status, communication information, input data format, or output data format of the first intelligent agent change, the processor 100 can also receive update requests for the first intelligent agent. The update request may include updated information describing the first intelligent agent. The processor 100 can modify the first resource record based on the updated information and update the version information or update time in the first resource record. If the function field, scenario field, or organization field in the first registered domain name changes, the processor 100 can also modify the first registered domain name based on the update request and update the corresponding relationship stored in the memory 201 accordingly.

[0110] Through the above registration process, the device containing processor 100 can pre-obtain the correspondence between multiple intelligent agents, multiple registered domain names, and multiple resource records. Upon receiving the first query request, processor 100 can directly match the registered domain name based on the first query domain name, or match the resource record based on the multiple fields included in the first query domain name, thereby determining at least one target intelligent agent that matches the first query domain name.

[0111] S203: Processor 100 feeds back the parsing results, which include resource records of at least one target agent.

[0112] In this embodiment, after determining at least one target agent matching the first query domain name, the processor 100 can obtain resource records corresponding to each of the at least one target agent and generate a parsing result based on the obtained resource records. The processor 100 can feed back the parsing result to the sender of the first query request, so that the sender of the first query request can obtain relevant information about at least one target agent based on the parsing result, or further determine the agent for performing the target task from the at least one target agent.

[0113] The resolution result can be in the form of a DNS resolution response message. The processor 100 can write the query identifier from the first query request into the resolution response message to make the resolution result correspond to the first query request. The resolution result may also include the first query domain name, query type, response status, number of resource records, and at least one resource record of the target agent. When the query type of the first query request is TXT, the resource records included in the resolution result can be TXT records.

[0114] When the first query request is sent by a recursive DNS server, the processor 100 can send the resolution result back to that recursive DNS server. After receiving the resolution result, the recursive DNS server can send it back to the client device that initiated the query. The recursive DNS server can also cache resource records based on the cache validity period of the resource records in the resolution result. When a subsequent query request for the same domain name from the first query is received, and the cached resource record is still valid, the recursive DNS server can directly send back the cached resource record without sending another query request to the processor 100.

[0115] When the first query request is sent directly by the user terminal device, the processor 100 can directly feed back the parsing result to the user terminal device. For example, the first query request can be sent by a second intelligent agent on the user terminal device. After receiving the parsing result, the second intelligent agent can parse the resource records of at least one target intelligent agent, and select one or more target intelligent agents to establish a communication connection based on the capability information, operating status information, and communication information of each target intelligent agent.

[0116] When the number of at least one target agent is one, the processor 100 can directly include the resource record of that target agent in the parsing result. When the number of at least one target agent is greater than one, the processor 100 can sort the resource records of multiple target agents and include the resource records of multiple target agents in the parsing result according to the sorting result. By sorting multiple resource records, the sender of the first query request can obtain information of target agents that best match the query requirements.

[0117] Optionally, in the parsing results, the resource records of at least one target agent can be sorted based on one or more of the following: the degree of matching between at least one target agent and the first query domain name, the running status of at least one target agent, and the response speed of at least one target agent.

[0118] Specifically, the degree of matching between the target agent and the first query domain name can be determined based on the matching results between multiple fields in the first query domain name and the registered domain name or resource record corresponding to the target agent. For example, the processor 100 can determine the degree of matching between the target agent and the first query domain name based on one or more of the following: the degree of matching between the function field and capability information, the degree of matching between the scenario field and scenario information, and the degree of matching between the organization field and the organization to which it belongs. The higher the degree of matching between the target agent and the first query domain name, the higher the ranking of the resource record of the target agent can be in the parsing results.

[0119] The operational status of a target intelligent agent can include online status, maintenance status, offline status, or abnormal status. For example, the sorting priority corresponding to the online status can be higher than that corresponding to the maintenance status, and the sorting priority corresponding to the maintenance status can be higher than that corresponding to the offline or abnormal status. The processor 100 can determine the operational status of each target intelligent agent based on the operational status information in the resource records and prioritize the resource records of target intelligent agents that are currently able to provide services normally.

[0120] The response speed of the target agent can be determined based on its historical communication or access data. For example, the processor 100 can calculate the time interval between sending an access request to the target agent and receiving feedback from the target agent, and determine the average response time of the target agent based on multiple time intervals. The shorter the response time, the faster the target agent's response speed. The processor 100 can then prioritize the resource records of target agents with faster response speeds over those with slower response speeds.

[0121] In one implementation, the processor 100 can sort multiple target agents sequentially according to their matching degree, running state, and response speed. For example, the processor 100 can first sort multiple target agents from high to low matching degree; for multiple target agents with the same matching degree, sort them according to the priority corresponding to their running state; for multiple target agents with the same matching degree and running state, sort them from fast to slow response speed.

[0122] As another implementation, the processor 100 can set corresponding weights for matching degree, running status, and response speed, and determine the ranking parameters for each target agent based on the matching degree, running status, response speed, and their corresponding weights. The processor 100 can arrange the resource records of multiple target agents in descending order of the ranking parameters. The weights corresponding to different ranking indicators can be pre-configured or adjusted according to the application scenario corresponding to the first query request.

[0123] The processor 100 can include the resource records of all sorted target agents in the parsing result, or it can select a preset number of target agents from the sorted target agents and include only the resource records of the selected target agents in the parsing result. For example, the processor 100 can return the resource record of the first-ranked target agent, or it can return the resource records of multiple first-ranked target agents.

[0124] In this embodiment, the first query domain name can, for example, carry information about the target agent in different dimensions such as function, scenario, organization, and management domain name through multiple fields. This allows the processor 100 to directly match agents based on the first query domain name without relying on the private identifiers or dedicated interfaces of each agent platform, thus achieving a unified query entry point and significantly reducing cross-platform adaptation costs. Simultaneously, applying the agent resolution process to the Domain Name System (DNS) fully utilizes the existing hierarchical structure and distributed resolution mechanism of DNS, enabling query requests to be automatically routed to the corresponding server according to the existing recursive resolution process. This improves the standardization of the resolution process and system compatibility, and eliminates the need for additional dedicated communication protocols or resolution mechanisms, thereby simplifying the system architecture. Furthermore, by providing at least one resource record of the target agent through DNS feedback, the communication overhead and processing delay caused by multiple calls to different platform interfaces can be avoided, reducing redundant matching and invalid queries, and improving overall resolution efficiency. Moreover, since the first query domain name contains information about the target agent in different dimensions, the returned resolution results are more consistent with the query requirements, further improving the accuracy of agent matching.

[0125] Furthermore, embodiments of this application also provide an intelligent agent analysis device. See also... Figure 3 , Figure 3 This paper shows a schematic diagram of the structure of an intelligent agent analysis device according to an embodiment of the present application. Figure 3 The intelligent agent analysis device 300 shown includes:

[0126] The communication module 301 is used to receive a first query request for the domain name system. The first query request includes a first query domain name, which includes multiple fields, each of which is used to describe information about different dimensions of the intelligent agent to be queried.

[0127] The determining module 302 is used to determine at least one target agent that matches the first query domain name in response to the first query request;

[0128] The communication module 301 is also used to provide feedback on the parsing results, which include resource records of at least one target agent.

[0129] In one possible implementation, the determining module 302 is specifically used to determine at least one registered domain name that matches the first query domain name from a plurality of registered domain names, wherein the plurality of registered domain names correspond one-to-one with a plurality of smart agents, and at least one registered domain name corresponds to at least one target smart agent, and the registered domain name includes a plurality of fields; or, specifically used to determine at least one resource record that matches the first query domain name from a plurality of resource records, wherein at least one resource record corresponds to at least one target smart agent.

[0130] In one possible implementation, the communication module 301 is further configured to receive a first registration request before receiving a first query request for the Domain Name System, the first registration request including a first registered domain name and information for describing a first intelligent agent; the apparatus further includes an establishment module 303 configured to establish a correspondence between the first registered domain name and a first resource record, the first resource record including information for describing the first intelligent agent.

[0131] In one possible implementation, the multiple fields include one or more of a function field, a scenario field, an organization field, and a management domain field; the function field is used to indicate the functions provided by the agent, the scenario field is used to indicate the scenarios to which the agent is applicable, the organization field is used to indicate the organization to which the agent belongs, and the management domain field is used to indicate the domain name service for managing the agent.

[0132] In one possible implementation, the determining module 302 is further configured to determine whether the first registered domain name conforms to the domain name naming rules before establishing the correspondence between the first registered domain name and the first resource record; the establishing module 303 is specifically configured to establish the correspondence between the first registered domain name and the first resource record when the first registered domain name conforms to the domain name naming rules.

[0133] In one possible implementation, the number of at least one target agent is greater than 1, and in the parsing result, the resource records of at least one target agent are sorted based on one or more of the following: the degree of matching between at least one target agent and the first query domain name, the running status of at least one target agent, and the response speed of at least one target agent.

[0134] In one possible implementation, the resource record includes one or more of the following: agent version information, scene information, capability information, running status information, communication information for accessing the agent, input data format, and output data format.

[0135] In one possible implementation, the first query request is sent by a recursive domain name system server, or the first query request is sent by a user terminal device.

[0136] It should be noted that the information interaction and execution process between the modules and units of the above-mentioned device are based on the same concept as the method embodiment in this application, and the resulting technical effects are the same as those in the method embodiment in this application. For details, please refer to the description in the method embodiment shown above in this application, and it will not be repeated here.

[0137] Furthermore, embodiments of this application also provide a computing device. See also... Figure 4 , Figure 4 A schematic diagram of the hardware structure of a computing device according to an embodiment of this application is shown. Figure 4 As shown, computing device 400 may include processor 401 and memory 402.

[0138] The memory 402 is used to store computer programs;

[0139] The processor 401 is configured to execute the agent parsing method described in the above method embodiments according to the computer program.

[0140] In addition, this application embodiment also provides a computer-readable storage medium for storing a computer program for executing the agent parsing method described in the above method embodiment.

[0141] In addition, this application also provides a computer program product containing instructions that, when run on a computing device, causes the computing device to execute the intelligent agent parsing method described in the above method embodiments.

[0142] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that all or part of the steps in the methods of the above embodiments can be implemented by means of software plus a general-purpose hardware platform. Based on this understanding, the technical solution of this application can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as a read-only memory (ROM) / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network communication device such as a router) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0143] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The device embodiments described above are merely illustrative. Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0144] The above description is merely an exemplary implementation of this application and is not intended to limit the scope of protection of this application.

Claims

1. A method for parsing intelligent agents, characterized in that, The method includes: Receive a first query request for the Domain Name System, the first query request includes a first query domain name, the first query domain name includes multiple fields, the multiple fields are used to describe information of different dimensions of the intelligent agent to be queried; In response to the first query request, at least one target agent that matches the first query domain name is identified; Feedback analysis results, which include resource records of the at least one target agent.

2. The method according to claim 1, characterized in that, The determination of at least one target agent matching the first query domain name includes: Determine at least one registered domain name from a plurality of registered domain names that matches the first queried domain name, wherein the plurality of registered domain names correspond one-to-one with a plurality of intelligent agents, the at least one registered domain name corresponds to the at least one target intelligent agent, and the registered domain name includes the plurality of fields; or, At least one resource record is determined from a plurality of resource records that matches the first queried domain name, the at least one resource record corresponding to the at least one target agent.

3. The method according to claim 2, characterized in that, Before receiving the first query request for the Domain Name System, the method further includes: Receive a first registration request, the first registration request including a first registration domain name and information for describing a first intelligent agent; Establish a correspondence between the first registered domain name and the first resource record, wherein the first resource record includes the information used to describe the first intelligent agent.

4. The method according to any one of claims 1-3, characterized in that, The multiple fields include one or more of the following: functional fields, scenario fields, organization fields, and management domain field; The function field indicates the function provided by the agent, the scenario field indicates the scenario to which the agent is applicable, the organization field indicates the organization to which the agent belongs, and the management domain field indicates the domain name service for managing the agent.

5. The method according to claim 3, characterized in that, Before establishing the correspondence between the first registered domain name and the first resource record, the method further includes: Determine whether the first registered domain name conforms to the domain name naming rules; Establishing the correspondence between the first registered domain name and the first resource record includes: If the first registered domain name conforms to the domain name naming rules, a correspondence is established between the first registered domain name and the first resource record.

6. The method according to claim 1, characterized in that, The number of the at least one target agent is greater than 1. In the parsing result, the resource records of the at least one target agent are sorted based on one or more of the following: the degree of matching between the at least one target agent and the first query domain name, the running status of the at least one target agent, and the response speed of the at least one target agent.

7. The method according to claim 1, characterized in that, The resource record includes one or more of the following: agent version information, scene information, capability information, running status information, communication information for accessing the agent, input data format, and output data format.

8. The method according to claim 1, characterized in that, The first query request is sent by a recursive domain name system server, or the first query request is sent by a user terminal device.

9. An intelligent agent analysis device, characterized in that, The device includes: The communication module is used to receive a first query request for the Domain Name System, the first query request including a first query domain name, the first query domain name including multiple fields, the multiple fields being used to describe information of different dimensions of the intelligent agent to be queried; A determination module, configured to determine at least one target agent matching the first query domain name in response to the first query request; The communication module is also used to provide feedback on the parsing results, which include the resource records of the at least one target agent.

10. A computer program product containing instructions, characterized in that, When it is run on a computing device, it causes the computing device to perform the method as described in any one of claims 1 to 8.

Images