Intelligent agent processing method and device, equipment, medium and product

Abstract

The invention discloses an intelligent agent processing method, device and equipment, a medium and a product. The method comprises the following steps: in response to an event of releasing an agent, determining an agent deployment mode associated with the agent, and obtaining associated data of the agent; the associated data at least comprises operation environment information and tool configuration information of the intelligent agent; and obtaining an agent deployment constraint condition of the target service system, and packaging the associated data based on the agent deployment mode and the agent deployment constraint condition to obtain an application installation package corresponding to the agent, so as to deploy the application installation package to the target service system. According to the technical scheme provided by the invention, the resource consumption and the cost are reduced, the adaptability between the intelligent agent and the target service system is improved, and the effect of ensuring the stability and the safety of the operation of the intelligent agent in the target service system is achieved.

Description

Technical Field

[0001] This invention relates to the field of computer processing technology, and in particular to a method, apparatus, device, medium and product for processing intelligent agents. Background Technology

[0002] To ensure the security and stability of intelligent agent applications, the agents are typically designed and debugged in the development environment before being migrated to the production environment for official deployment. Before running the intelligent agent in the production environment, it needs to be deployed.

[0003] Currently, the typical way to deploy intelligent agents is to pre-deploy the same intelligent agent development platform in the production environment as in the development environment; then, after designing and debugging the intelligent agent on the intelligent agent development platform in the development environment, the static definition file of the intelligent agent is exported and migrated to another intelligent agent development platform in the production environment.

[0004] However, this method of repeatedly building complete intelligent agent development platforms not only wastes resources and increases maintenance costs, but also easily leads to the models or plugins that intelligent agent applications rely on being incompatible with the production environment, resulting in poor stability of intelligent agents in the production environment, or even problems that prevent them from running. Summary of the Invention

[0005] This invention provides a method, apparatus, device, medium, and product for processing intelligent agents, thereby reducing resource consumption and costs, improving the adaptability between intelligent agents and target business systems, ensuring the stability and security of intelligent agents running in target business systems, and meeting the security and compliance requirements for isolating development and production environments.

[0006] According to one aspect of the present invention, an intelligent agent processing method is provided, the method comprising:

[0007] In response to an event that releases an agent, the system determines the agent deployment method associated with the agent and obtains the associated data of the agent; wherein the associated data includes at least the agent's operating environment information and tool configuration information.

[0008] Obtain the agent deployment constraints of the target business system. Based on the agent deployment method and the agent deployment constraints, encapsulate the associated data to obtain the application installation package corresponding to the agent, and deploy the application installation package to the target business system.

[0009] According to another aspect of the present invention, an intelligent agent processing apparatus is provided, the apparatus comprising:

[0010] The data acquisition module is used to respond to an event that releases an intelligent agent, determine the deployment method of the intelligent agent associated with the intelligent agent, and acquire the associated data of the intelligent agent; wherein, the associated data includes at least the intelligent agent's operating environment information and tool configuration information;

[0011] The data processing module is used to obtain the agent deployment constraints of the target business system, and based on the agent deployment method and the agent deployment constraints, encapsulate the associated data to obtain the application installation package corresponding to the agent, so as to deploy the application installation package to the target business system.

[0012] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0013] At least one processor; and a memory communicatively connected to said at least one processor; wherein,

[0014] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the intelligent agent processing method according to any embodiment of the present invention.

[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the intelligent agent processing method according to any embodiment of the present invention.

[0016] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the intelligent agent processing method as described in any embodiment of the present invention.

[0017] The technical solution of this invention, in response to an event that publishes an intelligent agent, determines the intelligent agent deployment method associated with the intelligent agent and obtains the associated data of the intelligent agent. The associated data includes at least the intelligent agent's operating environment information and tool configuration information. The intelligent agent deployment constraints of the target business system are obtained. Based on the intelligent agent deployment method and intelligent agent deployment constraints, the associated data is encapsulated to obtain an application installation package corresponding to the intelligent agent, so as to deploy the application installation package to the target business system. This solves the problems of repeated environment construction, resource waste, high deployment costs and poor intelligent agent operation stability caused by strong dependence on the development platform runtime environment in the prior art. It realizes the dynamic determination of the intelligent agent deployment method matching the intelligent agent in response to an event that publishes an intelligent agent, and obtains associated data including at least the intelligent agent's operating environment information and tool configuration information. By using the agent deployment constraints and methods of the target business system, all related data are encapsulated and processed to generate an application installation package that is highly compatible with the target business system, self-contained, and compliant. This improves the compatibility between the agent and the target business system, while enabling the independent deployment and operation of the agent by simply deploying the application installation package on the target business system. This reduces resource consumption and costs, and ensures the stability and security of the agent application in the target business system.

[0018] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, 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 of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of an intelligent agent processing method provided according to an embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of association data for characterizing an intelligent agent provided according to an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of a process for characterizing the switching of the current version of an intelligent agent, provided according to an embodiment of the present invention;

[0023] Figure 4 This is a flowchart illustrating the intelligent agent processing method provided according to an embodiment of the present invention;

[0024] Figure 5 This is a flowchart of an intelligent agent processing method provided according to an embodiment of the present invention;

[0025] Figure 6 This is a schematic diagram for representing an information publishing page according to an embodiment of the present invention;

[0026] Figure 7 This is a flowchart of an intelligent agent processing method provided according to an embodiment of the present invention;

[0027] Figure 8 This is a schematic diagram of the structure of an intelligent agent processing device according to an embodiment of the present invention;

[0028] Figure 9 This is a schematic diagram of the structure of an electronic device that implements the intelligent agent processing method of the present invention. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0030] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] It should be noted that the collection, gathering, updating, analysis, processing, use, transmission, and storage of user personal information involved in the technical solution disclosed herein all comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken to prevent unauthorized access to user personal information data and to maintain user personal information security and network security. It should also be noted that the collection, gathering, updating, analysis, processing, use, transmission, and storage of user personal information involved in the technical solution disclosed herein are all conducted with the user's knowledge and consent, and comply with relevant privacy protection regulations.

[0032] Before introducing this technical solution, we can first describe the application scenario. The technical solution provided in this embodiment can be applied to any business scenario that requires the deployment of an intelligent agent. Here, an intelligent agent refers to a combination of hardware and software or a software module that can autonomously perceive environmental information, process information according to preset rules or autonomous decision-making logic, and execute specific actions to achieve a goal.

[0033] To meet diverse business needs, developers typically design and implement various types of intelligent agents, such as, but not limited to, chat assistants, workflows, and dialogue streams. After development, these agents need to be further integrated and deployed into actual business systems to provide intelligent service capabilities.

[0034] For security and stability reasons, enterprises or organizations typically strictly differentiate and isolate system environments at different stages, including but not limited to development, testing, and production environments. Currently, the development of agent applications generally relies on dedicated agent development platforms, which are usually deployed only in the development environment. This means that developers can only design, debug, and run agents within the platform of the current development environment; if it is necessary to deploy agents to other environments (such as production environments), it is often necessary to redeploy a complete agent development platform in the target environment, and then upload the agent DSL (Domain Specific Language) files exported from the development environment to the platform for release and operation.

[0035] For example, after an agent is designed and debugged on an agent development platform in the development environment, the current practice for deploying it to the production environment is usually to deploy an identical agent development platform in the production environment and import the agent as a DSL file. However, this approach of repeatedly building an agent development platform results in resource waste and high deployment costs. Furthermore, the agent in the production environment may experience functional abnormalities, performance degradation, or even complete failure due to inconsistencies in dependencies (such as large language model versions, plugin interfaces, database connections, network policies, etc.) compared to the development environment. When adapting or upgrading the platform in the production environment, downtime is often required, causing agent service interruptions and impacting upstream and downstream business systems that rely on the agent, thus posing risks.

[0036] To address the aforementioned issues, this invention provides a method for processing intelligent agents: by packaging the intelligent agent application into a lightweight installation package that can be installed and run independently, it eliminates the strong dependence on the original intelligent agent development platform. This installation package embeds all the runtime environment required by the intelligent agent, including behavioral logic, tool call interfaces, memory management modules, model adaptation layers, and necessary dependency configurations, enabling it to self-containedly start and execute in the target environment.

[0037] Furthermore, once the agent application is designed in the development environment, it can be exported as a standardized installation package with a single click and directly deployed to a production environment that complies with security standards. This eliminates the need to repeatedly deploy the entire development platform, achieving lightweight delivery. This not only reduces deployment costs and resource consumption but also improves deployment efficiency and the stability of agent operation. Alternatively, the agent installation package can be stored in a preset location, allowing the production environment server to install the agent package from the preset location as needed.

[0038] Figure 1 This is a flowchart of an intelligent agent processing method according to an embodiment of the present invention. This embodiment is applicable to any situation requiring the deployment of intelligent agents. The method can be executed by an intelligent agent processing device, which can be implemented in hardware and / or software and can be configured in a computing device. Figure 1 As shown, the method includes:

[0039] S110. In response to the event of publishing an agent, determine the agent deployment method associated with the agent and obtain the associated data of the agent.

[0040] The event of deploying an agent refers to the operational instruction that triggers the agent to move from the development or testing phase to the deployment phase. The agent deployment method refers to the specific form in which the agent is deployed to the target business system environment, including but not limited to: Docker image deployment, plug-in modular integration, Helm Chart package deployment, etc. A Docker image is a read-only template used to create Docker containers, containing all the file systems, dependencies, environment variables, and configurations required to run the agent application. A Helm Chart package is a standard packaging format on the Kubernetes platform used to define, install, and upgrade complex applications (including agents), consisting of a set of parameterized templates and metadata. Associated data refers to the complete set of information bound to the agent to support its correct operation, including at least runtime environment information and tool configuration information. Tool configuration information refers to the specific configuration of external capabilities (such as large language models, plugins, APIs, services, etc.) that the agent can call, including but not limited to tool name, calling interface address (such as API endpoint or function entry), parameter format, permission scope, timeout settings, Reis connection, MySQL connection, usage, etc.

[0041] Runtime environment information refers to all the hardware and software information required to ensure the intelligent agent operates as expected in the target business system environment; that is, the runtime environment on which the intelligent agent depends. The runtime environment refers to the entire dynamic execution phase of an intelligent agent application, from "being started or loaded" to "completing the task or being terminated," and the underlying environment supporting the normal operation of this phase. This includes, but is not limited to: operating system, network access interface, the intelligent agent program itself, algorithm model, functional components, knowledge base, hardware resources, scheduling module, data interaction interface, error handling mechanism, I / O interface, thread scheduling module, resource management module (such as CPU and memory limits), human-computer interaction components, DSL Handler (Domain-Specific Language Handler, used to parse, verify, convert, or execute intelligent agent definition information written in DSL form), LLM / Plugin Handler (Large Language Model / Plugin Handler), and various other data. Large Language Model and Plugin Processor: This is used to call the Large Language Model (LLM) for reasoning, generation or decision-making, as well as to manage and execute external tools (i.e. "plugins"), such as API calls, database queries, computation functions, etc., and to coordinate the data flow between the LLM and plugins (e.g., using LLM output as plugin input, or feeding plugin results back to the LLM for the next step of reasoning).

[0042] In this embodiment, the event of releasing an agent can be detected in real time or periodically. Optionally, the event of releasing an agent includes, but is not limited to, at least one of the following: a user clicks the release button for the agent; the automated build / test / release pipeline completes and generates a deliverable agent artifact; the development status of the agent is updated from "under development" to "development complete"; a deployment instruction for the agent is received from an external system; the agent version number changes and meets the preset release strategy; a scheduled release task is triggered; all test cases associated with the agent pass verification; or the agent definition information is marked as allowed to be released after an approval process.

[0043] When an event triggering the deployment of an agent is detected, the system responds to the event and resolves the agent's associated deployment method. One approach is to query the agent's metadata, retrieve predefined deployment tags (such as "containerized deployment" or "embedded integration"), and determine the deployment method based on these tags. Another approach is to dynamically analyze the agent's characteristics (such as GPU dependency or low-latency response requirements) and the target business system's capabilities (such as whether the production environment supports container orchestration, has a specific security gateway, or a Kubernetes cluster) upon event triggering, and automatically select the most suitable deployment method based on the joint analysis results. For example, if the target business system runs in a cloud-native environment, a microservice container mode is preferred; if its infrastructure is based on Kubernetes, the Helm Chart package can be used as the deployment method.

[0044] Furthermore, the associated data for the agent can be retrieved from the agent configuration library, which includes structured runtime environment information and tool configuration information. For example, see [link to example]. Figure 2 The associated data includes API interface addresses, LLM configurations, plugin configurations, DSL Handlers (LM / Plugin Handlers), Reis connections, MySQL connections, the UI (i.e., the human-computer interaction runtime component), and Agent Runtime (the extracted intelligent agent runtime environment). Based on this associated data and the determined intelligent agent deployment method, the corresponding application installation package is generated.

[0045] S120. Obtain the agent deployment constraints of the target business system. Based on the agent deployment method and agent deployment constraints, encapsulate the associated data to obtain the application installation package corresponding to the agent, and deploy the application installation package to the target business system.

[0046] The target business system refers to the specific production or operating environment where intelligent agents need to be deployed, such as, but not limited to: customer service systems, data analysis platforms, automated operation and maintenance platforms, intelligent marketing engines, risk control systems, human resource management systems, supply chain collaboration platforms, medical auxiliary diagnostic systems, financial transaction monitoring systems, content moderation platforms, intelligent customer service assistants, enterprise knowledge base Q&A systems, IoT device management platforms, office automation tools, etc. Intelligent agent deployment constraints refer to the various restrictive requirements imposed by the target business system on intelligent agent deployment. The application installation package refers to the intelligent agent software package generated after packaging, which can be independently distributed and deployed. Its content varies depending on the functions, dependencies, and runtime environments of different intelligent agents; the installation package exported by each intelligent agent may differ in structure and configuration.

[0047] In this embodiment, a query can be proactively initiated to the configuration management interface of the target business system in advance or in real time to obtain the intelligent agent deployment constraints defined therein; alternatively, intelligent agent deployment constraints can be passively received from the target business system. Optionally, intelligent agent deployment constraints include, but are not limited to: security constraints (such as prohibiting external network access), resource limitation constraints (such as maximum memory usage, CPU quota, and maximum concurrent connection count), compliance constraints (such as data localization and privacy protection requirements), dependency version compatibility constraints (such as only supporting specific large model versions, specifying standard indicator interfaces, specifying Python or Node.js runtime versions, and requiring exposure of standard indicator interfaces), callable tool constraints (such as only allowing the use of specific large model names, and only supporting registered internal plugins or interfaces), network isolation constraints (such as only allowing access to specific subnets or service endpoints), authentication and authorization constraints (such as forcing the use of unified identity authentication login), storage access permission constraints (such as prohibiting writing to local disks), log and monitoring specification constraints (such as requiring access to a unified log platform), high availability constraints (such as requiring support for multi-replica deployment), and container runtime limitation constraints (such as prohibiting privileged mode and requiring operation as a non-root user).

[0048] After obtaining the agent deployment constraints of the target business system, the associated data of the agents can be adapted based on these constraints. For example, if the target business system prohibits calls to external large model services, the model call endpoints in the associated data are automatically replaced with internal private model service addresses; if the target business system requires all external communication to go through an internal gateway, the API routes in the associated data are rewritten to point to compliant proxy entry points. After completing the above adaptation, the adjusted associated data can be packaged according to the packaging logic corresponding to the preset agent deployment method (such as Docker containerized deployment, Helm Chart deployment, etc.) to generate an application installation package that conforms to the target business system specifications. This application installation package embeds all the runtime environment, dependency configurations, and adapted tool and model call information required by the agent application. It has the capabilities of self-containment, self-adaptation, and self-running, and can be directly installed and started in the target business system with one click, without manual intervention, secondary debugging, or reliance on the original agent development platform, thereby achieving secure, efficient, and compliant agent delivery and deployment.

[0049] To ensure compatibility between the agent application installation package and the target business system, the process of encapsulating the agent's associated data based on the agent deployment method and constraints to obtain the application installation package corresponding to the agent can be achieved by adjusting the tool configuration information in the associated data based on the agent deployment constraints to obtain the adjusted associated data. Finally, the adjusted associated data can be packaged based on the agent deployment method to obtain the application installation package corresponding to the agent.

[0050] The adjusted associated data refers to the adapted data obtained by modifying, filtering, or redirecting the tool configuration information based on the original associated data and the deployment constraints of the target business system.

[0051] In practice, the tool configuration information in the associated data can be compared item by item with the agent deployment constraints using a deployment engine or configuration adaptation algorithm to identify any configuration items that do not comply with the agent deployment constraints. If such configuration items exist, they are automatically adjusted according to the agent deployment constraints: for example, if a tool in a configuration item needs to call an external public network API, but the target business system prohibits external network access, the tool will be automatically disabled or replaced with an internal service interface with equivalent functionality; if a tool in a configuration item requires a high level of privilege, but the target business system follows the principle of least privilege, its access scope can be automatically downgraded or sensitive operation permissions can be removed. After adjusting all configuration items that do not comply with the agent deployment constraints, the adjusted associated data will include the adapted tool configuration information and the unaffected runtime environment information.

[0052] If the default agent deployment method is Docker containerization, the adjusted association data can be packaged to generate a standard container image application installation package, ensuring that it can be directly deployed and run on the target business system's container platform. If the default agent deployment method is Helm Chart package deployment, the adjusted association data can be embedded into the Helm Chart template and values ​​file to generate a Helm Chart application installation package that conforms to the Kubernetes specification. This allows it to be deployed to the target business system's Kubernetes cluster with one click via Helm commands, and automatically applies the adapted tool configuration information.

[0053] It's important to note that adjustments to tool configuration information can be performed not only during the deployment phase but also upfront during agent development or release preparation. For example, after identifying the target business system and its deployment constraints, constraints matching the target business system environment can be activated in advance. This involves replacing generic placeholders in the original tool configuration information (such as {{search_api}}) with specific endpoints (such as internal search engine API addresses) or authentication methods that conform to the target business system. In this way, when responding to subsequent agent release events, the acquired related data will already be the version adapted to the target business system environment, eliminating the need for conversion during deployment.

[0054] The technical solution provided in this embodiment first precisely and automatically adjusts the tool configuration information according to the intelligent agent deployment constraints, and then combines the adjusted associated data with the selected intelligent agent deployment method to perform structured packaging. This not only ensures the functional integrity and security of the intelligent agent in the target business system, but also improves the automation level of deployment and cross-environment adaptability. The resulting application installation package has high self-consistency, compliance, and portability, effectively avoiding operational anomalies caused by tool call violations, dependency mismatches, permission out-of-bounds, or network policy conflicts, thereby ensuring the stable, secure, and reliable operation of the intelligent agent in the target business system environment.

[0055] To achieve standardized distribution, secure management, and efficient deployment of intelligent agent applications, the application installation package can be stored in a shared library before being deployed to the target business system. This allows the target business system to retrieve the application installation package from the shared library as needed and run the intelligent agent based on the application installation package.

[0056] Among them, the shared library refers to a centralized storage and distribution platform used to uniformly manage the installation packages of multiple intelligent agent applications. It supports version control, access permission management, metadata indexing and on-demand downloading, and can be deployed in the enterprise's internal network or private cloud environment to ensure that data does not leave the domain and meets security and compliance requirements.

[0057] In practice, after generating the application installation package, it can be uploaded to a unified shared library along with corresponding metadata tags. These metadata tags include, but are not limited to, agent name, version number, applicable business scenarios, compatible deployment methods (such as Docker images or Helm Charts), security level, dependency model type, and review status. When the target business system needs to deploy an agent, it can initiate a query request to the shared library through its built-in installation agent module. Based on its own business needs, environmental capabilities, and security policies, it can automatically match and download the most suitable application installation package and complete the installation. In addition, an interactive interface can be provided, allowing the target business system's operations and maintenance personnel or administrators to browse the agent download list, select the required application installation package for download and installation, and achieve flexibility in agent application installation package installation.

[0058] After installation, the target business system can automatically start the intelligent agent instance based on the runtime environment and tool configuration information embedded in the application installation package, enabling it to execute preset tasks, call adapted tools, interact with users or other systems, and continuously provide intelligent services. The entire process requires no manual intervention, achieving end-to-end automation from acquisition and installation to operation.

[0059] The technical solution provided in this embodiment, by uniformly storing application installation packages in a shared library and allowing target business systems to obtain and run them on demand, not only avoids version confusion, security risks, and operational burdens caused by repeated construction, manual transfer, or local copying of installation packages by various business systems, but also achieves centralized governance of intelligent agent versions, fine-grained control of permissions, and consistency in deployment behavior. Therefore, enterprises can quickly, reliably, and traceably inject intelligent agent capabilities into various business scenarios while meeting security isolation, audit compliance, and multi-environment isolation requirements, thereby improving the delivery efficiency, operational stability, and overall operational quality of intelligent services.

[0060] Furthermore, to ensure the security and compliance of intelligent agent applications, access control policies can be registered simultaneously when the application installation package is stored in the shared library. For example, only specific business systems, designated security domains, or services with specific roles can be allowed to pull the installation package. Before downloading, the target business system must submit identity credentials and a declaration of deployment intent (such as target environment type, usage description, etc.) to the shared library. The shared library only authorizes the download of the application installation package after verifying its legitimacy and permissions. The advantage of this approach is that it can effectively prevent highly sensitive or high-privilege intelligent agents from being misused or deployed to unauthorized environments, ensuring that they only run in compliant and controlled systems. After the target business system completes local deployment based on the downloaded installation package, it can run the intelligent agent correctly and securely based on the associated data (such as runtime environment information and tool configuration information) encapsulated within it.

[0061] Typically, some production-level business services have high requirements for system stability, requiring continuous 24 / 7 (e.g., 7x24) uninterrupted service and reliable access to agent services. This necessitates that agent applications have high availability deployment capabilities, ensuring uninterrupted business services during agent upgrades.

[0062] To ensure business continuity for target systems, once a new application installation package is stored in the shared repository, it can be marked as available for testing and only accessible to select target systems. After verification and stability, it can be upgraded to production-ready status, making it available to all authorized target systems. Target systems can periodically poll the shared repository or subscribe to change notifications. Upon detecting a new version, the update process is automatically triggered, including downloading the new installation package, stabilizing the new installation package's processing tasks, disabling the old version's agent instance, and enabling the new agent.

[0063] In this embodiment, the intelligent agent is run based on the application installation package in the target business system, including: if there is a historical version of the intelligent agent running in the target business system, the agent engine distributes tasks to the current version of the intelligent agent and the historical version of the intelligent agent in the application installation package, so that when the task processing attributes of the current version of the intelligent agent meet the preset conditions, the historical version of the intelligent agent is switched to the current version of the intelligent agent.

[0064] In this context, "historical version intelligent agent" refers to an older version of the intelligent agent instance that has been deployed and is currently providing services in the target business system. "Current version intelligent agent" refers to a new version of the intelligent agent instance newly deployed to the target business system via the application installation package, awaiting verification or switching. The agent engine is a coordination component deployed in the target business system, responsible for receiving external task requests and dynamically distributing tasks to the current or historical version of the intelligent agent according to preset strategies. "Task processing attributes" refers to metrics used to characterize the intelligent agent's performance, such as, but not limited to, task success rate, response latency, resource consumption, error rate, and user satisfaction. "Preset conditions" refers to predefined thresholds or rules used to determine whether the current version of the intelligent agent has the capability for full takeover.

[0065] In practical implementation, when a historical version of the intelligent agent already exists in the target business system, and the current version of the intelligent agent is newly deployed, the agent engine can automatically enter a dual-version coexistence mode. Initially, the agent engine can route most tasks to the historical version of the intelligent agent to ensure service stability, while simultaneously allocating a certain percentage (e.g., 5%) of task traffic to the current version of the intelligent agent for canary testing. During the intelligent task processing of the current version of the intelligent agent, the agent engine continuously collects the task processing attributes of the current version of the intelligent agent and compares them with preset conditions. Once all or some task processing attributes continuously meet the preset conditions, the agent engine can automatically switch all task traffic to the current version of the intelligent agent and deactivate the historical version of the intelligent agent, completing a smooth upgrade.

[0066] Optionally, the preset conditions include, but are not limited to, at least one or more of the following combinations: the success rate of tasks for a continuous preset duration is higher than the first threshold, the average response time is lower than the second threshold, the error rate per unit time is lower than the third threshold, resource consumption (such as CPU or memory usage) is stable within the fourth threshold range, user satisfaction is higher than the fifth threshold, the coverage of set business scenarios reaches the sixth threshold, no serious faults occur for more than the seventh duration, health probes pass checks continuously, no specific type of abnormal pattern appears in the logs, the success rate of calls to dependent services is higher than the eighth threshold, and the benchmark indicators of the agent under gray-scale traffic are better than those of historical versions.

[0067] It's important to note that the agent engine can adaptively adjust the proportion of task traffic allocated to the current version of the agent based on real-time business needs, the volume of received tasks, and the task processing attributes of the current version. For example, during high-load periods, it prioritizes using historically superior agent versions, while increasing the task traffic proportion for the current version during off-peak periods. If the current version of the agent performs exceptionally well in a specific task type, it will be prioritized for that type of task. The final switch is only executed after the agent engine confirms that the current version of the agent stably meets preset conditions across all business scenarios. The advantage of this setup is that it ensures the reliability of the target business system service and the efficiency of new version verification while avoiding the potential risks of a global switch due to local metrics meeting targets.

[0068] For example, see Figure 3 It can receive tasks sent by clients (such as web pages, apps, API callers, etc.) based on a proxy engine (such as Nginx proxy, agent version routing), and distribute the tasks to two independent agent instance clusters in the service layer (i.e., the target business system layer), including cluster 1 (V1) and cluster 2 (V2). Cluster 1 (V1) is used to run the old version of the agent (i.e., the historical version of the agent). Cluster 2 (V2) is used to run the new version of the agent (i.e., the current version of the agent).

[0069] Cluster 1 (V1) deploys historical versions of Agent Implementation 1 and Agent Implementation 2. Cluster 2 (V2) deploys newer versions of Agent Implementation 1 and Agent Implementation 2. When the task processing attributes of the new version of Agent Implementation 1 meet preset conditions, the historical version of Agent Implementation 1 can be switched to. Similarly, when the task processing attributes of the new version of Agent Implementation 2 meet preset conditions, the historical version of Agent Implementation 2 can be switched to. Simultaneously, during the task processing by the agent implementations, the processing results are sent to the storage layer. The storage layer includes various storage architectures such as MySQL, Redis, and S3 / NFS.

[0070] The technical solution provided in this embodiment supports the installation and operation of multiple versions of intelligent agents, ensuring uninterrupted service during agent upgrades. Furthermore, by introducing a proxy engine, it implements dynamic task traffic distribution and conditional smooth switching upgrades between the current and historical versions of the intelligent agent based on task processing attributes. This effectively avoids the risks of service instability or functional degradation that may result from direct replacement, ensuring business continuity while providing ample real-world validation for the new version of the intelligent agent. Therefore, the target business system can safely and efficiently iterate intelligent agent capabilities in complex production environments, improving the stability of intelligent operation.

[0071] Next, the technical solutions provided in the embodiments of the present invention will be explained in terms of process. Taking the execution of the technical solutions provided in the embodiments of the present disclosure in a development environment as an example, the development environment integrates the developed intelligent agent and the runtime environment of the intelligent agent. It is necessary to deploy the intelligent agent developed in the development environment to the production environment of the target business system.

[0072] When dealing with the interaction between development and production environments, please refer to, for example... Figure 4 The flowchart shown illustrates this. Its implementation can be as follows: Develop Agent 1, Agent 2, and Agent 3 in the development environment. Then, encapsulate the associated data for Agent 1, Agent 2, and Agent 3 respectively, resulting in application installation package 1 for Agent 1, application installation package 2 for Agent 2, and application installation package 3 for Agent 3. By exporting the agent application into an independently running application installation package file, the application installation package, as a deployment medium, can be deployed to various target business systems in other environments, independent of the development environment. The application installation package contains all the necessary environments for the current agent to run independently, including a complete runtime environment, plugin configuration information, LLM configuration information, and configuration information for the third-party middleware used.

[0073] Application installation packages 1, 2, and 3 can be deployed to the production environment as needed. The production environment includes target business system 1, target business system 2, and target business system 3. It should be noted that the target business systems within the same production environment can be divided as needed by the production environment administrator. This embodiment does not impose such limitations.

[0074] The technical solution provided in this embodiment determines the deployment method of the intelligent agent associated with the intelligent agent in response to the event of publishing the intelligent agent, and obtains the associated data of the intelligent agent. The associated data includes at least the intelligent agent's operating environment information and tool configuration information. The intelligent agent deployment constraints of the target business system are obtained. Based on the intelligent agent deployment method and intelligent agent deployment constraints, the associated data is encapsulated to obtain the application installation package corresponding to the intelligent agent, so as to deploy the application installation package to the target business system. This solves the problems of repeated environment construction, resource waste, high deployment cost and poor intelligent agent operation stability caused by strong dependence on the development platform runtime environment in the prior art. It realizes the dynamic determination of the intelligent agent deployment method matching the intelligent agent in response to the event of publishing the intelligent agent, and obtains the associated data including at least the intelligent agent's operating environment information and tool configuration information. By using the agent deployment constraints and methods of the target business system, all related data are encapsulated and processed to generate an application installation package that is highly compatible with the target business system, self-contained, and compliant. This improves the compatibility between the agent and the target business system, and enables the independent deployment and operation of the agent by simply deploying the application installation package on the target business system. This reduces deployment complexity, resource consumption, and costs, and ensures the stability and security of the agent application in the target business system.

[0075] Figure 5 This is a flowchart of an intelligent agent processing method according to an embodiment of the present invention. Based on the foregoing embodiments, step "S110" is further refined. Specific implementation details can be found in the technical solution of this embodiment. Technical terms that are the same as or corresponding to those in the above embodiments will not be repeated here.

[0076] like Figure 5 As shown, the method specifically includes the following steps:

[0077] S210. In response to the event of publishing the intelligent agent, display the information publishing page; the information publishing page includes a deployment method editing control.

[0078] The information publishing page refers to the interactive interface presented to the user after responding to the agent deployment event, used to edit or confirm information related to agent deployment. The deployment method editing control is a visual component used to display and edit the agent's deployment method, allowing users to view, select, or customize the agent's deployment method. For example, the deployment method editing control includes multiple optional deployment methods such as "containerized service," "Helm Chart package," or "plugin embedding."

[0079] In this embodiment, upon detecting an event that triggers the release of an agent, an information release page can be automatically navigated to or popped up. The deployment method editing control on this page can present the deployment methods supported by the released agent in a preset display format. Additional explanatory information can be added to each optional deployment method, such as its applicable scenarios and resource requirements (e.g., whether it relies on Kubernetes). Optionally, the preset display format includes, but is not limited to, any form such as: dropdown menus, radio button groups, configuration forms, lists, card layouts, etc. This improves the intuitiveness and efficiency of user operations, allowing users to flexibly adjust the selected deployment method in the deployment method editing control according to the actual environment of the target business system.

[0080] Furthermore, the information publishing page can be pre-associated with the agent to be published, and a reasonable default deployment method can be pre-populated in the information publishing page based on the agent's metadata (such as historical deployment records, dependency model types, tool call characteristics, etc.). Going further, when loading the information publishing page, the environmental characteristics of the target business system (such as whether it has a Kubernetes cluster, whether it allows access to external networks, whether it supports container runtime, etc.) can be analyzed to dynamically filter incompatible deployment method options or highlight and recommend the most suitable deployment method. Based on this, even if users view all available options through the deployment method editing control, visual prompts (such as disabled status, warning icons, or text descriptions) can clearly identify which deployment methods may violate security policies, exceed resource capabilities, or cause compatibility issues. This setup fully preserves the user's control over deployment decisions while effectively reducing the risk of deployment failure, operational anomalies, or security violations due to misselection or information asymmetry, thereby improving the reliability of the overall publishing process and the user experience.

[0081] S220. In response to triggering a first editing operation on the deployment method editing control, determine the agent deployment method corresponding to the first editing operation.

[0082] The first editing operation refers to the effective interactive behavior performed by the user on the deployment method editing control, which is used to express their selection or definition of the agent's deployment method. For example, the first editing operation may include: clicking a preset deployment method option (such as "containerized deployment"), selecting a deployment method by dragging and dropping, entering a voice command to specify the deployment type, filling in custom deployment parameters, or uploading / referencing an external deployment template file, etc.

[0083] In this embodiment, when a user clicks a preset deployment method option in the deployment method editing control, this click is identified as the first editing operation. At this time, the deployment method identifier associated with the clicked deployment method option (such as "docker-container" or "helm-chart") can be recorded as the current agent's agent deployment method. Alternatively, when a user references or uploads a deployment template file in the deployment method editing control, this reference or upload operation can also be identified as the first editing operation. In this case, the deployment template file can be parsed to extract the current agent's agent deployment method.

[0084] For example, an information publishing page can be developed in advance on an existing intelligent agent development platform. Based on this information publishing page, an installation package can be exported for a single intelligent agent application; that is, the intelligent agent application can be exported as a standalone installation package. A diagram of the information publishing page can be found here. Figure 6 When exporting an agent, you can select either Docker (a tool for defining and running multi-container Docker applications) or Helmchart in the deployment method editing control.

[0085] S230, Obtain the associated data of the agent; the associated data also includes agent definition information.

[0086] Among them, agent definition information refers to normative data used to describe the behavior, task objectives, interaction logic and operation structure of an agent in a high-level semantic manner. It is expressed in the form of a domain-specific language (DSL) and can be carried by YAML files with a fixed syntax format.

[0087] In this embodiment, the agent definition information is a set of metadata that completely describes the static attributes and dynamic behavioral intentions of the agent in a structured and declarative manner. Its content may include, but is not limited to: the agent's name, version, functional description, input / output format specifications, memory management strategy, execution flow node description information, system prompts, tool binding relationships, inference logic structure, security permission declaration, dependency model requirements, and other static configuration data.

[0088] When acquiring the associated data of an agent, its agent definition information can also be acquired simultaneously, so that it can be packaged into the application installation package. When the target business system installs the application package, if it contains agent definition information, it can be used to support subsequent auditing, debugging, or visual reconstruction operations. For example, during the export process, the visual node flowchart of the agent can be automatically converted into a text-based DSL code; this DSL code can adopt a syntax tree structure, explicitly defining the agent name, the functional description of each node, the connection relationships between nodes, and node attributes (such as timeout settings and retry policies).

[0089] However, in real-world scenarios, there might be situations where the target business system has already deployed a historical version of the agent application installation package, while the new version's agent definition information is completely identical to the historical version. In this case, to avoid redundancy, when packaging the new version's application installation package, you can choose not to include the agent definition information, only retaining the minimum configuration required for operation (such as runtime environment information and tool configuration information). Conversely, if the agent definition information differs between the two versions, the updated definition information must be packaged into the application installation package to ensure that the target business system can correctly identify and run the behavioral logic of the new version's agent.

[0090] To improve publishing efficiency, reduce storage overhead, and enhance the accuracy of user decisions, the information publishing page can also include prompts for exporting agent definitions. Before displaying the information publishing page, the current definition information to be exported and the already exported definition information can be compared to determine the appropriate agent definition export prompt.

[0091] The agent definition export prompt message refers to the suggestive text used to explain the current change status of the agent definition information, export recommendations, or potential impacts. For example, the agent definition export prompt message could be: "The agent definition content has been modified; it is recommended to re-export" or "Consistent with the last exported version; existing installation packages can be reused." The definition information currently to be exported refers to the latest agent definition information at the current export stage, derived from the latest configuration in the development environment. Exported definition information refers to a snapshot of the definition information contained in a previous successful export of this agent; this can be stored in a version management system or installation package metadata as a comparison benchmark.

[0092] In this embodiment, before displaying the information publishing page, the current definition information to be exported can be obtained from the agent development environment, and the most recently successfully exported definition information for that agent can be retrieved from the version repository. By comparing the content differences between the current definition information to be exported and the previously exported definition information field by field, such as whether node attributes have changed, whether the tool list has been added or deleted, and whether the agent name or target description has been adjusted, it is determined whether any differences exist. If differences exist, a suggestion to export the agent definition is generated, such as "Definition information has been updated; it is recommended to export the new version to ensure consistency." If there are no differences, a suggestion not to export the agent definition is generated, such as "Definition information has not changed; the existing deployment configuration can be used." The agent definition export suggestion is clearly displayed on the information publishing page to help users quickly determine whether the agent definition information needs to be re-exported.

[0093] For example, see [link to example]. Figure 6 The information publishing page can display a prompt message for exporting agent definitions: "(!) Prompt: The current version of the definition information to be exported is different from the version of the definition information you exported last time. It is recommended to export."

[0094] It's worth noting that during the comparison between the current definition information to be exported and the already exported definition information, the type of difference can be further identified. For example, if the difference is a non-functional content difference such as comments, spaces, or formatting adjustments, it can be considered a "non-functional change," and a prompt message such as "The definition content has not changed substantially" can be generated for exporting the agent definition. If the difference involves critical changes such as node attributes, it can be considered a high-impact change, and a prompt message such as "Critical definitions have been modified and must be re-exported to avoid runtime anomalies" can be generated for exporting the agent definition. This hierarchical prompt mechanism based on semantic impact helps users more accurately assess the necessity of exporting and prevents production environment failures caused by ignoring important modifications.

[0095] Furthermore, it allows for the comparison of currently pending export definitions with already exported definitions, while also linking code commit history, approval status, and test results. For example, if the currently pending export definition has been modified but has not yet passed testing, a message can be displayed stating, "The new definition has not been verified; exporting it may pose a risk." Conversely, if the modifications have passed full pipeline verification, a message can be displayed stating, "The new definition is ready; exporting and deployment are recommended." Such prompts ensure the quality of agent releases.

[0096] The technical solution provided in this embodiment achieves proactive perception and effective reminders of changes in the agent's status by intelligently comparing the current definition information to be exported with the already exported definition information before displaying the information publishing page, and dynamically generating accurate and context-sensitive agent definition export prompts accordingly. This not only effectively avoids deployment mismatches, functional degradation, or operational anomalies caused by inconsistent definitions, but also prevents redundant export operations when the definition has not changed, saving storage and computing resources. Thus, while ensuring system stability, it improves the efficiency of the agent delivery process.

[0097] To improve the convenience and controllability of users configuring whether to export agent definition information, the information publishing page also includes a definition export editing control; before displaying the information publishing page, in response to triggering a second editing operation on the definition export editing control, the agent definition export method corresponding to the second editing operation is determined.

[0098] The agent definition export method indicates whether to export the agent's definition information, that is, whether the agent's definition information is included in the application installation package. The definition export editing control is a visual interactive component used to control whether to export agent definition information. For example, it can be a checkbox, toggle button, or dropdown option, allowing users to specify whether to include the structured definition information of the agent when generating the application installation package. The second editing operation refers to the specific interactive behavior performed by the user on the definition export editing control, such as checking the "Export Definition" option, switching the export switch to the "On" state, or selecting "Do Not Export Definition" from the dropdown menu.

[0099] In this embodiment, the definition export editing control can be presented as a binary switch, defaulting to a preset state (e.g., "Do not export"), or dynamically set to an initial state based on previously generated agent definition export prompts. When the user performs a second editing operation (e.g., switches to "Export") before displaying the information publishing page, the agent definition export method is determined to be "Export definition information". In the subsequent application installation package generation, the complete agent definition information (e.g., a YAML-formatted DSL file) is embedded in the metadata directory of the installation package, facilitating the target business system's correct understanding of the agent. If the user does not enable export or explicitly selects "Do not export", the agent definition export method is determined to be "Do not export definition information". In this case, the generated application installation package only contains the necessary runtime environment information and tool configuration information, without exposing design-level definition details, thus balancing security and lightweight requirements.

[0100] For example, see [link to example]. Figure 6 When exporting an agent, you can select whether to export it together or not in the export definition editing control.

[0101] To further ensure the stability and rationality of agent deployment, when an agent definition export prompt message suggests exporting (e.g., "Definition has been updated, it is recommended to export again"), but the definition export editing control is still in the "do not export" state, the agent definition export prompt message can be highlighted through preset visual enhancement mechanisms (such as highlighting the prompt area, flashing icons, semi-transparent overlay, or pop-up confirmation dialog boxes) to attract the user's attention, prompt them to carefully evaluate whether the current choice is reasonable, and avoid the new version of the agent behaving abnormally or lacking functionality in the target business system due to the omission of key definition updates.

[0102] The technical solution provided in this embodiment introduces a definition export editing control in the information publishing page and responds to the user's second editing operation to dynamically determine the intelligent agent definition export method. This achieves fine-grained and configurable control over whether intelligent agent definition information is distributed with the installation package, thereby improving the flexibility of intelligent agent delivery.

[0103] Based on the determined agent definition export method, obtain the agent's associated data, including: when the agent definition export method is the common export method, obtain the agent definition information, runtime environment information, and tool configuration information; when the agent definition export method is the non-export method, obtain the agent's runtime environment information and tool configuration information.

[0104] In this embodiment, the agent definition export method is divided into "common export method" (i.e., exporting definition information) and "no export method" (i.e., exporting only the minimum configuration required for operation). When the agent definition export method is "common export method," three types of information can be completely retrieved from the agent metadata repository: agent definition information, runtime environment information, and tool configuration information. These are all treated as associated data and uniformly incorporated into the subsequent application installation package packaging process. If the agent definition export method is "no export method," only runtime environment information and tool configuration information are extracted. This method ensures that the installation package content is concise and secure, follows the user-defined export strategy, and effectively reduces storage overhead and transmission burden.

[0105] The technical solution provided in this embodiment dynamically determines whether to obtain agent definition information based on the agent definition export method, and combines it with runtime environment information and tool configuration information to form associated data. This achieves precise control and flexible adaptation of the agent's delivered content, meets the compliance requirements of lightweight, security and least privilege principles in the production environment, and ensures the correctness of agent functions and operational stability.

[0106] S240. Obtain the agent deployment constraints of the target business system. Based on the agent deployment method and agent deployment constraints, encapsulate the associated data to obtain the application installation package corresponding to the agent, and install the application installation package to the target business system.

[0107] The technical solution provided in this embodiment displays an information publishing page containing a deployment method editing control when responding to an event that releases an agent. This brings the choice of agent deployment method to the publishing entry point, enabling an explicit declaration of deployment intent. Furthermore, by responding to the user's first editing operation on the deployment method editing control, the agent deployment method is determined, improving the user's transparency of the deployment process. This effectively avoids rework, configuration errors, or operational anomalies caused by unclear or mismatched deployment methods, enhancing the end-to-end delivery efficiency and reliability of agents from development to deployment.

[0108] As an optional embodiment of the above embodiments, specific application scenario examples are provided to enable those skilled in the art to further understand the technical solutions of the embodiments of the present invention. Specifically, please refer to the following detailed content.

[0109] See Figure 7 In the development environment's agent development platform, you can select the agent you want to export. When exporting an agent, you can choose the agent deployment method associated with it (including Docker image and Helm chart package deployment methods). Furthermore, you can export the agent's definition information, runtime environment loading, tool configuration information, and other associated data. Based on the agent deployment method and production environment deployment constraints, the associated data is encapsulated to obtain the application installation package corresponding to the agent.

[0110] When deploying an agent in a production environment, you can obtain the agent's application installation package as needed. Deploy it to the target business system and configure middleware information (such as database users). Start the application installation package and configure the tools required for its runtime (such as components, plugins, large models, etc.). Once completed, the agent can be accessed, enabling it to run automatically.

[0111] The technical solution provided in this embodiment develops an intelligent agent application on an intelligent agent development platform in the development environment and exports it as an application installation package. This enables independent installation and deployment in the production environment, facilitating upgrades and switching between multiple intelligent agent versions and ensuring uninterrupted online intelligent agent services. It allows for rapid integration of the intelligent agent application installation package into target business systems or other software, simplifying installation and deployment with low resource consumption. While enabling rapid delivery of the exported intelligent agent as an installation package to business systems, it ensures the compatibility between the deployed intelligent agent and the target business system, guaranteeing the stability and security of the intelligent agent application within the target business system, and simultaneously meeting the security and compliance requirements for isolating development and production environments.

[0112] Figure 8 This is a schematic diagram of the structure of an intelligent agent processing device according to an embodiment of the present invention. Figure 8 As shown, the device includes a data acquisition module 310 and a data processing module 320.

[0113] The data acquisition module 310 is used to respond to the event of publishing the intelligent agent, determine the intelligent agent deployment method associated with the intelligent agent, and acquire the associated data of the intelligent agent; wherein the associated data includes at least the intelligent agent's operating environment information and tool configuration information; the data processing module 320 is used to acquire the intelligent agent deployment constraints of the target business system, and based on the intelligent agent deployment method and the intelligent agent deployment constraints, encapsulate the associated data to obtain an application installation package corresponding to the intelligent agent, so as to deploy the application installation package to the target business system.

[0114] The technical solution of this embodiment determines the deployment method of the intelligent agent associated with the intelligent agent in response to the event of publishing the intelligent agent, and obtains the associated data of the intelligent agent. The associated data includes at least the intelligent agent's operating environment information and tool configuration information. The intelligent agent deployment constraints of the target business system are obtained. Based on the intelligent agent deployment method and intelligent agent deployment constraints, the associated data is encapsulated to obtain the application installation package corresponding to the intelligent agent, so as to deploy the application installation package to the target business system. This solves the problems of repeated environment construction, resource waste, high deployment cost and poor intelligent agent operation stability caused by strong dependence on the development platform runtime environment in the prior art. It realizes the dynamic determination of the intelligent agent deployment method matching the intelligent agent in response to the event of publishing the intelligent agent, and obtains the associated data including at least the intelligent agent's operating environment information and tool configuration information. By using the agent deployment constraints and methods of the target business system, all related data are encapsulated and processed to generate an application installation package that is highly compatible with the target business system, self-contained, and compliant. This improves the compatibility between the agent and the target business system, while enabling the independent deployment and operation of the agent by simply deploying the application installation package on the target business system. This reduces resource consumption and costs, and ensures the stability and security of the agent application in the target business system.

[0115] Optionally, based on the above-described device, the data acquisition module 310 includes:

[0116] The information publishing page display unit is used to display the information publishing page in response to events from the publishing agent; the information publishing page includes a deployment method editing control;

[0117] The deployment method determination unit is used to determine the agent deployment method corresponding to the first editing operation in response to triggering a first editing operation on the deployment method editing control.

[0118] Optionally, based on the above-described device, the information publishing page may further include a defined export editing control; the data acquisition module 310 may further include:

[0119] The definition export method determination unit is used to determine the agent definition export method corresponding to the second editing operation in response to triggering a second editing operation on the definition export editing control; wherein, the agent definition export method is used to characterize whether to export the agent definition information of the agent;

[0120] Optionally, based on the above-described device, the associated data may further include agent definition information; the data acquisition module 310 may further include:

[0121] The first unit is used to obtain the agent definition information, runtime environment information and tool configuration information of the agent when the agent definition export method is the common export method;

[0122] The second unit is used to obtain the operating environment information and tool configuration information of the agent when the agent definition export method is set to non-export mode.

[0123] Optionally, based on the above-described device, the information publishing page includes a prompt message for exporting the intelligent agent definition; before displaying the information publishing page, the device further includes:

[0124] The export prompt information display unit is used to compare the current definition information to be exported and the definition information already exported of the agent to determine the agent definition export prompt information.

[0125] Based on the above-mentioned device, optionally, the data processing module 320 includes:

[0126] An information adjustment unit is used to adjust the tool configuration information in the associated data based on the deployment constraints of the intelligent agent, so as to obtain the adjusted associated data;

[0127] The application installation package determination unit is used to package the adjusted associated data based on the agent deployment method to obtain an application installation package corresponding to the agent.

[0128] Based on the above-mentioned device, optionally, the data processing module 320 includes:

[0129] The installation package storage unit is used to store the application installation package in a shared library, so that the target business system can obtain the application installation package from the shared library and run the intelligent agent based on the application installation package.

[0130] The target business system is configured to, if a historical version of the intelligent agent is running in the target business system, distribute tasks to the current version of the intelligent agent in the application installation package and the historical version of the intelligent agent based on the agent engine, so as to switch the historical version of the intelligent agent to the current version of the intelligent agent when the task processing attributes of the current version of the intelligent agent meet the preset conditions.

[0131] The intelligent agent processing device provided in the embodiments of the present invention can execute the intelligent agent processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0132] Figure 9This is a schematic diagram of the structure of an electronic device implementing the intelligent agent processing method of the embodiments of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0133] like Figure 9 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory 12 or a random access memory 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the read-only memory 12 or a computer program loaded from storage unit 18 into the random access memory 13. The random access memory 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, read-only memory 12, and random access memory 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0134] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0135] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as agent processing methods.

[0136] In some embodiments, the agent processing method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via read-only memory 12 and / or communication unit 19. When the computer program is loaded into random access memory 13 and executed by processor 11, one or more steps of the agent processing method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to execute the agent processing method by any other suitable means (e.g., by means of firmware).

[0137] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0138] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0139] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0140] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0141] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0142] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0143] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from read-only memory 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of the embodiments of the present invention.

[0144] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the intelligent agent processing method provided in any embodiment of this invention.

[0145] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0146] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0147] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for processing intelligent agents, characterized in that, include: In response to an event that releases an agent, the system determines the agent deployment method associated with the agent and obtains the associated data of the agent; wherein the associated data includes at least the agent's operating environment information and tool configuration information. Obtain the agent deployment constraints of the target business system. Based on the agent deployment method and the agent deployment constraints, encapsulate the associated data to obtain the application installation package corresponding to the agent, and deploy the application installation package to the target business system.

2. The method according to claim 1, characterized in that, The step of responding to an event that publishes an agent, and determining the agent deployment method associated with that agent, includes: In response to an event that triggers the release of an intelligent agent, an information release page is displayed; the information release page includes a deployment method editing control. In response to triggering a first editing operation on the deployment method editing control, the deployment method of the agent corresponding to the first editing operation is determined.

3. The method according to claim 2, characterized in that, The information publishing page also includes a defined export editing control; Before displaying the information publishing page, the method further includes: In response to triggering a second editing operation on the definition export editing control, a smart agent definition export method corresponding to the second editing operation is determined; wherein, the smart agent definition export method is used to characterize whether to export the smart agent definition information; Accordingly, the associated data also includes agent definition information; obtaining the associated data of the agent includes: When the agent definition is exported using the common export method, the agent definition information, runtime environment information, and tool configuration information of the agent are obtained. When the agent definition export method is set to non-export, the agent's runtime environment information and tool configuration information are obtained.

4. The method according to claim 2, characterized in that, The information publishing page includes a prompt message for exporting the intelligent agent definition; Before displaying the information publishing page, the method further includes: The current definition information to be exported and the already exported definition information of the agent are compared to determine the agent definition export prompt information.

5. The method according to claim 1, characterized in that, The process of encapsulating the associated data of the intelligent agent based on the agent deployment method and the agent deployment constraints to obtain an application installation package corresponding to the intelligent agent includes: Based on the deployment constraints of the intelligent agent, the tool configuration information in the associated data is adjusted to obtain the adjusted associated data; Based on the aforementioned agent deployment method, the adjusted associated data is packaged to obtain an application installation package corresponding to the agent.

6. The method according to claim 1, characterized in that, Deploying the application installation package to the target business system includes: The application installation package is stored in a shared library so that the target business system can obtain the application installation package from the shared library and run the intelligent agent based on the application installation package.

7. The method according to claim 6, characterized in that, The process of running the intelligent agent based on the target business system according to the application installation package includes: If the target business system is running a historical version of the intelligent agent, then the agent engine distributes tasks to the current version of the intelligent agent in the application installation package and the historical version of the intelligent agent, so that when the task processing attributes of the current version of the intelligent agent meet the preset conditions, the historical version of the intelligent agent is switched to the current version of the intelligent agent.

8. An intelligent agent processing device, characterized in that, include: The data acquisition module is used to respond to an event that releases an intelligent agent, determine the deployment method of the intelligent agent associated with the intelligent agent, and acquire the associated data of the intelligent agent; wherein, the associated data includes at least the intelligent agent's operating environment information and tool configuration information; The data processing module is used to obtain the agent deployment constraints of the target business system, and based on the agent deployment method and the agent deployment constraints, encapsulate the associated data to obtain the application installation package corresponding to the agent, so as to deploy the application installation package to the target business system.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the agent processing method according to any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the agent processing method as described in any one of claims 1-7.

Images